OBJECTIVE ASSESSMENT OF RISK TENDENCY21_EN_REV

THE OBJECTIVE ASSESSMENT OF RISK TENDENCY AS A PERSONALITY DIMENSION

Víctor J. Rubio* José Santacreu José Manuel Hernández

University Autónoma of Madrid

*Correspondence to author: Dr. Víctor J. Rubio. Facultad de Psicología, Universidad Autónoma de Madrid, 28049, Madrid (Spain). e-mail: [email protected]

The objective assessment of the risk tendency as a personality dimension

1. Theoretical background Is now the best moment for selling the stokes, or would it be better to hold on a little bit longer? The traffic light is nearly red; do you have enough time to cross over? Everyone will agree on the fact that in these situations risk is involved. However, a unique definition of risk is far from being universally accepted (Nicholson, Soane, Fenton-O’Creevy & Willman, 2005). Following Yates & Stone (1992), the risk construct entails three critical elements: potential losses, the significance of losses, and the uncertainty of losses. We would also add another previous (implicit) element: a decision-making context. Risk regards to a situation in which a person has several alternatives for acting not all of them produce outcomes as satisfactory as desired. Thus, the worth of each available alternative is the result of the risk and some other considerations that give value (success or failure) to the decision. Approaches to the study of risk taking behavior (RTB) have also depicted the diversity of conceptualizations of risk. For instance, the traditional Decision Theory approach to the study of RTB has been focused on the characteristics of the situations, ignoring interpersonal differences and studying average behavior. Some other approaches, however, have focused on cross-situational consistency of individuals’ risk tendency. These approaches have also produced differences in assessment issues. For instance, one of the most common research paradigms has looked at choice behavior in experimental situations using lotteries or games. The traditional study entails showing a set of multi-outcome lotteries with the same expected value but different prize distribution. Each lottery involves drawing one ticket (see Lopes, 1987, for her proposal of experimental lotteries). Usually, subjects will choose the preferred lottery on which to bet when they are presented in pairs. Another format is that individuals have to price the lottery (to choose the smallest amount of money for which they would sell the chance of playing each lottery).

1

V. Rubio, J. Santacreu, J.M. Hernández

Even though the risk assessment procedures mentioned below are task based, they are not free of problems. Firstly, they would not be considered as objective tests due to the lack of accomplishment of the main characteristic regarding this way of assessing Cattell posed,, namely, the absence of “motivational distortion” (Cattell, 1958; Cattell & Warburton, 1967). Secondly, they have been questioned as a representation of how individuals behave in naturalistic situations (Wagenaar, 1988). Thirdly, none of them has been designed in a context of a theory of personality. In most of the cases individual differences have been vaguely embraced as risk attitudes that are defined in terms of the curvature -concave (risk averse) or convex (risk seeker)- of the utility function. At most, a few authors have connected individual differences to motivational aspects, such as Atkinson and his colleagues (Atkinson, 1957, 1983; Atkinson & Litwin, 1960), or Schneider and Lopes (1986). There are some other approaches that are more focused on personality aspects of RTB that have produced significant instruments. One of the first is the Kogan-Wallach Choice Dilemmas Questionnaire (Kogan & Wallach, 1960, 1964; Wallach, Kogan & Bem, 1962). It consists of 12 situations representing a choice dilemma between a risky and a safe course of action. Subjects should select the probability level for the risk alternative success that would make it sufficiently attractive to be chosen. These authors also developed the Chance Bets Instrument (Kogan & Wallach, 1964) which consists of 66 pairs of dice bets varying the probabilities of winning and losing, and the amounts of money to be won or lost (with the same expected value EV=0). Individuals could choose the bet they preferred to play. More recently Weber, Blais & Betz (2002) have developed a risk attitude scale that consists of 8item subscales in four content domains (health/safety, ethical, social, and recreational risks) and two 4-item subscales (investment and gambling) for financial risk taking. Nicholson et al (2005) have also developed another instrument which assesses the reported frequency of risk behaviors in six different domains (recreation, health, career, finance, safety and social).

2


Probably, the most well-known instrument for assessing RTB is the Zuckerman’s Sensation Seeking Scale (SSS, Zuckerman, 1979; Zuckerman, Eysenck & Eysenck, 1978). The scale consists of four subscales (Thrill and adventure seeking, Experience seeking, Disinhibition, and Boredom susceptibility). Its widespread use has produced an amount of psychometric and cross-cultural data and it has become the most frequent instrument for assessing risk taking. However, the major problem the assessment of RTB has to tackle is the fact that there is a lack of high cross-situational consistency among measures and domains of risk tendency (Bromiley & Curley, 1992, MacCrimmon & Wehrung, 1986; Schoemaker, 1990; Slovic, 1964; Soane & Chmiel, 2005; Weber, et al., 2002; Weber & Milliman, 1997). Currently, even though inventories present better interconsistency indexes, it would be, in part, due to the use of equivalent questionnaire items (Zuckerman, 1979b). 2. Three tests for an objective personality assessment of risk-taking behavior The authors have developed three computerised task-based tests for an objective assessment of the risk tendency construct: The Betting Dice Test (BDT), The Roulette Test (RT), and The Crossing the Street Test (CtST) (Arend, Botella, Contreras, Hernández & Santacreu, 2003; Hernández, Rubio, Revuelta y Santacreu, in press; Rubio, Santacreu & Hernandez, 2004; Santacreu, Rubio, & Hernández, in press, Santé & Santacreu, 2001). The BDT consists of a task in which individuals have to bet on one alternative out of four in order to estimate the result of the sum of two dice: More than 4, More than 7, More than 9, a straight bet on Number 12 (see Figure 1). Each alternative is associated to a prize: 1, 2, 5, and 30 points, respectively. Subjects are encouraged to obtain as many points as possible. Individuals have to make their bet but are not informed about the numbers of trials (10) nor the results supposedly obtained for each of their bets. It is assumed that the choice of an option with a higher probability (though a lower prize) is a more conservative choice than the

3


choice of a highly improbable alternative. Thus, the lower the score, the more conservative the risk tendency is. The RT is similar to the BDT. In this case subjects have to bet on one out of four options of the game of roulette: Thirty Numbers (from 1 to 30), Fifteen Numbers (even numbers from 2 to 30), Six Numbers (from 31 to 36), and a straight bet (17) (see Figure 2). Prizes, instructions, lack of feedback, system messages, and scoring are the same as above. The CtST consists of a task in which individuals should decide where to cross a pedestrian from one side of a road to another in order to reach a chemist (see Figure 3). The point is that the pedestrian should get to the chemist as quickly as possible but without causing a car crash or being knocked down. The pedestrian is moved by clicking over the arrow keys displayed and he should get to the chemist on the other side of the road in front of the left side of the screen, which is the starting position. In the beginning, cars run from the left to the right and the arrows are not active. After 10 seconds, the road is “veiled”. Thus, no car can be seen. Then, subjects have to decide where the pedestrian will cross bearing in mind the fact that the closer to the left corner of the screen they decide to cross, the less they will be able to see if a car is coming. To the contrary, the further they are from the left corner, the longer they will spend. It is assumed that participants are greater risk-takers the closer to the left side of the screen that they decide to cross. Subjects have 10 trials and they are not informed about any sort of accident they might have produced. Risk score is equal to the average of the distance from the right side of the screen in a horizontal axis. PLEASE INSERT FIGURES 1, 2, 3 The tests have shown satisfactory internal consistency as well as an acceptable temporal stability, as shown in Table 1. PLEASE INSERT TABLE 1 3. Empirical validation of the tests

4


Three pieces of research have been carried out in order to determine the validity of the tests proposed. Firstly, 233 university graduates were assessed using the three tests in order to establish the relationship among them (Rubio, Contreras, Hernandez & Santacreu, submitted). The results show that the correlation between the two betting tests were r = .787 (p < 0.01). The correlation between the CtSTand the BDT and the RT were r = .569 and r = .508, respectively (p < 0.01). These results support the idea of the convergent validity of the tests. Secondly, a broad study was carried out in order to test the feasibility of the risk tests to discriminate the guessing tendency individuals show in a multiple option test as a criterion for validating the instruments (Rubio, Hernandez & Santacreu, submitted). 4,966 applicants to an ab initio air traffic control (ATC) training course were assessed using a test battery which included, among others, the RT and the CtST, as well as a four-option test about basic knowledge of ATC. The applicants were informed that the score of this test was going to be the sum of the right responses minus 1/3rd of the errors. The guessing tendency was estimated according to the number of errors and omissions that individuals showed in the multipleoption test. Subjects were classified in five groups from highly conservatives to high risk takers, according to the risk test scores. A discriminant analysis was carried out using the two risk tests as discriminant variables and the guessing tendency scores as the classification variable. Results showed a significant discrimination among groups for both variables. Finally, a predictive validation study was carried out (Santacreu, et al., in press). For this, 6,123 applicants for an ab initio ATC training course were assessed using a test battery which included the BDT and the CtST as well as tests for assessing another personality and aptitude dimensions. 95 applicants were selected for the ATC training program which consisted of five modules: Basic Knowledge (BK), Area Control Unit (ACU), Approach Control Unit (ApCU), Aerodrome Control Unit (AeCU), and No-ATS unit (NATS). The final scores of these five modules were used as criteria to test the predictive validity of the

5


instruments. Results showed that the range restriction corrected R of the complete test battery was reasonably high for the five modules. Regarding the risk tests, the BDT showed significant standardized β for ACU and ApCU. The CtST showed significant coefficients for ACU, ApCU, AeCU, and NATS. The corrected correlations between the BDT and CtST scores, and the training program module scores are shown in Table 2. PLEASE INSERT TABLE 2 4. Discussion Risk tendency looks at a construct which awakens more interest than agreement. The problem of how to measure it is directly related to how it is understood. In this chapter, we have assumed that risk refers to a situation in which the person has to choose from different alternatives, each one having potential losses or gains characterized by their uncertainty. It is thought that a person who opts for a less probable alternative than the others would be called riskier. From this point of view, can we identify whether one person is more prone to a risk decision than another? Moreover, can we predict who will behave as a risk taker in future situations? To answer this question demands previously looking into whether or not risk contexts are equivalent. In other words, is there a cross-situational consistency in RTB? Research does not provide conclusive evidence about that, though we can find classes of behavior which tend to covariate. One explanation could be that situations impose most of the variance of behavior. Thus, an individual who usually makes risky business decisions would not bet on a lottery. Some authors (Santacreu, Rubio & Hernandez, 2004) have suggested that personality might be responsible for individuals’ behavior if the situation does not impose a particular “correct” response (for instance, if the others expect someone to behave as riskier in business or if the previous results when doing so had been highly profitable). The assessment of the idiosyncratic way a person has when facing an uncertain situation demands a context in

6


which individuals do not have any experience or any particular response is particularly reinforced. 5. Final appraisal Despite what has been said, RTB has been mainly assessed by means of questionnaires, like almost all personality dimensions. In them, people are asked about how prone they are to be involved in activities associated to risk or to what extent they observe speed restrictions when driving. At most, people are asked about how many traffic violations they have committed in the past month or if they have been involved in a car crash. This way of assessment presents serious restrictions. Firstly, self-report measures are subjected on voluntary and non-voluntary biases (Kubinger, 2002;, Viswesvaran & Ones, 1999). On the other hand, there are limitations in predicting future behavior. This is particularly true, the broader the personality trait is (Paunonen, Haddock, Forsterling & Keinonen, 2003). The individuals’ self-report based assessment of personality dimensions contrasts with the important RTB research tradition of using task-based tests. Based on this tradition and on the idea that what people say about themselves does not necessarily coincide with what they really do, the authors have developed a set of objective task-based tests for assessing RTB that have shown their feasibility as well as adequate psychometric properties which allow us to propose them as an alternative to the traditional assessment of this construct.

7


References Arend, I., Botella, J., Contreras, M.J., Hernández, J.M. & Santacreu, J. (2003). A betting dice test to study the interactive style of risk-taking behavior. The Psychological Record, 53, 217-230. Atkinson, J.W. (1957). Motivational determinants of risk-taking behavior. Psychological Review, 64, 359-372. Atkinson, J.W. & Litwin, G.H. (1960). Achievement motive and test anxiety conceived as motive to approach success and motive to avoid failure. Journal of Abnormal and Social Psychology, 60, 52-63. Bromiley, P. & Curley, S.P. (1992). Individual differences in risk taking. In J.F. Yates (Ed.), Risk-taking behaviour. Chichester, UK: John Wiley. Cattell, R.B. (1958). What is “objective” in “objective personality tests”? Journal of Consulting Psychology, 5, 285-289. Cattell, R.B. & Warburton, F.W. (1967). Objective Personality and Motivation Tests. Urbana: University of Illinois Press. Hernández, J.M. Rubio, V.J., Revuelta, J. & Santacreu, J. (in press). A procedure for estimating intra-subject consistency of behavior. Educational and Psychological Measurement. Kogan, N. & Wallach, M.A. (1960). Certainty of judgment and the evaluation of risk. Psychological Report, 6, 207-213. Kogan, N. & Wallach, M.A. (1964). Risk taking. A study in cognition and personality. New York: Holt, Rinehart and Winston. Kubinger, K.D. (2002). On faking personality inventories. Psychology Science, 44, 10-16.

8


Lopes, L.L. (1987). Between hope and fear: The psychology of risk. Advances in Experimental Social Psychology, 20, 255-295. MacCrimmon, K.R. & Wehrung, D.A. (1986). Taking risks. The management of uncertainty. New York: Free Press. Nicholson, N. Soane, E., Fenton-O’Creevy, M. & Willman, P. (2005). Personality and domain-specific risk taking. Journal of Risk Research, 8, 157-176. Paunonen, S.V., Haddock, G., Forsterling, F. & Keinonen, M. (2003). Broad versus narrow personality measures and the prediction of behaviour across cultures. European Journal of Personality, 17, 413-433. Rubio, V.J., Contreras, M.J., Hernandez, J.M. & Santacreu, J. (submitted). The assessment of the interactive style risk-tendency. Spanish Journal of Psychology. Rubio, V.J., Hernandez, J.M. & Santacreu, J. (submitted). Can we predict risk taking behaviour? Two tasks for predicting guessing tendencies in a multiple option task. Personality and Individual Differences. Rubio, V.J., Santacreu, J. & Hernández, J.M. (2004). The objective assessment of personality: An alternative to self-report based assessment. Análisis y Modificación de Conducta, 30, 827-840. Santacreu, J., Rubio, V.J. & Hernandez, J.M. (2004). Evaluación objetiva de la personalidad. Análisis y Modificación de Conducta, 30, 803-825. Santacreu, J., Rubio, V.J. & Hernández, J.M. (in press). The objective assessment of personality: Cattell’s T-data revisited and more. Psychology Science. Santé, L. & Santacreu, J. (2001): La eficacia (o la suerte) como moduladora en la evaluación del estilo interactivo: Tendencia al riesgo. Acta Comportamentalia, 9, 163-168.

9


Schneider, S.L. & Lopes, L.L. (1986). Reflection in preferences under risk: Who and when may suggest why. Journal of Experimental Psychology: Human Perception and Performance, 12, 535-548. Schoemaker, P.J.H. (1990). Are risk preferences related across payoff domains and response modes? Management Science, 36, 1451-1463. Slovic, P. (1964). Assessment of risk taking behavior. Psychological Bulletin, 61, 330-333. Soane, E. & Chmiel, N. (2005). Are risk preferences consistent? The influence of decision domain and personality. Personality and Individual Differences, 38, 1781-1791. Viswesvaran, C. & Ones, D.S. (1999). Meta-analyses of fakability estimates: Implications for personality measurement. Educational & Psychological Measurement, 59, 197-210. Wagenaar, W.A. (1988). Paradoxes of gambling behavior. London: Erlbaum. Wallach, M.A., Kogan, N. & Bem, D.J. (1962). Group influence on individual risk taking. Journal of Abnormal and Social Psychology, 65, 75-86. Weber, E.U. & Milliman, R. (1997). Perceived risk attitudes: relating risk perceptions to risky choice. Management Science, 43, 122-143. Weber, E.U., Bleis, A.R. & Betz, N. (2002). A domain-specific risk attitude scale: Measuring risk perceptions and risk behaviours. Journal of Behavioral Decision Making, 15, 263-290 Yates, J.F. & Stone, E.R. (1992). The risk construct. In J.F. Yates (Ed.), Risk-taking behaviour. Chichester, UK: John Wiley. Zuckerman, M. (1979). Sensation seeking: Beyond the optimal level of arousal. Hillsdale, NJ: Erlbaum. Zuckerman, M. (1979b). Trait, states, situations, and uncertainty. Journal of Behavioral Assessment, 1, 43-54.

10


Zuckerman, M., Eysenck, S. B. G., & Eysenck, H. J. (1978). Sensation seeking in England and America: Cross-cultural, age, and sex comparisons. Journal of Consulting and Clinical Psychology, 46,139-149.

11


Table 1: Internal consistency and temporal stability of the risk-taking objective tests

Cronbach’s α Test-retest

BDT

RT

CtST

.835

.815

.966

.600 (one day)

.430 (one year)

Not available

BDT: Betting Dice Test. RT: Roulette Test. CtST: Crossing the Street Test

Table 2: Corrected correlations between Risk-taking tests and the Air Traffic Control training course scores Training Course units Basic Knowledge Area Control Approach Control Aerodrome Control No-ATS

BDT -.229 -.295 -.165 -.089 -.078

CtST -.200 -.082 -.403 -.193 -.153

BDT: Betting Dice Test. RT: Roulette Test. CtST: Crossing the Street Test

12


Figure 1: The test screen of the Betting Dice Test

13


Figure 2: The test screen of the Roulette Test

14


Figure 3: The test screen of the Crossing the Street Test

15

OBJECTIVE ASSESSMENT OF RISK TENDENCY21_EN_REV

Recommend Documents