An evaluation of the Choice Dilemma Questionnaire as a measure of risk-taking propensity
Main Article Content
The Choice Dilemma Questionnaire (CDQ) has been extensively used in risk taking studies due to its ease of administration and its ability to consistently produce risky shifts. However despite its wide use, it has been subject to a number of criticisms. This study attempts to see if the CDQ score is factorially pure or if it is made up of subscales. Responses from 202 students were factor analyzed and it was found that 4 factors emerged. It is concluded that risk-taking is a multidimensional concept and the different types of risks need to be taken into account if we are to use risk-taking propensity as a personality measure while studying individual differences.
Important decisions take place under conditions of uncertainty and risk. This has prompted research on decision making under risk and attracted academicians and practitioners from such diverse fields as economics (Knight 1971), psychology (Kahneman & Tversky, 1979), finance (Jarret, 1990), and management (Bowman, 1980). Despite the advances in quantitative techniques, we are unable to reduce decision making into a routine function, and we value good decision-makers. The use of personality tests in selection and evaluation points to the nexus between the personality traits of the decision maker and the decisions made. Therefore a number of researchers have focused their attention on the risk-taking propensity of managers and entrepreneurs (MacCrimmon & Wehrung, 1984; Sunder & Kamalanabhan, 1993).
Though a variety of instruments and techniques have been used by researchers (Chaubey, 1974; Kogan & Wallach, 1964; Montgomery & Landers, 1974; Singh, 1989), the Choice Dilemma Questionnaire (CDQ) stands out in the research on risk-taking propensity. The CDQ has been used in studies encompassing a wide range of research interests such as the risky shift phenomenon (Burnstein, Miller, Vinokur, Katz, & Crowley, 1971), risk value hypothesis (Brown, 1965; Levinger & Schnider, 1969), the influence of age on risk-taking (Vroom & Pahl, 1971), and cultural effects on risk-taking (Carlson & Davis, 1971). The instrument as developed by Stoner (1961) and popularized by Kogan and Wallach (1964) has been widely used for measuring the general risk-taking propensity of individuals. The instrument has also been used in studies in an attempt to identify differences between groups of managers and entrepreneurs (Brockhaus, 1980). Despite its popularity, the CDQ has been criticized on various counts. Shaver and Scott (1991) comment as follows:
“The CDQ is being put to uses for which it was not designed. As employed by Kogan and Wallach and legions of other psychologists, the CDQ was the vehicle to measure changes in expressed levels of riskiness that occurred as a consequence of group discussion. The CDQ never underwent the sort of reliability and validity testing that would be required for a real personality measure. To use such a measure as the index of a relatively enduring personality trait is a serious methodological error”.
MacCrimmon and Wehrung (1984) pointed to the following deficiency in the CDQ: “In the usual choice dilemmas, the subject is asked to advise 12 different individuals in highly dissimilar settings. Each of these situations implies a situation for which a person has no experience”.
Cartwright (1971), in a review of studies which used CDQ, noted that a person’s rank on the CDQ score does not necessarily correspond to his/her rank on even a majority of items. Though Higbee (1971) noted that others have found CDQ scores reflecting behavior in simple betting tasks, he raised questions as to whether the CDQ scores reflect behavior in complex decision situations. Vroom and Pahl (1971), using a subset of the items in the CDQ, reported that the intercorrelation between the items is poor.
The measure of risk-taking as expressed by the CDQ is the sum of scores of the individual on the 12 items of the CDQ. It is possible that not all the items on the CDQ measure the same construct, and that the various criticisms of the Instrument are a result of our misplaced trust in a single measure of the total score as the measure of risk-taking propensity. This could account for the fact that risk-taking propensity as measured by the CDQ has failed to differentiate between entrepreneurs and managers (Brockhaus, 1980) despite the importance attached to risk-taking in entrepreneuring. Support for this is found in the article by Liles (1974) who listed different types of risks faced by an entrepreneur.
In this study we are operating on the premise that risk-taking is a multidimensional concept and a person can exhibit different risk-taking propensities on different types of risks. This can explain the inconclusive, and many conflicting, results on the risk-taking behavior of various groups (Sexton & Bowman, 1985). We propose to carry out factor analysis of the CDQ scores to identify different factors or subscales in the instrument.
Method
Two hundred and twelve undergraduate students from three colleges in Madras volunteered to participate in the study, and they were asked to complete the CDQ questionnaire. Of the responses received, 202 were found usable, as the remaining were either incomplete or contained ambiguous answers. The mean age of the respondents was 21.5 years.
Results and Discussion
The means and standard deviations of the scores obtained by the participants on each of the items and the total are given in Table 1, and it can be seen that the mean scores on the items range from 5.3 to 7.5 and the standard deviations range from 2.4 to 3.1. Multivariate analysis of variance of the scores (repeated measures design MANOVA) shows that the difference between means of the items is significant at p < 0.05, lending support to our premise that people differ in their risk-taking propensity depending on the type of risk.
Table 1. Mean and Standard Deviation of Item Scores
Intercorrelations of the items were calculated and these are shown in Table 2. The item-correlations are low, and are significant for only 15 of the 66 intercorrelations.
Table 2. Item Intercorrelations
It is therefore probable that all the items in the CDQ do not represent the same dimension. A factor analysis of the scores was carried out to check how many factors emerged. The results of principal component analysis are shown in Table 3, and it is seen that four factors accounting for 49.5% of the variance can be extracted with eigenvalues above 1.
Table 3. Extraction 1 for Analysis 1, Principal-Components Analysis (PC)
The factors were then subjected to orthogonal rotation using varimax, and the loading of the items on the four factors is shown in Table 4.
Table 4. Rotated Factor Matrix
Note. Minimum loadings of ± .35 and above are shown in the table.
As only 49.5% of the variance is explained by the four factors, it was decided to extract more factors to see if this could improve interpretation. It was found that 72% of the variance is explained if 7 factors are extracted, and 78% of the variance is explained when eight factors are extracted, respectively. However, the results of the analysis by increasing the number of factors to 7 or 8 did not help interpretability.
The results of factor analysis show that the CDQ has a minimum of four scales. Scale 1, represented by factor 1, we labeled the role perception/ prestige risk. It represents the loss of face associated with one’s not living up to the expectations of the role. Scale 2, represented by a factor we called the commitment risk is where one risks losing his/her freedom of change for a desirable, but risky, outcome. Scale 3, represented by factor 3, we called the career risk, wherein a stable and acceptable career is risked in the hope of achieving a more desirable career. Scale 4, represented by factor 4, we called the physical and mental well-being risk. Here one stands to lose his/her peace of mind and/or health if they are not successful. These labels are only a preliminary attempt to name the different types of risks represented by the different factors. The correlation of each item with the corresponding factor (highest loading) and the correlation of the item with the total CDQ scores are shown in Table 5. It can be seen that item correlations with the subscales (factors) are higher than the item correlations with the total CDQ scores lending support to the multi-dimensionality concept of risk-taking propensity.
Table 5. Correlation Between Total CDQ Scores and Each Factor Loading
Though the risk-taking behavior of managers, (MacCrimmon & Wehrung, 1984), entrepreneurs (Brockhaus, 1980), and gamblers (Phillips, 1972) have been extensively studied, most of the researchers have adopted a unidimensional approach. This could account for the many conflicting results obtained in research on risk-taking. MacCrimmon and Wehrung (1990) used a portfolio of measures for assessing risk-taking and found by factor analysis that six factors emerged. The results of this study also support the multi-dimensionality concept of risk-taking. Questions have been raised on the validity and reliability of the CDQ scores, and the results of the study show that both the validity and the reliability of the test would be higher if we were to recognize the fact that the items represent different types of risks, and develop scales accordingly. We conclude that, while a general risk-taking propensity may be of use in studying the risky shift phenomenon, it would be preferable to consider risk-taking propensities based on the particular type of risk in studying individual differences.
References
Bowman, E. H. (1980). A risk and return paradox for strategic management. Cambridge, MA: Massachusetts Institute of Technology.
Brockhaus, R. H. (1980). Risk taking propensity of entrepreneurs. Academy of Management Journal, 23(6) 509-520.
Brown, R. (1965). Social psychology. New York: Free Press.
Burnstein, E., Miller, H., Vinokur, A., Katz, S., & Crowley, J. (1971). Risky shift is eminently rational. Journal of Personality and Social Psychology, 20(3), 462-471.
Carlson, J. A,. & Davis, N. C. (1971). Cultural values and the risky shift. Journal of Personality and Social Psychology, 20(3), 3921-3929.
Cartwright, D. (1971). Risk-taking by individuals and groups. An assessment of research employing choice dilemmas. Journal of Personality and Social Psychology, 20, 361-378.
Chaubey, N. P. (1974). Effect of age on expectancy of success and on risk-taking behavior. Journal of Personality and Social Psychology, 29(6), 774-778.
Higbee, K. L. (1971). Expression of “Walter Mitty-ness” in actual behavior. Journal of Personality and Social Psychology, 20(3), 416-422.
Jarret. J. (1990). Business decisions and managerial risk: A note on decision analysis approach. Management Science, 18(1), 95-99.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263-291.
Knight, F. H. (1971). Risk, uncertainty and profit. Chicago: University of Chicago Press.
Kogan, N., & Wallach, A. (1964). Risk taking. New York: Holt Rhinehart & Winston.
Levinger, G., & Schnider, D. J. (1969). Test of a risk is a value hypothesis. Journal of Personality and Social Psychology, 11, 165-170.
Liles, P. R. (1974). Who are the entrepreneurs? Business Topics, 22(1), 5-14.
MacCrimmon, K. R., & Wehrung, D. A. (1984). The risk-in-basket. Journal of Business, 57(3), 367-387.
MacCrimmon, K. R., & Wehrung, D. A. (1990). Characteristics of risk-taking executives. Management Science, 36(4), 422-435.
Montgomery, G. T., & Landers, W. F. (1974). Transmission of risk-taking through modelling at two age levels. Psychological Reports, 34, 1187-1196.
Phillips, R. (1972). Probability preferences of gamblers and non-gamblers. Psychological Reports, 31, 652-654.
Sexton, D., & Bowman, N. (1985). Personality inventory for potential entrepreneurs: Evaluation of a modified JPI/PRF-E test instrument. Proceedings Entrepreneurship Research Conference, 513-529.
Shaver, G. K., & Scott, L. R. (1991). Person and process choice: The psychology of new venture creation. Entrepreneurship theory and practice, Winter, 23-45.
Singh, S. (1989). Personality characteristics, work values, and life style of fast and slow progressing small scale industrial entrepreneurs. The Journal of Social Psychology, 129(6), 801-805.
Sunder, D. L., & Kamalanabhan, T. J. (1993). Reaction to risk: Are entrepreneurs different from others? Paper presented at the XXX annual conference of IAAP at Jodhpur, India.
Stoner, J. A. (1961). A comparison of individual and group decisions involving risk. Unpublished master’s thesis, Sloan School of Management, Massachusetts Institute of Technology, Cambridge.
Vroom, V. K., & Pahl, B. (1971). Relationship between age and risk-taking among managers. Journal of Applied Psychology, 55(5), 399-405.
Bowman, E. H. (1980). A risk and return paradox for strategic management. Cambridge, MA: Massachusetts Institute of Technology.
Brockhaus, R. H. (1980). Risk taking propensity of entrepreneurs. Academy of Management Journal, 23(6) 509-520.
Brown, R. (1965). Social psychology. New York: Free Press.
Burnstein, E., Miller, H., Vinokur, A., Katz, S., & Crowley, J. (1971). Risky shift is eminently rational. Journal of Personality and Social Psychology, 20(3), 462-471.
Carlson, J. A,. & Davis, N. C. (1971). Cultural values and the risky shift. Journal of Personality and Social Psychology, 20(3), 3921-3929.
Cartwright, D. (1971). Risk-taking by individuals and groups. An assessment of research employing choice dilemmas. Journal of Personality and Social Psychology, 20, 361-378.
Chaubey, N. P. (1974). Effect of age on expectancy of success and on risk-taking behavior. Journal of Personality and Social Psychology, 29(6), 774-778.
Higbee, K. L. (1971). Expression of “Walter Mitty-ness” in actual behavior. Journal of Personality and Social Psychology, 20(3), 416-422.
Jarret. J. (1990). Business decisions and managerial risk: A note on decision analysis approach. Management Science, 18(1), 95-99.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263-291.
Knight, F. H. (1971). Risk, uncertainty and profit. Chicago: University of Chicago Press.
Kogan, N., & Wallach, A. (1964). Risk taking. New York: Holt Rhinehart & Winston.
Levinger, G., & Schnider, D. J. (1969). Test of a risk is a value hypothesis. Journal of Personality and Social Psychology, 11, 165-170.
Liles, P. R. (1974). Who are the entrepreneurs? Business Topics, 22(1), 5-14.
MacCrimmon, K. R., & Wehrung, D. A. (1984). The risk-in-basket. Journal of Business, 57(3), 367-387.
MacCrimmon, K. R., & Wehrung, D. A. (1990). Characteristics of risk-taking executives. Management Science, 36(4), 422-435.
Montgomery, G. T., & Landers, W. F. (1974). Transmission of risk-taking through modelling at two age levels. Psychological Reports, 34, 1187-1196.
Phillips, R. (1972). Probability preferences of gamblers and non-gamblers. Psychological Reports, 31, 652-654.
Sexton, D., & Bowman, N. (1985). Personality inventory for potential entrepreneurs: Evaluation of a modified JPI/PRF-E test instrument. Proceedings Entrepreneurship Research Conference, 513-529.
Shaver, G. K., & Scott, L. R. (1991). Person and process choice: The psychology of new venture creation. Entrepreneurship theory and practice, Winter, 23-45.
Singh, S. (1989). Personality characteristics, work values, and life style of fast and slow progressing small scale industrial entrepreneurs. The Journal of Social Psychology, 129(6), 801-805.
Sunder, D. L., & Kamalanabhan, T. J. (1993). Reaction to risk: Are entrepreneurs different from others? Paper presented at the XXX annual conference of IAAP at Jodhpur, India.
Stoner, J. A. (1961). A comparison of individual and group decisions involving risk. Unpublished master’s thesis, Sloan School of Management, Massachusetts Institute of Technology, Cambridge.
Vroom, V. K., & Pahl, B. (1971). Relationship between age and risk-taking among managers. Journal of Applied Psychology, 55(5), 399-405.
Table 1. Mean and Standard Deviation of Item Scores
Table 2. Item Intercorrelations
Table 3. Extraction 1 for Analysis 1, Principal-Components Analysis (PC)
Table 4. Rotated Factor Matrix
Note. Minimum loadings of ± .35 and above are shown in the table.
Table 5. Correlation Between Total CDQ Scores and Each Factor Loading
Appreciation is due to reviewers including
Professor Purnima Mathur
Indian Institute of Technology
New Delhi.
T. J. Kamalanabhan, Faculty of Management, Universiti Telekom, Melaka-75450, Malaysia.