Choice Dilemmas Questionnaire

Abstract

The Choice Dilemmas Questionnaire (CDQ), originally conceptualized and developed by social psychologists Michael A. Wallach and Nathan Kogan (1959, 1961; Kogan & Wallach, 1964), is one of the most historically significant psychometric instruments in the study of behavioral decision-making, risk tolerance, and small group dynamics. Consisting of 12 realistic narrative vignettes, the instrument presents respondents with hypothetical decision scenarios in which a protagonist must choose between an attractive but uncertain course of action (Alternative X) and a secure, default course of action offering a modest, guaranteed outcome (Alternative Y). Respondents serve as advisory decision-makers, indicating the minimum probability of success (ranging from 1 in 10 to 9 in 10, or an outright refusal to recommend the gamble) they would demand before counseling the protagonist to select the risky course.

Psychometrically, the CDQ was operationalized to assess an individual’s failure deterrence value—the subjective threshold at which the anticipated negative utility of failure overrides the prospective benefits of success. Although originally scored as a unidimensional composite index of generalized risk disposition, subsequent psychometric evaluations have demonstrated that the CDQ is multidimensional, reflecting distinct risk domains including financial-vocational security, physical safety, social approval, and interpersonal commitment. Internal consistency estimates across literature vary widely (Cronbach’s alpha typically ranging between .50 and .75), reflecting deliberate contextual variance across vignettes. The CDQ served as the empirical engine behind the seminal discovery of the “risky shift” and subsequent theories of group polarization, making it an indispensable landmark in experimental social psychology, organizational behavior, and behavioral economics.

Keywords

Choice Dilemmas Questionnaire, CDQ, risky shift, group polarization, Nathan Kogan, Michael A. Wallach, failure deterrence, decision making under risk, diffusion of responsibility, social comparison theory, persuasive arguments theory, subjective probability, risk tolerance, risk perception.

Authors

The Choice Dilemmas Questionnaire was developed through the collaborative research program of Michael A. Wallach and Nathan Kogan during the late 1950s and early 1960s, culminating in their landmark monograph, Risk Taking: A Study in Cognition and Personality (1964).

• Nathan Kogan, Ph.D. (1926–2013) was a prominent cognitive and developmental psychologist. He served for decades as a Senior Research Psychologist at the Educational Testing Service (ETS) in Princeton, New Jersey, and later held the position of Professor of Psychology at the Graduate Faculty of the New School for Social Research. Kogan’s work focused extensively on cognitive styles, creative thinking across the lifespan, and the intersection of cognitive processes with personality dynamics.
• Michael A. Wallach, Ph.D. (1933–2019) was an influential personality and social psychologist who served as Professor of Psychology at Duke University. His scholarship examined human creativity, psychometrics, modes of thinking in young children, and the social determinants of risk-taking and ethical choices.

Institutional affiliations during the formative development of the CDQ included the Age Center of New England, Harvard University, the Massachusetts Institute of Technology, and the Educational Testing Service. Inquiries regarding historical research archives are typically managed through university repository collections and the published scientific record.

Purpose

The primary purpose of the Choice Dilemmas Questionnaire is to quantify an individual’s behavioral willingness to endorse risky options when confronting probabilistic real-world dilemmas, as well as to evaluate how social interaction modulates individual judgment. Wallach and Kogan designed the CDQ to overcome the ecological limitations of conventional laboratory utility measures of their era, which relied almost exclusively on abstract monetary lotteries, dice rolls, or artificial card gambles. Such paradigms frequently failed to capture the multifaceted affective, social, and moral pressures that characterize everyday human risk-taking.

Specifically, the CDQ seeks to:

• Measure Failure Deterrence: Gauge the quantitative threshold of probability an individual requires to overcome the dread or negative valence of potential failure. Higher probability requirements indicate pronounced failure deterrence (conservative risk disposition), whereas willingness to proceed under slim probabilities indicates low failure deterrence (bold risk propensity).
• Investigate Social Influence and Group Dynamics: Provide standardized stimulus materials to study shift phenomena in dyadic and small-group discussions. Historically, James Stoner (1961) used Wallach and Kogan’s unpublished materials to demonstrate that group decisions systematically diverged from pre-discussion individual averages, yielding the “risky shift” effect.
• Examine Cognitive-Personality Correlates: Differentiate individuals along hypothesized personality dimensions—such as generalized test anxiety, defensiveness, need for achievement, and cognitive rigidity—to determine whether risk preference is a trait-like orientation or a context-dependent state.
• Facilitate Organizational and Advisory Research: Model how professionals advise clients or subordinates under conditions of uncertainty, bridging the gap between direct personal vulnerability and delegated advisory decision-making.

Psychological Construct

The core construct evaluated by the Choice Dilemmas Questionnaire is Failure Deterrence Value within Risky Decision Making. In classical decision theory (von Neumann & Morgenstern, 1944), choices are conceptualized as functions of objective probability and outcome utility. Wallach and Kogan posited that human decision-makers do not operate merely as mathematical calculators; rather, subjective decision thresholds are governed by the psychological weight assigned to the prospect of acute failure relative to the psychological value of triumphant success.

Dimensions Embedded in the CDQ Scenarios

Although the CDQ was initially conceptualized as a single aggregate index of generalized risk tolerance, psychometric and experimental scrutiny revealed that the 12 vignettes represent several distinct life domains:

• Vocational and Career Risk: Evaluated in Item 1 (Mr. A leaving an established electronics firm for an unstable startup) and Item 8 (Mr. H choosing between medical school and an uncertain career as a concert pianist). Here, the construct involves trading guaranteed professional security and pensions for potentially elite achievement versus total career dislocation.
• Somatic and Physical Safety: Exemplified in Item 2 (Mr. B undergoing a dangerous open-heart operation to cure a debilitating ailment) and Item 9 (Mr. J escaping a brutal prisoner-of-war camp where capture means execution). In these dilemmas, the downside failure involves mortal catastrophe (death or execution), invoking deep existential dread and activating distinct biological survival mechanisms.
• Financial and Capital Speculation: Evaluated in Item 3 (Mr. C investing a modest inheritance in volatile new-product stock versus safe blue-chip bonds) and Item 5 (Mr. E expanding an industrial corporation into a politically unstable foreign nation vulnerable to nationalization). The construct centers on capital risk, prospective monetary yield, and economic self-preservation.
• Competitive Strategy and Gamesmanship: Reflected in Item 4 (Mr. D calling a risky football play to win the championship versus settling for a tie) and Item 7 (Mr. G executing a deceptive, high-risk chess maneuver against an elite grandmaster). The construct taps into tactical audacity, competitive status, and zero-sum outcomes.
• Academic and Intellectual Ambition: Captured in Item 6 (Mr. F selecting a world-renowned doctoral chemistry program with high attrition versus a lower-tier program with guaranteed degree completion) and Item 11 (Mr. L spending five post-doctoral years investigating an unresolved, revolutionary physics problem versus publishing incremental, safe papers). The construct measures intellectual risk-taking and the pursuit of elite scientific prestige.
• Socio-Political and Relational Commitment: Measured in Item 10 (Mr. K sacrificing financial stability to run for Congress in a contested minority district) and Item 12 (Mr. M proceeding with marriage to Miss T despite emerging ideological and interpersonal incompatibilities). The construct reflects public exposure, interpersonal friction, and long-term socio-emotional risk.

Theoretical Framework

The Choice Dilemmas Questionnaire is situated at the intersection of several pivotal theoretical movements in 20th-century psychology:

1. Expected Utility and Subjective Probability

Classical economics assumes that rational actors maximize subjective expected utility. However, the CDQ reverses the standard econometric paradigm: rather than fixing probabilities and eliciting monetary equivalents (certainty equivalents), it fixes the qualitative payoffs (the attractive Alternative X versus the status-quo Alternative Y) and asks the participant to identify the exact subjective probability parameter ($p$) at which the expected utilities equalize. In this framework:

Utility(Alternative Y) = p × Utility(Success of X) + (1 – p) × Utility(Failure of X)

The respondent’s selected probability cutoff directly reveals their subjective weighting of the failure condition. A high probability requirement (e.g., 9 in 10) signals an asymmetrical dread of failure, whereas a low probability requirement (e.g., 1 in 10 or 3 in 10) reflects high optimism, sensation-seeking, or extreme utility placed on the upside.

2. The Risky Shift and Group Polarization

The historical fame of the CDQ is inextricably tied to the discovery of the “risky shift.” In master’s thesis research at MIT’s Sloan School of Management, James Stoner (1961) utilized Wallach and Kogan’s items to test whether group decision-making diluted individual boldness. Contrary to standard managerial dogma of the era—which held that committees are innately cautious, conservative, and compromise-driven—Stoner discovered that post-discussion group consensus shifted significantly toward greater risk than the pre-discussion average of individual members.

Wallach, Kogan, and Bem (1962, 1964) verified this phenomenon in extensive laboratory experiments and initially formulated the Diffusion of Responsibility hypothesis: within a deliberating group, the psychological burden of potential failure is distributed across all members, diminishing individual accountability and enabling bolder choices.

Subsequent theorists proposed competing explanatory mechanisms that broadened the construct into Group Polarization (Myers & Lamm, 1976):

• Social Comparison Theory (Normative Influence): Proposed by Roger Brown (1965), this view argues that individuals perceive risk-taking as a socially valued trait in Western cultures. During deliberation, individuals discover that peers are at least as daring as themselves, prompting social competition to demonstrate equal or superior courage.
• Persuasive Arguments Theory (Informational Influence): Formulated by Vinokur and Burnstein (1974), this model posits that group shifts are determined by the cognitive balance of novel, persuasive arguments generated during discussion. Because society favors growth, ambition, and courage in career, sports, and ambition dilemmas (Items 1, 4, 6, 7), the pool of pro-risk arguments naturally outweighs conservative arguments. Conversely, in somatic dilemmas where life itself is at stake (such as Item 2, Mr. B’s heart surgery, or Item 12, marriage instability), discussion often produces a cautious shift, because arguments stressing survival and emotional security dominate.

Validity

Psychometric evaluations of the Choice Dilemmas Questionnaire have yielded extensive evidence regarding its construct, convergent, discriminant, and predictive validity, alongside important psychometric caveats.

Construct and Predictive Validity

The construct validity of the CDQ is historically demonstrated by its sensitivity to experimental manipulation in small-group decision-making studies. Across hundreds of replications throughout the 1960s and 1970s (Cartwright, 1971; Dion, Baron, & Miller, 1970; Pruitt, 1971), the CDQ reliably predicted post-deliberation choice shifts. Specific items consistently produce predictable vectors of shift:

• Items 1, 3, 4, 6, 7, and 11 reliably elicit robust risky shifts following group discussion ($p < .001$), confirming that cultural values favoring enterprise, courage, and scholarly dedication drive informational exchange.
• Items 2 (the heart surgery) and 12 (the marriage) reliably generate cautious shifts, validating the hypothesis that when the cost of failure entails irreversible physical demise or catastrophic relational failure, the deterrence value operates in the opposite direction.

Convergent and Discriminant Validity

Studies examining the convergent validity of the CDQ with other risk-taking measures have highlighted both strengths and theoretical limitations:

• Correlations with Objective Gambling Tasks: Early investigations by Kogan and Wallach (1964) identified modest positive correlations ($r pprox .20$ to $.35$) between CDQ scores and willingness to bet on low-probability lotteries in lab settings. However, subsequent psychometric reviews by Paul Slovic (1962, 1972) demonstrated that risk-taking is predominantly domain-specific rather than a monolithic, unitary personality trait. CDQ scores often showed near-zero correlations with real-world physical risk-taking (e.g., extreme sports) or monetary wagers with real financial consequences.
• Relation to Trait Anxiety and Defensiveness: Kogan and Wallach (1964) established meaningful discriminant interactions: among individuals low in defensiveness (measured via the Marlowe-Crowne Social Desirability Scale) and low in test anxiety (Mandler-Sarason Test Anxiety Questionnaire), CDQ scores demonstrated stable cognitive associations with analytical problem-solving and divergent thinking. For highly defensive, anxious subjects, CDQ scores fluctuated dramatically, reflecting ego-protective maneuvers rather than authentic risk calculus.
• Contrast with Modern Domain-Specific Instruments: Contemporary validation comparisons indicate that while the CDQ shares conceptual overlap with the financial and ethical subscales of modern instruments like the Domain-Specific Risk-Taking (DOSPERT) scale (Weber, Blais, & Betz, 2002), the CDQ measures prescriptive advisory risk (advising a third party) rather than first-person behavioral risk, conferring unique discriminant properties.

Reliability

Reliability analyses of the CDQ reflect the complex psychometric nature of scenario-based dilemmas:

Internal Consistency

Because the 12 scenarios were deliberately written to encompass diverse human activities—ranging from cardiovascular medical surgery to college football tactics, scientific research trajectories, and wartime escape—the inter-item correlations are inherently modest. Published studies report overall Cronbach’s alpha coefficients typically ranging from .53 to .71 (Cartwright, 1971; Kogan & Wallach, 1964; Myers & Arenson, 1972). Split-half reliability coefficients, corrected by the Spearman-Brown formula, historically hovered around $.55$ to $.68$ for male cohorts and $.50$ to $.65$ for female cohorts.

When the scale is decomposed into sub-domains (e.g., isolating vocational/achievement dilemmas from physical/somatic dilemmas), internal consistency estimates within homogeneous sub-clusters improve moderately ($r pprox .65$ to $.78$). The modest overall alpha is widely recognized in psychometric literature not as a measurement flaw, but as evidence of the multidimensional, domain-specific nature of real-life decision-making under risk.

Test-Retest Stability

Temporal stability assessments have demonstrated satisfactory test-retest reliability across non-intervention control intervals:

• Across short intervals (1 to 3 weeks), individual test-retest correlations range between $r = .75$ and $r = .82$, indicating that individual baseline risk thresholds remain relatively stable in the absence of group interaction or contextual framing manipulations.
• Across extended periods (several months), test-retest coefficients attenuate to approximately $r = .55$ to $.65$, reflecting longitudinal shifts in personal circumstances, maturity, and life experiences.

Factor Analysis

Although early researchers computed a single sum or mean across all 12 items to represent a global “risk score,” comprehensive Exploratory Factor Analyses (EFA) and subsequent Confirmatory Factor Analyses (CFA) have consistently refuted unidimensionality (Cartwright, 1971, 1973; Slovic, 1972).

Factor Structure and Loadings

Principal components analyses with varimax rotation typically extract three or four distinct, interpretable factors accounting for roughly 52% to 64% of total variance:

• Factor 1: Vocational, Ambition, and Achievement Dilemmas (Accounting for ~24% of variance):
  • Item 1 (Mr. A: Electrical Engineer / Startup Job): Loadings typically range between .62 and .74.
  • Item 6 (Mr. F: Prestigious Chemistry Graduate School): Loadings range between .58 and .71.
  • Item 8 (Mr. H: Concert Pianist vs. Physician): Loadings range between .55 and .68.
  • Item 11 (Mr. L: Long-term Physics Problem): Loadings range between .64 and .76.

  This primary factor captures standard aspirational career risks where failure entails missed career advancements but no physical harm.

• Factor 2: Somatic Integrity and Mortal Peril (Accounting for ~16% of variance):
  • Item 2 (Mr. B: Heart Surgery): Strong positive loading (.70 to .82).
  • Item 9 (Mr. J: Prisoner-of-War Escape): Strong positive loading (.65 to .78).

  This factor represents survival-threat dilemmas where a failed gamble results in physical mortality or immediate execution, producing pronounced failure deterrence.

• Factor 3: Strategic Competition and Tactical Risk (Accounting for ~12% of variance):
  • Item 4 (Mr. D: Football Last-Second Play): Loadings between .60 and .75.
  • Item 7 (Mr. G: Deceptive Chess Move): Loadings between .58 and .72.

  Characterized by discrete, rule-governed zero-sum competitive environments where outcome feedback is instantaneous.

• Factor 4: Interpersonal, Political, and Financial Vulnerability (Accounting for ~10% of variance):
  • Item 3 (Mr. C: Stock Market Speculation): Loadings between .45 and .62.
  • Item 5 (Mr. E: Foreign Plant Expansion): Loadings between .48 and .65.
  • Item 10 (Mr. K: Political Campaign): Loadings between .51 and .67.
  • Item 12 (Mr. M: Marrying Miss T): Moderate loadings (.40 to .60), often displaying secondary cross-loadings.

Model Fit and Modern Psychometric Appraisal

Modern CFA investigations confirm that a hierarchical or correlated four-factor model provides vastly superior fit over a single-factor model (e.g., Comparative Fit Index $[CFI] > .92$, Root Mean Square Error of Approximation $[RMSEA] < .06$ for the multidimensional structure, versus $CFI < .70$ for the unifactorial model). Psychometricians strongly recommend reporting subscale or item-level scores rather than collapsing the CDQ into a single aggregated total.

Instrument / Measurement Tool

• Test Type: Advisory behavioral decision-making inventory / Scenario-based psychometric questionnaire.
• Format: Paper-and-pencil or computer-administered narrative vignettes presenting hypothetical decision scenarios.
• Item Count: 12 standardized dilemmas (Mr. A through Mr. M; the letter “I” is omitted to avoid participant confusion with the numeral 1 or the first-person pronoun).
• Response Scale: Six-point quasi-probabilistic discrete scale per item. The respondent marks the lowest acceptable probability of success required to endorse the risky alternative:
  • 1 in 10 (10% chance of success; scored as 1)
  • 3 in 10 (30% chance of success; scored as 3)
  • 5 in 10 (50% chance of success; scored as 5)
  • 7 in 10 (70% chance of success; scored as 7)
  • 9 in 10 (90% chance of success; scored as 9)
  • Refusal option (“Place a check here if you think [the protagonist] should not [choose the risky alternative] no matter what the probabilities”; scored as 10 or excluded according to protocol).
• Administration Time: Approximately 15 to 25 minutes for individual self-administration; 45 to 60 minutes when combined with dyadic or small-group consensus discussions.
• Target Population: Adults and adolescents (reading level: high school and above).
• Scoring Rules:
  • Each item is assigned a numerical value corresponding to the odd’s numerator: 1, 3, 5, 7, 9. The refusal option is coded as 10 (representing a required probability of 10 in 10 or 1.00), though some historical investigators coded refusal as an omission or assigned it a score of 10.5.
  • Composite Mean / Total Score: Summing the 12 items yields a theoretical score range of 12 to 120. Lower total scores represent higher risk tolerance (lower failure deterrence), while higher total scores indicate lower risk tolerance (higher failure deterrence / conservatism).
  • Domain Sub-Scores: Researchers typically compute distinct averages for vocational-achievement items (1, 6, 8, 11), somatic/survival items (2, 9), competitive/tactical items (4, 7), and socioeconomic/interpersonal items (3, 5, 10, 12).

Permissions & Fee and Test Year

• Test Year: Developed in 1959–1961; widely disseminated in journal publications (Wallach, Kogan, & Bem, 1962) and fully published in monograph format in 1964.
• Original Publisher / Copyright Holder: Holt, Rinehart & Winston (1964 monograph: Risk Taking: A Study in Cognition and Personality). The original scale items are cataloged under APA PsycTESTS (DOI: 10.1037/t00821-000).
• Permissions and Fee: The Choice Dilemmas Questionnaire is considered an open historical psychometric tool for academic, educational, and non-commercial research purposes. Researchers are expected to appropriately cite Wallach, Kogan, and subsequent primary literature. For commercial adaptation or inclusion in monetized commercial platforms, formal copyright clearance should be confirmed with the respective corporate successors or copyright holders.

References

• Brown, R. (1965). Social psychology. Free Press.
• Cartwright, D. (1971). Risk taking by groups: A study of new common sense. Journal of Personality and Social Psychology, 20(3), 361–378. https://doi.org/10.1037/h0031911
• Cartwright, D. (1973). Determinants of scientific progress: The case of the risky shift. American Psychologist, 28(3), 222–231. https://doi.org/10.1037/h0034509
• Dion, K. L., Baron, R. S., & Miller, N. (1970). Why do groups make riskier decisions than individuals? In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 5, pp. 305–377). Academic Press. https://doi.org/10.1016/S0065-2601(08)60094-1
• Kogan, N., & Wallach, M. A. (1964). Risk taking: A study in cognition and personality. Holt, Rinehart & Winston. https://doi.org/10.1037/t00821-000
• Myers, D. G., & Arenson, S. J. (1972). Enhancement of dominant risk tendencies in group discussion. Psychological Reports, 30(2), 615–623. https://doi.org/10.2466/pr0.1972.30.2.615
• Myers, D. G., & Lamm, H. (1976). The group polarization phenomenon. Psychological Bulletin, 83(4), 602–627. https://doi.org/10.1037/0033-2909.83.4.602
• Pruitt, D. G. (1971). Choice shifts in group discussion: An introduction to the field. Journal of Personality and Social Psychology, 20(3), 339–360. https://doi.org/10.1037/h0031922
• Slovic, P. (1962). Convergent validation of risk taking measures. Journal of Abnormal and Social Psychology, 65(1), 68–71. https://doi.org/10.1037/h0043818
• Slovic, P. (1972). Information processing, situation specificity, and generality of risk-taking behavior. Journal of Personality and Social Psychology, 22(1), 128–134. https://doi.org/10.1037/h0032608
• Stoner, J. A. F. (1961). A comparison of individual and group decisions involving risk (Unpublished master’s thesis). Massachusetts Institute of Technology, Cambridge, MA.
• Vinokur, A., & Burnstein, E. (1974). Effects of partially shared persuasive arguments on group-induced shifts: A group-problem-solving approach. Journal of Personality and Social Psychology, 29(3), 305–315. https://doi.org/10.1037/h0036010
• von Neumann, J., & Morgenstern, O. (1944). Theory of games and economic behavior. Princeton University Press.
• Wallach, M. A., Kogan, N., & Bem, D. J. (1962). Group influence on individual risk taking. Journal of Abnormal and Social Psychology, 65(2), 75–86. https://doi.org/10.1037/h0044376
• Wallach, M. A., Kogan, N., & Bem, D. J. (1964). Diffusion of responsibility and level of risk taking in groups. Journal of Abnormal and Social Psychology, 68(3), 263–274. https://doi.org/10.1037/h0042457
• Weber, E. U., Blais, A. R., & Betz, N. E. (2002). A domain-specific risk-attitude scale: Measuring risk perceptions and risk behaviors. Journal of Behavioral Decision Making, 15(4), 263–290. https://doi.org/10.1002/bdm.414

13. Items of the Scale (Questionnaire)