A Nudge in a New Direction: Integrating Behavioral Economic Strategies Into Suicide Prevention Work

Abstract

Three commonly used behavioral economic strategies were tested to investigate their utility in suicide prevention and mental health initiatives. Study 1 used a social norms nudge to potentially increase the people who accessed an online suicide risk factor intervention via e-mail (N = 14,792). E-mails containing the social norm nudge were 164% more likely to click on the link relative to those who received the e-mail without the nudge. Study 2 used item count technique to better estimate suicidal ideation compared with direct questioning methods endorsed by two groups of online participants (N = 787). No difference between groups was found. Study 3 used framing techniques to understand if participants (N = 787) were more likely to access online coping skills when framed as being able to help others who may go through a suicidal crisis rather than themselves. Findings indicated more participants accessed the coping skills when framed as having utility for helping other people going through a suicidal crisis.

Keywords

behavioral economics nudge item count technique suicide

Suicide prevention faces challenges such as low treatment-seeking rates, questionable accuracy from self-disclosure methods, and effectively disseminating evidenced-based skills and practices to the public (Anestis & Green, 2015; King et al., 2015; Luoma, Martin, & Pearson, 2002). Strategies, foundational concepts, and techniques often used in behavioral economics could be helpful in addressing these important problems. Behavioral economics has yielded many techniques and insights for how to change behavior. This field integrates research from economics (Thaler & Sunstein, 2008), social psychology (Cialdini, 2003), and cognitive psychology (Tversky & Kahneman, 1981). The current study is the first, to our knowledge, to integrate behavioral economic concepts into mental health initiatives and suicide prevention messaging. We investigate the utility of extremely cost- and time-effective behavioral economic strategies that have the ability to be widely implemented into a number of different research and clinical settings related to suicide prevention work.

Theoretical Background and Current Studies

Nudges

Behavioral economics takes psychological phenomena and behavioral assumptions into account (e.g., cognitive biases) when attempting to understand and improve human economic and decision-making behaviors. This fundamental shift in perspective has produced improvements in helping increase desired behaviors by anticipating common cognitive and emotional responses that affect human decision making (Thaler & Benartzi, 2004). One technique called nudges has been successful in providing highly scalable, cost-effective interventions that can alter decision-making outcomes. Nudges have had positive impacts in areas such as increasing retirement savings (Carroll, Choi, Laibson, Madrian, & Metrick, 2009), energy conservation (Allcott, 2011), and influenza vaccinations (Milkman, Beshears, Choi, Laibson, & Madrian, 2011). Despite the impact that nudges have had in other fields, mental health—and suicide prevention in particular—has not utilized these techniques to help overcome long-standing challenges.

One well-known nudge is using social norms (e.g., injunctive norms, descriptive norms) to help correct misinformation to influence attitude and behavior change (Borsari & Carey, 2003; Cialdini, 2003). Social norms are thought to have a powerful impact on attitudes and behavior because humans have a strong desire to be liked by others (Cialdini & Goldstein, 2004). Because of these norms, individuals often strive not to deviate too far from how others act or what others perceive to be appropriate. Social norms have been seen to predict behavior patterns such as drinking among fraternity and sorority members, with those who perceived people drinking more (descriptive norm) as well as those who perceived greater acceptability around drinking (injunctive norm) predicting baseline and future drinking behavior (Larimer, Turner, Mallett, & Geisner, 2004). Although using social norms nudges experimentally (i.e., using social norms to influence behavior) has not been used in suicide research, one study found that people holding elevated attempt and ideation descriptive norms correlated with higher rates of suicide-related thoughts and behaviors (STBs) among adolescents (Reyes-Portillo, Lake, Kleinman, & Gould, 2018). This finding is similar to alcohol studies in that people who have exaggerated views regarding the frequency of such actions/behaviors are more likely to experience them. Because displaying social norms nudges has been seen to affect future behaviors (Perkins, 2002), it is plausible that using social norms could increase help-seeking behaviors for STBs.

Item count technique

Because of the strong need and desire to belong and feel accepted by others (Baumeister & Leary, 1995), cognitive biases (e.g., social desirability bias; Krumpal, 2013) and fear of stigma (Barney, Griffiths, Jorm, & Christensen, 2006) can often interfere with accurate and truthful reporting of socially undesirable behaviors. Item count technique (ICT; Droitcour et al., 1991)—also referred to as unmatched count technique and list randomization—is routinely used in economics but has not been used in psychopathology research. ICT randomizes two groups to answer a short list of statements and asks people to report how many, but not which, statements are true. The first group receives several (three to seven usually) non-key items with low variance (e.g., “I have been upset, angry, or sad at least once in the past year”), and the other group receives the exact same questions and one additional sensitive question (e.g., “Have you ever thought about killing yourself?”). Subtracting the means for the true statements can give a more accurate account of prevalence than direct reporting. One meta-analysis found that in 48 comparisons, 63% of the estimates for socially undesirable behavior were significantly larger when elicited through ICT (Holbrook & Krosnick, 2010). This technique has been used with multiple sensitive attitudes and behaviors, such as shoplifting (Tsuchiya, Hirai, & Ono, 2007), anti-gay hate crimes (Rayburn, Earleywine, & Davidson, 2003b), hate victimization (Rayburn et al., 2003a), and sexual risk behaviors (LaBrie & Earleywine, 2000).

Framing strategies

Framing manipulations are used to present an option that highlights either positive or negative aspects that ultimately affect the attractiveness of the option (Tversky & Kahneman, 1981). Framing strategies have been used as nudges in numerous different studies involving health care (O’Connor, 1989), tax compliance (Hasseldine, Hite, James, & Toumi, 2007), and work performance (Hossain & List, 2012). Using framing strategies could improve the dissemination of suicide prevention information because engagement rates with routinely distributed online suicide prevention resources have been low (King et al., 2015). There are several reasons such materials might not be used, one being stigma surrounding STBs (Barney et al., 2006). Framing strategies targeting less stigmatized risk factors or disorders that are often associated with the main treatment goal have seen higher rates of engagement, for instance, targeting anxiety or mood disorders instead of substance use disorders (Blanco et al., 2008; Capron, Bauer, Madson, & Schmidt, 2018). For this reason, framing suicide prevention skills by highlighting less stigmatized affiliated risk factors, such as anxiety (Cougle, Keough, Riccardi, & Sachs-Ericsson, 2009), rather than STBs could help increase treatment-seeking behaviors. In addition, using framing strategies may help diminish the effects of optimism bias (i.e., the tendency for humans to underrate their chances of risk for themselves; Sharot, 2011) on treatment-seeking behaviors. As a byproduct, this technique may increase dissemination of prevention and resource materials to a wider audience, ultimately increasing effective peer support in the community. Suicide prevention organizations such as the National Strategy for Suicide Prevention and Suicide Prevention Resource Center have called for peer support interventions to play an increased role in suicide prevention efforts (Reed, 2013). For example, because people perceive that others are at higher risk for STBs and are more likely to need help, these individuals may be more willing to be exposed to an intervention for the purpose of helping others rather than themselves (Sharot, 2011). Framing dissemination in this way could increase the amount of people who access the information through bypassing stigmatization or guilt about accessing the information and circumventing optimism bias.

Aims and hypotheses

The following are three pilot studies aimed at implementing commonly used techniques in behavioral economic literature to mental health work and suicide prevention messaging. In Study 1, we hypothesized that individuals receiving a nudge in an e-mail informing them of a free online mental health intervention to reduce a commonly cited risk factor for suicide (anxiety; Cougle et al., 2009; Khan, Leventhal, Khan, & Brown, 2002) will access the intervention significantly more often than those who received a link and e-mail without the nudge. In Study 2, we hypothesized that individuals who receive indirect questioning method (ICT) would yield a larger estimate of past-year suicidal ideation (SI) compared with those directly asked about past-year SI. In Study 3, we hypothesized that individuals asked if they would be interested in learning skills for how to help others going through a suicidal crisis will click on a resource link significantly more often than those who are asked if they would be interested in learning skills to help themselves during a suicidal crisis.

Study 1

Method

Two campus-wide e-mails were sent to undergraduate students (N = 14,792) at a midsize university in the southern United States during Suicide Prevention Week asking them to participate in a free online mental health intervention and access a list of mental health resources. Most recent demographic information for this university stated that the average student age is 25.79 years old, mostly female (n = 9,344, 63.17%), and mainly identifies as White (n = 9,308, 62.93%) or Black (n = 4,033, 27.26%).

Around Suicide Prevention Week (mid-September), two separate nudge messages were sent within campus-wide e-mail mailouts. The messages were sent to undergraduate students in a sequential design to investigate possible differences in how wording could affect the amount of time an online intervention was accessed. These mailout e-mails are sent by the university at least once per week, every week, throughout the year and contain many links and blocks of information for various campus updates (e.g., power outages, weather warnings), activities (e.g., tickets for sporting events), and other miscellaneous information (e.g., research opportunities). For the intervention, one e-mail contained a social norms nudge (one injunctive norm, one descriptive norm) and offered students a link to a brief online intervention to treat a commonly cited risk factor for suicide (i.e., anxiety) as well as national and local resources for STBs (e.g., National Suicide Prevention Lifeline, campus mental health services). The second e-mail was sent two weeks later and did not contain a nudge but had a link to the same intervention and services. Please see Figure 1 for the exact wording of each e-mail. We chose to explicitly reference anxiety in the context of suicide directly because anxiety is a consistent correlate of suicide outcomes (Cougle et al., 2009), is more common than suicidal thoughts (American Psychiatric Association, 2013; Curtin, Warner, & Hedegaard, 2016), and may be less stigmatizing than STBs (Blanco et al., 2008). All study recruitment and procedures were approved by the first author’s Institutional Review Board.

Fig. 1.

Campus-wide e-mails for Study 1 on social norms nudges.

Results

Because of the methodological design, we were unable to ascertain how many people accessed the e-mails sent out. Within the entire sample, 91 participants of the 14,792 responded (< 1%) to the e-mail, but the data for the amount of students who read/opened the university mailout e-mail or read the nudge message specifically are unavailable. Results showed that individuals receiving the nudge were 164% more likely to click on the resource link when the e-mail contained the nudge wording compared with those who did not receive the e-mail with the nudge wording (n = 57 vs. 34). A z-test comparing the proportions of these two groups revealed that the two proportions were significantly different, z = 2.1, p = .033.

Study 2

Method

Study recruitment and procedures were approved by the first author’s Institutional Review Board. Participants were 800 adults in the United States recruited from a popular survey website (Amazon’s Mechanical Turk [Mturk], http://www.mechanicalturk.com; Buhrmester, Kwang, & Gosling, 2011) from which participants can receive compensation for engaging in research. The study was advertised as a research project seeking participants to complete a 5-minute survey regarding health history and behaviors. Participants were compensated $0.25 for completing this study. Qualifications to be eligible for this study were restricted to participants who were at least 18 years of age, lived within the United States, and had a completion rate of 95% for other MTurk surveys. Participants were given a link to one of two Qualtrics surveys and after giving consent, completed a brief online study.

A total of 13 participants were excluded from analyses for incorrectly answering a validation question (n = 13). The remaining 787 participants were randomized to receive direct assessment of past-year suicidal ideation (Group A; n = 395) or ICT assessment (Group B; n = 392). The remaining participants were mainly female (n = 500, 63.5%). The majority of participants self-identified as being White (n = 567, 72.0%), 6.9% as Asian (n = 54), 8.8% as Black (n = 69), 5.5% as Hispanic (n = 43), 4.0% as Native American, 3.4% coming from mixed parents (n = 27), and 2.9% as “Other” (n = 23). The age of the participants ranged from 18 to 76 years old, with a mean age of 36.84 (SD = 11.90) years old. The two groups did not differ statistically in age, race/ethnicity, or sex.

Indirect self-report measure

Item count technique

ICT is an indirect questioning procedure that is used to estimate the proportion of people engaging in a culturally sensitive behavior (Droitcour et al., 1991). This method asks one group of people a short list of low variance items that does not contain the sensitive behavior; a second group reports on the same questions as well as an additional item regarding a sensitive behavior (Droitcour et al., 1991; Karlan & Zinman, 2011). In this study, our sensitive behavior was past-year SI. The first survey group was asked to indicate how many items, but not which, are true for them from a list of four innocuous statements concerning health-related behaviors and history. These questions included the following four items: “In the last year, I have been upset, angry, or sad”; “In the past, I have been diagnosed with Hutchinson-Gilford Porgeria”; “I have access to some sort of medical care (e.g., hospital, dentist office, medical clinic)”; and “I have access to the Internet.” The second group received the same four statements, plus one additional statement regarding SI: “I have had thoughts of killing myself at least once in the past year.” The first group was directly asked this same question through a different measure later in the survey, whereas the second group was not (see Fig. 2).

Fig. 2.

Item count technique (ICT; Droitcour et al., 1991) questions for Group A and Group B in Study 2.

Direct self-report measure

Suicide Behaviors Questionnaire–Revised

Item 2 of the Suicide Behaviors Questionnaire–Revised (SBQ-R; Osman et al., 2001) was used as a direct measure of SI in the past year. This item assesses frequency of SI over the past 12 months from never to very often (5 or more times). Scores were dichotomized as representing no versus any level of SI in the past year to match that of the ICT statement.

Results

Results showed that in Group A, 30.4% (n = 395) reported SI within the past year from direct questioning via the SBQ-R. By subtracting the ICT mean of Group B (M = 2.929) from Group A (M = 3.232), it leaves an estimate of the proportion for how many people within the sample receiving the additional statement endorsed the sensitive behavior item (Group B = 30.3%). These results suggest that individuals had similar rates of SI in each and did not feel more comfortable endorsing with preference to any single method (z = .001, p = .992).

Study 3

Method

The same methodology and sample from Study 2 were used for Study 3. After answering the questions from Study 2, Group A (n = 395) was asked if they would be interested in clicking on a link that could provide them with resources and skills to help themselves when they were experiencing a suicidal crisis. It read: “Would you be interested in learning how you could help yourself navigate thoughts of suicide or times of distress? If so, please click on the following link.” Group B (n = 392) was given the exact same link but asked if they would be interesting in learning skills so that they could help others when they were experiencing a suicidal crisis. It read: “Would you be interested in learning how you could help friends, family members, co-workers, etc., navigate thoughts of suicide or times of distress? If so, please click on the following link.”

Results

Results showed that individuals receiving wording that asked them to help others in Group B (54 out of 392) were 167% (odds ratio = .63, 95% confidence interval = [.40, .98]) more likely to click on the resource link compared with those who were asked to help themselves navigate an emotional crisis in Group A (36 out of 395). A z-test comparing the proportions of these two groups revealed that the two proportions were significantly different, z = 1.98, p = .048. Descriptive statistics revealed that those receiving the wording for helping others stayed on the link page only slightly longer on average (M = 8.014 s vs. M = 7.801 s) but had a much wider range for staying on the link page (218.14 vs. 63.44). The demographic characteristics between those who clicked and those who did not were not statistically different in age, t(785) = −1.13, p = .26; sex, χ(1) = .06, p = .81; or race/ethnicity, χ(6) = 9.42, p = .15

Discussion

The current series of studies tested various messaging and assessment strategies rooted in behavioral economic theory to improve engagement with mental health and suicide prevention resources as well as suicide risk assessment. Study 1 demonstrated that the simple presentation of injunctive and descriptive norms about the social acceptability and efficacy of suicide-related professional treatment increased engagement with a risk factor associated with suicide (i.e., anxiety). Similarly, Study 3 demonstrated that the presentation of online suicide prevention resources as a means to help someone else who may be experiencing STBs increased engagement with these materials compared with when they were presented as a way to help oneself. Finally, Study 2 demonstrated that the utilization of ICT to assess the experience of recent SI may not increase accurate online assessment as predicted.

The promising results of Studies 1 and 3 have important potential implications for suicide prevention messaging to promote professional help seeking. Prior research suggests that many individuals who die by suicide do not seek mental health treatment within the months or even year of their death (Downs & Eisenberg, 2011; Luoma et al., 2002). Because organizations such as the American Foundation for Suicide Prevention have developed programs to facilitate treatment seeking in adults experiencing STBs (King et al., 2015), the incorporation of nudge messaging within programs may be a simple strategy for increasing engagement. Specifically, recruitment for these programs could incorporate messaging regarding how socially acceptable peers find treatment seeking and that treatment is often effective as a means to reduce barriers of stigma and concerns regarding treatment efficacy.

Although subjective and descriptive norm nudge messaging may increase engagement, it is certainly possible that not all individuals will be relieved of stigmatizing attitudes about suicide through this intervention. Study 3 demonstrated that individuals may be more likely to engage with online suicide prevention resources if its introduction is framed as a way of helping others who may experience STBs versus helping oneself. These results point to a potential way of not only increasing an individual’s engagement with suicide prevention resources as a means of increasing individual help seeking but also providing a way to increase peer support for others experiencing STBs. The use of other-oriented messaging to increase initial engagement with suicide prevention resources may be quite important because previous research indicates that few adults (8% of a total sample) offered free online suicide screening and suicide prevention resources actually engage with these services (King et al., 2015). The incorporation of other-oriented messaging may be a highly scalable, time-effective, and cost-effective way of engaging individuals with suicide prevention resources as a means to increase their support of peers struggling with STBs and can be tailored for any population (e.g., military).

Results from Study 2 were counter to our hypothesis that individuals in the ICT group would enhance online assessment efforts for past-year suicidal ideation estimates. Despite the null results showing no statistical difference between direct and indirect reporting methods, there are many important implications from the findings regarding ICT. Multiple studies have shown ICT to be an effective way to reduce social desirability bias and yield higher estimates of sensitive behaviors (e.g., sexual risk behaviors; LaBrie & Earleywine, 2000). In addition, ICT has outperformed other indirect measurements for sensitive behaviors (Coutts & Jann, 2008) and offers a simpler administration than these measurements (Droitcour et al., 1991). Because of the evidence suggesting that ICT is effective as well as literature showing that anonymous online questionnaires eliminate bias and give more accurate data, an important finding from our results is that the data gathered from routinely used anonymous online studies about suicide-related variables are likely truthful, accurate, and much less affected by social desirability bias. Given the similar percentages from the ICT technique and direct questioning method, we can also infer that like other sensitive behaviors, the ICT procedures are appropriate for use with STBs. Future studies might benefit from using ICT when accuracy of epidemiological data is paramount and individual suicide risk assessment is not the main focus of research. The ICT could provide more accurate incidence and prevalence rates and be especially powerful in populations, such as the military, in which there is a strong motivation to deny or conceal STBs (Anestis & Green, 2015) and/or individuals must report suicide-related information but are unable to remain anonymous in providing that data.

The implications for broad mental health and suicide prevention initiatives of the results of these three studies should be considered in tandem with study limitations. First, Studies 1 and 3 assessed engagement with suicide prevention resources as the primary dependent variable, not actual treatment seeking or engagement with professional services. Research suggests that increased engagement with online suicide prevention information is related to increases in treatment seeking (King et al., 2015); however, this study did not directly test this relationship. Without evidence that the nudges studied induce behavioral change in treatment seeking and engagement with mental health resources, it is possible using nudges has implications for increasing the visibility of suicide prevention messaging but not initiatives that directly prevent suicide. In addition, previous research demonstrates that individuals who have previously engaged with mental health care are more likely to demonstrate intent to seek treatment in the future (Niederkrotenthaler, Reidenberg, Till, & Gould, 2014). Without the assessment of this confound, study results cannot dismiss the possibility that changes in engagement with the suicide prevention link in Studies 1 and 3 were due to history of mental health treatment. Other similar confounds could include personally held stigma about suicide and suicide literacy (Batterham, Calear, & Christensen, 2013).

For Studies 2 and 3, there is the possibility of carryover effects (i.e., the participant may have felt less anonymous regarding his or history of suicide and thus felt reluctant to click on resources aimed at helping themselves). However, given the findings regarding ICT in Study 2, evidence suggests that people are equally likely to endorse suicide-related thoughts regardless of whether direct or indirect questioning is used in an online study. The ICT does not specify endorsed items (e.g., past-year ideation); therefore, we are unable to test whether these effects were present in our study. Regarding Study 1, only 91 students (< 1% of the total sample) from Study 1 clicked on the link for the intervention. However, the primary purpose of the social norms nudge was not to assess how well the intervention would be received in an undergraduate population but to show that adding a few extra words can significantly increase the return on a very brief and cost-effective investment. The assessment strategy employed in Study 2 to quantify the experience of recent SI may not be ideal as well. Research demonstrates that single item measurement of suicide-related constructs may overestimate clinical severity and prevalence (Millner, Lee, & Nock, 2015). Future research testing the use of ICT in the assessment of suicide risk would benefit from adopting both online and in-person assessment as well as the use of psychometrically validated semi-structured interview and self-report measures of STB severity and frequency.

Considerations regarding the sample representativeness should also be taken into account. Given the decision to conduct a campus-wide recruitment for Study 1, it is difficult to understand the representativeness of the study group and thus the generalizability of study results. In addition, because only the total amount of link clicks was collected per group, it is impossible with the current study design to understand who clicked on the link and who did not. The demographics of Studies 2 and 3 are generally comparable to U.S. demographics of citizens over the age of 18; however, our sample may have included an overrepresentation of those who identify with a female gender. Follow-up controlled investigations should seek to replicate and extend results in more representative samples.

In sum, results from three pilot studies suggest that behavioral economic strategies could be viable techniques for increasing treatment-seeking behavior and dissemination of suicide prevention skills and gathering more accurate prevalence rates. Future research will need to implement these techniques in different settings and populations to tailor the effectiveness of each strategy. The rationale for this future work is supported by the low-cost, highly scalable nature of the messaging initiatives employed and the simple incorporation of ICT in already established suicide risk assessment protocols. As experts call for highly scalable public health interventions to prevent suicide (World Health Organization, 2012), the current study provides initial support that short and inexpensive behavioral economic strategies could be beneficial components of these broader prevention strategies.

Footnotes

Action Editor

Kelly L. Klump served as action editor for this article.

Author Contributions

B. W. Bauer developed the study concept. B. W. Bauer, R. P. Tucker, and D. W. Capron, contributed to the study design. Testing and data collection were performed by B. W. Bauer B. W. Bauer. performed the data analyses and interpretation. B. W. Bauer and R. P. Tucker drafted the manuscript. D. W. Capron provided critical revisions. All the authors approved the final manuscript for submission.

ORCID iD

Brian W. Bauer

Declaration of Conflicting Interests

The author(s) declared that there were no conflicts of interest with respect to the authorship or the publication of this article.

References

Allcott

(2011). Social norms and energy conservation. Journal of Public Economics, 95, 1082–1095.

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author.

Anestis

M. D.

Green

B. A.

(2015). The impact of varying levels of confidentiality on disclosure of suicidal thoughts in a sample of United States National Guard personnel. Journal of Clinical Psychology, 71, 1023–1030.

Barney

L. J.

Griffiths

K. M.

Jorm

A. F.

Christensen

(2006). Stigma about depression and its impact on help-seeking intentions. Australian & New Zealand Journal of Psychiatry, 40, 51–54.

Batterham

P. J.

Calear

A. L.

Christensen

(2013). Correlates of suicide stigma and suicide literacy in the community. Suicide and Life-Threatening Behavior, 43, 406–417.

Baumeister

R. F.

Leary

M. R.

(1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497–529.

Blanco

Okuda

Wright

Hasin

D. S.

Grant

B. F.

Liu

S.-M.

Olfson

(2008). Mental health of college students and their non-college-attending peers: Results from the National Epidemiologic Study on Alcohol and Related Conditions. Archives of General Psychiatry, 65, 1429–1437. doi:10.1001/archpsyc.65.12.1429

Borsari

Carey

K. B.

(2003). Descriptive and injunctive norms in college drinking: A meta-analytic integration. Journal of Studies on Alcohol, 64, 331–341.

Buhrmester

Kwang

Gosling

S. D.

(2011). Amazon’s Mechanical Turk: A new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1), 3–5.

10.

Capron

D. W.

Bauer

B. W.

Madson

M. B.

Schmidt

N. B.

(2018). Treatment seeking among college students with comorbid hazardous drinking and elevated mood/anxiety symptoms. Substance Use & Misuse, 53, 1041–1050.

11.

Carroll

G. D.

Choi

J. J.

Laibson

Madrian

B. C.

Metrick

(2009). Optimal defaults and active decisions. The Quarterly Journal of Economics, 124, 1639–1674.

12.

Cialdini

R. B.

(2003). Crafting normative messages to protect the environment. Current Directions in Psychological Science, 12, 105–109.

13.

Cialdini

R. B.

Goldstein

N. J.

(2004). Social influence: Compliance and conformity. Annual Review of Psychology, 55, 591–621.

14.

Cougle

J. R.

Keough

M. E.

Riccardi

C. J.

Sachs-Ericsson

(2009). Anxiety disorders and suicidality in the National Comorbidity Survey-Replication. Journal of Psychiatric Research, 43, 825–829.

15.

Coutts

Jann

(2011). Sensitive questions in online surveys: Experimental results for the randomized response technique (RRT) and the unmatched count technique (UCT). Sociological Methods & Research, 40, 169–193.

16.

Curtin

S. C.

Warner

Hedegaard

(2016). Increase in suicide in the United States, 1999–2014. https://www.ncbi.nlm.nih.gov/pubmed/27111185 NCHS Data Brief, 241, 1–8.

17.

Downs

M. F.

Eisenberg

(2012). Help seeking and treatment use among suicidal college students. Journal of American College Health, 60, 104–114.

18.

Droitcour

Caspar

R. A.

Hubbard

M. L.

Parsley

T. L.

Visscher

Ezzati

T. M.

(1991). The item count technique as a method of indirect questioning: A review of its development and a case study application. In Biemer

P. P.

Groves

R. M.

Lyberg

L. E.

Mathiowetz

N. A.

Sudman

(Eds.), Measurement errors in surveys (pp. 185–210). New York, NY: Wiley.

19.

Hasseldine

Hite

James

Toumi

(2007). Persuasive communications: Tax compliance enforcement strategies for sole proprietors. Contemporary Accounting Research, 24(1), 171–194.

20.

Holbrook

A. L.

Krosnick

J. A.

(2010). Social desirability bias in voter turnout reports: Tests using the item count technique. Public Opinion Quarterly, 74, 37–67.

21.

Hossain

List

J. A.

(2012). The behavioralist visits the factory: Increasing productivity using simple framing manipulations. Management Science, 58, 2151–2167.

22.

Isaac

Elias

Katz

L. Y.

Belik

S. L.

Deane

F. P.

Enns

M. W.

. . . Swampy Cree Suicide Prevention Team (12 members) 8. (2009). Gatekeeper training as a preventative intervention for suicide: A systematic review. The Canadian Journal of Psychiatry, 54, 260–268.

23.

Karlan

D. S.

Zinman

(2012). ICT for sensitive behavior: An application for measuring use of loan proceeds. Journal of Development Economics, 98, 71–75.

24.

Khan

Leventhal

R. M.

Khan

Brown

W. A.

(2002). Suicide risk in patients with anxiety disorders: A meta-analysis of the FDA database. Journal of Affective Disorders, 68, 183–190.

25.

King

C. A.

Eisenberg

Zheng

Czyz

Kramer

Horwitz

Chermack

(2015). Online suicide risk screening and intervention with college students: A pilot randomized controlled trial. Journal of Consulting and Clinical Psychology, 83, 630–636.

26.

Krumpal

(2013). Determinants of social desirability bias in sensitive surveys: A literature review. Quality & Quantity, 47, 2025–2047.

27.

LaBrie

J. W.

Earleywine

(2000). Sexual risk behaviors and alcohol: Higher base rates revealed using the unmatched-count technique. Journal of Sex Research, 37, 321–326.

28.

Larimer

M. E.

Turner

A. P.

Mallett

K. A.

Geisner

I. M.

(2004). Predicting drinking behavior and alcohol-related problems among fraternity and sorority members: Examining the role of descriptive and injunctive norms. Psychology of Addictive Behaviors, 18, 203–212.

29.

Luoma

J. B.

Martin

C. E.

Pearson

J. L.

(2002). Contact with mental health and primary care providers before suicide: A review of the evidence. American Journal of Psychiatry, 159, 909–916.

30.

Milkman

K. L.

Beshears

Choi

J. J.

Laibson

Madrian

B. C.

(2011). Using implementation intentions prompts to enhance influenza vaccination rates. Proceedings of the National Academy of Sciences, 108, 10415–10420.

31.

Millner

A. J.

Lee

M. D.

Nock

M. K.

(2015). Single-item measurement of suicidal behaviors: Validity and consequences of misclassification. PLoS One, 10(10), Article e0141606.

32.

Niederkrotenthaler

Reidenberg

D. J.

Till

Gould

M. S.

(2014). Increasing help-seeking and referrals for individuals at risk for suicide by decreasing stigma: The role of mass media. American Journal of Preventive Medicine, 47, S235–S243.

33.

O’Connor

A. M

. (1989). Effects of framing and level of probability on patients’ preferences for cancer chemotherapy. Journal of Clinical Epidemiology, 42, 119–126.

34.

Osman

Bagge

C. L.

Gutierrez

P. M.

Konick

L. C.

Kopper

B. A.

Barrios

F. X.

(2001). The Suicidal Behaviors Questionnaire-Revised (SBQ-R): Validation with clinical and nonclinical samples. Assessment, 8, 443–454.

35.

Perkins

H. W.

(2002). Social norms and the prevention of alcohol misuse in collegiate contexts. Journal of Studies on Alcohol, 14, 164–172.

36.

Rayburn

N. R.

Earleywine

Davison

G. C.

(2003a). Base rates of hate crime victimization among college students. Journal of Interpersonal Violence, 18, 1209–1221.

37.

Rayburn

N. R.

Earleywine

Davison

G. C.

(2003b). An investigation of base rates of anti-gay hate crimes using the unmatched-count technique. Journal of Aggression, Maltreatment & Trauma, 6, 137–152.

38.

Reed

(2013). Advancing peer support in suicide prevention. Retrieved from https://www.sprc.org/news/advancing-peer-support-suicide-prevention

39.

Reyes-Portillo

J. A.

Lake

A. M.

Kleinman

Gould

M. S.

(2018). The relation between descriptive norms, suicide ideation, and suicide attempts among adolescents [Advance online publication]. Suicide and Life-Threatening Behavior. doi:10.1111/sltb.12446

40.

Sharot

(2011). The optimism bias. Current Biology, 21, R941–R945.

41.

Thaler

R. H.

Benartzi

(2004). Save more tomorrow: Using behavioral economics to increase employee saving. Journal of Political Economy, 112(S1), S164–S187.

42.

Thaler

R. T.

Sunstein

C. R.

(2008). Nudge: Improving decisions about health, wealth, and happiness. New Haven, CT: Yale University Press.

43.

Tsuchiya

Hirai

Ono

(2007). A study of the properties of the item count technique. Public Opinion Quarterly, 71, 253–272.

44.

Tversky

Kahneman

(1981). The framing of decisions and the psychology of choice. Science, 211, 453–458.

45.

World Health Organization. (2012). Public health action for the prevention of suicide: A framework. Retrieved from http://apps.who.int/iris/bitstream/handle/10665/75166/9789241503570_eng.pdf;jsessionid=88067E82104D4FE4A62CB08C923C9457?sequence=1