Resumo:
Propsito do Trabalho:
The purpose of this paper is to assess the accountants� cognitive abilities and identify which demographic variables are associated with such a performance. In order to reach this, we used a simple three item test for cognitive abilities developed by Frederick (2005) - CRT - on a sample of 11,719 Brazilian accountants to assess their tendency to override a preponderant response alternative that is incorrect and to engage in further reflection that leads to the correct response. Hence, we investigated which demographic characteristics are the most associated with a more reflective (or intuitive) behavior.
Frederick (2005); Toplak, West and Stanovich (2011) and other authors form the Cognitive Science literature have shown evidences that the CRT is a more potent predictor of performance on a wide sample of tasks from the heuristics-and-biases literature than measures of cognitive ability, thinking dispositions, and executive functioning. Therefore, in this paper, we rely exclusively on the CRT to assess accountants' cognitive abilities.
Then, CRT is recognized as a predictor of judgment and decision making, that is of special interest in the context of the adoption of principles-oriented accounting standards, such as the IFRS, because the interpretation of such principles requires more intensive cognitive efforts than rules-based accounting standards. In the absence of any relevant reference or empirical evidence to support such a statement, we present three anecdotal evidences that compare the accounting policies generally accepted in Brazil until 2010 with those adopted as a consequence of the implementation of the IFRSs.
Base da plataforma terica:
Psychology and Cognitive Science are mature fields of study, and their interactions with Finance and Economics are also mature (see FREDERICK, 2005; KAHNEMAN, 2011). There are also applications of the psychological research into law (KOROBKIN & ULEN, 2000), and accounting (BELKAOUI, 2004; ASHTON; ASHTON, 1995; BONNER, 2008).
Behavioral Accounting (BA) �emphasizes the relevance of accounting information to decision making as well as the individual and group behavior caused by the communication of this information. Accounting is assumed to be action oriented; that is, its purpose is to influence action, or behavior, directly through the informational content of the message conveyed and indirectly through the behavior of accountants. Because accounting is considered a behavioral process, behavioral accounting can be described as the application of behavioral science to Accounting� (BELKAOUI, 1989, p. xi). Consequently, BA is open for many different researching methods, for example: interviews (BEATTIE et al, 2001; BEATTIE et al, 2012), surveys (GOVINDARAJAN, 1984; HARTMANN; MASS, 2011), and experiments (LIBBY, 1975; ROSMAN, 2011).
On the other hand, Judgment and Decision Making (JDM) research, a sub-field of BA, focus on the behavior of individuals, and intensively uses surveys and experiments, where individuals are the respondents or subjects.
Indeed, we used Frederick�s cognitive reflection test (CRT) to assess the cognitive abilities of Brazilian accountants.
CRT is criteria to measure how impulsively or more reflectively people make decisions. The CRT consists of three questions to which the intuitive (i.e., impulsive and spontaneous) answer is wrong. Correct answers indicate a higher degree of reflectivity and deliberate thinking. The quintessential item from the CRT was first discussed by Kahneman and Frederick (2002) in an article that reframed the heuristics-and-biases literature in terms of the concept of attribute substitution.
The intuitive answers are attributed to the �system 1� process of decision making, because it is the first answer that respondents� cognition suggests; therefore, system 1 is considered �fast�. If such an answers is not identified being wrong, the �system 2� is not activated. However, if respondent reflectively tests whether the first answer is wrong and deliberately think about the problem again, she might find the correct answer; therefore, system 2 is considered �slow� (TOPLAK; WEST; STANOVICH, 2011).
Mtodo de investigao:
Data were collected via a longer questionnaire developed by some of the coauthors of this article in behalf of the Brazilian Association of Accountants (Conselho Federal de Contabilidade, hereafter CFC). Such a longer questionnaire was entitled �Professional Accountants Profile, edition 2012/13� (CFC, 2013).
A total of 12,578 accountants answered the questionnaire. As a check against respondents answering randomly, we eliminated the observations of respondents who answered unreasonable answers for the three CRT questions, and who did not answer their age or income. After eliminations, the analyzed sample was comprised by 11,719 accountants.
To assess the relationship among dependent variable (sum of the Cognitive Reflection Test scores) and independent variables (gender, age, register type, formal education level and monthly income) the most suitable technic of multivariate data analysis seems to be ordered logistic regression, in which, for each unit increased in a given independent variable, the CRT score (dependent variable) is expected to change by its respective coefficient, if the other variables are held constant. On the other hand, it provides another interpretation approach through the odds ratio , which is especially useful if the independent variable is a dummy. In our specific case, the probability of a higher cognitive score (3) versus lower (0) is odds ratio value times higher (lower) for the group that assumes value 1 (one) than for the other. For instance, if the odds ratio result for gender is 0.5, then the probability of an individual owning a high CRT score (equal to 3) is 0.5 times lower if it is a woman (according to Table 5, if participant is woman the gender variable assumes the value 1, as described in Table 3), while the other variables are held constant. Still, if the odds ratio result for gender is 1.5, then the probability of an individual owning a high CRT score (3) is 1.5 times higher (not lower) if it is a woman.
Resultados, concluses e suas implicaes:
Findings support the Cognitive Science literature, that higher test scores on the CRT are correlated with gender (male) and age (young). Additionally, we identified that high CRT scores are indeed correlated with the educational level (post-graduate), income (high-income, or well-succeeded), and the type of accountants registration at the CFC (bachelor accountants, not technicians). Indeed, either based on the analysis of proportions and on the ordered logistic regression�s odds ratios, the results suggest that (i) young, (ii) male, (iii) post-graduated, (iv) high-income and (v) those professionals registered at CFC as bachelor accountants make more reflective decisions than (i) mature, (ii) female, (iii) not bachelor, (iv) low-income and (v) professionals registered as technicians. According to Kahneman and Frederick (2002), Frederick (2005) and Kahneman (2011), intuitive answers are attributed to the fast �system 1�, while reflective answers are attributed to the slow �system 2�. Individuals that are more intuitive use to rely more frequently on the answers provided by system 1. Therefore, the results show that mature, female, low-income professionals registered as technicians tend to be more confident on their system 1, than the other group of accountants.
Results also show that Brazilian professional accountants scored lower than college students from Canada (TOPLAK; WEST; STANOVICH, 2011), Germany (OECHSSLER; ROIDER; SCHMITZ, 2009; HOPPE; KRUSTERER, 2011) and the U.S.A. (FREDERICK, 2005).
The findings invite accounting researchers to further investigate the impacts of such a behavior on judgment and decision making in accounting related issues, such as the interpretation of accounting standards and its implementation on the recognition, classification and measurement of transactions. Indeed, further research on judgment and decision making in accounting could investigate if the CRT score is associated with heuristic-bias and specific decision in the accounting context. For example: (i) the interpretation of verbal probabilities (DOUPNIK; RICHTER, 2003 and 2004; DOUPNIK; RICCIO, 2006) is associated with the accountants� cognitive abilities?; (ii) the classification of items in accordance with vague standards (PENNO, 2008; CARDOSO; AQUINO, 2010) is associated with the accountants� cognitive abilities?; (iii) the identification of impression management (MERKL-DAVIES; BRENNAN, 2011; JONES, 2011) is associated with the accountants� cognitive abilities?.
Further international research could investigate if the findings presented in this paper are similar to those that could be obtained in other jurisdictions, especially in countries which environment are similar to the Brazilian environment. For such an assessment, the frameworks presented by Gray (1988) or Nobes (1998) could be of great use.
Researchers interested in �Accounting and Gender� could investigate the association of CRT scores and the fact that women hired into public accounting were not progressing up the ranks as expected (DALTON; HILL; RAMSAY, 1997).
Researchers interested in �Accounting Education� could further investigate the association of CRT scores and the formal education; and those interested in �Accounting Career� could investigate the association of CRT scores and individual�s income.
Additionally, researchers not interested in Accounting, but in Cognitive Sciences, could analyze the impact of respondents� recruitment strategy on the CRT scores. Notice that, different from the majority of the researches that assessed CRT score recruited college students and offered them a monetary reward, we recruited professional accountants and did not offer them any monetary reward. The CRT scores reported in this study are lower than the score presented in the reviewed literature.
Referncias bibliogrficas:
HOPPE, Eva I.; KUSTERER, David J. Behavioral biases and cognitive reflection. Economic Letters, v. 110, p.97-100, 2011.
FREDERICK, Shane. Cognitive Reflection and Decision Making. Journal of Economic Perspectives, v. 19 (1), p. 25-42, Fall 2005.
KAHNEMAN, Daniel. Thinking, fast and slow. New York: Farrar, Straus and Giroux. 499pp. 2011.
KAHNEMAN, Daniel; FREDERICK, Shane. Representativeness revisited: attribute substitution in intuitive judgment. In T. Gilovich, D. Griffin & D. Kahneman (Eds.), Heuristics and biases: The psychology of intuitive judgment. (pp. 49-81). New York: Cambridge University Press. 2002.
OECHSSLER, J�rg; ROIDER, Andreas; SCHMITZ, Patrick W. Cognitive abilities and behavioral biases. Journal of Economic Behavior & Organization. v. 72, p. 147-152, 2009.
SILVA, Jarbas S. Se um taco e uma bola custam R$110, e o taco custa R$100 a mais do que a bola, quanto custa a bola? Thesis. Master in Administration, Funda��o Getulio Vargas. Rio de Janeiro: FGV. 72 p. 2005.
TOPLAK, Maggie, E.; WEST, Richard F.; STANOVICH, Keith E. The cognitive reflection test as a predictor of performance on heuristics-and-biases tasks. Memory & Cognition. v.39 (7), p.1275-1289. 2011.
|
