Dorien Bangma

FDM AND NORMAL AGING | 47 three) will be performed to determine the effects of age on these measures of cognition. For both step two and three, we controlled for demographic variables (i.e., gender, years of education, employment status and annual year income) and a Bonferroni correction was applied to control for alpha-error growth in multiple testing (i.e., p < .0038 and p < .017 were considered significant for step two and three, respectively). When significant results were found in step two and three, additional hierarchical regression analyses (step four) were performed to investigate the effects of age and FDM while controlling for demographic variables and standard measures of cognition. A significant reduction in the relation between age and FDM, as determined by the Sobel test, is indicative of a mediating effect of cognition. Finally, correlation analyses (Pearson) were used to determine the coherence between the FDM tests and to further explore the relation between FDM tests and standard neuropsychological measures. Bonferroni correction was applied to control for alpha-errors and p < .001 was considered significant. Results Age effects on FDM No significant effects of age were found on the total scores of the FCAI-NL and FDMI and on the application of financial decision-making styles (FDS; Table 3.2). However, with regard to the FCAI-NL years of education ( b = 0.79 [0.35, 1.23], p = .001) as well as gross annual year income ( b = 1.4 [0.19, 2.60], p = .023) were found to be significant predictors. Number of years of education was also a significant predictor of the total score of the FDMI ( b = 0.11 [0.04, 0.19], p = .003) and years of education and gross annual year income were both significant predictors of the ‘intuitive style’ of the FDS ( b = -0.167 [-0.30, -0.03], p = .015 and b = -0.432 [-0.81, -0.06], p = .025, respectively). None of the other FDS styles (i.e., rational, dependent, avoidant and spontaneous) could be predicted by any of the control variables. Furthermore, no significant effects of age, or of any of the control variables, were found on the FDM-I/D, neither for deliberative decisions nor for intuitive decisions (Table 3.2). The number of years of education significantly predicted the CDR performance ( b = 0.17 [0.98, 0.25], p = < .001). However, age was found to be a significant additional predictor of the performance on the CDR, explaining 15.1% of variance; a finding that is supported by both the internal and external validation analyses (Table 3.2). Negative effects of age were also found on the IGT. The total netscore of the IGT was significantly predicted by age, explaining 7.1% of variance. Specifically, the second, fourth and fifth trial of the IGT were significantly predicted by age (Table 3.2). No significant effects of age were found for the first and third trial of the IGT and the control variables did not predict the performance on the IGT. The internal validation analyses partly supported these results. However, the results could not be replicated in the external validation analyses (Table 3.2). Gender and gross annual year income were found to explain 27.2% of variance of scores on the IBQ ( b = -5.55 [-8.32, -2.78], p < .001 and b = 1.20 [0.15, 2.25], p = .026, respectively). Women had lower IBQ scores, indicating a stronger impulsive buying tendency, than men. However,