Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/university-of-california-press/what-diverse-samples-can-teach-us-about-cognitive-vulnerability-to-Gwqiz8uqxI
- Publisher
- University of California Press
- eISSN
- 2474-7394
- DOI
- 10.1525/collabra.71346
- Publisher site
- See Article on Publisher Site
Abstract
Haeffel, G. J., Burke, H., Vander Missen, M., & Brouder, L. (2023). What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression. Collabra: Psychology, 9(1). https://doi.org/10.1525/collabra.71346 Clinical Psychology What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression 1 1 1 1 Gerald J. Haeffel , Hugh Burke , Marissa Vander Missen , Lily Brouder Department of Psychology, University of Notre Dame, Notre Dame, IN, USA Keywords: cognitive vulnerability, hopelessness theory of depression, Nepal, Honduras, Black, generalizability, theory building https://doi.org/10.1525/collabra.71346 Collabra: Psychology Vol. 9, Issue 1, 2023 Tests of generalizability can diversify psychological science and improve theories and measurement. To this end, we conducted five studies testing the cognitive vulnerability to depression hypothesis featured in the hopelessness theory of depression: Study 1 was conducted with Honduran young adults (n = 50); Study 2 was conducted with Nepali adults (n = 34); Study 3 was conducted with Western hemisphere adults (n = 104); Study 4 was conducted with Black U.S. adults (n = 119); and Study 5 was conducted with U.S. undergraduates (n = 110). Results showed that cognitive vulnerability could be measured reliably in diverse populations and the distribution of vulnerability scores was similar for all samples. However, the tendency to generate negative inferences about stress had different implications for depression depending on sample; the association between cognitive vulnerability and depressive symptoms did not generalize to Honduran and Nepali participants. It is now necessary to understand why a negative cognitive style confers risk for depression in some contexts but not others (e.g., is it issues related to measurement, theory, or both). The results also suggest that understanding and reducing the global burden of depression will require more than simply “translating” existing cognitive measures and theories to other countries. Psychological scientists and the samples they use tend search often requires travel, money, and additional re- to be White and WEIRD (Western, Educated, Industrialized, search personnel. There can also be language barriers and Rich, and Democratic; Clancy & Davis, 2019; Henrich et al., differences in access to technology (e.g., computers, inter- 2010). According to Thalmayer and colleagues (2021), only net, etc.). 11% of the world’s population is represented in the top psy- Another factor that may be preventing scientists from chology journals with 60% percent of authors and samples testing their theories in more diverse samples is fear that specific to the United States. Further, only 3-6% of all pub- the original findings will not generalize. Although it is un- lished research articles on mental health include partici- realistic to assume that all theories will apply equally well pants from low and middle-income countries (El Khoury et to all people and in all contexts, there seems to be a grow- al., 2021; Gallegos et al., 2013; Haque & Kamal, 2019; Patel ing intolerance and stigmatization of failed replications. & Sumathipala, 2001). Given such limited representation, This is probably a result of psychology’s replication crisis it is not surprising that the validity and robustness of psy- (e.g., Open Science Collaboration, 2015), the prevalence of chological theories have been called into question. questionable research practices (e.g., p-hacking, HARKing, There are a number of reasons for why the scientists and and piecemeal publication; Simmons et al., 2011), and re- research samples in psychological science are White and cent high-profile cases of scientific misconduct (e.g., Giner- WEIRD. First, there is systemic racism in who is admitted Sorolla, 2017; Ledford, 2010; Murray, 2002). to doctoral programs and hired as faculty at research uni- It is critical to incentivize and value research on gener- versities (e.g., Baumgartner et al., 2010; D’Augelli & Her- alizability (Haeffel & Cobb, 2022). This work can diversify shberger, 1993; Gildersleeve et al., 2011) Second, it is far psychological research, which makes for better science easier to recruit and run studies with convenience samples (e.g., Meadon & Spurrett, 2010; Medin & Lee, 2012; Plaut, (e.g., college students) than more representative samples. 2010). Further, focusing on generalizability may help curb It is not merely a matter of swapping white and WEIRD researchers’ tendency to overstate the generalizability and samples for non-white and non-WEIRD samples; this re- real-world implications of their findings (e.g., Rad et al., a Correspondence concerning this article should be addressed to Gerald Haeffel, Corbett Hall, University of Notre Dame, Notre Dame, IN 46556. E-mail: ghaeffel@nd.edu What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression 2018). If researchers know that their theories will be tested the life span (see Romens et al., 2008 for review). It can in diverse samples, then they may be more conservative be thought of as one’s “native language” for interpreting with their conclusions. Finally, studies that test general- life stress (Haeffel & Kaschak, 2019). Research has largely izability can inform measurement and constrain theories. supported the cognitive vulnerability hypothesis featured Replication attempts in diverse samples are going to expose in the hopelessness theory (Abramson et al., 1999; Hankin problems with current theories and measures, which will et al., 2005; Hong et al., 2006; Metalsky et al., 1987; Swend- require scientists to rethink and reformulate them. As sci- sen, 1997). Over 50 published studies have shown that it entists, we likely learn more from a failed replication than is possible to reliably assess individual differences in cog- a successful replication because scientific progress comes nitive vulnerability and that those differences precede and from disagreement (i.e., we learn more from being wrong predict future depressive symptoms and disorders (see Ha- than from being right; Haeffel, 2022; Popper, 1959). effel et al., 2008 for review). It is important to underscore that if a study does not As is the case with much of psychological science, a lim- replicate in a more diverse sample, then it does not mean itation of research on hopelessness theory’s cognitive vul- that the original study was invalid or the result of scientific nerability hypothesis (negative attributional style) is that misconduct. Instead, it means that the theory and findings studies have largely focused on college students from the may only apply to a specific sample or within a specific con- United States (Haeffel et al., 2008). Conducting studies with text. For example, a study testing a theory about marriage undergraduates makes sense because they are at the peak perceptions in American college students may not general- age for developing depression and experience increased ize to other cultures around the globe (or even older adults levels of interpersonal stress (Hankin et al., 1998). However, in the U.S.), but that does not invalidate the work in col- it is now time to take the next logical step in this area of lege students. It simply constrains the theory and conclu- research and examine the generalizability of the theory to sions that can be made; namely, that the results only ap- other populations. This is particularly important for under- ply to American college students. The next step, then, is standing depression as it is a global disorder affecting over to understand why the theory does not apply to other pop- 300 million people around the world (World Health Organi- ulations and revise it. If researchers are to embrace more zation, 2017). rigorous tests of their theories, then the field must accept To address this gap in the literature, we conducted 5 the idea that being wrong can facilitate scientific progress. studies. Studies 1 and 2 were conducted with Honduran and Clearly, a replicable knowledge base is a stalwart of science, Nepali adults, respectively. These were opportunity sam- but refutations are often more informative than confirma- ples; they were chosen because they represent understud- tions. Failed replications push researchers to be more so- ied populations and because we were had established re- phisticated and specific in their theorizing. It forces them lationships with locals at the study sites. Study 1 was to specify the conditions for when a theory will and will not conducted in a sample of Latino young adults from an or- apply. This will lead to stronger theories and a more inclu- phanage in the Republic of Honduras. Honduras has some sive and global psychological science. of the highest rates of depression in the world, yet it has The purpose of this research is to test the generalizabil- received among the least empirical attention by depression ity of hopelessness theory’s cognitive vulnerability hypoth- researchers (Gallegos et al., 2013; Gibbons, 2020; Thal- esis (Abramson et al., 1989) in five unique samples from mayer et al., 2021; VandenBos & Winkler, 2015). The rate of around the globe. We chose to focus on the hopelessness adolescent depression in the Republic of Honduras is 21% theory of depression because of it is defined by a well-ar- (Central American Refugee Health Profiles, 2017) whereas in ticulated theory, has strong empirical support, and there the United States it is approximately 13% (NIMH, 2017). is a psychometrically validated measure of cognitive vul- Study 2 was conducted with a small sample of Nepali nerability (Haeffel et al., 2008). According to the theory adults. Nepal has a population of about 30 million, and it (Abramson et al., 1989), some people have a cognitive vul- is one of the most economically disadvantaged countries in nerability that puts them at heightened risk for developing South Asia. Nearly 90% of the population lives in rural ar- depressive symptoms and depressive disorders. Cognitive eas, and “psychiatric disorders make up approximately 11% vulnerability refers to the tendency of a person to generate of the total amount of disease and illness in the country” overly negative inferences about the cause, consequences, (Hall et al., 2016, p. 278). According to Luitel and colleagues and self-worth implications of stressful life events. Specif- (2015), the mental health care situation in Nepal is “dire.” ically, when faced with a stressful life event, an individual Studies 3-5 recruited samples living in the Western Hemi- with high levels of cognitive vulnerability is likely to: (a) at- sphere. Study 3 was conducted with a sample of adults from tribute the event to stable and global causes; (b) view the Western countries. Study 4 was conducted with a sample of event as likely to lead to other negative consequences; and Black adults from the United States. Study 5 was conducted (c) construe the event as implying that he or she is unwor- with a U.S. sample of college students. thy or deficient. Individuals who generate these three types We had three primary hypotheses. First, we hypothe- of negative inferences are hypothesized to be at greater risk sized that it would be possible to reliably detect individual for depression than people who do not generate these types differences in cognitive vulnerability and depressive symp- of inferences. toms in all five samples. Specifically, we predicted the Cognitive vulnerability tends to solidify during early translated (Spanish and Nepali) measures of cognitive vul- adolescence, and then shows trait-like stability throughout nerability and depressive symptoms would demonstrate Collabra: Psychology 2 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression good reliability as operationalized by measures of internal For example, in populations with chronically high levels of consistency (i.e., alpha and Macdonald’s omega); based on stress (e.g., poverty, war, etc.), people may experience de- prior research, we expected alpha and MacDonald’s omega pressive symptoms regardless of their cognitive vulnera- levels to exceed .80 for the measure of cognitive vulnerabil- bility level. In other words, when stress is overwhelming, ity and the two measures of depressive symptoms. Second, it may push any individual (even those with low levels we hypothesized that cognitive vulnerability would show of cognitive vulnerability) over the threshold into depres- a normal distribution of scores in all five samples. These sion. In this case, cognitive vulnerability will not be as hypotheses are based on the assumption that generating highly correlated with depressive symptoms (e.g., titration causal attributions for events in one’s life is a universal model; Abramson et al., 1997). There is preliminary evi- human process. In other words, the tendency to generate dence from studies testing similar cognitive vulnerability interpretations of stressful life events that have negative constructs showing that cultural context matters for risk for implications for one’s future and self-worth is present to depression. For example, Auerbach and colleagues (2010) a greater or lesser degree in all humans. Finally, we hy- found that the interaction effects of perceived control and pothesized that cognitive vulnerability would be moder- life stress on depressive symptoms was different for Cana- ately positively correlated (ranging from .3-.6) with depres- dian and Chinese adolescents (see Cohen et al., 2013 for sive symptoms (regardless of depressive symptom measure similar findings in a Chinese adolescent sample for negative used) as found in prior research (see Haeffel et al., 2008 for attachment cognitions). These results indicate that it may review). be presumptuous to hypothesize the results of the current All three of our hypotheses assume that cognitive vul- study will directly map onto prior research using American nerability is a universal human construct that past research college students. using Western and white samples will generalize to (behave Method similarly in) to other cultures and populations. This as- sumption is partially supported by preliminary work in- Participants vestigating related vulnerability constructs such as those Study 1- Honduran Young Adults. Participants were 50 featured in Beck’s theory (1979; dysfunctional attitudes) young adults (26 females, 24 males) living in a Honduras and Nolen-Hoeksema’s ruminative response theory (1991; orphanage approximately one hour outside of the capital brooding). Research on these theories found that the re- city, Tegucigalpa. Participants ranged in age from 18-27 (M sults reported in U.S. samples generalized to populations = 24, SD = 3). The Honduran orphanage provides living ac- living in the United Arab Emirates, Egypt, Asia and Europe commodations, medical care, food, and water for approxi- (Beshai et al., 2016; Chahar Mahali et al., 2020; Thomas & mately 300 orphans. Schooling is required during the week, Altareb, 2012). and upon entering adolescence, each student is also given However, there is also reason to hypothesize that cogni- the opportunity to learn a “trade” (shoemaking, electricity, tive vulnerability may not generalize to more diverse pop- etc.) and receive higher education. Data was collected dur- ulations. Research shows that a person’s level of cognitive ing May 2018. vulnerability is influenced by his or her early social envi- Study 2 – Nepali Adults. Participants were 34 native ronment. One particularly important predictor of cognitive Nepali adults (29 females, 5 males) ranging in age from vulnerability levels is the direct feedback adolescents re- 22-55 (M = 42, SD = 8). They were receiving services from ceive about stress from their parents, peers, and teachers Koshish, which is part of a “a national self-help organiza- (Alloy et al., 2001; Garber & Flynn, 2001; Mezulis et al., tion working in the field of mental health in Nepal.” Clients 2006; Peterson & Seligman, 1984; Stark et al., 1996). Cul- of Koshish (mainly women and girls) are often homeless, tural norms may also influence the types of feedback ado- estranged from their families, and/or victims of sexual/ lescents receive about stressful life events. In more indi- physical abuse. All clients are screened at intake for acute vidualistic cultures, adolescents might be more likely to episodes of mental illness and tend to stay at the resident receive feedback related to one’s worth and individual fu- home an average of three months. The Koshish transit ture. However, in more collectivistic cultures, feedback may home houses approximately thirty people at a time. The focus on the role of the greater social context in interpret- participants in this study were involved in self-help therapy ing the stressful event (Knyazev et al., 2017). This means groups, which Koshish runs for clients with chronic mental that there may be some cultures in which cognitive vulner- health problems. Data was collected during July 2019. ability levels are lower on average. Study 3 – Western Hemisphere Adults. Participants were Further, the cultural context may affect the degree to 104 adults (65 females, 39 males) from the Western hemi- which cognitive vulnerability and depression levels are as- sphere who were recruited via an online participant plat- sociated. In research using undergraduates, cognitive vul- form (Prolific; Palan & Schitter, 2018). Participants ranged nerability scores are normally distributed, and stress tends in age from 18-61 (M = 36) and resided in a variety of West- to be acute (rather than chronic). In these studies, high ern countries including Canada, Ireland, Scotland, England, levels of cognitive vulnerability combine with the greater and the United States. Participants were homogenous with stress occurrences to predict depressive symptoms. How- regard to race and ethnicity with 95% self-reporting as Cau- ever, in non-White and non-WEIRD samples, it is possible casian, 3% African American, 4% “other”, 1 % Hispanic, that cultural and geographic differences in stress will affect and 1% Asian. One hundred thirteen participants started how cognitive vulnerability is associated with depression. Collabra: Psychology 3 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression the study, but 9 participants were removed for not com- asked to imagine the events happening to themselves; they pleting at least one of the measures. All participants in- then are asked to write down what they believe to be the cluded in the analyses completed at least 95% of the items cause of the event. Likert scale ratings (1-7) are then made on all measures. Participants write-in answers on the CSQ for the three vulnerability dimensions featured in the hope- were examined for repetitions and unusual/non-sensical lessness theory of depression: stability and globality; prob- answers; no participants were excluded for these reasons. able consequences of each event; and the self-worth impli- Data was collected during January 2020. cations of each event. An individual’s CSQ score is his/her Study 4 – Black U.S. Adults. Participants were 119 adults average Likert scale rating across these three dimensions (70 females, 49 males) from the United States who were re- (stability and globality, consequences, and self-worth char- cruited via an on-line participant platform (Prolific). Partic- acteristics) for the 6 hypothetical negative life events. Com- ipants ranged in age from 18-66 (M = 31), were all from the posite score ranges from 1 to 7, with higher scores reflect- United States, and all self-identified as Black. One hundred ing greater levels of cognitive vulnerability to depression. twenty participants started the study, but 1 participant was The CSQ has good internal consistency, reliability, and va- removed for not completing at least one of the measures. lidity (see Haeffel et al., 2008 for review; Haeffel & Howard, All participants included in the analyses completed at least 2010). The CSQ was translated to Spanish for use with the 95% of the items on all measures. Participants write-in an- Honduran young adults with the help of the administra- swers on the CSQ were examined for repetitions and un- tors of the Honduran orphanage as well as the University usual/non-sensical answers; no participants were excluded of Notre Dame Romance Languages department. The final for these reasons. Data was collected during October 2019. translated version of the CSQ was evaluated and approved Study 5 – U.S. Undergraduates. Participants were 110 by administrators of the orphanage. undergraduates (79 females, 31 males) from a private, mid- The Nepali sample was administered a single-scenario sized university in the Midwestern United States who were CSQ called the Particular Inference Questionnaire (PIQ; recruited through the University’s online volunteer partic- Haeffel, 2011; Metalsky et al., 1987). This short measure of ipant pool. Participants ranged in age from 18-23 (M = 19) vulnerability was used because of time constraints and fea- and the ethnicity of the sample was representative of the sibility issues related to translating a longer measure from race and ethnicity of the University - 78% Caucasian, 13% English to Nepali. The PIQ uses a single idiographic sce- African American, 2% Asian-American, 2% Native Amer- nario about which participants make Likert ratings (1-7) on ican, and 4% “Other.” One hundred fifteen participants the three vulnerability dimensions featured in the hope- started the study, but five participants were removed for lessness theory of depression: stability and globality; prob- not completing at least one of the measures. All partici- able consequences of each event; and the self-worth im- pants included in the analyses completed at least 95% of plications of each event. Like the CSQ, an individual’s PIQ the items on all measures. Data was collected between No- score is his/her average rating across these three dimen- vember 2019 and January 2020. sions (stability and globality, consequences, and self-worth characteristics) for the single scenario. The PIQ was trans- Power Analysis lated from English to Nepali with the help of a paid native speaking translator. The translated version of the PIQ was Analyses were not focused on null hypothesis testing, then evaluated and approved by three separate KOSHISH but rather on the distribution of scores, reliability estimates staff, including the head of program development. The PIQ for measures, and a simple bivariate correlation. A post hoc was also administered to the Western adult, Black adult, power analysis using G*Power showed that a sample size and college student samples so that we could make direct of 34 participants was needed to detect a one-tailed (pos- comparisons to the Nepali sample. itive) bivariate correlation of medium effect size (r = ~.4) Depressive Symptoms. Two measures of depressive with power of .80 and alpha level of alpha = 05. The medium symptoms were used. The Honduran adolescent, Western effect size was based on the correlation sizes between cog- Adult, Black adult, and college student samples (Studies 1, nitive vulnerability and depressive symptoms found in prior 3, 4 and 5) were administered the Beck Depression Inven- work (correlations typically range between .3 and. 6; see tory (BDI; Beck et al., 1979). The BDI is a 21 item self-report Haeffel et al., 2008 for review). Note that the Honduran questionnaire. Scores are created by summing the items and Nepali samples were not sufficiently powered to detect (range 0-63), with higher scores indicating greater levels of small effects. depressive symptoms. The BDI has demonstrated high inter- nal consistency (coefficient alpha is typically greater than Measures .8), good test-retest reliability (r= .60-.83 for nonpsychi- atric samples), and validity with both college and psychi- Cognitive Vulnerability. Two measures were used to atric samples (see Beck et al., 1988, for review). The BDI measure cognitive vulnerability (as featured in the hope- was translated to Spanish for use with the Honduran young lessness theory of depression; Abramson et al., 1989). The adults with the help of the administrators of the Honduran Honduran young adults, Western adults, Black U.S. adults, orphanage as well as the University of Notre Dame Ro- and U.S. college students (Studies 1, 3, 4 and 5) were ad- mance Language department. The translated version of the ministered the Cognitive Style Questionnaire (CSQ; Haeffel BDI was then evaluated and approved by administrators of et al., 2008). Participants are presented with 6 hypothetical the orphanage. negative events (3 achievement and 3 interpersonal) and Collabra: Psychology 4 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression The Nepali sample was administered the anhedonic sub- formation about our work to the local administrators by scale of the Mood and Anxiety Symptom Questionnaire presenting short talks and discussions about cognition and (MASQ; Watson et al., 1995). The MASQ is a 90-item self- depression. We then consulted with the local leaders to de- report questionnaire that assesses general depressive and termine how best to conduct the study. As part of this col- specific anhedonic symptoms of depression based on the laboration, the local team was involved in the study design tripartite theory of anxiety and depression (Clark & Wat- and measure translations. We believe that this approach led son, 1991). The anhedonic subscale contains 22 items that to assessments that more closely matched the local dialect assess symptoms hypothesized to be specific to depression and had increased face validity. The drawback of this ap- (e.g., low positive affect, loss of pleasure in daily activities). proach is that we did not use an existing measure validated The MASQ has demonstrated good reliability and validity by prior work. (e.g., Watson et al., 1995). The anhedonic subscale was All procedures for all five studies were approved by the translated from English to Nepali with the help of a paid University of Notre Dame human subject review board as native speaking translator. The translated version of the well as by each on-site location (Honduran orphanage and MASQ was then evaluated and approved by three separate Koshish, respectively). The complete data set and other ma- KOSHISH staff, including the head of program develop- terials can be found here: https://osf.io/umg9p/. No data ment. The MASQ anhedonic subscale was also adminis- were excluded from the data set or the analyses. All mea- tered to the Western adult, Black adult, and college student sures administered to participants are described in the sec- samples so that we could make direct comparisons to the tions below and included in the data set and analyses. Nepali sample. Results Procedure Hypothesis 1: Cognitive Vulnerability Can Be Measured Reliably All five studies used a cross-sectional design in which measures of cognitive vulnerability and depressive symp- The measure of cognitive vulnerability (CSQ) exhibited toms were administered in a single study session. Measures strong psychometric properties in the four samples in were administered in-person for Studies 1 and 2 (Honduran which it was administered. Coefficient alpha and McDon- and Nepali samples, respectively). In Study 1 (Honduran ald’s omega were greater than .80 in the four samples in sample), native Spanish-speakers were on site during the which it was administered (see Figure 1; Honduran sample questionnaire session to ensure an appropriate testing en- [alpha = .81; McDonald’s omega = .83], Western adult sam- vironment and to answer any questions. The questionnaires ple [alpha = .89; McDonald’s omega = .89], Black adult sam- were administered during three classroom sessions, with ple [alpha = .92; McDonald’s omega = .93], and the under- the first two sessions consisting of twenty students and the graduate sample [alpha = .89; McDonald’s omega = .88]). third consisting of ten students. In Study 2 (Nepali sam- In contrast, the single-scenario measure of cognitive ple), native Nepali speakers were on site during the ques- vulnerability (PIQ) exhibited low reliability estimates in the tionnaire session to ensure an appropriate testing environ- four samples in which it was administered. Coefficient al- ment and to answer any questions. The questionnaires were pha and McDonald’s omega ranged between .45 and .60 (see administered across five group sessions with 5-9 people in Figure 1; Nepali sample [alpha = .45; McDonald’s omega = session. Studies 3-5 (Western adults, Black U.S. adults, and .60], Western adult sample [alpha = .52; McDonald’s omega U.S. undergraduates, respectively) were conducted online. = .60], Black adult sample [alpha = .48; McDonald’s omega Studies 3 and 4 used the online platform Prolific to re- = .59], and the undergraduate sample [alpha = .57; McDon- cruit participants and were paid approximately $9.50 per ald’s omega = .60]). The low reliability was most likely due hour. We chose Prolific because samples tend to be more di- to the small number of items on the scale relative to the verse, are more scientifically naïve, more honest, and pro- CSQ. vide higher quality data than other popular online plat- Both measures of depressive symptoms also exhibited forms (Peer et al., 2017, 2022). Study 5 used the University’s strong psychometric properties in all five samples (see Fig- online extra credit psychology participant pool. Partici- ure 1). Consistent with prior research, results showed that pants were given course credit for their participation. the BDI had strong reliability estimates in the four samples For the Honduran study, we chose to create new Spanish in which it was administered: Honduran sample (alpha = versions of the CSQ and depressive symptom measures .91; McDonald’s omega = .92), the Western adult sample rather than use existing translations for several reasons (alpha = .92; McDonald’s omega = .92), the Black adult sam- (note, we did not find Nepali versions of these measures ple (alpha = .93; McDonald’s omega = .93), and the under- and, thus, had to create our own versions). The first reason graduate sample (alpha = .91; McDonald’s omega = .91). is that our partnerships with the Honduran and Nepali The anhedonic subscale of the MASQ also exhibited strong sights were not primarily for research. The researchers from reliability in the four samples in which it was administered: our lab were volunteers at these sites and were there to the Nepali sample (alpha = .89; McDonald’s omega = .89), serve the needs of the local volunteer organizations. The the Western adult sample (alpha = .93; McDonald’s omega research studies were secondary to the volunteer effort. = .93), the Black adult sample (alpha = .95; McDonald’s When were afforded the opportunity to conduct research, omega = .95), and the undergraduate sample (alpha = .93; we wanted to ensure the studies were collaborative and not McDonald’s omega = .93). directive. To this end, the research members provided in- Collabra: Psychology 5 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression Figure 1. Distribution of cognitive vulnerability and depressive symptom scores as a function of study. Cognitive vulnerability scores, when measured with the Hypothesis 2: Cognitive Vulnerability Scores Will PIQ, were significantly greater in the Nepali adults (M = Be Normally Distributed 4.52, SD = 1.04) than in Western adults (M = 3.67, SD = .1.16; Cognitive vulnerability scores, as operationalized by t = 3.67, p =.002), Black U.S. adults (M = 3.49, SD = 1.26; tukey both the CSQ and PIQ, tended to be normally distributed t = 4.51, p < .001), and U.S. undergraduates (M = 3.63, tukey (see Figures 1 and 2). The p-values for Shapiro-Wilk test SD = 1.13; t = 3.89, p < .001). tukey of normality were not significant in any of the samples for In contrast to the results for cognitive vulnerability, the the CSQ. These results are consistent with the hypothesis distribution of depressive symptom scores tended to be that cognitive vulnerability, as measured by the CSQ, is a skewed (see Figure 1). Samples also differed on levels of de- continuous construct present to a greater or lesser degree pressive symptoms (see Figure 2). When measured with the in most people; however, more research using diverse pop- BDI, U.S. undergraduates (M = 8.41, SD = 7.78) reported sig- ulations from around the world is needed before making nificantly lower levels of depressive symptoms than Hon- definitive conclusions. Tests of normality for the PIQ also duran young adults (M = 13.97, SD = 10.93; t = 3.33, p tukey confirmed a normal distribution with one exception. The =.005), Western hemisphere adults (M = 12.51, SD = 9.73; Shapiro-Wilk test for the PIQ was significant (p = .02) for t = 3.06, p =.01), and Black U.S. adults (M = 11.95, SD tukey the Western adult sample, which indicates non-normality. = 10.98; t = 2.73, p =.03). When measured with anhe- tukey Although the distribution of scores were highly similar donic scale on the MASQ, Western adults (M = 67.20, SD = for all five samples, overall levels of cognitive vulnerability 18.52) had significantly greater levels of depressive symp- varied by population (i.e., the distribution shifted). As toms than U.S. undergraduates (M = 58.50, SD = 16.87; t = shown in Figure 2, cognitive vulnerability scores, as mea- 3.49, p =.003). There were no other significant differ- tukey sured by the CSQ, were significantly lower in the Honduran ences among the four samples (Nepali adults [M = 61.13, SD sample (M = 3.44, SD = .93) than in Western adults (M = = 13.95; Black U.S. adults [M = 64.53, SD = 20.11]). 4.12, SD = .92; t = -4.22, p < .001), Black U.S. adults tukey Hypothesis 3: Cognitive Vulnerability Will Be (M = 3.85, SD = 1.02; t = -2.61, p =.046), and U.S. un- tukey dergraduates (M = 4.28, SD = .84; t =-5.11, p < .001). Correlated with Depressive Symptoms tukey CSQ scores in undergraduates were not significantly dif- The strength of the association between cognitive vul- ferent than those of Western hemisphere adults (t = -1.05, nerability and depressive symptoms was consistent across p =.72), but tended to be greater than those reported tukey Western adult, Black U.S. adult, and U.S. undergraduate by Black U.S. Adults (t = 3.23, p =.007). Western hemi- tukey samples (i.e., WEIRD participants; See Figure 3). For these sphere adults and Black U.S. Adults did not differ on CSQ three populations, we found significant modest correlations scores (t = 2.12, p = .15) tukey between the BDI and CSQ ranging from .53 to .59 (cor- Collabra: Psychology 6 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression Figure 2. Box plots (means) and scatter plots (participants scores) for cognitive vulnerability and depressive symptoms as a function of study. relations ranged from .48-.49 when measuring depressive Similarly, the moderate association between cognitive vul- symptoms using the Anhedonic subscale of the MASQ; nerability and depressive symptoms found in college stu- tests of significance between correlations found no differ- dents did not replicate in the Nepali adult sample. There ences for the three samples, all p values > .29). was not a significant association between the PIQ and an- There was greater variation in the association between hedonic subscale of the MASQ (r = -.11, p = .52; this value cognitive vulnerability and depressive symptoms when cog- was significantly weaker than found in the WEIRD samples nitive vulnerability was measured using the PIQ. The corre- ( p <.001 for all correlation comparisons). lations ranged from .31 (Western adults) to .54 (Black U.S. Discussion adults) with most of the correlations hovering around .40. The unusually low correlation in Western adults (.31) was The people who might benefit most from psychological partially due to an outlier, which when removed, increased science (and from whom we might learn the most) are the the correlation to .37. The size of the correlations found least likely to be considered in our theories and studies. The in these three populations were similar to those found in purpose of the current research was to add to the 3-6% of prior research with U.S. college students (correlations typ- published research articles on mental health that include ically r = ~.4; Haeffel et al., 2008) and not significantly dif- participants from low and middle-income countries (Patel ferent from one another (tests of significance between cor- & Sumathipala, 2001), and to illustrate how tests of gener- relations found all p values > .29). alizability can advance science. Specifically, we tested the In contrast, the association between cognitive vulnera- generalizability of hopelessness theory’s cognitive vulner- bility and depressive symptoms did not generalize to Hon- ability hypothesis in five unique samples from around the duran young adults and Nepali adults (non-WEIRD sam- world. ples). In the Honduran young adults, the correlation The results corroborated some aspects of hopelessness between cognitive vulnerability (CSQ) and depressive theory, but also revealed aspects of the theory that may symptoms (BDI) was .30, which was weaker (but not sta- not generalize. In support of the theory, results showed tistically significantly different; p = .11) than that found in that individual differences in cognitive vulnerability could the three WEIRD populations (in which correlations ranged be measured reliably (with the CSQ) in diverse people sam- from .53 to .59). Moreover, it appears that the small, but ples. Indicators of internal reliability (alpha and McDon- significant correlation in Honduran young adults, was due ald’s omega) were similar in strength as those reported in to a single outlier (see Figure 3). The correlation between prior studies using college samples (Haeffel et al., 2007, cognitive vulnerability and depressive symptoms was no 2008). This finding suggests that the Spanish translation of longer significant (r = .12, p = .42) when the outlier was re- the CSQ can be used to measure individual differences in moved from the data; the correlation without the outlier this construct. Notably, the Spanish translation of the BDI was significantly weaker in strength than the correlation in the other samples (p <.01 for all correlation comparisons). Collabra: Psychology 7 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression Figure 3. Correlations among measures cognitive vulnerability and depressive symptoms as a function of study. * p <.05, **p<.01, ***p<.001 and Nepali translation of the MASQ also exhibited strong lower) yet reported twice the level of depressive symptoms. psychometric properties. This is the opposite of what one would predict based on the Further, the distributions of vulnerability scores were cognitive vulnerability hypothesis. Further, the small sig- highly similar in all five populations. Cognitive vulnerabil- nificant correlation between cognitive vulnerability and de- ity was a normally distributed continuous construct present pressive symptoms found in the Honduran group was no to a greater or lesser degree in all five samples. In other longer significant when removing an outlier. Similarly, in words, generating inferences about the stability and glob- the Nepali adults, there was not a significant correlation ality of life stress was something that all participants were between cognitive vulnerability and depressive symptoms. able to do. This work is consistent with recent theorizing Although Nepali adults had among the highest levels of (Haeffel & Kaschak, 2019) proposing that cognitive vulner- cognitive vulnerability, they did not report higher levels of ability can be thought of as one’s native language for inter- depressive symptoms. preting life stress. These results “disagree” with prior research testing However, results also revealed important differences hopelessness theory’s cognitive vulnerability hypothesis among samples. The association between cognitive vulner- (Popper, 1959), and suggest that the generalizability of the ability and depressive symptoms did not generalize. Cor- theory and or the measurement of its constructs may be roborating prior research, there was a moderately sized (r limited to specific populations. In other words, it may not = ~.4 - .5) positive correlation between these constructs be possible to simply translate existing measures in this in samples of Western adults, Black U.S. adults, and U.S. area and expect to find the same results in more diverse undergraduates that was relatively robust to differences in populations. The most parsimonious explanation for the measurement (e.g., CSQ vs. PIQ, BDI vs. MASQ). However, current findings is that the cognitive vulnerability hypoth- the association between cognitive vulnerability and depres- esis does not hold in these populations. In other words, the sive symptoms did not hold in Honduran young adults and theory only applies to specific cultures or contexts. How- Nepali adults. It is noteworthy that Honduran young adults ever, alternative explanations (e.g., methodological differ- reported significantly lower levels of cognitive vulnerability ences such as those discussed below) must be ruled out be- than U.S. undergraduates (about .5 of a standard deviation fore concluding that cognitive vulnerability to depression Collabra: Psychology 8 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression (as conceptualized in the hopelessness theory) is largely a lations, translating and testing commonly used measures, Western phenomenon. etc.; Cumming, 2014; Sakaluk, 2016). One explanation for the discrepant results is with regard This research also had limitations, the most prominent to the role of stress. The cognitive vulnerability hypothesis being the use of a cross-sectional design and small sample follows a diathesis-stress model (Abramson et al., 1989), sizes in two of the studies. The cross-sectional design did which predicts that vulnerability and stress combine to pre- not allow for conclusions about temporal precedence. Thus, dict depression. Typically, this interaction is assumed to be it remains possible that the association between cognitive synergistic whereby high levels of cognitive vulnerability vulnerability and depressive symptoms could emerge over a combine with high levels of stress to predict depression prospective longitudinal time frame (although researchers (Lewinsohn et al., 2001). The synergistic interaction pat- would still need to explain why the cross-sectional associ- tern cannot explain why Honduran young adults with low ations do not appear to generalize to all populations). Re- levels of cognitive vulnerability reported high levels of de- latedly, we did not include a measure of stress, which is pressive symptoms as they should be highly resistant to needed to determine if the titration model can account for depression. However, a titration model (Abramson et al., the non-replications found in our research. Determining 1989, 1997) of vulnerability-stress interactions can explain the pattern of the vulnerability-stress interaction will be an these findings. In the titration model, even low “doses” of important avenue for future research testing the generaliz- stress are sufficient to trigger depression in highly vulnera- ability of hopelessness theory. ble individuals whereas high doses of stress can trigger de- The sample sizes for the Honduran and Nepali studies pression in both low and highly vulnerable individuals. This were small and not powered to detect correlations smaller means that under conditions of high stress, cognitive vul- than those reported in prior work. This underscores the dif- nerability should have a weak association with depression ficulty in doing generalizability research; in our case, this symptoms because everyone is at risk. But, under condi- included short time frames for assessment and the avail- tions of low stress, only cognitively vulnerable individuals ability of people on-site who could help with the trans- should be at risk for depression. lation of measures and answering participant questions. The titration framework can help explain the weak as- It was not possible to run 300 Honduran young adults or sociation between cognitive vulnerability and depressive Nepali women because our partnership sites did not even symptoms in Honduran young adults. Despite having rela- have access to that many people. That said, the analyses tively low levels of cognitive vulnerability, Honduran young were largely focused on the distribution of scores and a adults still had high levels of depressive symptoms. This simple bivariate correlation rather than more sophisticated makes sense if Honduran young adults in an orphanage set- analyses requiring large sample sizes for null-hypothesis ting were also experiencing high levels of stress as both vul- testing. Prior to running the current set of studies, we had nerable and non-vulnerable individuals would be at risk for not given a great deal of thought to how to test theories depression. without searching for p < .05. However, alternative options The lack of an association between cognitive vulnerabil- exist. For example, Sakaluk (2016) argues that science ity and depressive symptoms (see scatter plot in Figure 3) should “explore small, confirm big.” According to this ap- in the Nepali sample is more difficult to understand. The proach, scientists start with small scale studies that test the Nepali sample had significantly higher levels of cognitive hypothesis of interest, but then work to corroborate those vulnerability (as measured by the PIQ) than the other popu- effects with larger data sets. Similarly, Lakens (2021) ar- lations yet did not report greater levels of depressive symp- gues that underpowered studies can still provide informa- toms. According to titration model (and synergistic model) tion about potential effect sizes. The caveat to these other this is possible if participants were not experiencing stress approaches is that they require a strong theory (e.g., highly or had a reduction in stress. In this specific case, the ser- specific hypotheses), an established knowledge base, and vices provided by Koshish (mental health treatment and valid and reliable measures. In our case, the hopelessness housing) may have resulted in significant stress relief. theory and its accompanying measure, the CSQ, have been This study had both strengths and limitations. A major tested extensively in the literature (Liu et al., 2015). The strength of the current research is the use of five unique cognitive vulnerability hypothesis has been tested cross- samples, including both WEIRD and non-WEIRD partici- sectionally (Just et al., 2001; Oliver et al., 2007), longitu- pants. Further, we demonstrated that it is possible to create dinally (e.g., Hankin et al., 2004; Lakdawalla et al., 2007; translated measures of cognitive vulnerability (e.g., Span- Russell et al., 2014), and in behavioral high-risk designs ish version of the CSQ; Nepali version of the PIQ) that have (e.g., Abramson et al., 1999; Alloy et al., 2000, 2006). There similar psychometric properties as their English counter- are decades of research (see reviews by Haeffel et al., 2008 parts. Finally, this work shows that it is possible to falsify and Liu et al., 2015) showing how cognitive vulnerability hypotheses and contribute new knowledge without relying behaves; we know it is normally distributed, has moderate solely on null hypothesis testing. Rejecting the null is nei- cross-sectional correlations with depressive symptoms, ther necessary nor sufficient for acquiring new knowledge predicts depressive symptoms longitudinally, and has (Meehl, 1978). Good theories lend themselves to a variety strong temporal stability (see Haeffel et al., 2008 for re- of examinations that do not always require p<.05 (e.g., ex- view). The CSQ and mood measures have been used exten- amining distributions of scores, plotting bivariate corre- sively in past research, which means it is possible to com- Collabra: Psychology 9 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression pare the distributions of scores on the CSQ and correlation sults. The results also suggest that the tendency to gener- plots across studies. ate negative inferences about stress has different implica- Nonetheless, more research is needed before definitive tions for depression depending on culture. This means that conclusions can be made about the generalizability of researchers must now try to understand why negative cog- hopelessness theory’s cognitive vulnerability hypothesis. nitions confer risk for depression in some contexts but not Based on the current findings, we can conclude that “some- others. thing is going on” regarding the generalizability of the hopelessness theory. Our results did not replicate/general- ize to at least some non-Western populations when trans- Contributions lating existing measures and using the same basic admin- istration techniques. The question is “why?” Are the G.H., H.B, M.V.M contributed to conceptualization and discrepant results due to a problem with the theory or design. a problem with the methodology (i.e., confounds such as G.H., H.B, M.V.M contributed to acquisition of data. measurement issues [translated vs. English] or differences G.H., H.B, M.V.M, L.B. contributed to analysis and inter- in familiarity with answering questionnaires)? Future re- pretation of data. search will need to disentangle why and how the differences G.H., H.B, M.V.M, L.B. contributed to drafts and revisions between samples emerged. In the meantime, it still useful of the article. to publish these kinds of discrepant findings even if we do G.H. supervised and mentored the student authors on not fully understand why the differences arise. These dis- the manuscript. agreements make us question our theories as well as the All authors approved the submitted version for publica- generalizability of our methods. This two-step approach is tion. used in a variety of areas in science including psychother- apy research. There are hundreds of studies published every Acknowledgements year demonstrating that Treatment A is more efficacious than Treatment B; however, we often do not understand We would like to thank Nuestros Pequeños Hermanos In- why that is (or why a any specific treatment works). ternational and Koshish for allowing us to work with their That said, the obstacles to understanding non-replica- residents. We would also like to thank Andrea Alatorre for tions in diverse samples are not insignificant. It is often im- her comments on an earlier version of the manuscript. This possible to use the same exact methodologies used for con- research was presented as a poster at the 2021 Association venience samples. For example, our Nepali women were not for Psychological Science Conference (recipient of the RISE Prolific users. And it would make little sense to run West- Award and SSCP Global Mental Health Award). ern adults by sitting on the floor with them and the paper and pencil questionnaires. These kinds of methodological Funding Information differences may be inevitable. Thus, it may be necessary to take a multi-method approach in future research. For ex- The research was funded, in part, by UROP grants from ample, in this case, the next step is to gather further in- the University of Notre Dame. Funding for Study 4 was pro- formation (e.g., via qualitative interviews) about cultural vided by a Race and Ethnicity research grant from the Uni- norms and participants’ understanding of these constructs versity of Notre Dame. (both vulnerability and depressive symptoms). This infor- mation may provide clues about the relevance and under- Competing Interests standing of these constructs in specific cultures and how to We declare no conflicts of interest. better adapt the language of the measures to tap the con- structs of interest. Data Accessibility Statement In conclusion, there is now a spotlight on systemic racism and a growing choir of voices requesting greater di- Measures and participant data can be found on this pa- versity in psychological science (Lewis, 2021). Tests of gen- per’s OSF project page: https://osf.io/umg9p/ eralizability using diverse samples can help to solve this problem as well as lead to better theories. These results Submitted: August 05, 2022 PST, Accepted: February 14, 2023 show that we cannot simply “transport” the same measures PST and methods to different samples and expect the same re- This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International License (CCBY-4.0). View this license’s legal deed at http://creativecommons.org/licenses/by/4.0 and legal code at http://creativecom- mons.org/licenses/by/4.0/legalcode for more information. Collabra: Psychology 10 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression References Abramson, L. Y., Alloy, L. B., & Hogan, M. E. (1997). Central American Refugee Health Profiles. (2017). Centers Cognitive/Personality subtypes of depression: for Disease Control and Prevention. https://www.cd Theories in search of disorders. CTAR, 21, 247–265. c.gov/immigrantrefugeehealth/profiles/central-ameri can/index.html Abramson, L. Y., Alloy, L. B., Hogan, M. E., Whitehouse, W. G., Donovan, P., Rose, D. T., Panzarella, C., & Chahar Mahali, S., Beshai, S., Feeney, J. R., & Mishra, S. Raniere, D. (1999). Cognitive vulnerability to (2020). Associations of negative cognitions, depression: Theory and evidence. Journal of Cognitive emotional regulation, and depression symptoms Psychotherapy, 13(1), 5–20. https://doi.org/10.1891/0 across four continents: International support for the 889-8391.13.1.5 cognitive model of depression. BMC Psychiatry, 20(1). https://doi.org/10.1186/s12888-019-2423-x Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1989). Hopelessness depression: A theory-based subtype of Clancy, K. B. H., & Davis, J. L. (2019). Soylent Is People, depression. Psychological Review, 96(2), 358–372. http and WEIRD Is White: Biological Anthropology, s://doi.org/10.1037/0033-295x.96.2.358 Whiteness, and the Limits of the WEIRD. Annual Review of Anthropology, 48(1), 169–186. https://doi.or Alloy, L. B., Abramson, L. Y., Hogan, M. E., Whitehouse, g/10.1146/annurev-anthro-102218-011133 W. G., Rose, D. T., Robinson, M. S., Kim, R. S., & Lapkin, J. B. (2000). The Temple-Wisconsin Cognitive Clark, L. A., & Watson, D. (1991). Tripartite model of Vulnerability to Depression (CVD) Project: Lifetime anxiety and depression: Psychometric evidence and history of Axis I psychopathology in individuals at taxonomic implications. Journal of Abnormal high and low cognitive vulnerability to depression. Psychology, 100(3), 316–336. https://doi.org/10.1037/ Journal of Abnormal Psychology, 109(3), 403–418. http 0021-843x.100.3.316 s://doi.org/10.1037/0021-843x.109.3.403 Cohen, J. R., Hankin, B. L., Gibb, B. E., Hammen, C., Alloy, L. B., Abramson, L. Y., Tashman, N. A., Berrebbi, Hazel, N. A., Ma, D., Yao, S., Zhu, X. Z., & Abela, J. R. D. S., Hogan, M. E., Whitehouse, W. G., Crossfield, A. Z. (2013). Negative attachment cognitions and G., & Morocco, A. (2001). Developmental origins of emotional distress in mainland Chinese adolescents: cognitive vulnerability to depression: Parenting, a prospective multiwave test of vulnerability-stress cognitive, and inferential feedback styles of the and stress generation models. Journal of Clinical Child parents of individuals at high and low cognitive risk & Adolescent Psychology, 42(4), 531–544. https://doi.o for depression. Cognitive Therapy and Research, 25(4), rg/10.1080/15374416.2012.749787 397–423. https://doi.org/10.1023/a:1005534503148 Cumming, G. (2014). The new statistics: Why and how. Alloy, L. B., Abramson, L. Y., Whitehouse, W. G., Hogan, Psychological Science, 25(1), 7–29. https://doi.org/10.1 M. E., Panzarella, C., & Rose, D. T. (2006). 177/0956797613504966 Prospective incidence of first onsets and recurrences D’Augelli, A., & Hershberger, R. (1993). Lesbian, gay, of depression in individuals at high and low cognitive and bisexual youth in community settings: Personal risk for depression. Journal of Abnormal Psychology, challenges and mental health problems. American 115, 145–156. https://doi.org/10.1037/0021843X.11 Journal of Community Psychology, 21(4), 421–448. http 5.1.145 s://doi.org/10.1007/bf00942151 Auerbach, R. P., Eberhart, N. K., & Abela, R. Z. (2010). El Khoury, J., Kanj, R., Adam, L., Kanj, R., Hajaig, A. J., Cognitive vulnerability to depression in Canadian Haddad, F., & El Helou, R. C. (2021). The and Chinese adolescents. Journal of Abnormal Child contribution of authors from low- and middle- Psychology, 38(1), 57–68. https://doi.org/10.1007/s108 income countries to top-tier psychiatric literature. 02-009-9344-y European Science Editing, 47, 72187. https://doi.org/1 Baumgartner, L. M., Johnson-Bailey, J., Lund, C. L., & 0.3897/ese.2021.e72187 Colin, S. A. J. (2010). Racism and white privilege in Gallegos, J., Linan-Thompson, S., Stark, K., & adult education graduate programs: Admissions, Ruvalcaba, N. (2013). Preventing childhood anxiety retention, and curricula. New Directions for Adult and and depression: Testing the effectiveness of a Continuing Education, 2010(125), 27–40. https://doi.or schoolbased program in Mexico. Psicología Educativa, g/10.1002/ace.360 19(1), 37–44. https://doi.org/10.5093/ed2013a6 Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Garber, J., & Flynn, C. (2001). Vulnerability to Cognitive therapy of depression. Guilford. depression in childhood and adolescence. In R. E. Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Ingram & J. M. Price (Eds.), Vulnerability to Psychometric properties of the Beck Depression psychopathology: Risk across the lifespan (pp. Inventory: Twenty-five years of evaluation. Clinical 175–225). The Guilford Press. Psychology Review, 8(1), 77–100. https://doi.org/10.10 Gibbons, J. L. (2020). What we can learn about 16/0272-7358(88)90050-5 multiculturalism from Latin American psychology. In Beshai, S., Dobson, K. S., Adel, A., & Hanna, N. (2016). S. Safdar, C. Kwantes, & W. Friedlmeier (Eds.), Wiser A cross-cultural study of the cognitive model of world with multiculturalism: Proceedings from the 24th depression: Cognitive experiences converge between Congress of the International Association for Cross- Egypt and Canada. PLOS ONE, 11(3), e0150699. http Cultural Psychology. https://scholarworks.gvsu.edu/ia s://doi.org/10.1371/journal.pone.0150699 ccp_papers/282 Collabra: Psychology 11 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression Gildersleeve, R. E., Croom, N. N., & Vasquez, P. L. Hankin, B. L., Fraley, R. C., & Abela, J. R. Z. (2005). (2011). “Am I going crazy?!”: A Critical Race Analysis Daily Depression and Cognitions About Stress: of Doctoral Education. Equity & Excellence in Evidence for a Traitlike Depressogenic Cognitive Education, 44(1), 93–114. https://doi.org/10.1080/106 Style and the Prediction of Depressive Symptoms in a 65684.2011.539472 Prospective Daily Diary Study. Journal of Personality and Social Psychology, 88(4), 673–685. https://doi.org/ Giner-Sorolla, R. (2017). Expression of concern 10.1037/0022-3514.88.4.673 regarding Förster and Denzler (2012). Journal of Experimental Social Psychology, 73, 289–299. https://d Haque, M., & Kamal, Z. (2019). Challenges of oi.org/10.1016/j.jesp.2017.04.002 publishing in high impact medical journals from lower-and middle-income countries. Bangladesh Haeffel, G. J. (2011). After further deliberation: Journal of Medical Science, 18(4), 671–674. https://do Cognitive vulnerability predicts changes in event- i.org/10.3329/bjms.v18i4.42868 specific negative inferences for a poor midterm grade. Cognitive Therapy and Research, 35(4), 285–292. http Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The s://doi.org/10.1007/s10608-010-9298-y Weirdest People in the World? SSRN Electronic Journal. https://doi.org/10.2139/ssrn.1601785 Haeffel, G. J. (2022). Psychology needs to get tired of winning. Royal Society Open Science, 9(6), 220099. htt Hong, R. Y., Gwee, K., & Karia, M. (2006). The role of ps://doi.org/10.1098/rsos.220099 event-specific pessimistic inferences in the etiological chain of hopelessness depression. Haeffel, G. J., Abramson, L. Y., Brazy, P. C., Shah, J. Y., Personality and Individual Differences, 41(6), Teachman, B. A., & Nosek, B. A. (2007). Explicit and 1119–1129. https://doi.org/10.1016/j.paid.2006.04.01 implicit cognition: A preliminary test of a dual- process theory of cognitive vulnerability to depression. Behaviour Research and Therapy, 45(6), Just, N., Abramson, L. Y., & Alloy, L. B. (2001). Remitted 1155–1167. https://doi.org/10.1016/j.brat.2006.09.003 depression studies as tests of thecognitive vulnerability hypotheses of depression onset: A Haeffel, G. J., & Cobb, W. R. (2022). Tests of critique of conceptual analysis. Clinical Psychology generalizability can diversify psychology and improve Review, 21(1), 63–83. https://doi.org/10.1016/s0272-7 theories. Nature Reviews Psychology, 1(4), 186–187. ht 358(99)00035-5 tps://doi.org/10.1038/s44159-022-00039-x Knyazev, G. G., Kuznetsova, V. B., Savostyanov, A. N., & Haeffel, G. J., Gibb, B. E., Metalsky, G. I., Alloy, L. B., Dorosheva, E. A. (2017). Does collectivism act as a Abramson, L. Y., Hankin, B. L., Joiner, T. E., Jr., & protective factor for depression in Russia? Personality Swendsen, J. D. (2008). Measuring cognitive and Individual Differences, 108, 26–31. https://doi.org/ vulnerability to depression: Development and 10.1016/j.paid.2016.11.066 validation of the cognitive style questionnaire. Clinical Psychology Review, 28(5), 824–836. https://do Lakdawalla, Z., Hankin, B. L., & Mermelstein, R. (2007). i.org/10.1016/j.cpr.2007.12.001 Cognitive theories of depression in children and adolescents: A conceptual and quantitative review. Haeffel, G. J., & Howard, G. S. (2010). Self-report: Clinical Child and Family Psychology Review, 10(1), Psychology’s four-letter word. American Journal of 1–24. https://doi.org/10.1007/s10567-006-0013-1 Psychology, 123(2), 181–188. https://doi.org/10.5406/ amerjpsyc.123.2.0181 Lakens, D. (2021). Sample size justification. https://doi.or g/10.31234/osf.io/9d3yf Haeffel, G. J., & Kaschak, M. P. (2019). Rethinking how we think about cognitive interventions for Ledford, H. (2010). Harvard probe kept under wraps: depression: An example from research on second- researchers call for the release of findings of the Marc language acquisition. Clinical Psychological Science, Hauser misconduct investigation. Nature, 466(7309), 7(1), 68–76. https://doi.org/10.1177/21677026187 908–909. https://doi.org/10.1038/466908a Hall, S. E., Watson, T. S., Kellums, M. L., & Kimmel, J. Lewinsohn, P. M., Joiner, T. E., & Rohde, P. (2001). (2016). Mental health needs and resources in Nepal. Evaluation of cognitive diathesis-stress models in International Journal of Culture and Mental Health, predicting major depressive disorder in adolescents. 9(3), 278–284. https://doi.org/10.1080/17542863.201 Journal of Abnormal Psychology, 110(2), 203–215. http 6.1185135 s://doi.org/10.1037/0021-843x.110.2.203 Hankin, B. L., Abramson, L. Y., Miller, N., & Haeffel, G. Lewis, N. A., Jr. (2021). Can we achieve “equality” when J. (2004). Cognitive vulnerability-stress theories of we have different understandings of its meaning? depression: Examining affective specificity in the How contexts and identities shape the pursuit of prediction of depression versus anxiety in three egalitarian goals. Psychological Inquiry, 32(3), prospective studies. Cognitive Therapy and Research, 155–164. https://doi.org/10.1080/1047840x.2021.197 28(3), 309–345. https://doi.org/10.1023/b:cotr.000003 1441 1805.60529.0d Liu, R. T., Kleiman, E. M., Nestor, B. A., & Cheek, S. M. Hankin, B. L., Abramson, L. Y., Moffitt, T. E., Silva, P. A., (2015). The hopelessness theory of depression: A McGee, R., & Angell, K. E. (1998). Development of quarter century in review. Clinical Psychology: Science depression from preadolescence to young adulthood: and Practice, 22, 345–365. Emerging gender differences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107(1), 128–140. https://doi.org/10.1037/0021-843x.1 07.1.128 Collabra: Psychology 12 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression Luitel, N. P., Jordans, M. J., Adhikari, A., Upadhaya, N., Peer, E., Brandimarte, L., Samat, S., & Acquisti, A. Hanlon, C., Lund, C., & Komproe, I. H. (2015). Mental (2017). Beyond the Turk: Alternative platforms for health care in Nepal: current situation and crowdsourcing behavioral research. Journal of challenges for development of a district mental Experimental Social Psychology, 70, 153–163. https://d health care plan. Conflict and Health, 9(1), 3. https://d oi.org/10.1016/j.jesp.2017.01.006 oi.org/10.1186/s13031-014-0030-5 Peer, E., Rothschild, D., Gordon, A., Evernden, Z., & Meadon, M., & Spurrett, D. (2010). It’s not just the Damer, E. (2022). Data quality of platforms and subjects – there are too many WEIRD researchers. panels for online behavioral research. Behavior Behavioral and Brain Sciences, 33(2–3), 104–105. http Research Methods, 54(4), 1643–1662. https://doi.org/1 s://doi.org/10.1017/s0140525x10000208 0.3758/s13428-021-01694-3 Medin, D. L., & Lee, C. D. (2012). Diversity Makes Better Peterson, C., & Seligman, M. E. (1984). Causal Science. Association for Psychological Science explanations as a risk factor for depression: Theory Presidential Column. and evidence. Psychological Review, 91(3), 347–374. ht tps://doi.org/10.1037/0033-295x.91.3.347 Meehl, P. E. (1978). Theoretical risks and tabular asterisks: Sir Karl, Sir Ronald, and the slow progress Plaut, V. C. (2010). Diversity Science: Why and How of soft psychology. Journal of Consulting and Clinical Difference Makes a Difference. Psychological Inquiry, Psychology, 46(4), 806–834. https://doi.org/10.1037/0 21(2), 77–99. https://doi.org/10.1080/1047840100367 022-006x.46.4.806 6501 Metalsky, G. I., Halberstadt, L. J., & Abramson, L. Y. Popper, K. R. (1959). The Logic of Scientific Discovery. (1987). Vulnerability to depressive mood reactions: Hutchinson of London. https://doi.org/10.1063/1.306 Toward a more powerful test of the diathesis–stress 0577 and causal mediation components of the Rad, M. S., Martingano, A. J., & Ginges, J. (2018). reformulated theory of depression. Journal of Toward a psychology of Homo sapiens: Making Personality and Social Psychology, 52(2), 386–393. http psychological science more representative of the s://doi.org/10.1037/0022-3514.52.2.386 human population. Proceedings of the National Mezulis, A. H., Hyde, J. S., & Abramson, L. Y. (2006). Academy of Sciences, 115(45), 11401–11405. https://d The developmental origins of cognitive vulnerability oi.org/10.1073/pnas.1721165115 to depression: Temperament, parenting, and negative Romens, S. E., Abramson, L. Y., & Alloy, L. B. (2008). life events in childhood as contributors to negative High and Low Cognitive Risk For Depression: cognitive style. Developmental Psychology, 42(6), Stability From Late Adolescence to Early Adulthood. 1012–1025. https://doi.org/10.1037/0012-1649.42.6.1 Cognitive Therapy and Research, 33(5), 480–498. http s://doi.org/10.1007/s10608-008-9219-5 Murray, B. (2002). Research fraud needn’t happen at all. Russell, A., Haeffel, G. J., Hankin, B. L., Maxwell, S. E., APA Monitor. http://www.apa.org/monitor/feb02/frau & Perera, R. A. (2014). Moving beyond main effects: d.aspx A data analytic strategy for testing complex theories National Institute of Mental Health. (2017). U.S. of clinical phenomena. Clinical Psychology: Science Department of Health and Human Services. https://w and Practice, 21(4), 385–397. https://doi.org/10.1111/ ww.nimh.nih.gov/health/topics/depression/index.sht cpsp.12083 ml Sakaluk, J. K. (2016). Exploring small, confirming big: Nolen-Hoeksema, S., & Morrow, J. (1991). A prospective An alternative system to The New Statistics for study of depression and posttraumatic stress advancing cumulative and replicable psychological symptoms after a natural disaster: The 1989 Loma research. Journal of Experimental Social Psychology, 66, Prieta earthquake. Journal of Personality and Social 47–54. https://doi.org/10.1016/j.jesp.2015.09.013 Psychology, 61(1), 115–121. https://doi.org/10.1037/0 Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011). 022-3514.61.1.115 False-Positive Psychology: Undisclosed Flexibility in Oliver, J. M., Murphy, S. L., Ferland, D. R., & Ross, M. J. Data Collection and Analysis Allows Presenting (2007). Contributions of the cognitive style Anything as Significant. Psychological Science, 22(11), questionnaire and the dysfunctional attitude scale to 1359–1366. https://doi.org/10.1177/09567976114176 measuring cognitive vulnerability to depression. 32 Cognitive Therapy and Research, 31(1), 51–69. http Stark, K. D., Schmidt, K. L., & Joiner, T. E. Jr. (1996). s://doi.org/10.1007/s10608-006-9067-0 Cognitive triad: Relationship to depressive Open Science Collaboration. (2015). Estimating the symptoms, parents’ cognitive triad, and perceived reproducibility of psychological science. Science, 349. parental messages. Journal of Abnormal Child https://doi.org/10.1126/science.aac4716pmid:263154 Psychology, 24(5), 615–631. https://doi.org/10.1007/bf 43 01670103 Palan, S., & Schitter, C. (2018). Prolific.ac – A subject Swendsen, J. D. (1997). Anxiety, depression, and their pool for online experiments. Journal of Behavioral and comorbidity: An experience sampling test of the Experimental Finance, 17, 22–27. https://doi.org/10.10 helplessness-hopelessness theory. Cognitive Therapy 16/j.jbef.2017.12.2004 and Research, 21(1), 97–114. https://doi.org/10.1023/ a:1021872410824 Patel, V., & Sumathipala, A. (2001). International representation in psychiatric journals: A survey of 6 leading journals. British Journal of Psychiatry, 168(1), 406–409. https://doi.org/10.1037/amp0000622 Collabra: Psychology 13 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression Thalmayer, A. G., Toscanelli, C., & Arnett, J. J. (2021). Watson, D., Clark, L. A., Weber, K., Assenheimer, J. S., The neglected 95% revisited: Is American psychology Strauss, M. E., & McCormick, R. A. (1995). Testing a becoming less American? American Psychologist, tripartite model: II. Exploring the symptom structure 76(1), 116–129. https://doi.org/10.1037/amp0000622 of anxiety and depression in student, adult, and patient samples. Journal of Abnormal Psychology, Thomas, J., & Altareb, B. (2012). Cognitive vulnerability 104(1), 15–25. https://doi.org/10.1037/0021-843x.10 to depression: an exploration of dysfunctional 4.1.15 attitudes and ruminative response styles in the United Arab Emirates. Psychology and Psychotherapy, World Health Organization. (2017). https://www.who.in 85(1), 117–121. https://doi.org/10.1111/j.2044-8341.2 t/news-room/fact-sheets/detail/depression 011.02015.x VandenBos, G. R., & Winkler, J. M. (2015). An analysis of the status of journals and research in psychology from Latin America. Psicologia: Reflexão e Crítica, 28(Suppl 1), 82–93. https://doi.org/10.1590/1678-715 3.20152840012 Collabra: Psychology 14 What Diverse Samples Can Teach Us About Cognitive Vulnerability to Depression Supplementary Materials Peer Review Communication Download: https://collabra.scholasticahq.com/article/71346-what-diverse-samples-can-teach-us-about-cognitive- vulnerability-to-depression/attachment/148671.docx?auth_token=i-l3gY4EFz5CkhNFms_j Collabra: Psychology
Journal
Collabra Psychology – University of California Press
Published: Mar 1, 2023
Keywords: cognitive vulnerability; hopelessness theory of depression; Nepal; Honduras; Black; generalizability; theory building