Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Does Poor Neighbourhood Reputation Create a Neighbourhood Effect on Employment? The Results of a Field Experiment in the UK:

Does Poor Neighbourhood Reputation Create a Neighbourhood Effect on Employment? The Results of a... There are substantial variations in labour market outcomes between neighbour- hoods. One potential partial explanation is that residents of some neighbourhoods face discrimination from employers. Although studies of deprived areas have recorded resident perceptions of discrimination by employers and negative employer perceptions of certain areas, until now there has been no direct evidence on whether employers treat job applicants differently by area of residence. This paper reports a unique experiment to test for a neighbourhood reputation effect involving 2001 applications to 667 real jobs by fictional candidates nominally resident in neighbour- hoods with poor and bland reputations. The experiment found no statistically signif- icant difference in employer treatment of applicants from these areas, indicating that people living in neighbourhoods with poor reputations did not face ‘postcode dis- crimination’ in the labour market, at the initial selection stage. Introduction This article explores neighbourhood effects labour markets. It focuses on exploring one as an explanation for variations in employ- pathway as an explanation: ‘stigma’, or poor ment rates between neighbourhoods within neighbourhood reputation. It exploits a Rebecca Tunstall is in the Centre for Housing Policy, University of York, Heslington, York, YO10 6DT, UK. Email: becky.tunstall@york.ac.uk. Anne Green is in the Institute for Employment Research, University of Warwick, Social Studies Building, Coventry, CV4 7AL, UK. Email: Anne.Green@warwick.ac.uk. Ruth Lupton and Katie Bates are in the Centre for the Analysis for Social Exclusion, London School of Economics, London, UK. Email: r.lupton@lse.ac.uk and katie.bates@lse.ac.uk. Simon Watmough is in the European Institute, Fiesole, Firenze, Italy. Email: simon.watmough@ eui.eu. 0042-0980 Print/1360-063X Online 2013 Urban Studies Journal Limited DOI: 10.1177/0042098013492230 764 REBECCA TUNSTALL ET AL. field experiment methodology, supported 2006). This paper focuses on variations by contextual qualitative and quantitative between neighbourhoods within LLMs. In evidence, through case studies of three local the UK, the National Equality Panel (2010) labour markets (LLMs) and nine neigh- highlighted that in the most deprived tenth bourhoods in England and Wales. It focuses of neighbourhoods only 55 per cent of on one outcome: employment. The hypoth- adults were employed compared with more esis it aims to test is that than 80 per cent in the least deprived half of neighbourhoods nation-wide. Residents of neighbourhoods with poor repu- Concerns about uneven employment tations fare worse when applying for relatively rates between neighbourhoods and the (rel- low-skilled jobs than residents of neighbour- atively) low absolute employment rates in hoods in the same labour market with better some areas have been part of the motivation reputations, all other things being equal. for successive generations of urban policy in the UK and elsewhere. Closing gaps in If verification is found for this hypothesis, neighbourhood worklessness was a key it is evidence of a neighbourhood effect objective of the National Strategy for operating through the ‘stigma’ pathway Neighbourhood Renewal (in England) from (although it is possible that neighbourhood 1998 to 2008 (Amion Consulting, 2010), effects might also be operating through the Working Neighbourhoods Fund (in other pathways). It would suggest that England) (Dewson et al., 2007) and the City neighbourhood effects contribute to some Strategy Pathfinders (in Great Britain) extent to variation in employment rates (Green and Adam, 2011). Areas with high between neighbourhoods. unemployment and low employment tend The article begins by examining varia- to overlap with areas with poor reputations. tions in employment rates at neighbour- hood level and discussing explanations for such variations. The methodology used for Explaining Variations in the particular field experiment which is the Employment Rates between focus of the article is then discussed. Next Neighbourhoods within Labour the results of the experiment are presented. Markets The article ends with a discussion of what There are three broad categories of expla- the results mean for the role of the ‘stigma’ nation for variation within labour markets: neighbourhood effects pathway. first, skills mismatch, largely as a result of residential sorting; secondly, spatial mis- match; and thirdly, neighbourhood effects. Variations in Employment Rates These explanations are not exclusive and between Neighbourhoods within could operate together for additive or inter- Labour Markets acting effects. There is longstanding academic and policy First, residential sorting and changes in interest in the existence and persistence of skills requirements can result in spatial geographical variations in employment, concentrations of individuals who are at economic inactivity and unemployment greater risk of non-employment due to rates. These concerns encompass a range of poor skills, lack of qualifications, ill health geographies—from regional, to LLM and and lack of recent work experience neighbourhood scales (Green and Owen, (Houston, 2005; Green and Owen, 2006). POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 765 Hence, spatial variations in employment institutional, with 14 sub-categories. Under and non-employment rates are a function ‘institutional’ mechanisms, he lists local of spatial differences in individual charac- institutional resources, the behaviour of teristics and a mismatch between the skills local market actors and private services, and or attributes required or preferred by stigmatisation. Thus, the behaviour of employers and those held by people in cer- employers and neighbourhood stigma are tain neighbourhoods. Skills mismatches just two of numerous potential mechanisms. could be reduced by improving individuals’ relative skills (and other characteristics). Secondly, the spatial mismatch hypoth- Explaining Variations in esis explains variations in non-employment Employment Rates between through geographical disparities between Neighbourhoods within LLMs the location of jobs and potential employ- through Neighbourhood Effects ees (see Kain, 1968; Holzer, 1991; Houston, Much research into neighbourhood effects 2005). Spatial mismatch can develop has used multivariate analysis of quantitative through residential sorting, the impact of data on the characteristics of, and outcomes housing policy and uneven geographical for, residents of different neighbourhoods impacts of economic restructuring. For (Kleinhans, 2004; Joseph et al., 2007; Galster, instance in a case study of an inner London 2007). Qualitative research (for example, borough, Watt (2003) described how many with employers and intermediaries), local authority tenants were in effect mar- although sometimes overlooked, is particu- ooned over a 30-year period as the immedi- larly important in generating hypotheses and ate area lost almost all its manual in developing and testing ideas about path- employment. Over time, those in employ- ways (Lupton, 2003a). Another approach ment travelled longer distances to work. with potential is the use of randomised con- Thirdly, when applied in relation to trol trials or experimental methods. employment outcomes, the neighbourhood Each of these methods is independently effects hypothesis suggests that spatial varia- valid. However, combinations of methods tions in non-employment are not reducible offer the potential to confirm and accumu- to compositional effects (i.e. characteristics late evidence. While econometric analysis is of residents) but rather that there is an addi- broadly seen as the ‘gold standard’ within tional area related effect on employment (or neighbourhood effects research, in the field other outcomes) that results from spatial of employer preferences, this kind of data concentrations of individuals who are at analysis is seen as secondary to experimental greater risk of non-employment or who studies (Riach and Rich, 2004). Research have other disadvantages (Syrett, 2008), or into neighbourhood effects so far has tended from other characteristics of the area (for to explore neighbourhood effects without example, Lupton, 2003a; van Ham et al., focusing on specific mechanisms (Galster, 2012). Potential pathways for neighbour- hood effects may be either endogenous (i.e. 2012). This paper explores neighbourhood internal to the neighbourhood) or due to the relationship between the neighbourhood effects as an explanation for variations in employment rates between neighbourhoods and outside people or institutions. In a com- prehensive typology of mechanisms, Galster within labour markets. It focuses on explor- (2012) prefers the categories of social inter- ing one pathway as an explanation: ‘stigma’, active, environmental, geographical and or poor neighbourhood reputation. It 766 REBECCA TUNSTALL ET AL. exploits a field experiment methodology, residents. These reputations may affect the supported by contextual qualitative and behaviour of all these groups towards the quantitative evidence, through case studies neighbourhoods—for example in decisions to provide services or to employ residents. of three LLMs and nine neighbourhoods in England and Wales. It focuses on one out- Existing residents are likely to be aware of these poor external reputations and this in come: employment. itself may affect their attitudes and beha- The hypothesis it aims to test is that viours, including, for example, decisions to apply for jobs and their behaviour in the Residents of neighbourhoods with poor repu- job application process. Poor reputation tations fare worse when applying for relatively may also affect the provision of private and low-skilled jobs than residents of neighbour- hoods in the same labour market with better public services (for example, Christie and Rolfe, 1992). reputations, all other things being equal. Over 30 years ago, McGregor (1977) used multivariate analysis, now established If verification is found for this hypothesis, it is evidence of a neighbourhood effect as the core approach to investigating neigh- bourhood effects, to explore the role of operating through the ‘stigma’ pathway stigma in explaining variations in employ- (although it is possible that neighbourhood ment rates between neighbourhoods within effects might also be operating through labour markets through neighbourhood other pathways). It would suggest that effects. He examined unemployment rates neighbourhood effects contribute to some and unemployment duration for men in extent to variation in employment rates Scotland, including Ferguslie Park, a neigh- between neighbourhoods. bourhood with a poor reputation, control- ling for age, skill, industry, marital status, length of time in previous job, preference Evidence for the Role of for light, medium or heavy work, prefer- Neighbourhood Reputation in ences for local or more distant work, and Explaining Variations in preferences for day or any hours, and con- Employment Rates through cluded that Neighbourhood Effects Evidence that some neighbourhoods have although the Ferguslie Park sample would worse reputations than others is ubiquitous experience relatively high unemployment in urban studies (for example, Wilson, duration no matter where they lived, the fact 1987; Tilly et al., 2001; Waquant, 1993; of their residence in Ferguslie Park signifi- Hastings and Dean, 2003; de Souza Briggs cantly adds to the disadvantages associated et al., 2010). Neighbourhoods with poor with their individual characteristics reputations often are the same neighbour- (McGregor, 1977, p. 311). hoods that have low rates of employment and high rates of non-employment and To date, no other studies appear to have unemployment. This brief review focuses used such techniques to explore the stigma on evidence from the UK. pathway for neighbourhood effects on Relatively poor reputations may be held employment. by outsiders including the general public, Studies using other methods have added key public- and private-sector decision- important information. Although employer makers and service providers, or potential statements may not fully reflect practice and POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 767 do not preclude unconscious preferences, Nottingham, Green et al. (1991, p. 273) several studies provide insights into poten- found evidence of a preference amongst tial neighbourhood preferences from an some employers for local workers, regard- employer perspective (Pager and Quillian, less of neighbourhood reputation: ‘‘[they] 2005). Writing about four US metropolitan preferred their employees to be locally areas, Tilly et al. noted that based, since this was thought to make them more reliable’’. Zenou (2002) and Lupton each manager forms his or her own mental (2003b) uncovered similar preferences. This map of . [their] environment. Employer’s suggests a potential ‘spatial match’ effect or maps, in turn, have important effects on the positive neighbourhood effect for people in labor market (Tilly et al., 2001, p. 304). neighbourhoods close to employment. Qualitative research can contribute to A majority of employers in this study understanding of how pathways for neigh- thought there were systematic differences bourhood effects may work and the circum- between urban and suburban workers stances in which they may not operate. (across neighbourhoods) within a labour Nunn et al. (2010) suggested that neigh- market, although in the US context it bourhood effects via area reputation should be noted that spatial differences are required ‘local knowledge’ amongst employ- also heavily racialised. Negative attitudes ers, discretion and ability to deviate outside about particular groups of workers could automated or rigid processes amongst those feed into spatial mismatch, if they affect involved in recruitment. Like the exercise of employers’ location decisions and the pat- other forms of employer preference or dis- tern of employment opportunities. crimination, this may have been influenced In the UK, some studies have found at by the state of the labour market, with eco- least some self-reported preference for indi- nomic circumstances offering more choice viduals not from areas with poor reputa- for employers resulting in more screening. tions (for example, Hastings and Dean, Another source of information is inter- 2003; Aleksandraviciene et al., 2005). views with job applicants and those working Interviews with employers, with recent with them.Thisisaweaker source,asthese experience of recruitment in less skilled informants have only indirect insight into occupations in various cities in Great employer preferences and no oversight of Britain, about their selection and recruit- aggregate patterns. However, numerous stud- ment procedures suggested that employers ies have recorded the belief that ‘postcode dis- preferred workers from areas without poor crimination’ is taking place in the UK (for reputations, but the effect was only at the example, Lawless and Smith, 1998; SEU, margin and in specific conditions (Nunn 1998; Taylor, 1998; Fieldhouse, 1999; Roberts, et al., 2010). Employers generally talked of 1999; Dean and Hastings, 2000; Speak, 2000; individuals’ personal characteristics as the Mellor, 2002; Hastings and Dean, 2003; main factor in recruitment decisions. Taylor, 2003; Aleksandraviciene et al., 2005; Address could be an issue, but this mainly Sanderson, 2006; Green and White, 2007; related to travel-to-work considerations, Fletcher, 2007; Dewson et al., 2007; Bates particularly where jobs involved anti-social et al., 2007; Fletcher et al., 2008) and beyond hours. A few indications of a lack of will- (for example, in France (Waquant, 1993; ingness to employ applicants from certain Recchia, 2008), in Australia (Atkinson and areas, or an intention to subject them to Jacobs, 2008) and in the US (Tilly et al., special attention, emerged. In their study of 2001)). More specifically, some social housing 768 REBECCA TUNSTALL ET AL. tenants interviewed in four different areas Houston summed up the situation for the about employment in 2008 thought that there UK as follows: ‘‘direct evidence of ‘post- was postcode discrimination against them code discrimination’ is difficult to find’’ (Fletcher et al., 2008). In a study of working- (2005, p. 229). The pattern appears similar class women in Scotland and northern at least for France, the US and Australia. England, Taylor (2003, par. 7.4) reported: ‘‘Several . explained their long-term unem- ployment by having specific, devalued post- Exploring the Role of codes’’. Bates et al. (2007) found at least one Neighbourhood Reputation in such allegation in a study of employment in Variations in Employment Rates rural areas. A study of high unemployment through Neighbourhood Effects areas in Coventry reported Using an Experimental Method Given the difficulties establishing direct evi- Although they could not easily always pro- dence of neighbourhood effects linked to a vide evidence for it, many project workers particular pathway through multivariate anal- felt that ‘postcode discrimination’ against ysis, or through interviews alone, the study people from areas of high unemployment who are stereotyped as unreliable workers used an experimental method, supported by contextual quantitative and qualitative undoubtedly operated (Aleksandraviciene et al., 2005, pp. 64–65). research. This permits testing the hypothesis set out earlier, through establishing In the evaluation of the 12 Working (1) variations in employment rates between Neighbourhood Pilots covering areas of neighbourhoods; high unemployment and inactivity, inter- (2) variations in reputation between viewees said neighbourhoods; (3) employers’ awareness of neighbour- Employers tar everyone with the same brush. hood reputations, attitudes to different It’s just not fair. The area definitely goes neighbourhoods, assessment of neigh- against you [and] if you put [this area] as bourhood reputation as a factor in deci- your address on your application that puts sion-making; and employers off (Dewson et al., 2007, p. 32). (4) variations in success in applying for employment amongst residents of neigh- In a rare example of direct evidence, albeit bourhoods with different reputations, all second-hand and for a single case, Speak other factors being equal (same jobs; (2000) reported that a resident from Benwell equivalent applicant characteristics). in Newcastle upon Tyne had been inter- viewed for a job but was not given it and was There is very limited experimental evidence told: ‘‘It’s not you, we think you’d do the job on neighbourhood effects. There are ethical, fine. but if you live in [Benwell] you either know a villain or you are a villain’’. legal, political and cost problems in creating In summary, existing evidence from the experimental policy design in urban and UK on the role of stigma in explaining var- social policies (Stafford et al., 2001). In a iations in employment rates between neigh- few cases policy design has enabled rando- bourhoods within labour markets through mised control trial methods in housing and neighbourhood effects was described urban policy, most notably the Moving to recently as ‘‘thin’’ (Dewson et al., 2007). Opportunity (MTO) programme for the POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 769 Table 1. Case study LLMs Case study labour Approximate population JSA claimant rate, Rank by unemployment market name of working age, 2010 2010 (percentage) rate out of UK TTWAs, 2010 (weighted by population of working age) Weak 250,000 5.7 Highest 5 per cent Medium 500,000 4.1 Not highest 5 per cent but highest 20 per cent Strong 250,000 3.2 Lowest 40 per cent Source: NOMIS (www.nomisweb.co.uk; accessed March 2011); based on 2007 mid-year population estimates. relocation of public housing residents in the 1994). It focuses on employer behaviour. US (de Souza et al., 2010). This is amongst Applicant behaviour—for example in response the most quoted studies in reviews of evi- to belief about employer prejudice—may also dence on neighbourhood effects. There has play a role in explaining employment varia- been some use of randomised assignment to tions between areas. different treatment groups in welfare-to- work interventions (Eardly and Thompson, Labour Market and Neighbourhood Case 1997; Kornfeld et al., 1999; Rangarajan and Studies Novak, 1999; Walker, 2000; Stratford et al., 2005; Purdon et al., 2006; Burns et al., 2007). Three LLMs across the UK were selected (see Table 1). The criteria for selecting them were that The Field Experiment Method — each should have at least some neigh- The experimental method we used was the bourhoods with markedly poor local ‘correspondence test’, involving sending out reputations; multiple applications to real jobs, differing as — each should be large enough to generate faraspossibleonlyin terms of thevariable sufficient vacancies but small enough being tested; here, the reputation of the appli- that local neighbourhoods within them cant’s neighbourhood. This has been used might be known to employers; before in studies of employer preferences or — they should include ‘weak’, ‘medium’ and discrimination in employment by ethnicity, ‘strong’ LLMs in terms of unemployment age, gender and disability status (Jowell and rates, in case competition for vacancies Prescott-Clarke, 1970; Riach and Rich, 2004; affected the neighbourhood effects; and Pager, 2007; McGinnity et al., 2009; Wood — each should have low minority ethnic et al., 2009). It has some inevitable limitations: populations, to rule out any effects it cannotbeusedtoresearchthatpartofthe from employer preference for different labour market where jobs are not formally ethnic groups (Fieldhouse, 1999; Tilly advertised, or the parts of the recruitment pro- et al., 2001). cess that involve face-to-face interactions. There is evidence that recruitment through Three neighbourhoods were selected in informal networks tends to mean recruitment each LLM using the following criteria, based of people similar to existing staff (Marsden, on desk research on local identity and 770 REBECCA TUNSTALL ET AL. reputations; analyses of secondary sources public identified the worst reputation two on deprivation levels, ethnic mix and acces- or three neighbourhoods with remarkable sibility indicators; and field visits— consistency. This was used to confirm the including street interviews (discussed in selection of the ‘poor reputation’ and ‘bland more detail later) which were used to con- reputation’ neighbourhoods from desk firm the poor reputation and bland reputa- research. In each of the three LLM areas, tion of shortlisted areas just over half of those interviewed (53 per cent) thought that the ‘poor reputation’ — two particularly deprived neighbour- case study neighbourhoods were undesir- hoods with well-established ‘poor’ able places to live. The same proportion reputations (five out of the six neigh- thought it was ‘very’ or ‘fairly likely’ that bourhoods were in the top 5 per cent of local employers would look less favourably Indices of Multiple Deprivation [IMD] on people from these or similar areas. Only scores andone wasin the top15per a minority (23 per cent) thought that it was cent) and one not particularly deprived ‘very’ or ‘fairly unlikely’ that employers neighbourhood (all three neighbour- would look unfavourably on applicants hoods were in the middle 40–60 per from these neighbourhoods. cent of neighbourhoods on IMD scores) with a ‘bland’ reputation, to Interviews with employers serve as a comparator; To provide insights into employers’ aware- — each had relatively strong local identities, ness of neighbourhood reputations, atti- linked to identifiers including neigh- tudes to different neighbourhoods, bourhood name, main street name or assessment of neighbourhood reputation as postcode; a factor in decision-making, and to confirm — each had relatively small minority ethnic our understanding of employers’ recruit- populations (the White British popula- tion share was around 97 per cent in the ment methods and the realism of the appli- cations used in the experimental method neighbourhoods in the ‘weak’ LLM, 93 per cent in the ‘medium’ and 90 per cent (discussed later), we interviewed 14 employers and 11 labour market intermedi- in the ‘strong’ LLM); — all were at a similar distance and public aries (i.e. employment agencies and job see- kers’ advisors). Employers interviewed were transport travel time from the city centre, to limit (although not exclude) based in the three case study LLMs and had recently recruited people to one or more of any effect of employer preference for more accessible employees (Green the job types examined in the experiment (see later). They encompassed different et al., 1991; Zenou, 2002; Lupton, 2003b; Nunn et al., 2010). establishment sizes and industries, predo- minantly from the private sector but also from the public and voluntary sectors. Field Visits and Street Interviews All employers and labour market inter- The field visits included interviews with 81 mediaries interviewed were aware of certain members of the public, providing what neighbourhoods with poor reputations. appears to be the first, although small scale, Again, there was consistency between inter- evidence of the perceptions of the general viewees and their views overlapped closely public on potential neighbourhood reputa- with the assessment of the general public tion effects. In each LLM, members of the and desk research. Three of the 11 labour POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 771 market intermediaries thought that neigh- One of the employers interviewed in the bourhood reputation effects on employ- medium LLM admitted to ‘thinking twice’ ment might exist. One intermediary in the about applicants from some neighbourhoods medium LLM reported some cases of appli- cant perceptions of neighbourhood reputa- There are areas where you sort of think, tion effects hmm, you know, not too sure about that . where you have a large number of unem- There have been instances where people have ployed people, where you have council said: ‘Because of where I live, I won’t get a job. accommodation. They won’t trust me’. Some of the employers stated preferences Another, from the weak LLM, said that he on other grounds such as distance to work, believed in neighbourhood reputation clothing and accent. effects strong enough to promote action The Experiment We used to take address off applications . we remove address to remove a stigma. The jobs search and application strategy In the strong LLM, another said of fictional job applicants. In the experi- mental phase of the study, we searched for Employers would say ‘no’ to the idea that and applied to real jobs in the three LLMs they discriminate on area—but privately the that fitted the following criteria answer is ‘yes’. — jobs advertised on www.direct.gov.uk, gumtree.com and aggregator sites; This respondent subsequently significantly amended the point, ‘‘it’s not necessarily — advertised August 2010-June 2011; — the job location appeared to be within about prejudice, but the majority of people have barriers to entry’’. While this evidence the TTWA boundary; is strong enough to support the hypothesis — selected job types (office admin, cleaner, that neighbourhood reputation effects security guard, sales assistant, accounts might form a partial and subsidiary expla- clerk, kitchen hand and chef jobs). nation for variations in employment rates, These job types did not require degrees, and encourages further exploration, it is not vocational qualifications or substantial in itself strong enough to support a hypoth- experience, but were prevalent enough esis that they are a major explanation. to give a good overview of the low- and medium-skilled labour market and Almost all employers stated unambiguously that they looked equally on applications from included jobs principally held by men, by women and either gender. all areas. However, one private-sector employer from the strong LLM explicitly showed aware- ness of the potential for discrimination, but From this subset, we selected jobs that met claimed that he would not apply it in practice, the requirements of the experimental at least up to interview stage method I am not going to be that prejudiced on areas — for which main recruitment and selec- . So you get ‘em in, and see what they’re like tion decision-maker appeared to be . (emphasis added). based in the LLM, and was likely to be 772 REBECCA TUNSTALL ET AL. aware of local neighbourhood However, the personas were intended to reputations; represent people who would be relatively — could be applied to without face-to- attractive candidates for jobs that required face or phone contact (i.e. via email, limited education and skills. The aim was upload to website or post); and to limit the number of cases in which none — for which the closing date was at least of the three applicants was successful, 2–3 days hence, to allow time to make which would hinder investigation of dis- three applications that did not arrive crimination, while ensuring applications simultaneously. were realistic. To obviate any employer preference for other characteristics, all three applicants for any one job had the The jobs applied to. Between August same gender. All had names chosen to 2010 and June 2011, 2001 applications avoid signalling minority ethnicity (see were made to 667 jobs (i.e. three applica- Wood et al., 2009, for further discussion). tions in each case). There were 197 applica- All were in their early 20s. tions to office admin jobs, 139 to retail One of the three applicants for each job jobs, 97 to chef jobs, 75 to cleaner jobs, 74 appeared to be living in each of the three to kitchen hand jobs, 73 to accounts clerk local neighbourhoods selected. The appli- jobs and 12 to security guard jobs, reflect- cant address, including a fictional numbered ing the incidence of vacancies and varia- home in a real major street in the area likely tions in the extent to which jobs fitted to be well known, the area name and the experimental constraints. Of the jobs postcode were prominently stated at the top applied for, 246 were in the strong LLM, of each CV. Applicants differed in their 261 in the medium LLM and 159 in the exact qualifications and work experience, weak LLM; 76 per cent of jobs applied for and CV typeface and layout used, in order to did not offer a traditional full-time ‘9am– maintain as much similarity between candi- 5pm’ work schedule. This largely reflects dates as possible without raising the suspi- the type of jobs available via www.direct. cions of employers. Addresses in each of the gov.uk in the three LLMs. Fifty-four per three neighbourhoods were allocated ran- cent of the jobs for which wage data were domly to the three prepared CVs and cover- available paid at the minimum wage level. ing letters as the final stage of the process in application to every job. CVs’ contents were The characteristics of the fictional job placed on one of three CV templates with applicants. ‘Personas’ were created for different fonts and layouts, which were each of the job types applied to, with fic- made similarly attractive. Age, qualifications tional names, addresses, dates of birth, edu- and work experience were rotated between cational and work histories and real phone CV templates at intervals throughout the numbers and email addresses, which were experiment. This random allocation ensured used as the basis of CVs and covering let- that any differences in employers’ responses ters. These were then slightly tailored to to candidates living in different neighbour- individual jobs. It is possible that employ- hoods could not be attributed to their age, ers might be more likely to express neigh- the style of their CV, exact qualifications or bourhood reputation discrimination minor differences in work experience. against less attractive candidates. This has Most other applicant characteristics were been found in the case of ethnic discrimi- largely determined by job type and were nation (Dovidio and Gartner, 2000). intended to create candidates for the various POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 773 jobs who were promising but not exception- advertisements were first posted on the ally overqualified. In 286 (43 per cent) of Internet. More than half were sent within the jobs, all three applicants were male; in three days of the advert first appearing. the remainder, all were female. Jobs were Recording the results for the fictional allocated by gender in accordance with the candidates. After sending off three gender profile of employment revealed by matched applications to any one job, we secondary data sources. The slight predomi- monitored specially established email and nance of female applicants in the experi- voicemail accounts to receive employer ment overall reflects the availability of job responses. We were not able to monitor types advertised and the availability of jobs any employer responses that might have which we could apply to using our methods. been made by post to applicants’ home Applicants were allocated academic and addresses, but employer and intermediary vocational qualifications of the types com- interviews suggested that communication monly held by young people in the LLMs by post was exceptional. As the addresses and relevant to the post they were applying used were false ones, any employer writing for. All of the applicants had continual work to one of our candidates would have had records since they had completed education mail returned and could then have tried and some were given Saturday job experi- another means to get in touch, but we did ence before this. not learn of any such experiences. Wherever a job required a car and a clean Some 620 applications (31 per cent) driving licence, they were given to all three resulted in responses from potential employ- applicants, and likewise to candidates for ers. When we received a positive response, any jobs with unsocial hours in order to dis- we responded as fast as possible via email to tinguish between any discrimination by withdraw the applicant from consideration, employers based on ‘travelability’ rather stating that the candidates had already than discrimination against people from accepted another offer or that their circum- areas with poor reputations. Hence the can- stances had changed. Where more than one didates had a higher rate of private transport application for the same job received a posi- mobility than might be typical for appli- tive response from employers, the style and cants for these types of jobs. content of our responses to employers were varied. One month after the application, we stopped monitoring for responses. Making applications. The majority of The experiment was concerned with pos- applications took the form of a CV or a CV itive responses at the first selection stage. and covering letter. Covering letters were Thus the experimental results cannot be used in many cases, even if they were not applied directly to subsequent stages, such explicitly demanded by employers, in order as interviews, or to the selection process as to maximise positive response. Over 85 per a whole. cent of applications were made via the Internet. We used local assistants to put the remainder of applications in the post, to Results of the Field Experiment achieve local postmarks. In most of these cases, we avoided sending all the applica- ‘First-stage’ Positive Response Rates tions on the same day in order to reduce employer suspicion. In most cases, applica- A total of 17 per cent of the 2001 applica- tions were made very rapidly after tions received one of a range of first-stage 774 REBECCA TUNSTALL ET AL. positive responses. This share of positive to all candidates should not be treated as responses is higher than that achieved in positive evidence of non-discrimination by another recent experimental study (Wood employers. In the second case, all met some et al., 2009). This may reflect differences in minimum standard and, while all appear to the mix of jobs applied for, the LLMs studied have been preferred over any other real or the quality of experimental applications. candidates applying, there is no sign of pre- Three applicants were offered a post right ference between them. away. Most first-stage positive responses did In 475 (71 per cent) of the 667 jobs applied not lead directly or with any certainty to job for, none of the three candidates received a offers. Thirteen per cent of applicants were first-stage positive response. In the remaining invited to meet employers and 2 per cent were 192 (29 per cent) of jobs, employers gave a asked for further information—for example, positive response to one or more of our can- what days and hours they might be able to didates for the same post. These are the cases work. Thus fewer than one in five of the used to explore employer preferences and dis- experimental candidates got through the first crimination, following Bovenkerk (1992) and round of selection for jobs that required rela- Wood et al. (2009). tively limited skills and experience, and which generally paid close to the minimum wage. ‘First-stage’ Positive Response Rates Where Young people with fewer labour market Employer Showed A Preference for One or advantages would be likely to experience More Candidates lower first-stage positive response rates when In accordance with the experimental metho- applying for the same sorts of jobs. Thirteen dology deployed, in each of the 192 sets of per cent of applications received one of a applications with at least one first-stage pos- range of responses we classified as negative, itive response, one of the applications was including an acknowledgement of application for a candidate from a bland reputation but no further correspondence, or notice they neighbourhood. Of these, 140 (62.5 per had been unsuccessful; 69 per cent of applica- cent) applications received a first-stage posi- tions received no response of any kind. Previous studies using an experimental tive response (Table 2). The other two appli- cations in each of the 192 sets were from method to test for discrimination in employment have discussed several possible candidates from poor reputation neighbour- hoods, totalling 384 applications. Of these, methods of conceptualising discrimination and analysing results (for example, Riach 230 (59.9 per cent) applications received a positive response. The 2.6 percentage point and Rich, 2004). The key issue is how to deal with jobs in which none of the candi- difference between the success percentages for the two neighbourhood types provides a dates received a positive response, and those in which all of the candidates did. measure of aggregate net ‘postcode discrimi- nation’. However, the net preference was The first case suggests none met some min- imum standard, or that no appointment small and it was not statistically significant at was actually made, and offers no sign of the 1 per cent, 5 per cent or 10 per cent level. preference between the experimental candi- In each of the three LLMs, there was a dates. Wood et al. (2009) who conducted difference between the positive response the most recent experiment of this type in rates for the neighbourhood types, with the UK, and Bovenkerk (1992), who has applications from the bland reputation prepared a manual for the conduct of these neighbourhood having a slightly higher tests and analyses, argue that non-response positive response rate. The level of net POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 775 Table 2. Employers’ preference for applicants from different neighbourhood types Total sets of applications (A) Positive response: (B) Positive response: Net preference P-value with one or more bland reputation poor reputation percentage point positive responses neighbourhood neighbourhood difference (A-B) Number percent Number percent 192 120 62.5 230 59.9 2.6 0.1170 As there were twice as many poor reputation as bland reputation neighbourhood applications (384 compared with 192), the denominator for column B is the number of sets of applications with one or more positive responses multiplied by two (384). We used the two-sample z-test for proportions. Source: Experiment. preference found was highest in the reputations fare worse when applying for medium LLM, at 4.5 percentage points, but relatively low-skilled jobs than residents of in no area was it statistically significant. neighbourhoods in the same labour market We examined subsets of the results for with better reputations was not proven. evidence of ‘postcode discrimination’ in any It remains possible that, in a larger corre- particular job or employer type. In some spondence test study, the net preferences found would achieve statistical significance. cases, there was evidence of a small amount of net preference, but in no case was the dif- It also remains possible that poor neigh- bourhood reputation might create neigh- ference in positive responses between those from neighbourhoods with different reputa- bourhood effects on employment in parts of these LLMs or parts of the recruitment and tions statistically significant. In summary, we do not find statistically selection process that were outside the scope of this experiment, in particular for unpro- significant evidence that employers prefer those living in neighbourhoods with bland mising candidates and for jobs applied to reputations to those living in neighbour- face-to-face. hoods with poor reputations, in the case of As it stands, this is a different result from that found through multivariate anal- attractive candidates looking for work in selected jobs requiring limited education ysis by McGregor (1977). One interpreta- and skills, in three contrasting LLMs. There tion is that the earlier study’s method was not able to take account of all hidden vari- was some net preference for candidates from neighbourhoods with bland reputa- ables. Another, perhaps more important one, however, is that the market for male tions but it was small in size and not statis- tically significant. manual labour in Glasgow in the 1970s dif- fered from the market for low-skilled ser- vice employment for both genders across Discussion England and Wales in 2010–11. The size of the net preferences is small in The ‘Stigma’ Neighbourhood Effects relation to the differences in employment Pathway and Variation in Employment rates between neighbourhoods. It is not so between Neighbourhoods small in relation to the typical size of neigh- Based on this evidence, the hypothesis that bourhood effects, one explanation for dif- residents of neighbourhoods with poor ferences in employment rates, as revealed 776 REBECCA TUNSTALL ET AL. in empirical studies (for example, Lupton, employment rates, or the extent to which 2003a; Syrett, 2008; van Ham et al., 2012). direct experience is reported. In a study of young unemployed people Reviewing the Sources for the Ideas about in Newham in London, Roberts (1999) the ‘Stigma’ Neighbourhood Effects found that almost a third of young people Pathway from the most deprived parts of the bor- ough thought that employers were put off The idea that poor neighbourhood reputa- by the area in which they lived. The converse tion might provide a pathway through finding was that a majority of residents did which neighbourhood effects might operate not think that poor neighbourhood reputa- appears to have developed through qualita- tion might create neighbourhood effects on tive research which indicates variations in employment. The evaluation of Working reputation coinciding with variation in Neighbourhood Pilots found that just over employment, and the presence of beliefs one in ten residents thought that ‘employ- about neighbourhood effects amongst at ers don’t want to employ local people’ least some residents and LLM intermedi- (Dewson et al., 2007). The converse is that aries. One of the irreplaceable functions of almost nine in ten did not agree with this open-ended qualitative work is to explore idea. In both cases, the context of the new areas of social enquiry and to generate research might have encouraged respon- hypotheses. However, not all such hypoth- dents to point to barriers to employment eses are the same and the strength of evi- other than their own characteristics and dence behind them, and the indications of behaviour. Like Roberts (1999) and the prevalence and salience of the processes Dewson et al. (2007), we found that only a they identity may vary. minority of interviewees supported the Most existing qualitative studies which neighbourhood reputation effects hypoth- have produced evidence of potential or per- esis. Members of the public were most ceived neighbourhood reputation effects on likely to think that neighbourhood reputa- employment have not been focused on tion effects on employment might exist, neighbourhood reputation effects (Lawless but they were also the group least likely to and Smith, 1998; SEU, 1998; Taylor, 1998; have direct experience or evidence of Fieldhouse, 1999; Roberts, 1999; Dean and neighbourhood reputation effects. Hastings, 2000; Speak, 2000; Mellor, 2002; It is possible that neighbourhood effects Taylor, 2003; Aleksandraviciene et al., 2005; may be specific not only to national or Dewson et al., 2007; Sanderson, 2006; neighbourhood contexts, but also to partic- Green and White, 2007; Fletcher, 2007; ular time-periods. Over the past five years, Fletcher et al., 2008). Many have not asked employers even for low-skilled, low-paid explicitly or directly about neighbourhood and manual work, have switched from reputation effects, but have recorded refer- paper and mail to electronic applications. ences that emerged in open-ended enquiries This practice was reflected amongst employ- or as interesting by-products of research ers interviewed for this project, with the par- into other issues. Thus they have tended to record mentions of potential neighbour- tial exception of those employing kitchen hands. This has meant that circumventing hood reputation effects without taking into account their prevalence or salience as any poor reputation neighbourhood effect by using a false postal address is both virtu- potential explanations for individual employment status or neighbourhood ally costless and practically redundant. POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 777 Conclusions existing evidence on neighbourhood effects suggests that the size of the effects is rela- In summary, this paper has found evidence tively small compared with differences in to contradict the hypothesis employment rates between neighbour- hoods, the results of this work do not rule Residents of neighbourhoods with poor repu- out a possible contribution of area reputa- tations fare worse when applying for relatively tion to neighbourhood effects on employ- low-skilled jobs than residents of neighbour- ment rates. hoods in the same labour market with better reputations, all other things being equal. Acknowledgements The result suggests that relatively well- The authors would like to thank all those who qualified candidates from areas with poor gave time to be interviewed, and two anonymous referees of this paper. reputations should not fear postcode dis- crimination, at least up until the interview stage of the recruitment process. It remains possible that poor neighbourhood reputa- Funding tion might create neighbourhood effects on The authors would like to thank the Joseph employment in parts of these LLMs or Rowntree Foundation for funding the research parts of the recruitment and selection pro- on which this article is based. cess that were outside the scope of this experiment. However, centralised and elec- Notes tronic recruitment may be reducing the scope for these effects in large parts of the 1. We used official travel-to-work areas labour market. (TTWAs) (Coombes and Bond, 2008) as the spatial units from which to select LLMs for On the basis of this evidence, there is no the experiment. argument for policy interventions, includ- 2. The www.direct.gov.uk website advertises all ing policies to reduce sorting or to address vacancies notified to Jobcentre Plus (the this neighbourhood effect pathway more Public Employment Service in Great directly, on the grounds of neighbourhood Britain), estimated to be 40 per cent of total effects on employment. Nonetheless, there vacancies, with higher coverage of may be other arguments for these policies lower-skilled vacancies. It is the single largest on area reputations; for example, that resi- source of job vacancies in the UK. dents of areas with poor reputations suffer discrimination in the provision of services, References or face unequal treatment or recognition. The paper demonstrates the value of Aleksandraviciene, I., Carpenter, M., Cleaver, P., investigating the sources of ideas about Hussain, Z. et al. (2005) The impact of com- neighbourhood effects pathways, and the munity based initiatives on unemployed men and women in Coventry: final report of sequel factors necessary for pathways to operate, as project on gender and class. Centre for Lifelong well as for testing pathways themselves. It Learning and Department of Sociology, Uni- demonstrates the value of experimental versity of Warwick. methods in exploration of neighbourhood Amion Consulting (2010) Evaluation of the effects. It also demonstrates how mixed national strategy for neighbourhood renewal: methods approaches may add to the value final evaluation. Department for Commu- of individual elements of research. As nities and Local Government, London. 778 REBECCA TUNSTALL ET AL. Atkinson, R. and Jacobs, K. (2008) Public housing Fletcher, D. R., Gore, T., Reeve, K., Robinson, D. in Australia: stigma, home and opportunity. et al. (2008) Social housing and worklessness: Research Paper No. 1, Housing and Commu- qualitative research findings. Research Report nity Research Unit, University of Tasmania. No. 521, Department for Work and Pensions, Bates, P., Carta, E., Dewson, S., Francis, R. and Corporate Document Services, Leeds. Pillai, R. (2007) Working in 21st century rural Galster, G. (2007) Should policy makers strive England: a scoping study. Commission for for neighborhood social mix? An analysis of Rural England, Institute for Employment the Western European evidence base, Hous- Studies, Brighton. ing Studies, 22(4), pp. 523–545. Bovenkerk, F. (1992) Testing discrimination in Galster, G. (2012) The mechanism(s) of neigh- natural experiments: a manual for interna- bourhood effects: theory, evidence, and policy tional comparative research on discrimination implications, in: M. van Ham, D. Manley, N. on the grounds of ‘race’ and ethnic origin. Inter- Bailey, L. Simpson and D. Maclennan (Eds) national Labour Office, Geneva. Neighbourhood Effects Research: New Perspec- Burns, T., Catty, J., Becker, T., Drake, R. E. tives, pp. 23–56. Dordrecht: Springer. et al. (2007) The effectiveness of supported Green, A. E. and Adam, D. (2011) City strategy: employment for people with severe mental final evaluation. Research Report No. 783, illness: a randomised control trial, Lancet, Department of Work and Pensions, Corpo- 370, pp. 1146–1152. rate Document Services, Leeds. Christie, I. and Rolfe, H. (1992) The TECs and Green, A. E. and Owen, D. W. (2006) The Geo- inner city regeneration, Policy Studies, 13, pp. graphy of Poor Skills and Access to Work. York: 54–63. Joseph Rowntree Foundation. Coombes, M. G. and Bond, S. (2008) Travel-to- Green, A. E. and White, R. J. (2007) Attachment work areas: the 2007 review. Office for to place: social networks, mobility and prospects National Statistics, London. of young people. Joseph Rowntree Foundation, Dean, J. and Hastings, A. (2000) Challenging York. images: housing estates, stigma and area regen- Green, F., Maguire, M., Roberts, B. and Wray, K. eration. Joseph Rowntree Foundation, York. (1991) Explaining persistent inner city unem- Dewson, S., Casebourne, J., Darlow, A., ployment: a case study of Nottingham, Local Bickerstaffe, T. et al. (2007) Evaluation of the Economy, 6, pp. 240–249. working neighbourhoods pilot: final report. Ham, M. van, Manley, D., Bailey, N., Simpson, L. Research Report No. 411, Department for and Maclennan, D. (2012) Neighbourhood Work and Pensions, Coporate Document effects research: new perspectives, in: M. van Services, Leeds. Ham, D. Manley, N. Bailey, L. Simpson and DfT (Department for Transport) (2011) Personal D. Maclennan (Eds) Neighbourhood Effects travel factsheet: commuting and business travel. Research: New Perspectives, pp. 1–22. Dor- Department for Transport, London. drecht: Springer. Dovidio, J. F. and Gartner, S. L. (2000) Aversive Hastings, A. and Dean, J. (2003) Challenging racism and selective decisions: 1989 and 1999, images: tackling stigma through estate regen- Psychological Science, 11(4), pp. 315–319. eration, Policy and Politics, 31, pp. 171–184. Eardley, T. and Thompson, M. (1997) Does case- Hirsch, D. and Moore, R. (2011) The living management help unemployed job seekers? a wage in the United Kingdom: building on success. review of the international evidence. Social Citizens UK, The Centre for Research in Policy Research Centre, Sydney. Social Policy and the Living Wage Foundation Fieldhouse, E. A. (1999) Ethnic minority (www.citizensuk.org/wp-content/uploads/2011/ unemployment and spatial mismatch: the 06/The-Living-Wage-in-the-United-Kingdom- case of London, Urban Studies, 36, pp. 1569– May-2011.pdf). 1596. Holzer, H. J. (1991) The spatial mismatch Fletcher, D. R. (2007) A culture of worklessness? hypothesis: what has the evidence shown?, Historical insights from the Manor and Park Urban Studies, 28, pp. 105–122. area of Sheffield, Policy and Politics, 35, pp. Houston, D. S. (2005) Employability, skills mis- 65–85. match and spatial mismatch in metropolitan POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 779 labour markets, Urban Studies, 42, pp. National Equality Panel (2010) An anatomy of 221–243. economic inequality in the UK. Government Joseph, M. L., Chaskin, R. J. and Webber, H. S. Equalities Office, London. (2007) The theoretical basis for addressing Nunn, A., Bickerstaffe, T., Hogarth, T., poverty through mixed-income development, Bosworth, D. et al. (2010) Postcode selection? Urban Affairs Review, 42(3), pp. 369–409. Employers’ use of area- and address-based Jowell, R. and Prescott-Clarke, P. (1970) Racial information shortcuts in recruitment decisions. discrimination and white-collar workers in Research Report No. 664, Department for Britain, Race and Class, 11, pp. 397–417. Work and Pensions, Corporate Document Kain, J. (1968) Housing segregation, negro Services, Leeds. unemployment and metropolitan segrega- Pager, D. (2007) The use of field experiments tion, Quarterly Journal of Economics, 82, for studies of employment discrimination: pp. 175–197. contributions, critiques, and directions for Kleinhans, R. (2004) Social implications of the future, Annals of the American Academy housing diversification in urban renewal: a of Political and Social Science, 609, pp. 104– review of recent literature, Journal of Housing 133. and Built Environment, 19, pp. 367–390. Pager, D. and Quillian, L. (2005) Walking the Kornfeld, R., Wood, M., Orr, L. and Long, D. talk? What employers say versus what they (1999) Impacts of the project network demon- do, American Sociological Review, 70, pp. 355– stration. Abt Associates Inc., Bethesda, MD. 380. Lawless, P. (1995) Inner-city and suburban Purdon, S., Stratford, N., Taylor, R., labour markets in a major English conurba- Natarajan, L. et al. (2006) Impacts of the job tion: processes and policy implications, retention and rehabilitation pilot. Research Urban Studies, 32, pp. 1097–1125. Report No. 342, Department for Work and Lawless, P. and Smith, Y. (1998) Poverty, Pensions, Corporate Document Services, inequality and exclusion in the contempo- Leeds. rary city, in: P. Lawless, R. Martin and S. Rangarajan, A. and Novak, T. (1999) The strug- Hardy (Eds) Unemployment and Social gle to sustain employment: the effectiveness of Exclusion: Landscapes of Labour Inequality, the post employment services demonstration. pp. 201–216. London: Jessica Kingsley. Mathematica Policy Research Inc., Princeton, Lupton, R. (2003a) Neighbourhood effects: can we NJ. measure them and does it matter? CASE Paper Recchia, F. (2008) Immigration, politics and vio- No. 73, London School of Economics. lence in urban France: between fiction and Lupton, R. (2003b) Poverty Street: The Dynamics facts, Information, Society and Justice, 2(1), of Neighbourhood Decline and Renewal. Bris- pp. 47–61. tol: Policy Press. Riach, P. A. and Rich, J. (2004) Deceptive field Marsden, P. V. (1994) The hiring process: experiments of discrimination: are they ethi- recruitment methods, American Behavioral cal?, Kyklos, 57(3), pp. 457–470. Scientist, 37(7), pp. 979–991. Roberts, F. (1999) Young people out of work in Mellor, R. (2002) Hypocritical city: cycles of south west Newham. Commentary Series No. urban exclusion, in: J. Peck and K. Ward 83, Centre for Institutional Studies, Univer- (Eds) Revolutionary City: Restructuring Man- sity of East London. chester, pp. 214–234. Manchester: Manche- Sanderson, I. (2006) Worklessness in deprived ster University Press. neighbourhoods: a review of the evidence. McGinnity, F., Nelson, J., Lunn, P. and Quinn, E. Department for Communities and Local (2009) Discrimination in recruitment: evidence Government, London. from a field experiment. The Equality Author- SEU (Social Exclusion Unit) (1998) Bringing ity and the Economic and Social Research Britain Together: A National Strategy for Institute, Dublin. Neighbourhood Renewal. Norwich: The Sta- McGregor, A. (1977) Intra-urban variations in tionery Office. unemployment: a case study, Urban Studies, Souza Briggs, X. de, Popkin, S. J. and Goering, J. 14, pp. 303–313. (2010) Moving to Opportunity: The Story of an 780 REBECCA TUNSTALL ET AL. American Experiment to Fight Ghetto Poverty. employers perceive neighbourhoods in four Oxford: Oxford University Press. urban labour markets, in: A. O’Connor, C. Speak, S. (2000) Barriers to lone parents’ employ- Tilly and L. T. Bobo (Eds) Urban inequality, ment: looking beyond the obvious, Local ch 6. New York: Russell Sage Foundation. Economy, 15, pp. 32–44. Walker, R. (2000) Learning if policy will work: Stafford, B., Greenberg, D. and Davis, A. (2001) the case of the new deal for disabled people, A review of the use of social experiments. Policy Studies, 21(4), pp. 313–332. Research report, Department of Social Secu- Waquant, L. J. D. (1993) Urban outcasts: stigma rity, London. and division in the black American ghetto Stratford, N., Farrell, C., Natarajan, L. and and the French urban periphery, International Lewis, J. (2005) Taking part in a randomised Journal of Urban and Regional Research, 17, control trial: a participants’ eye view of the job pp. 366–383. retention and rehabilitation pilot. Corporate Watt, P. (2003) Urban marginality and labour Document Services, Leeds. market restructuring: local authority tenants Syrett, S. (2008) Area effects: a review of the evi- and employment in an inner London Bor- dence prepared for CLG. Department of Com- ough, Urban Studies, 40, pp. 1769–1789. munities and Local Government, London. Wilson, W. J. (1987) The Truly Disadvantaged. Taylor, M. (1998) Combatting the social exclu- Chicago, IL: University of Chicago Press. sion of housing estates, Housing Studies, 13, Wood, M., Hales, J., Purdon, S., Sejersen, T. and pp. 819–832. Hayllar, O. (2009) A test for racial discrimina- Taylor, Y. (2003) Negotiation and navigation: an tion in recruitment practice in British cities. exploration of the spaces/places of working- Research Report No. 607, Department for class lesbians, Sociological Research Online, Work and Pensions, Corporate Document 9(1). Services, Leeds. Tilly, C., Moss, P., Kirschenman, J. and Zenou, Y. (2002) How do firms redline workers?, Kennelly, I. (2001) Space as a signal: how Journal of Urban Economics, 52, pp. 391–408. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Urban Studies: An International Journal of Research in Urban Studies SAGE

Does Poor Neighbourhood Reputation Create a Neighbourhood Effect on Employment? The Results of a Field Experiment in the UK:

Loading next page...
 
/lp/sage/does-poor-neighbourhood-reputation-create-a-neighbourhood-effect-on-va7YjZwOhf

References (84)

Publisher
SAGE
Copyright
Copyright © 2022 by Urban Studies Journal Limited
ISSN
0042-0980
eISSN
1360-063X
DOI
10.1177/0042098013492230
Publisher site
See Article on Publisher Site

Abstract

There are substantial variations in labour market outcomes between neighbour- hoods. One potential partial explanation is that residents of some neighbourhoods face discrimination from employers. Although studies of deprived areas have recorded resident perceptions of discrimination by employers and negative employer perceptions of certain areas, until now there has been no direct evidence on whether employers treat job applicants differently by area of residence. This paper reports a unique experiment to test for a neighbourhood reputation effect involving 2001 applications to 667 real jobs by fictional candidates nominally resident in neighbour- hoods with poor and bland reputations. The experiment found no statistically signif- icant difference in employer treatment of applicants from these areas, indicating that people living in neighbourhoods with poor reputations did not face ‘postcode dis- crimination’ in the labour market, at the initial selection stage. Introduction This article explores neighbourhood effects labour markets. It focuses on exploring one as an explanation for variations in employ- pathway as an explanation: ‘stigma’, or poor ment rates between neighbourhoods within neighbourhood reputation. It exploits a Rebecca Tunstall is in the Centre for Housing Policy, University of York, Heslington, York, YO10 6DT, UK. Email: becky.tunstall@york.ac.uk. Anne Green is in the Institute for Employment Research, University of Warwick, Social Studies Building, Coventry, CV4 7AL, UK. Email: Anne.Green@warwick.ac.uk. Ruth Lupton and Katie Bates are in the Centre for the Analysis for Social Exclusion, London School of Economics, London, UK. Email: r.lupton@lse.ac.uk and katie.bates@lse.ac.uk. Simon Watmough is in the European Institute, Fiesole, Firenze, Italy. Email: simon.watmough@ eui.eu. 0042-0980 Print/1360-063X Online 2013 Urban Studies Journal Limited DOI: 10.1177/0042098013492230 764 REBECCA TUNSTALL ET AL. field experiment methodology, supported 2006). This paper focuses on variations by contextual qualitative and quantitative between neighbourhoods within LLMs. In evidence, through case studies of three local the UK, the National Equality Panel (2010) labour markets (LLMs) and nine neigh- highlighted that in the most deprived tenth bourhoods in England and Wales. It focuses of neighbourhoods only 55 per cent of on one outcome: employment. The hypoth- adults were employed compared with more esis it aims to test is that than 80 per cent in the least deprived half of neighbourhoods nation-wide. Residents of neighbourhoods with poor repu- Concerns about uneven employment tations fare worse when applying for relatively rates between neighbourhoods and the (rel- low-skilled jobs than residents of neighbour- atively) low absolute employment rates in hoods in the same labour market with better some areas have been part of the motivation reputations, all other things being equal. for successive generations of urban policy in the UK and elsewhere. Closing gaps in If verification is found for this hypothesis, neighbourhood worklessness was a key it is evidence of a neighbourhood effect objective of the National Strategy for operating through the ‘stigma’ pathway Neighbourhood Renewal (in England) from (although it is possible that neighbourhood 1998 to 2008 (Amion Consulting, 2010), effects might also be operating through the Working Neighbourhoods Fund (in other pathways). It would suggest that England) (Dewson et al., 2007) and the City neighbourhood effects contribute to some Strategy Pathfinders (in Great Britain) extent to variation in employment rates (Green and Adam, 2011). Areas with high between neighbourhoods. unemployment and low employment tend The article begins by examining varia- to overlap with areas with poor reputations. tions in employment rates at neighbour- hood level and discussing explanations for such variations. The methodology used for Explaining Variations in the particular field experiment which is the Employment Rates between focus of the article is then discussed. Next Neighbourhoods within Labour the results of the experiment are presented. Markets The article ends with a discussion of what There are three broad categories of expla- the results mean for the role of the ‘stigma’ nation for variation within labour markets: neighbourhood effects pathway. first, skills mismatch, largely as a result of residential sorting; secondly, spatial mis- match; and thirdly, neighbourhood effects. Variations in Employment Rates These explanations are not exclusive and between Neighbourhoods within could operate together for additive or inter- Labour Markets acting effects. There is longstanding academic and policy First, residential sorting and changes in interest in the existence and persistence of skills requirements can result in spatial geographical variations in employment, concentrations of individuals who are at economic inactivity and unemployment greater risk of non-employment due to rates. These concerns encompass a range of poor skills, lack of qualifications, ill health geographies—from regional, to LLM and and lack of recent work experience neighbourhood scales (Green and Owen, (Houston, 2005; Green and Owen, 2006). POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 765 Hence, spatial variations in employment institutional, with 14 sub-categories. Under and non-employment rates are a function ‘institutional’ mechanisms, he lists local of spatial differences in individual charac- institutional resources, the behaviour of teristics and a mismatch between the skills local market actors and private services, and or attributes required or preferred by stigmatisation. Thus, the behaviour of employers and those held by people in cer- employers and neighbourhood stigma are tain neighbourhoods. Skills mismatches just two of numerous potential mechanisms. could be reduced by improving individuals’ relative skills (and other characteristics). Secondly, the spatial mismatch hypoth- Explaining Variations in esis explains variations in non-employment Employment Rates between through geographical disparities between Neighbourhoods within LLMs the location of jobs and potential employ- through Neighbourhood Effects ees (see Kain, 1968; Holzer, 1991; Houston, Much research into neighbourhood effects 2005). Spatial mismatch can develop has used multivariate analysis of quantitative through residential sorting, the impact of data on the characteristics of, and outcomes housing policy and uneven geographical for, residents of different neighbourhoods impacts of economic restructuring. For (Kleinhans, 2004; Joseph et al., 2007; Galster, instance in a case study of an inner London 2007). Qualitative research (for example, borough, Watt (2003) described how many with employers and intermediaries), local authority tenants were in effect mar- although sometimes overlooked, is particu- ooned over a 30-year period as the immedi- larly important in generating hypotheses and ate area lost almost all its manual in developing and testing ideas about path- employment. Over time, those in employ- ways (Lupton, 2003a). Another approach ment travelled longer distances to work. with potential is the use of randomised con- Thirdly, when applied in relation to trol trials or experimental methods. employment outcomes, the neighbourhood Each of these methods is independently effects hypothesis suggests that spatial varia- valid. However, combinations of methods tions in non-employment are not reducible offer the potential to confirm and accumu- to compositional effects (i.e. characteristics late evidence. While econometric analysis is of residents) but rather that there is an addi- broadly seen as the ‘gold standard’ within tional area related effect on employment (or neighbourhood effects research, in the field other outcomes) that results from spatial of employer preferences, this kind of data concentrations of individuals who are at analysis is seen as secondary to experimental greater risk of non-employment or who studies (Riach and Rich, 2004). Research have other disadvantages (Syrett, 2008), or into neighbourhood effects so far has tended from other characteristics of the area (for to explore neighbourhood effects without example, Lupton, 2003a; van Ham et al., focusing on specific mechanisms (Galster, 2012). Potential pathways for neighbour- hood effects may be either endogenous (i.e. 2012). This paper explores neighbourhood internal to the neighbourhood) or due to the relationship between the neighbourhood effects as an explanation for variations in employment rates between neighbourhoods and outside people or institutions. In a com- prehensive typology of mechanisms, Galster within labour markets. It focuses on explor- (2012) prefers the categories of social inter- ing one pathway as an explanation: ‘stigma’, active, environmental, geographical and or poor neighbourhood reputation. It 766 REBECCA TUNSTALL ET AL. exploits a field experiment methodology, residents. These reputations may affect the supported by contextual qualitative and behaviour of all these groups towards the quantitative evidence, through case studies neighbourhoods—for example in decisions to provide services or to employ residents. of three LLMs and nine neighbourhoods in England and Wales. It focuses on one out- Existing residents are likely to be aware of these poor external reputations and this in come: employment. itself may affect their attitudes and beha- The hypothesis it aims to test is that viours, including, for example, decisions to apply for jobs and their behaviour in the Residents of neighbourhoods with poor repu- job application process. Poor reputation tations fare worse when applying for relatively may also affect the provision of private and low-skilled jobs than residents of neighbour- hoods in the same labour market with better public services (for example, Christie and Rolfe, 1992). reputations, all other things being equal. Over 30 years ago, McGregor (1977) used multivariate analysis, now established If verification is found for this hypothesis, it is evidence of a neighbourhood effect as the core approach to investigating neigh- bourhood effects, to explore the role of operating through the ‘stigma’ pathway stigma in explaining variations in employ- (although it is possible that neighbourhood ment rates between neighbourhoods within effects might also be operating through labour markets through neighbourhood other pathways). It would suggest that effects. He examined unemployment rates neighbourhood effects contribute to some and unemployment duration for men in extent to variation in employment rates Scotland, including Ferguslie Park, a neigh- between neighbourhoods. bourhood with a poor reputation, control- ling for age, skill, industry, marital status, length of time in previous job, preference Evidence for the Role of for light, medium or heavy work, prefer- Neighbourhood Reputation in ences for local or more distant work, and Explaining Variations in preferences for day or any hours, and con- Employment Rates through cluded that Neighbourhood Effects Evidence that some neighbourhoods have although the Ferguslie Park sample would worse reputations than others is ubiquitous experience relatively high unemployment in urban studies (for example, Wilson, duration no matter where they lived, the fact 1987; Tilly et al., 2001; Waquant, 1993; of their residence in Ferguslie Park signifi- Hastings and Dean, 2003; de Souza Briggs cantly adds to the disadvantages associated et al., 2010). Neighbourhoods with poor with their individual characteristics reputations often are the same neighbour- (McGregor, 1977, p. 311). hoods that have low rates of employment and high rates of non-employment and To date, no other studies appear to have unemployment. This brief review focuses used such techniques to explore the stigma on evidence from the UK. pathway for neighbourhood effects on Relatively poor reputations may be held employment. by outsiders including the general public, Studies using other methods have added key public- and private-sector decision- important information. Although employer makers and service providers, or potential statements may not fully reflect practice and POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 767 do not preclude unconscious preferences, Nottingham, Green et al. (1991, p. 273) several studies provide insights into poten- found evidence of a preference amongst tial neighbourhood preferences from an some employers for local workers, regard- employer perspective (Pager and Quillian, less of neighbourhood reputation: ‘‘[they] 2005). Writing about four US metropolitan preferred their employees to be locally areas, Tilly et al. noted that based, since this was thought to make them more reliable’’. Zenou (2002) and Lupton each manager forms his or her own mental (2003b) uncovered similar preferences. This map of . [their] environment. Employer’s suggests a potential ‘spatial match’ effect or maps, in turn, have important effects on the positive neighbourhood effect for people in labor market (Tilly et al., 2001, p. 304). neighbourhoods close to employment. Qualitative research can contribute to A majority of employers in this study understanding of how pathways for neigh- thought there were systematic differences bourhood effects may work and the circum- between urban and suburban workers stances in which they may not operate. (across neighbourhoods) within a labour Nunn et al. (2010) suggested that neigh- market, although in the US context it bourhood effects via area reputation should be noted that spatial differences are required ‘local knowledge’ amongst employ- also heavily racialised. Negative attitudes ers, discretion and ability to deviate outside about particular groups of workers could automated or rigid processes amongst those feed into spatial mismatch, if they affect involved in recruitment. Like the exercise of employers’ location decisions and the pat- other forms of employer preference or dis- tern of employment opportunities. crimination, this may have been influenced In the UK, some studies have found at by the state of the labour market, with eco- least some self-reported preference for indi- nomic circumstances offering more choice viduals not from areas with poor reputa- for employers resulting in more screening. tions (for example, Hastings and Dean, Another source of information is inter- 2003; Aleksandraviciene et al., 2005). views with job applicants and those working Interviews with employers, with recent with them.Thisisaweaker source,asthese experience of recruitment in less skilled informants have only indirect insight into occupations in various cities in Great employer preferences and no oversight of Britain, about their selection and recruit- aggregate patterns. However, numerous stud- ment procedures suggested that employers ies have recorded the belief that ‘postcode dis- preferred workers from areas without poor crimination’ is taking place in the UK (for reputations, but the effect was only at the example, Lawless and Smith, 1998; SEU, margin and in specific conditions (Nunn 1998; Taylor, 1998; Fieldhouse, 1999; Roberts, et al., 2010). Employers generally talked of 1999; Dean and Hastings, 2000; Speak, 2000; individuals’ personal characteristics as the Mellor, 2002; Hastings and Dean, 2003; main factor in recruitment decisions. Taylor, 2003; Aleksandraviciene et al., 2005; Address could be an issue, but this mainly Sanderson, 2006; Green and White, 2007; related to travel-to-work considerations, Fletcher, 2007; Dewson et al., 2007; Bates particularly where jobs involved anti-social et al., 2007; Fletcher et al., 2008) and beyond hours. A few indications of a lack of will- (for example, in France (Waquant, 1993; ingness to employ applicants from certain Recchia, 2008), in Australia (Atkinson and areas, or an intention to subject them to Jacobs, 2008) and in the US (Tilly et al., special attention, emerged. In their study of 2001)). More specifically, some social housing 768 REBECCA TUNSTALL ET AL. tenants interviewed in four different areas Houston summed up the situation for the about employment in 2008 thought that there UK as follows: ‘‘direct evidence of ‘post- was postcode discrimination against them code discrimination’ is difficult to find’’ (Fletcher et al., 2008). In a study of working- (2005, p. 229). The pattern appears similar class women in Scotland and northern at least for France, the US and Australia. England, Taylor (2003, par. 7.4) reported: ‘‘Several . explained their long-term unem- ployment by having specific, devalued post- Exploring the Role of codes’’. Bates et al. (2007) found at least one Neighbourhood Reputation in such allegation in a study of employment in Variations in Employment Rates rural areas. A study of high unemployment through Neighbourhood Effects areas in Coventry reported Using an Experimental Method Given the difficulties establishing direct evi- Although they could not easily always pro- dence of neighbourhood effects linked to a vide evidence for it, many project workers particular pathway through multivariate anal- felt that ‘postcode discrimination’ against ysis, or through interviews alone, the study people from areas of high unemployment who are stereotyped as unreliable workers used an experimental method, supported by contextual quantitative and qualitative undoubtedly operated (Aleksandraviciene et al., 2005, pp. 64–65). research. This permits testing the hypothesis set out earlier, through establishing In the evaluation of the 12 Working (1) variations in employment rates between Neighbourhood Pilots covering areas of neighbourhoods; high unemployment and inactivity, inter- (2) variations in reputation between viewees said neighbourhoods; (3) employers’ awareness of neighbour- Employers tar everyone with the same brush. hood reputations, attitudes to different It’s just not fair. The area definitely goes neighbourhoods, assessment of neigh- against you [and] if you put [this area] as bourhood reputation as a factor in deci- your address on your application that puts sion-making; and employers off (Dewson et al., 2007, p. 32). (4) variations in success in applying for employment amongst residents of neigh- In a rare example of direct evidence, albeit bourhoods with different reputations, all second-hand and for a single case, Speak other factors being equal (same jobs; (2000) reported that a resident from Benwell equivalent applicant characteristics). in Newcastle upon Tyne had been inter- viewed for a job but was not given it and was There is very limited experimental evidence told: ‘‘It’s not you, we think you’d do the job on neighbourhood effects. There are ethical, fine. but if you live in [Benwell] you either know a villain or you are a villain’’. legal, political and cost problems in creating In summary, existing evidence from the experimental policy design in urban and UK on the role of stigma in explaining var- social policies (Stafford et al., 2001). In a iations in employment rates between neigh- few cases policy design has enabled rando- bourhoods within labour markets through mised control trial methods in housing and neighbourhood effects was described urban policy, most notably the Moving to recently as ‘‘thin’’ (Dewson et al., 2007). Opportunity (MTO) programme for the POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 769 Table 1. Case study LLMs Case study labour Approximate population JSA claimant rate, Rank by unemployment market name of working age, 2010 2010 (percentage) rate out of UK TTWAs, 2010 (weighted by population of working age) Weak 250,000 5.7 Highest 5 per cent Medium 500,000 4.1 Not highest 5 per cent but highest 20 per cent Strong 250,000 3.2 Lowest 40 per cent Source: NOMIS (www.nomisweb.co.uk; accessed March 2011); based on 2007 mid-year population estimates. relocation of public housing residents in the 1994). It focuses on employer behaviour. US (de Souza et al., 2010). This is amongst Applicant behaviour—for example in response the most quoted studies in reviews of evi- to belief about employer prejudice—may also dence on neighbourhood effects. There has play a role in explaining employment varia- been some use of randomised assignment to tions between areas. different treatment groups in welfare-to- work interventions (Eardly and Thompson, Labour Market and Neighbourhood Case 1997; Kornfeld et al., 1999; Rangarajan and Studies Novak, 1999; Walker, 2000; Stratford et al., 2005; Purdon et al., 2006; Burns et al., 2007). Three LLMs across the UK were selected (see Table 1). The criteria for selecting them were that The Field Experiment Method — each should have at least some neigh- The experimental method we used was the bourhoods with markedly poor local ‘correspondence test’, involving sending out reputations; multiple applications to real jobs, differing as — each should be large enough to generate faraspossibleonlyin terms of thevariable sufficient vacancies but small enough being tested; here, the reputation of the appli- that local neighbourhoods within them cant’s neighbourhood. This has been used might be known to employers; before in studies of employer preferences or — they should include ‘weak’, ‘medium’ and discrimination in employment by ethnicity, ‘strong’ LLMs in terms of unemployment age, gender and disability status (Jowell and rates, in case competition for vacancies Prescott-Clarke, 1970; Riach and Rich, 2004; affected the neighbourhood effects; and Pager, 2007; McGinnity et al., 2009; Wood — each should have low minority ethnic et al., 2009). It has some inevitable limitations: populations, to rule out any effects it cannotbeusedtoresearchthatpartofthe from employer preference for different labour market where jobs are not formally ethnic groups (Fieldhouse, 1999; Tilly advertised, or the parts of the recruitment pro- et al., 2001). cess that involve face-to-face interactions. There is evidence that recruitment through Three neighbourhoods were selected in informal networks tends to mean recruitment each LLM using the following criteria, based of people similar to existing staff (Marsden, on desk research on local identity and 770 REBECCA TUNSTALL ET AL. reputations; analyses of secondary sources public identified the worst reputation two on deprivation levels, ethnic mix and acces- or three neighbourhoods with remarkable sibility indicators; and field visits— consistency. This was used to confirm the including street interviews (discussed in selection of the ‘poor reputation’ and ‘bland more detail later) which were used to con- reputation’ neighbourhoods from desk firm the poor reputation and bland reputa- research. In each of the three LLM areas, tion of shortlisted areas just over half of those interviewed (53 per cent) thought that the ‘poor reputation’ — two particularly deprived neighbour- case study neighbourhoods were undesir- hoods with well-established ‘poor’ able places to live. The same proportion reputations (five out of the six neigh- thought it was ‘very’ or ‘fairly likely’ that bourhoods were in the top 5 per cent of local employers would look less favourably Indices of Multiple Deprivation [IMD] on people from these or similar areas. Only scores andone wasin the top15per a minority (23 per cent) thought that it was cent) and one not particularly deprived ‘very’ or ‘fairly unlikely’ that employers neighbourhood (all three neighbour- would look unfavourably on applicants hoods were in the middle 40–60 per from these neighbourhoods. cent of neighbourhoods on IMD scores) with a ‘bland’ reputation, to Interviews with employers serve as a comparator; To provide insights into employers’ aware- — each had relatively strong local identities, ness of neighbourhood reputations, atti- linked to identifiers including neigh- tudes to different neighbourhoods, bourhood name, main street name or assessment of neighbourhood reputation as postcode; a factor in decision-making, and to confirm — each had relatively small minority ethnic our understanding of employers’ recruit- populations (the White British popula- tion share was around 97 per cent in the ment methods and the realism of the appli- cations used in the experimental method neighbourhoods in the ‘weak’ LLM, 93 per cent in the ‘medium’ and 90 per cent (discussed later), we interviewed 14 employers and 11 labour market intermedi- in the ‘strong’ LLM); — all were at a similar distance and public aries (i.e. employment agencies and job see- kers’ advisors). Employers interviewed were transport travel time from the city centre, to limit (although not exclude) based in the three case study LLMs and had recently recruited people to one or more of any effect of employer preference for more accessible employees (Green the job types examined in the experiment (see later). They encompassed different et al., 1991; Zenou, 2002; Lupton, 2003b; Nunn et al., 2010). establishment sizes and industries, predo- minantly from the private sector but also from the public and voluntary sectors. Field Visits and Street Interviews All employers and labour market inter- The field visits included interviews with 81 mediaries interviewed were aware of certain members of the public, providing what neighbourhoods with poor reputations. appears to be the first, although small scale, Again, there was consistency between inter- evidence of the perceptions of the general viewees and their views overlapped closely public on potential neighbourhood reputa- with the assessment of the general public tion effects. In each LLM, members of the and desk research. Three of the 11 labour POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 771 market intermediaries thought that neigh- One of the employers interviewed in the bourhood reputation effects on employ- medium LLM admitted to ‘thinking twice’ ment might exist. One intermediary in the about applicants from some neighbourhoods medium LLM reported some cases of appli- cant perceptions of neighbourhood reputa- There are areas where you sort of think, tion effects hmm, you know, not too sure about that . where you have a large number of unem- There have been instances where people have ployed people, where you have council said: ‘Because of where I live, I won’t get a job. accommodation. They won’t trust me’. Some of the employers stated preferences Another, from the weak LLM, said that he on other grounds such as distance to work, believed in neighbourhood reputation clothing and accent. effects strong enough to promote action The Experiment We used to take address off applications . we remove address to remove a stigma. The jobs search and application strategy In the strong LLM, another said of fictional job applicants. In the experi- mental phase of the study, we searched for Employers would say ‘no’ to the idea that and applied to real jobs in the three LLMs they discriminate on area—but privately the that fitted the following criteria answer is ‘yes’. — jobs advertised on www.direct.gov.uk, gumtree.com and aggregator sites; This respondent subsequently significantly amended the point, ‘‘it’s not necessarily — advertised August 2010-June 2011; — the job location appeared to be within about prejudice, but the majority of people have barriers to entry’’. While this evidence the TTWA boundary; is strong enough to support the hypothesis — selected job types (office admin, cleaner, that neighbourhood reputation effects security guard, sales assistant, accounts might form a partial and subsidiary expla- clerk, kitchen hand and chef jobs). nation for variations in employment rates, These job types did not require degrees, and encourages further exploration, it is not vocational qualifications or substantial in itself strong enough to support a hypoth- experience, but were prevalent enough esis that they are a major explanation. to give a good overview of the low- and medium-skilled labour market and Almost all employers stated unambiguously that they looked equally on applications from included jobs principally held by men, by women and either gender. all areas. However, one private-sector employer from the strong LLM explicitly showed aware- ness of the potential for discrimination, but From this subset, we selected jobs that met claimed that he would not apply it in practice, the requirements of the experimental at least up to interview stage method I am not going to be that prejudiced on areas — for which main recruitment and selec- . So you get ‘em in, and see what they’re like tion decision-maker appeared to be . (emphasis added). based in the LLM, and was likely to be 772 REBECCA TUNSTALL ET AL. aware of local neighbourhood However, the personas were intended to reputations; represent people who would be relatively — could be applied to without face-to- attractive candidates for jobs that required face or phone contact (i.e. via email, limited education and skills. The aim was upload to website or post); and to limit the number of cases in which none — for which the closing date was at least of the three applicants was successful, 2–3 days hence, to allow time to make which would hinder investigation of dis- three applications that did not arrive crimination, while ensuring applications simultaneously. were realistic. To obviate any employer preference for other characteristics, all three applicants for any one job had the The jobs applied to. Between August same gender. All had names chosen to 2010 and June 2011, 2001 applications avoid signalling minority ethnicity (see were made to 667 jobs (i.e. three applica- Wood et al., 2009, for further discussion). tions in each case). There were 197 applica- All were in their early 20s. tions to office admin jobs, 139 to retail One of the three applicants for each job jobs, 97 to chef jobs, 75 to cleaner jobs, 74 appeared to be living in each of the three to kitchen hand jobs, 73 to accounts clerk local neighbourhoods selected. The appli- jobs and 12 to security guard jobs, reflect- cant address, including a fictional numbered ing the incidence of vacancies and varia- home in a real major street in the area likely tions in the extent to which jobs fitted to be well known, the area name and the experimental constraints. Of the jobs postcode were prominently stated at the top applied for, 246 were in the strong LLM, of each CV. Applicants differed in their 261 in the medium LLM and 159 in the exact qualifications and work experience, weak LLM; 76 per cent of jobs applied for and CV typeface and layout used, in order to did not offer a traditional full-time ‘9am– maintain as much similarity between candi- 5pm’ work schedule. This largely reflects dates as possible without raising the suspi- the type of jobs available via www.direct. cions of employers. Addresses in each of the gov.uk in the three LLMs. Fifty-four per three neighbourhoods were allocated ran- cent of the jobs for which wage data were domly to the three prepared CVs and cover- available paid at the minimum wage level. ing letters as the final stage of the process in application to every job. CVs’ contents were The characteristics of the fictional job placed on one of three CV templates with applicants. ‘Personas’ were created for different fonts and layouts, which were each of the job types applied to, with fic- made similarly attractive. Age, qualifications tional names, addresses, dates of birth, edu- and work experience were rotated between cational and work histories and real phone CV templates at intervals throughout the numbers and email addresses, which were experiment. This random allocation ensured used as the basis of CVs and covering let- that any differences in employers’ responses ters. These were then slightly tailored to to candidates living in different neighbour- individual jobs. It is possible that employ- hoods could not be attributed to their age, ers might be more likely to express neigh- the style of their CV, exact qualifications or bourhood reputation discrimination minor differences in work experience. against less attractive candidates. This has Most other applicant characteristics were been found in the case of ethnic discrimi- largely determined by job type and were nation (Dovidio and Gartner, 2000). intended to create candidates for the various POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 773 jobs who were promising but not exception- advertisements were first posted on the ally overqualified. In 286 (43 per cent) of Internet. More than half were sent within the jobs, all three applicants were male; in three days of the advert first appearing. the remainder, all were female. Jobs were Recording the results for the fictional allocated by gender in accordance with the candidates. After sending off three gender profile of employment revealed by matched applications to any one job, we secondary data sources. The slight predomi- monitored specially established email and nance of female applicants in the experi- voicemail accounts to receive employer ment overall reflects the availability of job responses. We were not able to monitor types advertised and the availability of jobs any employer responses that might have which we could apply to using our methods. been made by post to applicants’ home Applicants were allocated academic and addresses, but employer and intermediary vocational qualifications of the types com- interviews suggested that communication monly held by young people in the LLMs by post was exceptional. As the addresses and relevant to the post they were applying used were false ones, any employer writing for. All of the applicants had continual work to one of our candidates would have had records since they had completed education mail returned and could then have tried and some were given Saturday job experi- another means to get in touch, but we did ence before this. not learn of any such experiences. Wherever a job required a car and a clean Some 620 applications (31 per cent) driving licence, they were given to all three resulted in responses from potential employ- applicants, and likewise to candidates for ers. When we received a positive response, any jobs with unsocial hours in order to dis- we responded as fast as possible via email to tinguish between any discrimination by withdraw the applicant from consideration, employers based on ‘travelability’ rather stating that the candidates had already than discrimination against people from accepted another offer or that their circum- areas with poor reputations. Hence the can- stances had changed. Where more than one didates had a higher rate of private transport application for the same job received a posi- mobility than might be typical for appli- tive response from employers, the style and cants for these types of jobs. content of our responses to employers were varied. One month after the application, we stopped monitoring for responses. Making applications. The majority of The experiment was concerned with pos- applications took the form of a CV or a CV itive responses at the first selection stage. and covering letter. Covering letters were Thus the experimental results cannot be used in many cases, even if they were not applied directly to subsequent stages, such explicitly demanded by employers, in order as interviews, or to the selection process as to maximise positive response. Over 85 per a whole. cent of applications were made via the Internet. We used local assistants to put the remainder of applications in the post, to Results of the Field Experiment achieve local postmarks. In most of these cases, we avoided sending all the applica- ‘First-stage’ Positive Response Rates tions on the same day in order to reduce employer suspicion. In most cases, applica- A total of 17 per cent of the 2001 applica- tions were made very rapidly after tions received one of a range of first-stage 774 REBECCA TUNSTALL ET AL. positive responses. This share of positive to all candidates should not be treated as responses is higher than that achieved in positive evidence of non-discrimination by another recent experimental study (Wood employers. In the second case, all met some et al., 2009). This may reflect differences in minimum standard and, while all appear to the mix of jobs applied for, the LLMs studied have been preferred over any other real or the quality of experimental applications. candidates applying, there is no sign of pre- Three applicants were offered a post right ference between them. away. Most first-stage positive responses did In 475 (71 per cent) of the 667 jobs applied not lead directly or with any certainty to job for, none of the three candidates received a offers. Thirteen per cent of applicants were first-stage positive response. In the remaining invited to meet employers and 2 per cent were 192 (29 per cent) of jobs, employers gave a asked for further information—for example, positive response to one or more of our can- what days and hours they might be able to didates for the same post. These are the cases work. Thus fewer than one in five of the used to explore employer preferences and dis- experimental candidates got through the first crimination, following Bovenkerk (1992) and round of selection for jobs that required rela- Wood et al. (2009). tively limited skills and experience, and which generally paid close to the minimum wage. ‘First-stage’ Positive Response Rates Where Young people with fewer labour market Employer Showed A Preference for One or advantages would be likely to experience More Candidates lower first-stage positive response rates when In accordance with the experimental metho- applying for the same sorts of jobs. Thirteen dology deployed, in each of the 192 sets of per cent of applications received one of a applications with at least one first-stage pos- range of responses we classified as negative, itive response, one of the applications was including an acknowledgement of application for a candidate from a bland reputation but no further correspondence, or notice they neighbourhood. Of these, 140 (62.5 per had been unsuccessful; 69 per cent of applica- cent) applications received a first-stage posi- tions received no response of any kind. Previous studies using an experimental tive response (Table 2). The other two appli- cations in each of the 192 sets were from method to test for discrimination in employment have discussed several possible candidates from poor reputation neighbour- hoods, totalling 384 applications. Of these, methods of conceptualising discrimination and analysing results (for example, Riach 230 (59.9 per cent) applications received a positive response. The 2.6 percentage point and Rich, 2004). The key issue is how to deal with jobs in which none of the candi- difference between the success percentages for the two neighbourhood types provides a dates received a positive response, and those in which all of the candidates did. measure of aggregate net ‘postcode discrimi- nation’. However, the net preference was The first case suggests none met some min- imum standard, or that no appointment small and it was not statistically significant at was actually made, and offers no sign of the 1 per cent, 5 per cent or 10 per cent level. preference between the experimental candi- In each of the three LLMs, there was a dates. Wood et al. (2009) who conducted difference between the positive response the most recent experiment of this type in rates for the neighbourhood types, with the UK, and Bovenkerk (1992), who has applications from the bland reputation prepared a manual for the conduct of these neighbourhood having a slightly higher tests and analyses, argue that non-response positive response rate. The level of net POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 775 Table 2. Employers’ preference for applicants from different neighbourhood types Total sets of applications (A) Positive response: (B) Positive response: Net preference P-value with one or more bland reputation poor reputation percentage point positive responses neighbourhood neighbourhood difference (A-B) Number percent Number percent 192 120 62.5 230 59.9 2.6 0.1170 As there were twice as many poor reputation as bland reputation neighbourhood applications (384 compared with 192), the denominator for column B is the number of sets of applications with one or more positive responses multiplied by two (384). We used the two-sample z-test for proportions. Source: Experiment. preference found was highest in the reputations fare worse when applying for medium LLM, at 4.5 percentage points, but relatively low-skilled jobs than residents of in no area was it statistically significant. neighbourhoods in the same labour market We examined subsets of the results for with better reputations was not proven. evidence of ‘postcode discrimination’ in any It remains possible that, in a larger corre- particular job or employer type. In some spondence test study, the net preferences found would achieve statistical significance. cases, there was evidence of a small amount of net preference, but in no case was the dif- It also remains possible that poor neigh- bourhood reputation might create neigh- ference in positive responses between those from neighbourhoods with different reputa- bourhood effects on employment in parts of these LLMs or parts of the recruitment and tions statistically significant. In summary, we do not find statistically selection process that were outside the scope of this experiment, in particular for unpro- significant evidence that employers prefer those living in neighbourhoods with bland mising candidates and for jobs applied to reputations to those living in neighbour- face-to-face. hoods with poor reputations, in the case of As it stands, this is a different result from that found through multivariate anal- attractive candidates looking for work in selected jobs requiring limited education ysis by McGregor (1977). One interpreta- and skills, in three contrasting LLMs. There tion is that the earlier study’s method was not able to take account of all hidden vari- was some net preference for candidates from neighbourhoods with bland reputa- ables. Another, perhaps more important one, however, is that the market for male tions but it was small in size and not statis- tically significant. manual labour in Glasgow in the 1970s dif- fered from the market for low-skilled ser- vice employment for both genders across Discussion England and Wales in 2010–11. The size of the net preferences is small in The ‘Stigma’ Neighbourhood Effects relation to the differences in employment Pathway and Variation in Employment rates between neighbourhoods. It is not so between Neighbourhoods small in relation to the typical size of neigh- Based on this evidence, the hypothesis that bourhood effects, one explanation for dif- residents of neighbourhoods with poor ferences in employment rates, as revealed 776 REBECCA TUNSTALL ET AL. in empirical studies (for example, Lupton, employment rates, or the extent to which 2003a; Syrett, 2008; van Ham et al., 2012). direct experience is reported. In a study of young unemployed people Reviewing the Sources for the Ideas about in Newham in London, Roberts (1999) the ‘Stigma’ Neighbourhood Effects found that almost a third of young people Pathway from the most deprived parts of the bor- ough thought that employers were put off The idea that poor neighbourhood reputa- by the area in which they lived. The converse tion might provide a pathway through finding was that a majority of residents did which neighbourhood effects might operate not think that poor neighbourhood reputa- appears to have developed through qualita- tion might create neighbourhood effects on tive research which indicates variations in employment. The evaluation of Working reputation coinciding with variation in Neighbourhood Pilots found that just over employment, and the presence of beliefs one in ten residents thought that ‘employ- about neighbourhood effects amongst at ers don’t want to employ local people’ least some residents and LLM intermedi- (Dewson et al., 2007). The converse is that aries. One of the irreplaceable functions of almost nine in ten did not agree with this open-ended qualitative work is to explore idea. In both cases, the context of the new areas of social enquiry and to generate research might have encouraged respon- hypotheses. However, not all such hypoth- dents to point to barriers to employment eses are the same and the strength of evi- other than their own characteristics and dence behind them, and the indications of behaviour. Like Roberts (1999) and the prevalence and salience of the processes Dewson et al. (2007), we found that only a they identity may vary. minority of interviewees supported the Most existing qualitative studies which neighbourhood reputation effects hypoth- have produced evidence of potential or per- esis. Members of the public were most ceived neighbourhood reputation effects on likely to think that neighbourhood reputa- employment have not been focused on tion effects on employment might exist, neighbourhood reputation effects (Lawless but they were also the group least likely to and Smith, 1998; SEU, 1998; Taylor, 1998; have direct experience or evidence of Fieldhouse, 1999; Roberts, 1999; Dean and neighbourhood reputation effects. Hastings, 2000; Speak, 2000; Mellor, 2002; It is possible that neighbourhood effects Taylor, 2003; Aleksandraviciene et al., 2005; may be specific not only to national or Dewson et al., 2007; Sanderson, 2006; neighbourhood contexts, but also to partic- Green and White, 2007; Fletcher, 2007; ular time-periods. Over the past five years, Fletcher et al., 2008). Many have not asked employers even for low-skilled, low-paid explicitly or directly about neighbourhood and manual work, have switched from reputation effects, but have recorded refer- paper and mail to electronic applications. ences that emerged in open-ended enquiries This practice was reflected amongst employ- or as interesting by-products of research ers interviewed for this project, with the par- into other issues. Thus they have tended to record mentions of potential neighbour- tial exception of those employing kitchen hands. This has meant that circumventing hood reputation effects without taking into account their prevalence or salience as any poor reputation neighbourhood effect by using a false postal address is both virtu- potential explanations for individual employment status or neighbourhood ally costless and practically redundant. POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 777 Conclusions existing evidence on neighbourhood effects suggests that the size of the effects is rela- In summary, this paper has found evidence tively small compared with differences in to contradict the hypothesis employment rates between neighbour- hoods, the results of this work do not rule Residents of neighbourhoods with poor repu- out a possible contribution of area reputa- tations fare worse when applying for relatively tion to neighbourhood effects on employ- low-skilled jobs than residents of neighbour- ment rates. hoods in the same labour market with better reputations, all other things being equal. Acknowledgements The result suggests that relatively well- The authors would like to thank all those who qualified candidates from areas with poor gave time to be interviewed, and two anonymous referees of this paper. reputations should not fear postcode dis- crimination, at least up until the interview stage of the recruitment process. It remains possible that poor neighbourhood reputa- Funding tion might create neighbourhood effects on The authors would like to thank the Joseph employment in parts of these LLMs or Rowntree Foundation for funding the research parts of the recruitment and selection pro- on which this article is based. cess that were outside the scope of this experiment. However, centralised and elec- Notes tronic recruitment may be reducing the scope for these effects in large parts of the 1. We used official travel-to-work areas labour market. (TTWAs) (Coombes and Bond, 2008) as the spatial units from which to select LLMs for On the basis of this evidence, there is no the experiment. argument for policy interventions, includ- 2. The www.direct.gov.uk website advertises all ing policies to reduce sorting or to address vacancies notified to Jobcentre Plus (the this neighbourhood effect pathway more Public Employment Service in Great directly, on the grounds of neighbourhood Britain), estimated to be 40 per cent of total effects on employment. Nonetheless, there vacancies, with higher coverage of may be other arguments for these policies lower-skilled vacancies. It is the single largest on area reputations; for example, that resi- source of job vacancies in the UK. dents of areas with poor reputations suffer discrimination in the provision of services, References or face unequal treatment or recognition. The paper demonstrates the value of Aleksandraviciene, I., Carpenter, M., Cleaver, P., investigating the sources of ideas about Hussain, Z. et al. (2005) The impact of com- neighbourhood effects pathways, and the munity based initiatives on unemployed men and women in Coventry: final report of sequel factors necessary for pathways to operate, as project on gender and class. Centre for Lifelong well as for testing pathways themselves. It Learning and Department of Sociology, Uni- demonstrates the value of experimental versity of Warwick. methods in exploration of neighbourhood Amion Consulting (2010) Evaluation of the effects. It also demonstrates how mixed national strategy for neighbourhood renewal: methods approaches may add to the value final evaluation. Department for Commu- of individual elements of research. As nities and Local Government, London. 778 REBECCA TUNSTALL ET AL. Atkinson, R. and Jacobs, K. (2008) Public housing Fletcher, D. R., Gore, T., Reeve, K., Robinson, D. in Australia: stigma, home and opportunity. et al. (2008) Social housing and worklessness: Research Paper No. 1, Housing and Commu- qualitative research findings. Research Report nity Research Unit, University of Tasmania. No. 521, Department for Work and Pensions, Bates, P., Carta, E., Dewson, S., Francis, R. and Corporate Document Services, Leeds. Pillai, R. (2007) Working in 21st century rural Galster, G. (2007) Should policy makers strive England: a scoping study. Commission for for neighborhood social mix? An analysis of Rural England, Institute for Employment the Western European evidence base, Hous- Studies, Brighton. ing Studies, 22(4), pp. 523–545. Bovenkerk, F. (1992) Testing discrimination in Galster, G. (2012) The mechanism(s) of neigh- natural experiments: a manual for interna- bourhood effects: theory, evidence, and policy tional comparative research on discrimination implications, in: M. van Ham, D. Manley, N. on the grounds of ‘race’ and ethnic origin. Inter- Bailey, L. Simpson and D. Maclennan (Eds) national Labour Office, Geneva. Neighbourhood Effects Research: New Perspec- Burns, T., Catty, J., Becker, T., Drake, R. E. tives, pp. 23–56. Dordrecht: Springer. et al. (2007) The effectiveness of supported Green, A. E. and Adam, D. (2011) City strategy: employment for people with severe mental final evaluation. Research Report No. 783, illness: a randomised control trial, Lancet, Department of Work and Pensions, Corpo- 370, pp. 1146–1152. rate Document Services, Leeds. Christie, I. and Rolfe, H. (1992) The TECs and Green, A. E. and Owen, D. W. (2006) The Geo- inner city regeneration, Policy Studies, 13, pp. graphy of Poor Skills and Access to Work. York: 54–63. Joseph Rowntree Foundation. Coombes, M. G. and Bond, S. (2008) Travel-to- Green, A. E. and White, R. J. (2007) Attachment work areas: the 2007 review. Office for to place: social networks, mobility and prospects National Statistics, London. of young people. Joseph Rowntree Foundation, Dean, J. and Hastings, A. (2000) Challenging York. images: housing estates, stigma and area regen- Green, F., Maguire, M., Roberts, B. and Wray, K. eration. Joseph Rowntree Foundation, York. (1991) Explaining persistent inner city unem- Dewson, S., Casebourne, J., Darlow, A., ployment: a case study of Nottingham, Local Bickerstaffe, T. et al. (2007) Evaluation of the Economy, 6, pp. 240–249. working neighbourhoods pilot: final report. Ham, M. van, Manley, D., Bailey, N., Simpson, L. Research Report No. 411, Department for and Maclennan, D. (2012) Neighbourhood Work and Pensions, Coporate Document effects research: new perspectives, in: M. van Services, Leeds. Ham, D. Manley, N. Bailey, L. Simpson and DfT (Department for Transport) (2011) Personal D. Maclennan (Eds) Neighbourhood Effects travel factsheet: commuting and business travel. Research: New Perspectives, pp. 1–22. Dor- Department for Transport, London. drecht: Springer. Dovidio, J. F. and Gartner, S. L. (2000) Aversive Hastings, A. and Dean, J. (2003) Challenging racism and selective decisions: 1989 and 1999, images: tackling stigma through estate regen- Psychological Science, 11(4), pp. 315–319. eration, Policy and Politics, 31, pp. 171–184. Eardley, T. and Thompson, M. (1997) Does case- Hirsch, D. and Moore, R. (2011) The living management help unemployed job seekers? a wage in the United Kingdom: building on success. review of the international evidence. Social Citizens UK, The Centre for Research in Policy Research Centre, Sydney. Social Policy and the Living Wage Foundation Fieldhouse, E. A. (1999) Ethnic minority (www.citizensuk.org/wp-content/uploads/2011/ unemployment and spatial mismatch: the 06/The-Living-Wage-in-the-United-Kingdom- case of London, Urban Studies, 36, pp. 1569– May-2011.pdf). 1596. Holzer, H. J. (1991) The spatial mismatch Fletcher, D. R. (2007) A culture of worklessness? hypothesis: what has the evidence shown?, Historical insights from the Manor and Park Urban Studies, 28, pp. 105–122. area of Sheffield, Policy and Politics, 35, pp. Houston, D. S. (2005) Employability, skills mis- 65–85. match and spatial mismatch in metropolitan POOR NEIGHBOURHOOD REPUTATION AND EMPLOYMENT 779 labour markets, Urban Studies, 42, pp. National Equality Panel (2010) An anatomy of 221–243. economic inequality in the UK. Government Joseph, M. L., Chaskin, R. J. and Webber, H. S. Equalities Office, London. (2007) The theoretical basis for addressing Nunn, A., Bickerstaffe, T., Hogarth, T., poverty through mixed-income development, Bosworth, D. et al. (2010) Postcode selection? Urban Affairs Review, 42(3), pp. 369–409. Employers’ use of area- and address-based Jowell, R. and Prescott-Clarke, P. (1970) Racial information shortcuts in recruitment decisions. discrimination and white-collar workers in Research Report No. 664, Department for Britain, Race and Class, 11, pp. 397–417. Work and Pensions, Corporate Document Kain, J. (1968) Housing segregation, negro Services, Leeds. unemployment and metropolitan segrega- Pager, D. (2007) The use of field experiments tion, Quarterly Journal of Economics, 82, for studies of employment discrimination: pp. 175–197. contributions, critiques, and directions for Kleinhans, R. (2004) Social implications of the future, Annals of the American Academy housing diversification in urban renewal: a of Political and Social Science, 609, pp. 104– review of recent literature, Journal of Housing 133. and Built Environment, 19, pp. 367–390. Pager, D. and Quillian, L. (2005) Walking the Kornfeld, R., Wood, M., Orr, L. and Long, D. talk? What employers say versus what they (1999) Impacts of the project network demon- do, American Sociological Review, 70, pp. 355– stration. Abt Associates Inc., Bethesda, MD. 380. Lawless, P. (1995) Inner-city and suburban Purdon, S., Stratford, N., Taylor, R., labour markets in a major English conurba- Natarajan, L. et al. (2006) Impacts of the job tion: processes and policy implications, retention and rehabilitation pilot. Research Urban Studies, 32, pp. 1097–1125. Report No. 342, Department for Work and Lawless, P. and Smith, Y. (1998) Poverty, Pensions, Corporate Document Services, inequality and exclusion in the contempo- Leeds. rary city, in: P. Lawless, R. Martin and S. Rangarajan, A. and Novak, T. (1999) The strug- Hardy (Eds) Unemployment and Social gle to sustain employment: the effectiveness of Exclusion: Landscapes of Labour Inequality, the post employment services demonstration. pp. 201–216. London: Jessica Kingsley. Mathematica Policy Research Inc., Princeton, Lupton, R. (2003a) Neighbourhood effects: can we NJ. measure them and does it matter? CASE Paper Recchia, F. (2008) Immigration, politics and vio- No. 73, London School of Economics. lence in urban France: between fiction and Lupton, R. (2003b) Poverty Street: The Dynamics facts, Information, Society and Justice, 2(1), of Neighbourhood Decline and Renewal. Bris- pp. 47–61. tol: Policy Press. Riach, P. A. and Rich, J. (2004) Deceptive field Marsden, P. V. (1994) The hiring process: experiments of discrimination: are they ethi- recruitment methods, American Behavioral cal?, Kyklos, 57(3), pp. 457–470. Scientist, 37(7), pp. 979–991. Roberts, F. (1999) Young people out of work in Mellor, R. (2002) Hypocritical city: cycles of south west Newham. Commentary Series No. urban exclusion, in: J. Peck and K. Ward 83, Centre for Institutional Studies, Univer- (Eds) Revolutionary City: Restructuring Man- sity of East London. chester, pp. 214–234. Manchester: Manche- Sanderson, I. (2006) Worklessness in deprived ster University Press. neighbourhoods: a review of the evidence. McGinnity, F., Nelson, J., Lunn, P. and Quinn, E. Department for Communities and Local (2009) Discrimination in recruitment: evidence Government, London. from a field experiment. The Equality Author- SEU (Social Exclusion Unit) (1998) Bringing ity and the Economic and Social Research Britain Together: A National Strategy for Institute, Dublin. Neighbourhood Renewal. Norwich: The Sta- McGregor, A. (1977) Intra-urban variations in tionery Office. unemployment: a case study, Urban Studies, Souza Briggs, X. de, Popkin, S. J. and Goering, J. 14, pp. 303–313. (2010) Moving to Opportunity: The Story of an 780 REBECCA TUNSTALL ET AL. American Experiment to Fight Ghetto Poverty. employers perceive neighbourhoods in four Oxford: Oxford University Press. urban labour markets, in: A. O’Connor, C. Speak, S. (2000) Barriers to lone parents’ employ- Tilly and L. T. Bobo (Eds) Urban inequality, ment: looking beyond the obvious, Local ch 6. New York: Russell Sage Foundation. Economy, 15, pp. 32–44. Walker, R. (2000) Learning if policy will work: Stafford, B., Greenberg, D. and Davis, A. (2001) the case of the new deal for disabled people, A review of the use of social experiments. Policy Studies, 21(4), pp. 313–332. Research report, Department of Social Secu- Waquant, L. J. D. (1993) Urban outcasts: stigma rity, London. and division in the black American ghetto Stratford, N., Farrell, C., Natarajan, L. and and the French urban periphery, International Lewis, J. (2005) Taking part in a randomised Journal of Urban and Regional Research, 17, control trial: a participants’ eye view of the job pp. 366–383. retention and rehabilitation pilot. Corporate Watt, P. (2003) Urban marginality and labour Document Services, Leeds. market restructuring: local authority tenants Syrett, S. (2008) Area effects: a review of the evi- and employment in an inner London Bor- dence prepared for CLG. Department of Com- ough, Urban Studies, 40, pp. 1769–1789. munities and Local Government, London. Wilson, W. J. (1987) The Truly Disadvantaged. Taylor, M. (1998) Combatting the social exclu- Chicago, IL: University of Chicago Press. sion of housing estates, Housing Studies, 13, Wood, M., Hales, J., Purdon, S., Sejersen, T. and pp. 819–832. Hayllar, O. (2009) A test for racial discrimina- Taylor, Y. (2003) Negotiation and navigation: an tion in recruitment practice in British cities. exploration of the spaces/places of working- Research Report No. 607, Department for class lesbians, Sociological Research Online, Work and Pensions, Corporate Document 9(1). Services, Leeds. Tilly, C., Moss, P., Kirschenman, J. and Zenou, Y. (2002) How do firms redline workers?, Kennelly, I. (2001) Space as a signal: how Journal of Urban Economics, 52, pp. 391–408.

Journal

Urban Studies: An International Journal of Research in Urban StudiesSAGE

Published: Jun 24, 2013

There are no references for this article.