Detecting Applicant Faking With a Context-Specific Overclaiming Questionnaire
Abstract
Abstract: In the context of personnel selection, self-reports are often biased by social desirability. For example, applicants may overstate their knowledge to make a good impression on a potential employer. Overclaiming questionnaires (OCQs) offer a means to assess whether applicants claim to have knowledge that they do not have. Previous studies evaluating whether OCQs are capable of detecting faking in personnel selection contexts reported mixed results but did not take the fit between the content of OCQ items and the selection context into account. In the present study, we therefore tailored an OCQ to the specific application context and compared its performance to that of Residualized Individual Change Scores (RICS), a competing measure of faking based on an achievement motivation questionnaire. A total of 123 participants first answered the OCQ and the motivational questionnaire in a control condition without application context. The two measures were then completed again as part of a mock application process, and participants were asked to honestly report their faking behavior afterward. Participants exhibited more overclaiming in the application context than in the control condition. The OCQ and RICS scores predicted participants’ self-reported faking with comparable accuracy. These results suggest that OCQs can compete with other measures of faking if their content is appropriately tailored to the application context.
Social desirability bias is the tendency to respond in a way that will be regarded favorably by others rather than truthfully (e.g., Phillips & Clancy, 1972). This bias poses a serious threat to the validity of self-reports, especially in situations in which the respondent is being evaluated (Nederhof, 1985; Tourangeau & Yan, 2007). For instance, in personnel selection contexts, some applicants will likely overstate their ability or positive aspects of their personality to impress a potential employer (Klehe et al., 2012; Rosse et al., 1998; Roulin et al., 2016). To assess or reduce social desirability bias, several scales and methods have been proposed (e.g., Borkenau & Ostendorf, 1992; Musch et al., 2002; Nederhof, 1985; Paulhus, 1991; Stöber, 1999). However, social desirability scales have been shown to be of little – if any – use for measuring or controlling the influence of socially desirable responses in personnel selection contexts (e.g., Ellingson et al., 1999; Griffith & Peterson, 2008; Ones et al., 1996) as well as in low-demand situations (e.g., Borkenau & Ostendorf, 1992; Borkenau & Zaltauskas, 2009; Piedmont et al., 2000). In fact, these scales may (at least partly) assess trait variance of desirable characteristics such as honesty (e.g., Müller & Moshagen, 2019a; Zettler et al., 2015) rather than being indicators of response bias.
Overclaiming questionnaires (OCQs) appear to be a promising alternative that offer a means to capture self-enhancement by assessing the tendency to claim knowledge that actually does not exist (Paulhus et al., 2003). In OCQs, participants are presented with a list of items and asked to indicate how well they know each item, for example, on a 7-point scale. Unknown to the participants, some of these items are real (e.g., “culmination”) and some are made-up distractors that do not really exist (e.g., “antitagous”). Based on the principles of signal detection theory (Green & Swets, 1966; Paulhus et al., 2003), OCQs allow for the calculation of the OCQ bias, an index of an individual’s degree of self-enhancement. A more liberal response bias is therefore reflected in higher OCQ bias scores, which indicate a stronger tendency to overclaim.
One line of previous research has focused on placing OCQ bias scores into a nomological network. Early studies on overclaiming reported relationships with measures of self-enhancement such as impression management and narcissism (Paulhus et al., 2003; Randall & Fernandes, 1991). Also, Bensch et al. (2019) found overclaiming to be a distinct indicator of positive distortion in self-reports that is independent of other personality measures. In contrast, overclaiming was found to be widely independent of self-enhancement tendencies and faking ability (Goecke et al., 2020) as well as cheating in knowledge assessment (Steger et al., 2021). Similarly, Müller and Moshagen (2019a, 2019b) reported that OCQs are not capable of capturing self-serving response distortions. Instead, they argue that the cognitive processes of overclaiming might rather be automatic and unrelated to deliberate positive self-presentation, as they are similar to those underlying reconstruction biases in the hindsight bias (Müller & Moshagen, 2018).
Previous studies also investigated the usefulness of OCQs in personnel selection contexts. For example, Bing et al. (2011) reported that the prediction of study success by an ability and a personality test was improved when overclaiming was considered in a selection context. Specifically, overclaiming scores as a measure of faking could increase the validity of the personality test scores by suppressing unwanted error variance. Scores from an OCQ also allowed to differentiate between fake and real interviews about attitude, behavior, and personality (Kemper & Menold, 2014).
In a more critical evaluation, Feeney and Goffin (2015) compared an OCQ to Residualized Individual Change Scores (RICS; also known as Regression-Adjusted Difference Scores; Burns & Christiansen, 2011), the gold standard measure of faking according to the authors. In their study, RICS was established by repeatedly administering an extraversion questionnaire in two different contexts: an honest context and a scenario-based application context in which extraversion was a desired trait. The RICS measure was determined by regressing the extraversion score of each item in the application context on the extraversion score of the same item in the honest context. They took the regression residuals for all items and all individuals and calculated the RICS score for each participant as the average of the residuals across all items. Consequently, RICS is based on faking behavior operationalized as the variance in the application condition scores that could not be explained by the honest condition scores. Feeney and Goffin (2015) concluded that their OCQ performed worse than RICS at capturing applicant faking behavior because OCQ bias scores correlated significantly lower with self-reported faking tendencies. However, this finding must be qualified by the fact that the two competing measures differed greatly in their fit to the specific application context. In the study by Feeney and Goffin (2015), participants put themselves in the role of an applicant for a retail sales position. The extraversion questionnaire that was used for the RICS fit this application context well because extraversion is a desirable trait for a sales position (Vinchur et al., 1998). The OCQ, on the other hand, comprised items allegedly measuring general knowledge, which is hardly relevant for a retail sales position and can therefore be deemed a less desirable attribute in this context. Dunlop et al. (2020) have recently demonstrated that only an OCQ with job-relevant content as opposed to job-irrelevant content was indicative of faking in a job interview. Therefore, for a fair comparison of OCQ and RICS, the content-context fit must be taken into account. For example, Watts et al. (2021) tailored their OCQ to a sales job and reported that their OCQ was unrelated to future sales performance.
The central research question of the present study was: How well do OCQ scores versus RICS predict faking behavior when the content of both measures fits the application context well? To answer this question, we created OCQ and RICS instruments based on test content that was relevant to a fictitious scholarship application scenario. The RICS was based on an achievement motivation scale, and we expected that the scale (H1) as well as the OCQ scores (H2) would be elevated in the application context as compared to the honest context. The competing instruments were compared against an external criterion of faking behavior, following the approach by Feeney and Goffin (2015). We had no expectation as to which of the two methods would perform better and therefore posed the following undirected hypothesis: The predictive values of OCQ and RICS in predicting faking behavior differ (H3).
Method
Participants and Procedure
In total, 134 participants completed the experiment. For the analysis, 11 participants were excluded who had either failed the seriousness check (n = 2; Aust et al., 2013), failed the manipulation check (n = 8), or omitted a whole survey page (n = 1). In the final sample of 123 participants, 61% were female and 89% indicated German as their first language. Age was assessed as a categorical variable and was distributed as follows: 18–19 years: n = 8 (6.5%), 20–29 years: n = 72 (58.5%), 30–39 years: n = 12 (9.8%), 40–49 years: n = 4 (3.3%), 50–59 years: n = 19 (15.4%), 60–69 years: n = 6 (4.9%), ≥ 70 years: n = 2 (1.6%). As their highest educational attainment, 2 participants reported a lower secondary school leaving certificate (1.6%; referred to as “Volks-/Hauptschulabschluss” in the German educational system), 4 reported an intermediate secondary school leaving certificate (3.3%; “mittlere Reife”), 60 reported a high school diploma or specialized baccalaureate (48.8%; “Abitur” or “Fachabitur”), 51 reported a university degree (41.5%; “Studienabschluss”), and 6 reported a doctoral degree (4.9%; “Promotion”). Participants were recruited by students as part of a psychological assessment course on campus or in their private environment. Participants had to be at least 18 years old. Psychology students were excluded from participation to preclude prior knowledge of the concept of OCQs.
The study was conducted using an 11-page paper-pencil questionnaire under lab conditions, that is, the solutions could not be looked up on the Internet. After a short introduction, participants gave their informed consent. As an independent variable, we manipulated the context (honest vs. application) in which the competing instruments were administered within subjects. Similar to Feeney and Goffin (2015), participants were first instructed to answer as honestly and authentically as possible as part of a self-experience. The instructions read (the original German wording can be found in the Electronic Supplementary Material, ESM 1):
“On the following pages, we will ask you several questions. Please imagine now that you are participating in a self-experience study. You are eager to get accurate and unbiased feedback about your personality. The following tests are part of this self-experience. Answering these tests honestly is a prerequisite for your self-experience to be successful. So please answer the following questions as honestly and authentically as possible, and do not pretend.”
A manipulation check verified that participants read and understood the instructions by asking them to choose the correct statement from the following four statements: (1) “I am to present myself as negatively as possible in answering the questions on the following pages”; (2) “I am to respond to the questions on the following pages as if I were in a job application situation”; (3) “In answering the questions on the following pages, I am supposed to just tick anything without thinking about it” (4) (correct); “I am to respond to the questions on the following pages as they apply to me personally at the moment” (see ESM 1 for the original German wording). Eventually, participants completed an OCQ on knowledge about psychology, the Psychology Overclaiming Questionnaire (Psycho OCQ), and an achievement motivation test in the honest context. Then, the application scenario was introduced: Participants were asked to respond in a way that made a good impression and made them appear suitable for the highly endowed psychology scholarship. The instructions read (see ESM 1 for the original German wording):
“We will now present you with the same tests as before again. However, from now on, please imagine that you are studying psychology and applying for a highly endowed and prestigious scholarship. You desperately want to get this scholarship. The following tests are part of the application process. Doing well on these tests is a prerequisite for your application to be successful. So please answer the following questions in such a way that you have the best possible chance of receiving the scholarship.”
Again, participants completed the manipulation check before taking the Psycho OCQ and the achievement motivation test a second time. Residualized Individual Change Scores (RICS) were calculated based on the responses to the achievement motivation scale in the two different contexts. Both the Psycho OCQ and the achievement motivation test fit the scholarship application context well, and high scores were expected to be perceived as desirable for a successful application. After the application scenario ended, participants self-reported on their faking tendencies in the application context as an external criterion of faking behavior. Finally, participants were debriefed and thanked.
Measures
The Psychology Overclaiming Questionnaire (Psycho OCQ) was constructed to cover a wide range of knowledge about psychology and consisted of six domains: clinical psychology, general psychology, methodology, social psychology, biological psychology, and personality psychology. Each domain comprised six real items and two fake items, resulting in a total of 48 items. Items were identified with the help of psychology textbooks. Real items were selected such that participants without background knowledge of psychology would have a chance to classify them correctly, whereas, for experts in psychology, a correct classification would still be challenging (example for a real item: “Amygdala”). As lures, fictitious words were constructed that appear to belong to the field of psychology (example for a fake item: “Visnophobia”). Participants were asked to indicate how well they knew each item on a 7-point Likert scale from 1 = not at all to 7 = very well. The questionnaire is available from the authors upon request.
The Residualized Individual Change Score (RICS) was based on the German achievement motivation scale Leistungsmotivationsinventar-Kurzform (LMI-K; Schuler & Prochaska, 2001). The test covers a variety of dimensions of achievement motivation, for example, status orientation, confidence in success, commitment, and persistence. We assumed that participants would regard achievement motivation as a desirable trait in the application process and would indicate higher values in the application condition than in the honest condition, as was the case in numerous previous studies on faking in application contexts (e.g., Bing et al., 2011; Mueller-Hanson et al., 2003; Ziegler et al., 2007). It is common practice to select high-performing and achievement-motivated individuals as scholarship recipients. The LMI-K consists of 30 statements (e.g., “I get satisfaction from improving my own performance.”) and participants are asked to indicate their agreement with each statement on a 7-point scale ranging from 1 = complete rejection to 7 = complete agreement. For each item of the LMI-K, the score in the application context was regressed on the score in the honest context. The mean of the residuals from these regressions across all items yielded the RICS.
As an external criterion for faking behavior, we assessed participants’ faking tendencies in the application context using a 4-item scale adapted from Feeney and Goffin (2015). Participants reported their behavior on a 7-point scale from 1 = very infrequently to 7 = very frequently for the statements (1) “I improved my answers in the application situation”; (2) “In the application situation, I slightly understated my weaknesses”; (3) “In the application situation, I slightly exaggerated my strengths”; and (4) “I completely misrepresented myself in the application situation.” To form an external criterion of individual misrepresentation, the mean across all four items was calculated.
Results
For all analyses, a significance level of α = .05 was assumed. The internal consistency reliability of the scales was determined using McDonald’s ω (Hayes & Coutts, 2020). The reliability of the LMI-K in the honest and application context was ω = .92 and ω = .95, respectively. The reliability of the faking tendencies scale was ω = .81. We relied on signal detection theory to calculate the overclaiming index OCQ bias. When individuals indicate that they are familiar with a real item, it is counted as a hit (H), whereas claiming familiarity with a fake item is counted as a false alarm (FA). The location of the criterion, c = −.5 × [Φ−1(H) + Φ−1(FA)] (Stanislaw & Todorov, 1999; formula 7), reflects an individual’s threshold to respond that they are familiar with an item. Given that there is no clear cutoff for familiarity on the 7-point scale used in the present OCQ, we followed Paulhus et al. (2003) and computed indices for all possible cutoffs (i.e., ratings of 2, 3, 4, 5, 6, or 7 on the scale as reflecting familiarity) and averaged the resulting scores. As higher values of c are associated with less overclaiming, we reversed c to obtain an overclaiming measure (Paulhus et al., 2003), OCQ bias = −1 × c, such that higher values of the OCQ bias index reflected more overclaiming. In a similar way, a performance measure OCQ accuracy can be calculated as d’ = Φ−1(H) − Φ−1(FA) (Paulhus et al., 2003; Stanislaw & Todorov, 1999; formula 1) that quantifies how well participants are able to discriminate between real and fake items in the OCQ. The psychometric properties of the Psycho OCQ items are displayed in Table 1.
Expectedly, the experimental manipulation of the context had a large effect on both achievement motivation measured with the LMI-K and the OCQ bias scores. Participants indicated higher scores on the LMI-K in the application context than in the honest context (Figure 1), t(122) = 16.81, p < .001, d = 1.52 (H1). OCQ bias was also higher in the application context, M = 0.52, SD = 0.99, compared to the honest context, M = −0.95, SD = 0.41 (Figure 2), t(122) = 16.17, p < .001, d = 1.46 (H2). OCQ accuracy had a substantial variance in the honest context, M = 0.42, SD = 0.47, as well as in the application context, M = 0.37, SD = 0.39. The mean values did not differ between the two conditions, t(122) = 1.59, p = .114, d = 0.14, default BF01 = 2.92 (
; Morey & Rouder, 2022).
To assess the incremental effect of OCQ bias and RICS on faking behavior, we conducted multiple regression analyses with OCQ bias and RICS as predictors and faking behavior as the criterion (Table 2). The incremental effect of adding OCQ bias to a simple regression with RICS as a predictor results in a significant ΔR2 of .17, 90% CI [.07, .28], F(1, 120) = 31.18, p < .001. In the full model, both OCQ bias (β = .43) and RICS (β = .29) significantly predicted faking behavior even when controlling for the respective other variable. A Wald test conducted with the R package car (Fox & Weisberg, 2019) further indicated that the two predictors did not differ significantly in their predictive value, F(1, 120) = 1.27, p = .261 (H3). Results of an analysis similar to Feeney and Goffin (2015) in which each faking tendencies item is analyzed separately can be found in the supplementary material.
In addition to the frequentist approach reported above, we also conducted a Bayesian analysis to investigate whether there is support for the null hypothesis of H3. We specified a multivariate normal model and placed an informed LKJ prior distribution on the correlation matrix (Martin, 2021). We modeled the Fisher-z-transformed correlations with parameters ψ, for the average correlation of OCQ bias and RICS with faking behavior (with Logistic distribution L[0.70, 0.25] prior), and η, for the difference in correlations (with L[0.00, 0.55] prior, favoring small to medium differences). We used MCMC samples (Stan Development Team, 2023) to approximate the posterior distribution of η via logspline density estimation (Kooperberg, 2022). Finally, we estimated the Bayes factor relative to the null model positing equal correlations using the Savage-Dickey density ratio (Wagenmakers et al., 2010). The data lent moderate support for the null hypothesis that the correlation between OCQ bias and faking tendencies was comparable to the correlation between RICS and faking behavior, BF01 = 3.11 (H3). Assuming a difference exists, the posterior probability of this difference being larger than Δρ = ±0.1 (0.2, 0.3) was 0.479 (0.109, 0.011).
Discussion
The present study aimed to compare the utility of an OCQ bias measure in predicting applicant faking with that of Residualized Individual Change Scores (RICS). In a previous study, Feeney and Goffin (2015) considered RICS as an established measure of faking; their results suggested that RICS showed a stronger overlap with self-reports of faking behavior than an overclaiming measure. However, this study did not account for the better fit of the RICS measure to the application scenario. Extending on these findings, we aimed to compare the performance of OCQ bias and RICS in an application context in which the content of both measures fit equally well. The experimental manipulation of the context had the expected effect on the LMI-K, meaning that participants’ self-reported achievement motivation was substantially higher in an application situation as compared to an honest context (H1). This finding questions the applicability of achievement motivation scales in real-world applications, as the validity of the measurement seems to be threatened by the applicants’ positive self-presentation. The performance measure of the Psycho OCQ, OCQ accuracy, showed substantial variance and a mean value above zero in both the honest and the application context, strongly suggesting that participants were – to some extent – able to differentiate real from fake items. The internal consistency of the OCQ was good with ω > .93, apart from the fake items in the honest context for which an ω of .64 was obtained. The OCQ bias scores were also higher in the application context than in the honest context (H2). This suggests that in real-world applications, the OCQ bias is sensitive to applicants faking their answers, which is exactly what overclaiming questionnaires were created for. Both OCQ bias and RICS predicted participants’ self-reported faking behavior well and showed incremental validity over each other. In an exploratory analysis, OCQ bias significantly increased the explained variance in the external criterion when added to a simple regression with RICS as a predictor. OCQ bias and RICS did not differ with regard to their predictive accuracy according to the results of both a frequentist and a Bayesian approach (H3).
Extending on the findings of Feeney and Goffin (2015), our results thus suggest that the OCQ bias is just as good at detecting applicant faking behavior as the RICS measure, as long as the content-context fit of both measures is comparable. Crucially, however, applying RICS to real application situations is virtually impossible because applicants can hardly be given the same personality questionnaire twice and be expected to answer differently under “honest” and “application” instructions. Conveniently, OCQs do not require an experimental manipulation with a repeated administration of the same questionnaire. This is hence an important advantage of OCQs over RICS, allowing for OCQs to be employed in real-world applications.
Limitations
As a first limitation, we have to acknowledge that replicating our findings with other test materials, contexts, and samples is required. Currently, we cannot rule out that our conclusions are limited to the specific test content and application context of our study. Even more important is the potential caveat that the mock job application scenario in our study may not be comparable to real job applications. For example, actual applicants may be more apprehensive with respect to misrepresenting themselves in self-reports out of fear of negative consequences as compared to participants in a lab experiment being explicitly tasked with making a good impression. To this end, comparing tailored OCQs to RICS in high-stakes situations such as real personnel selections would be of great interest. Furthermore, the sample in the current study was fairly homogenous with the majority of participants being young and highly educated. Replications with more diverse samples are thus needed to clarify whether our findings generalize to populations with different demographic profiles.
A second limitation can be found in the external criterion used for measuring faking behavior in the current study, that is, the self-reported faking tendencies. An important condition for the validity of our results is that these self-reports allowed for a valid measurement of applicant faking. While our study design and the associations found with OCQ bias and RICS suggest reasonable validity of the faking tendencies, it cannot be ruled out that these self-reports were also contaminated by response biases such as social desirability and the need for consistency. That is, participants may tend to respond consistently with the previous task in the application context. Therefore, other validation criteria with a lower susceptibility to deliberate response distortions such as behavioral measures should be considered in future studies.
A third limitation could be due to response bias from answering the LMI-K and the OCQ twice in a row. Answering the same questionnaires in two different contexts may lead to hypothesis guessing and demand effects among participants. Some participants may tend to respond consistently across repeated administrations, whereas others may feel that they are expected to respond differently. Because the actual purpose of the OCQ is not apparent to the participants, the influence of bias on the OCQ scores should be small. Furthermore, while RICS always require repeated responses to the same questionnaire, OCQs usually require only a single administration. Therefore, the potential effects of a repeated presentation do not extend to practical applications of OCQs. Finally, it can be argued that hypothesis guessing also occurs in real job application contexts when participants speculate about the purpose of an inventory and then adjust their responses accordingly. Thus, even if some respondents provided biased responses in our study, they would likely have done so in real application situations as well.
Furthermore, the approach of condition-based regression analysis (CRA; Humberg et al., 2018a, 2018b) should be considered as an alternative to RICS in future studies (cf. Müller & Moshagen, 2019b, for an application of CRA in the context of overclaiming). Humberg et al. (2018a) criticize that common methods for measuring self-enhancement are unable to differentiate between the effects of self-enhancement and the effects of positivity of self-view. CRA allows for disentangling both constructs and promises more robust and unbiased empirical insights.
Conclusion
We conclude that a good content-context fit is an important requirement for the administration of OCQs, confirming the results of Dunlop et al. (2020). Ideally, measures used to assess participant faking should be tailored to the specific administration context, and studies comparing different measures of participant faking should ensure that all measures fit the specific context equally well. If these requirements are met, OCQs can keep up with competing measures of faking such as RICS. Therefore, OCQs remain a promising and innovative method for detecting applicant faking in job application contexts.
Electronic Supplementary Material
The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/1614-0001/a000411
References
(2013). Seriousness checks are useful to improve data validity in online research. Behavior Research Methods, 45(2), 527–535. https://doi.org/10.3758/s13428-012-0265-2
(2016). Berufsethische Richtlinien des Berufsverbandes Deutscher Psychologinnen und Psychologen e.V. und der Deutschen Gesellschaft für Psychologie e.V
[Professional ethical guidelines of the German Association of Psychologists and the German Psychological Society] . https://www.dgps.de/fileadmin/documents/Empfehlungen/berufsethische_richtlinien_dgps.pdf(2019). Teasing apart overclaiming, overconfidence, and socially desirable responding. Assessment, 26(3), 351–363. https://doi.org/10.1177/1073191117700268
(2011). Overclaiming as a measure of faking. Organizational Behavior and Human Decision Processes, 116(1), 148–162. https://doi.org/10.1016/j.obhdp.2011.05.006
(1992). Social desirability scales as moderator and suppressor variables. European Journal of Personality, 6(3), 199–214. https://doi.org/10.1002/per.2410060303
(2009). Effects of self‐enhancement on agreement on personality profiles. European Journal of Personality, 23(2), 107–123. https://doi.org/10.1002/per.707
(2011). Methods of measuring faking behavior. Human Performance, 24(4), 358–372. https://doi.org/10.1080/08959285.2011.597473
(2020). Liar! Liar! (when stakes are higher): Understanding how the overclaiming technique can be used to measure faking in personnel selection. Journal of Applied Psychology, 105(8), 784–799. https://doi.org/10.1037/apl0000463
(1999). Social desirability corrections in personality measurement: Issues of applicant comparison and construct validity. Journal of Applied Psychology, 84(2), 155–166. https://doi.org/10.1037/0021-9010.84.2.155
(2015). The Overclaiming Questionnaire: A good way to measure faking? Personality and Individual Differences, 82, 248–252. https://doi.org/10.1016/j.paid.2015.03.038
(2019). An R companion to applied regression (3rd ed.). Sage.
(2020). Testing competing claims about overclaiming. Intelligence, 81, Article
101470 . https://doi.org/10.1016/j.intell.2020.101470(1966). Signal detection theory and psychophysics. Wiley.
(2008). The failure of social desirability measures to capture applicant faking behavior. Industrial and Organizational Psychology, 1(3), 308–311. https://doi.org/10.1111/j.1754-9434.2008.00053.x
(2020). Use omega rather than Cronbach’s alpha for estimating reliability. But…. Communication Methods and Measures, 14(1), 1–24. https://doi.org/10.1080/19312458.2020.1718629
(2018a). Enhanced versus simply positive: A new condition-based regression analysis to disentangle effects of self-enhancement from effects of positivity of self-view. Journal of Personality and Social Psychology, 114(2), 303–322. https://doi.org/10.1037/pspp0000134
(2018b). Why condition-based regression analysis (CRA) is indeed a valid test of self-enhancement effects: A response to Krueger et al. (2017). Collabra: Psychology, 4(1), Article
26 . https://doi.org/10.1525/collabra.137(2014). Nuisance or remedy? The utility of stylistic responding as an indicator of data fabrication in surveys. Methodology, 10(3), 92–99. https://doi.org/10.1027/1614-2241/a000078
(2012). Responding to personality tests in a selection context: The role of the ability to identify criteria and the ideal-employee factor. Human Performance, 25(4), 273–302. https://doi.org/10.1080/08959285.2012.703733
(2022). logspline: Routines for logspline density estimation. https://CRAN.R-project.org/package=logspline
(2021, April 1). Stephen R. Martin, PhD [Blog]. Informative priors for correlation matrices: An easy approach. http://srmart.in/informative-priors-for-correlation-matrices-an-easy-approach/
(2022). BayesFactor: Computation of Bayes factors for common designs. https://CRAN.R-project.org/package=BayesFactor
(2003). Faking and selection: Considering the use of personality from select-in and select-out perspectives. Journal of Applied Psychology, 88(2), 348–355. https://doi.org/10.1037/0021-9010.88.2.348
(2018). Overclaiming shares processes with the hindsight bias. Personality and Individual Differences, 134, 298–300. https://doi.org/10.1016/j.paid.2018.06.035
(2019a). True virtue, self-presentation, or both? A behavioral test of impression management and overclaiming. Psychological Assessment, 31(2), 181–191. https://doi.org/10.1037/pas0000657
(2019b). Controlling for response bias in self-ratings of personality: A comparison of impression management scales and the overclaiming technique. Journal of Personality Assessment, 101(3), 229–236. https://doi.org/10.1080/00223891.2018.1451870
(2002). Ein Inventar zur Erfassung von zwei Faktoren sozialer Erwünschtheit
[An inventory for the assessment of two factors of social desirability] . Diagnostica, 48(3), 121–129. https://doi.org/10.1026//0012-1924.48.3.121(1985). Methods of coping with social desirability bias: A review. European Journal of Social Psychology, 15(3), 263–280. https://doi.org/10.1002/ejsp.2420150303
(1996). Role of social desirability in personality testing for personnel selection: The red herring. Journal of Applied Psychology, 81(6), 660–679. https://doi.org/10.1037/0021-9010.81.6.660
(1991).
Measurement and control of response bias . In J. P. RobinsonP. R. ShaverL. S. WrightsmanEds., Measures of personality and social psychological attitudes (Vol. 1, pp. 17–59). Academic Press.(2003). The over-claiming technique: Measuring self-enhancement independent of ability. Journal of Personality and Social Psychology, 84(4), 890–904. https://doi.org/10.1037/0022-3514.84.4.890
(1972). Some effects of “social desirability” in survey studies. American Journal of Sociology, 77(5), 921–940. https://doi.org/10.1086/225231
(2000). On the invalidity of validity scales: Evidence from self-reports and observer ratings in volunteer samples. Journal of Personality and Social Psychology, 78(3), 582–593. https://doi.org/10.1037/0022-3514.78.3.582
(1991). The social desirability response bias in ethics research. Journal of Business Ethics, 10(11), 805–817. https://doi.org/10.1007/978-94-007-4126-3_9
(1998). The impact of response distortion on preemployment personality testing and hiring decisions. Journal of Applied Psychology, 83(4), 634–644. https://doi.org/10.1037/0021-9010.83.4.634
(2016). A dynamic model of applicant faking. Organizational Psychology Review, 6(2), 145–170. https://doi.org/10.1177/2041386615580875
(2001). Leistungsmotivationsinventar: LMI; [Dimensionen berufsbezogener Leistungsorientierung]
[Achievement Motivation Inventory: LMI; [Dimensions of job-related achievement orientation]] . Hogrefe.. (2023). Stan modeling language users guide and reference manual, 2.21.8. https://mc-stan.org
(1999). Calculation of signal detection theory measures. Behavior Research Methods, Instruments, & Computers, 31(1), 137–149. https://doi.org/10.3758/BF03207704
(2021). Caught in the act: Predicting cheating in unproctored knowledge assessment. Assessment, 28(3), 1004–1017. https://doi.org/10.1177/1073191120914970
(1999). Die Soziale-Erwünschtheits-Skala-17 (SES-17): Entwicklung und erste Befunde zu Reliabilität und Validität
[The Social Desirability Scale-17 (SDS-17): Development and first findings on reliability and validity] . Diagnostica, 45(4), 173–177. https://doi.org/10.1026//0012-1924.45.4.173(2007). Sensitive questions in surveys. Psychological Bulletin, 133(5), 859–883. https://doi.org/10.1037/0033-2909.133.5.859
(1998). A meta-analytic review of predictors of job performance for salespeople. Journal of Applied Psychology, 83(4), 586–597. https://doi.org/10.1037/0021-9010.83.4.586
(2010). Bayesian hypothesis testing for psychologists: A tutorial on the Savage–Dickey method. Cognitive Psychology, 60(3), 158–189. https://doi.org/10.1016/j.cogpsych.2009.12.001
(2021). Impression management and overclaiming on job applications: Related to future sales performance? International Journal of Selection and Assessment, 29(1), 134–138. https://doi.org/10.1111/ijsa.12310
. (2013). World Medical Association Declaration of Helsinki: Ethical principles for medical research involving human subjects. JAMA, 310, 2191–2194. https://doi.org/10.1001/jama.2013.281053
(2015). Dishonest responding or true virtue? A behavioral test of impression management. Personality and Individual Differences, 81, 107–111. https://doi.org/10.1016/j.paid.2014.10.007
(2007). Fakability of different measurement methods for achievement motivation: Questionnaire, semi-projective, and objective. Psychology Science, 49(4), 291–307.
