- Email: [email protected]
Does impulsivity influence performance in reasoning?
Personality and Individual Differences 33 (2002) 1031–1043 www.elsevier.com/locate/paid
Does impulsivity influence performance in reasoning? Karl Schweizer* Abt. Methodenlehre, Evaluation und Forschungsmethoden, Psychologisches Institut, Johann Wolfgang Goethe-Universita¨t, Mertonstr. 17, 60054 Frankfurt a. M., Germany Received 19 July 2001; received in revised form 30 October 2001; accepted 30 November 2001
Abstract The influence of impulsivity on reasoning was investigated in a sample of 108 high school and university students. Impulsivity was represented by the PRF Impulsivity Scale [Jackson, D. N. (1974). Manual for the Personality Research Form (2nd ed.). Goshen: Research Psychologists Press.], the MMPI Impulsivity Scale [Gough, H. G. (1957). California psychological inventory manual. Palo Alto: Consulting Psychologists Press.] and the FPI Impulsiveness Scale [Fahrenberg, J., Hampel, R., & Selg, H. (1994). Das Freiburger Perso¨nlichkeitsinventar FPI. 6. Aufl. Go¨ttingen: Hogrefe.]. Reasoning was measured by the Figural Reasoning Scale (Horn, W., 1983) and the Numeric/Alphabetical Reasoning Scale [Horn, W. (1983). Leistungs-Pru¨fSystem. Go¨ttingen: Hogrefe.]. Furthermore, measures of neuroticism, extraversion, achievement orientation and surgency were applied. Substantial negative correlations of reasoning with impulsivity and neuroticism were observed. Structural equation modeling was applied to predict reasoning by means of an impulsivity composite. A path coefficient of 0.33 was observed. This result suggests the interpretation that a high degree of impulsivity impairs performance in completing reasoning tasks. # 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction Impulsivity is a trait which is ascribed to persons a major characteristic of which is that their behavior is to a considerable degree guided by impulses. Persons high in impulsivity prefer the immediate over the well-reasoned response. In such persons new stimuli are rather likely to interrupt the execution of well-conceived behavioral plans which are unrelated to the new stimuli. Although impulsive behavior seems to be easily identifiable, impulsivity is described rather differently. In their attempt to clarify the meaning of impulsivity Depue and Collins (1999) found that it is characterized as including various lower-order traits such as sensation seeking, novelty seeking, boldness, adventuresomeness, boredom susceptibility, unreliability, and unorderliness. Furthermore, in the Diagnostic and Statistical Manual for the Mental Disorders (DSM-IV; * Corresponding author. Tel.: +49-69-798-22894; fax: +49-69-798-23847. E-mail address: [email protected] (K. Schweizer). 0191-8869/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0191-8869(01)00209-4
1032
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
American Psychiatric Association, 1994) impulsivity is associated with various disorders such as impulse-control disorders, antisocial personality disorder, mania, and dementia. A look at the various models of the structure of personality can help to obtain a better image of impulsivity. The 1977 version of Eysenck and Eysenck’s (1977) three factor theory of personality includes four specific dimensions of impulsivity: narrow impulsiveness, risk-taking, non-planning, and liveliness. The first of these dimensions was shown to be associated with neuroticism and the other ones with extraversion. Irrespective of these dimensions, impulsivity led to a robust factor in an attempt to construct a single scale (S. B. G. Eysenck, Pearson, Easting, & Allsopp, 1985) which even showed cross-cultural consistency (S. B. G. Eysenck, Daum, Schugens, & Diehl, 1990). In Cloninger’s model of personality which includes four temperament dimensions impulsivity is an aspect of novelty seeking (Cloninger, Svrakic, & Przybeck, 1993). According to Cloninger impulsivity is the biology-based tendency to respond fast to novel stimuli. The three factor model of personality by Tellegen (1982) also includes the aspect of impulsivity. This model is constituted by the factors positive emotionality, negative emotionality and constraint. A low degree of constraint is closely associated with a high degree of impulsivity. Finally, the Five Factor Model of Personality is to be considered. Costa and McCrae (1992) note that the neuroticism and conscientiousness factors include impulsivity facets. The various representations of impulsivity led Whiteside and Lynam (2001) to investigate the structure of impulsivity. They applied a large number of impulsivity scales to a sample of over 400 young adults and identified four impulsivity factors by means of exploratory factor analysis. They labeled these factors urgency, (lack of) premeditation, (lack of) perseverance, and sensation seeking. Most of the intercorrelations between the corresponding scores were low to moderate indicating a considerable degree of heterogeneity among the facets of impulsivity. Thus, the conceptual variety associated with the various models of personality finds its explanation. Impulsivity is also regarded as a cognitive style (Kagan, 1965; Kagan, Moss, & Sigel, 1963) showing two extremes one of which is denoted impulsive style and the other one reflective style. The impulsive style is characterized by fast decisions of which the certainty of being correct is low whereas the reflective style denotes the preference for late decisions associated with a high degree of certainty. The cognitive style has been assessed by means of tests which require the detection of differences between similar figures, and it has served well in describing children’s behavior (Messer, 1976). The investigation of impulsive behavior in children is especially attractive because the children’s behavior reflects social norms to a lower degree than the adults’ behavior. Furthermore, it needs to be emphasized that impulsivity is considered as a temperament (Buss & Plomin, 1975; Cloninger et al., 1993). A temperament is defined as a characteristic style of behaving, which is hereditary (Allport, 1961). However, evidence concerning heredity of impulsivity is not unequivocal. Pedersen, Plomin, McClearn, and Friberg (1988) provided positive evidence whereas Buss and Plomin (1984) did not. Additional support for the assumption of heredity results from the observation of impulsive behavior in children since children’s behavior is less determined by experiences than the adults’ behavior. 1.1. Possible consequences of impulsivity for reasoning Since impulsivity tends to interrupt the execution of well-conceived behavioral plans, a disadvantageous influence on performance can be expected when the demands are complex.
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1033
Reasoning is a prominent instance of a complex mental activity. Therefore, a disadvantageous influence of impulsivity on performance in reasoning tasks and related tasks can be hypothesized. Impulsive behavior can cause inappropriate responding in completing reasoning tasks, i.e. responding exceptionally fast and without a sufficient degree of certainty. Accordingly, impulsivity was supposed to favor the selection of cognitive associations which are easily available over other associations (Stott, 1985) and also the neglect of available information (Brebner & Stough, 1995). Such behavioral effects should be especially disadvantageous in reasoning since reasoning usually requires the execution of many cognitive operations which are to be performed according to a specific algorithm. The exclusive concentration on the task, combined with the suppression of other behavioral impulses, is advantageous for the successful completion of a reasoning task if it can be sustained for a sufficiently long period of time. The processes which are associated with the completion of reasoning tasks interact with each other and, therefore, are described as mechanisms (Schweizer, 2000, 2001). The successful completion of a reasoning task is likely when all the operations of the algorithm are performed appropriately and within the time limit (Salthouse, 1993, 1996). Deviations from the algorithm reduce the likelihood of the successful completion of the task in impulsive persons and, thus, contributes to the other sources of possible failure. Apparently, in persons showing a high degree of impulsivity the aforementioned reasons provide a disadvantage. 1.2. Evidence concerning the assumed influence of impulsivity on reasoning Despite the reasonableness of all these considerations, the correlations reported in a study by Furnham, Forde, and Cotter (1998) were only of small to moderate size. In this study impulsivity was represented by the corresponding scale of the Eysenck Personality Profiles (Eysenck, Barrett, Wilson, & Jackson, 1992). These findings are contrasted by the results of a study by Phillips and Rabbitt (1995) who reported insignificant results only. In this study a behavioral measure was applied for the assessment of impulsivity instead of a personality measure. Consequently, the difference may be due to the different types of measure. The personality measure is based on longterm experiences whereas the behavioral measure reflects momentary behavior only. Impulsivity has not been given much consideration in the investigation of the relationship between personality and intelligence. However, since it has been regarded as a component of neuroticism (e.g. Eysenck et al., 1992) and reasoning is considered as a component of intelligence (see Snyderman & Rothman, 1987) there is the possibility to deduce a hypothesis concerning the relationship between impulsivity and reasoning from the findings of studies investigating the relationship between neuroticism and intelligence. The available evidence suggests the existence of a small to moderate negative relationship between neuroticism and intelligence. In a meta-analytic study by Ackerman and Heggestad (1997) weak but significant negative relationships between sales interpretable as components of neuroticism and various measures of intelligence are reported. Further indirect evidence is provided by a more recent study which reports the data of 30,000 persons (Austin, Hofer, Deary, & Eber, 2000). In this study neuroticism is shown to influence the relationship between different ability measures. There are also some studies relating impulsivity, intelligence and arousal level (e.g. Gupta, 1988; Humphreys & Revelle, 1984; Matthews, 1987; Revelle, Humphreys, Simon, & Gilliland, 1980). The authors of these studies expect an influence of the arousal level on the relationship between
1034
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
impulsivity and intelligence. These studies even provide evidence that caffeine administration and time of day have an effect on the relationship although the results are not consistent. Despite these findings, Brebner and Stough (1995) still regard these studies as insufficient for unravelling the theoretical connections between impulsivity, intelligence and arousal level. 1.3. Aims of the study The investigation of the relationship between impulsivity and reasoning in a more comprehensive way than in the previous studies was the main aim of this study. It was expected that impulsivity and reasoning correlate substantially. Three different scales which were expected to measure impulsivity were selected for the new study. The set of these scales could be assumed to provide a better representation of impulsivity than the individual scales included in other studies. However, these scales were not expected to represent all the factors of the Whiteside and Lynam (2001) model of impulsivity since only a subset of factors could be regarded as important with respect to reasoning. Furthermore, extraversion and neuroticism which were assumed to include facets of impulsivity were considered. Two further personality scales were added in order to have a more complete description of personality.
2. Method 2.1. Participants The sample of this study included 108 participants (82 women and 26 men). The mean age was 22.78 years with a standard deviation of 4.81. The lower limit for participation was 17 years of age. The participants were high school and university students. 2.2. Measures 2.2.1. Reasoning scales Reasoning was assessed by means of the two reasoning scales included in the Leistungs-Pru¨fSystem (LPS; Horn, 1983). The first scale required figural reasoning and the second scale numeric/alphabetical reasoning. Accordingly, the first scale is denoted Figural Reasoning Scale and the second one Numeric/Alphabetical Reasoning Scale. In completing the items of these scales the individuals had to identify the inherent logic of a series of figures, numbers or letters and apply this logic inductively. Each series was constructed systematically with one exception which had to be identified (e.g. aa aa aa a aa). The correlation between the two reasoning scales was 0.62 in the sample of this study. For some of the statistical investigations reported in the following sections the scores obtained by means of the two scales were combined in order achieve an overall reasoning score. 2.2.2. Impulsivity scales Impulsivity was measured by three scales: the impulsivity Scale of the Personality Research Form (PRF; Jackson, 1974; German version: Stumpf, Angleitner, Wieck, Jackson, & Beloch-Till,
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1035
1985), the impulsivity Scale (Gough, 1957) which was based on the items of the Minnesota Multiphasic Personality Inventory (MMPI; Hathaway & McKinley, 1951; German version: Spreen, 1963), and the impulsiveness scale of the Freiburger Perso¨nlichkeitsinventar (FPI; Fahrenberg, Hampel, & Selg, 1994). These scales included 16 (PRF), 20 (MMPI) and 12 (FPI) items. These scales are denoted PRF Impulsivity scale, MMPI Impulsivity Scale and FPI Impulsiveness Scale respectively. All the items included a dichotomous response format. 2.2.3. Neuroticism scale The emotionality scale of the Freiburger Perso¨nlichkeitsinventar (Fahrenberg et al., 1994) was selected for this purpose since this scale was originally constructed to represent the neuroticism dimension of the Eysencks’ PEN system (Eysenck & Eysenck, 1985). It is addressed FPI Neuroticism Scale. 2.2.4. Other personality scales Three further personality scales were selected for this study: the surgency scale of the 16 Personality Factors questionnaire (16PF; Cattell, Eber, & Tatsuoka, 1970; German version: Schneewind, Schro¨der, & Cattell, 1994), the extraversion and achievement orientation scales of the Freiburger Perso¨nlichkeitsinventar (Fahrenberg et al., 1994). The surgency scale was selected because of a large correlation ( 0.41) which was reported between this scale and the PRF impulsivity scale (Stumpf et al., 1985). The extraversion scale was considered because it was found to correlate with intelligence in other studies (Ackerman & Heggestad, 1997) and the achievement orientation scale because of its content which suggested a positive relationship with reasoning. Furthermore, the scales allowed us to obtain a more comprehensive description of personality. These scales are denoted 16PF Surgency Scale, FPI Extraversion Scale and FPI Achievement Orientation Scale. 2.3. Procedure The reasoning tests and personality scales were applied in the following order: Figural Reasoning Scale, Numeric/Alphabetical Reasoning Scale, PRF Impulsivity Scale, 16PF Surgency Scale, FPI scales, MMPI Impulsivity Scale. The items of the FPI scales were included into one item list. The participants were tested in small groups.
3. Results The means and standard deviations of the raw scores of the scales to measure reasoning and personality traits are included in Table 1. The means varied between 6.44 and 64.10 and the standard deviations between 2.51 and 7.44. The differences between the means reflected the differences between the numbers of items included in the individual scales. The reasoning scores were quite large. This was obvious from the upper limits of the scales, which were 40 in both cases. However, these large scores were not unusual since the participants showed a high educational level. The other statistics did not indicate that the sample differed considerably from the population.
1036
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
3.1. Effects of sex and age The reasoning and personality scores were investigated for effects due to sex or age by means of computing Pearson correlations. A correlation of 0.22 (P<0.05) was observed between sex and the Numeric/Alphabetical Reasoning Scale. This meant that being male was associated with a slightly higher average reasoning scores than being female. All the other correlations with the reasoning scales were close to zero. Neither one of the impulsivity scales was correlated with sex or age, nor the FPI Neuroticism Scale. The 16PF Surgency Scale was correlated with age (r= 0.24, P<0.05) whereas all the other personality scales were not related to sex or age. These results did not give reason to suspect that effects due to either sex or age would influence the relationships between the reasoning and trait scores. 3.2. The relationships of reasoning and personality traits The Pearson correlations between the reasoning scales and the trait scales are presented in Table 2.
Table 1 Means and standard deviations for the raw scores of the Reasoning and Personality Scales Scale
Mean
SD
Reasoning (overall) Figural Reasoning Scale Numeric/Alphabetical Reasoning Scale PRF Impulsivity Scale MMPI Impulsivity Scale FPI Impulsiveness Scale FPI Neuroticism Scale 16PF Surgency Scale FPI Extraversion Scale FPI Achievement Orientation Scale
64.10 31.59 32.51 9.34 8.67 6.86 7.16 10.93 6.74 6.44
7.44 4.35 3.92 3.05 3.79 3.26 3.45 3.53 3.44 2.51
Figural reasoning
Numeric/alphabetical reasoning
Table 2 Pearson correlations between Reasoning Scales and Personality Scales Personality Scales PRF Impulsivity Scale MMPI Impulsivity Scale FPI Impulsiveness Scale FPI Neuroticism Scale 16PF Surgency Scale FPI Extraversion Scale FPI Achievement Orientation Scale * P< 0.05 ** P< 0.01
Reasoning 0.215* 0.321** 0.132 0.300** 0.021 0.045 0.105
0.199* 0.283** 0.134 0.289** 0.011 0.025 0.087
0.187 0.295** 0.103 0.248** 0.053 0.058 0.103
1037
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
All the correlations of the reasoning and impulsivity scales were negative. Five of the nine correlations concerning this relationship reached the level of significance. All the correlations of the MMPI Impulsivity Scale and reasoning were significant. The PRF Impulsivity Scale showed two substantial correlations with reasoning. In contrast, neither one of the three correlations between the FPI Impulsiveness Scale and reasoning were significant. Furthermore, the FPI Neuroticism scale correlated substantially with all the reasoning scales. In contrast, neither one of the further personality scales, 16PF Surgency Scale, FPI Extraversion Scale, and FPI Achievement Orientation Scale, led to a substantial correlation with reasoning. All these further correlations were close to zero. 3.3. The relationships among the neuroticism and impulsivity scales The heterogeneity of the results observed for the impulsivity scales suggested the investigation of the relationships among the impulsivity scales. The FPI Neuroticism Scale was additionally considered since neuroticism was assumed to include facets of impulsivity as a component. The intercorrelations among these scales are presented in Table 3. It is obvious that the intercorrelations varied considerably. The lowest correlation was 0.145 between the PRF Impulsivity Scale and the FPI Impulsiveness Scale and the highest correlation 0.618 between the MMPI Impulsivity Scale and the FPI Neuroticism Scale. All the correlations were positive; but only five of the six correlations reached the level of significance. The correlation matrix suggested interesting rankings of relationships among the impulsivity scales: the PRF Impulsivity Scale showed a moderate relationship with the MMPI Impulsivity Scale and proved to be independent of the FPI Impulsiveness Scale. The MMPI Impulsivity Scale was correlated with both the PRF Impulsivity Scale and the FPI Impulsiveness Scale, and the correlations were of approximately equal size. The correlation of the FPI Impulsiveness Scale with the MMPI Impulsivity Scale was high, and the correlation with the PRF Impulsivity Scale insignificant. Furthermore, the correlations of the FPI Neuroticism Scale and the impulsivity scales differed considerably. Whereas the correlations with the MMPI Impulsivity Scale and the FPI Impulsiveness Scale were high, the correlation with the PRF Impulsivity Scale was small. The correlation of the FPI Neuroticism Scale and the MMPI Impulsivity Scale differed from the correlation of the FPI Neuroticism Scale and the PRF Impulsivity Scale [t(97)=4.45, P<0.01] and the correlation of the FPI Neuroticism Scale and the FPI Impulsiveness Scale from the correlation of the FPI Neuroticism Scale and the PRF Impulsivity Scale [t(97)=3.51, P<0.01). Moreover, the correlation Table 3 Intercorrelations and factor loadings of the Impulsivity and Neuroticism Scales Scale
(PRF)
Correlation Impulsivity (MMPI)
PRF Impulsivity Scale MMPI Impulsivity Scale FPI Impulsiveness Scale FPI Neuroticism Scale
1.000 0.338** 0.145 0.209*
1.000 0.411** 0.618**
* P< 0.05 ** P< 0.01
(FPI)
1.000 0.583**
Neuroticism
Factor loading
1.000
0.311 0.727 0.616 0.869
1038
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
of the MMPI Impulsivity Scale and the FPI Impulsiveness Scale differed from the correlation of the FPI Impulsiveness Scale and the PRF Impulsivity Scale [t(97)=2.50, P<0.02). Obviously, the three impulsivity scales did not show the high degree of homogeneity, that would be observable if conceptual similarities among the three scales were high. Instead, the large differences between the correlations suggested that the scales must really represent different facets of impulsivity. Furthermore, factor analysis according to the common factor model was computed in order to learn more about the structure of the scales. The PRF Impulsivity Scale, the MMPI Impulsivity Scale, the FPI Impulsiveness Scale and the FPI Neuroticism Scale were included in this analysis. The largest eigenvalue was 2.213. There was no further eigenvalue larger than one so that no more than one component should be extracted from the data. The factor loadings are included in the last column of Table 3. They varied between 0.311 and 0.869. Apparently, the FPI Neuroticism Scale was most characteristic of this factor and the PRF Impulsivity Scale least characteristic. These loadings suggested homogeneity among the FPI Neuroticism Scale, the MMPI Impulsivity Scale and the FPI Impulsiveness Scale. In contrast, the PRF Impulsivity Scale seemed to share only a moderate amount of variance with the other scales. One can speculate whether the inclusion of a further indicator which would be similar to the PRF Impulsivity Scale would have increased the number of factors to two.1 The heterogeneity of the correlations with reasoning, the heterogeneity of the intercorrelations among the impulsivity scales and the results of factor analysis gave a reason to investigate the contents of the scale items for differences between the scales. This investigation of the items of the PRF Impulsivity Scale revealed descriptions of persons who were spontaneous and gave way to impulses and, additionally, descriptions of the contrary of well-controlled behavior. However, the items of this scale gave hardly any reason to suspect the presence of corresponding emotional or motivational states. In contrast, the investigation of the items of the MMPI Impulsivity Scale revealed the almost complete absence of descriptions of impulsive behavior. However, the contents of several items suggested the presence of emotional and motivational states which could be regarded as a precondition of impulsive behavior. These were states which could be assumed to prevent or disrupt the long-term concentration on a specific task. Some items suggested a high degree of similarity between this scale and the factors of premeditation and urgency identified by Whiteside and Lynam (2001). The investigation of the FPI Impulsiveness Scale showed the prevalence of another type of content. The items of this scale included emotional words related to annoyance or aggressiveness. Furthermore, behavior which seemed to be related to the states of annoyance or aggressiveness was described. This behavior was presented as the result of unfavorable stimuli. In sum, the investigation of the contents of the items revealed differences which supported the hypothesis that the three impulsivity scales represented different facets of impulsivity. As a result of this investigation, it was concluded that conceptual heterogeneity of the scales are the sources of the differences between the correlations. 3.4. Models for the prediction of reasoning Finally, structural equation modeling was applied in order to investigate the prediction of reasoning by means of the MMPI Impulsivity Scale, the PRF Impulsivity Scale, the FPI Impulsiveness 1
An anonymous reviewer contributed this interpretation of the result of factor analysis.
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1039
Scale and the FPI Neuroticism Scale. The FPI Neuroticism Scale was excluded in a later step because this scale could be expected to represent other characteristics of personality additionally. Three models were constructed. The goodness-of-fit and the parameters of these models were estimated by means of LISREL (Jo¨reskog & So¨rbom, 1996). The first model was the two-component model. This model was constructed because the results of factor analysis could tentatively be interpreted as suggesting two factors corresponding to two factors obtained by Whiteside and Lynam (2001): the MMPI Impulsivity Scale, the FPI Impulsiveness Scale and the FPI Neuroticism Scale on the one hand and the PRF Impulsivity Scale on the other hand.1 Accordingly, this model included two independent latent variables. The first one was represented by the MMPI Impulsivity Scale, the FPI Impulsiveness Scale and the FPI Neuroticism Scale and the second one by the PRF Impulsivity Scale. The Figural Reasoning Scale and the Numeric/Alphabetical Reasoning Scale served as manifest variables of the dependent latent variable which was denoted reasoning. The investigation of the goodness-of-fit of this model revealed a Chi-square of 10.13 (df=8), a RMSEA of 0.059, a GFI of 0.97, an AGFI of 0.91, a NFI of 0.94, a NNFI of 0.97 and a CFI of 0.98. All these measures of goodness-of-fit indicated an acceptable or good degree of agreement between data and model. The first path coefficient relating the first latent variable to reasoning was 0.29 (t= 2.15, P<0.05) indicating a moderate degree of influence of impulsivity on reasoning. In contrast, the second path coefficient relating the second latent variable to reasoning was 0.12 (t= 1.24, n.s.). This result suggested the lack of such a link. It was interesting to find that the value of the first path coefficient was lower than the value of the correlation between the MMPI Impulsivity Scale and reasoning. The second model was the one-component model. It was constructed to include one independent and one dependent latent variable. The MMPI Impulsivity Scale, the PRF Impulsivity Scale, the FPI Impulsiveness Scale and the FPI Neuroticism Scale were selected as manifest variables of the independent latent variable which was denoted impulsivity. Again, the Figural Reasoning Scale and the Numeric/Alphabetical Reasoning Scale served as manifest variables of the dependent latent variable. The investigation of the goodness-of-fit of this model revealed a Chi-square of 12.88 (df=9), a RMSEA of 0.063, a GFI of 0.96, an AGFI of 0.91, a NFI of 0.93, a NNFI of 0.97 and a CFI of 0.98. All these measures of goodness-of-fit indicated an acceptable or good degree of agreement between data and model. A substantial path coefficient of 0.33 (t= 3.25, P<0.01) relating impulsivity to reasoning was observed. Since the two models were nested, it was possible to identify the best model by means of the Chi-square difference. The two models differed by a Chi-square difference of 1.95 (df=1). Since the Chi-square table did not indicate that the difference reached the level of significance, preference had to be given to the more parsimonious model. The one-component model was the more parsimonious model since it included only one independent latent variable instead of two independent latent variables. However, since this model was heterogeneous insofar as it included one neuroticism scale beside several impulsivity scales, a third model called pure-impulsivity model was constructed. This model included the MMPI Impulsivity Scale, the PRF Impulsivity Scale and the FPI Impulsiveness Scale as independent latent variables and the Figural Reasoning Scale and the Numeric/ Alphabetical Reasoning Scale as dependent latent variables. A graphical representation of this model, which includes the estimates, is given in Fig. 1.
1040
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
Fig. 1. Pure-impulsivity model for the prediction of reasoning.
A Chi-square of 3.98 (df=5), a RMSEA of 0.000, a GFI of 0.99, an AGFI of 0.96, a NFI of 0.96, a NNFI of 1.02 and a CFI of 1.00 were obtained. An excellent degree of agreement between data and model was indicated by all these measures of goodness-of-fit. The path coefficient relating impulsivity to reasoning was 0.33 (t= 2.89, P<0.01). This coefficient suggested a moderate degree of influence of impulsivity on reasoning. Since this model showed the same degree of efficiency in predicting reasoning as the one-component model but was more homogeneous, it had to be selected as the most appropriate model.
4. Discussion The results of this study support the assumption that impulsivity influences reasoning. Negative correlations of small to moderate size between scales measuring impulsivity and reasoning were observed. This finding is in agreement with results obtained in investigating related concepts. Furthermore, it was interesting to find considerable differences between the sizes of the correlations for the three impulsivity scales. Differences between the contents of the items of the three scales are presumably responsible for the differences in correlation. The existence of such differences is even suggested by the findings of the Whiteside and Lynam (2001) study. The highest correlations were found for the MMPI Impulsivity Scale some items of which indicated the presence of emotional and motivational states related to impulsivity or its opposite. The other scales emphasized either the presence of instances of impulsive behavior or the dominance of annoyance and aggressiveness. The importance of the MMPI Impulsivity Scale may be due to several reasons. Firstly, the trait which is measured by this scale is not obvious to the respondent since hardly any one of the items describes an instance of impulsive behavior. This characteristic prevents respondents from describing themselves as impulsive intentionally. This may occasionally happen because impulsive behavior is regarded as desirable in some social groups. Secondly, items selected in giving preference to observations over theoretical considerations in the construction of a scale may integrate aspects of impulsivity, which are not apparent, but are as important as the most obvious instances of impulsive behavior.
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1041
Furthermore, it was interesting to find that the pure-impulsivity model and the one-component model provided better predictions of reasoning than the two-component model as was obvious from the path coefficients. This was presumably due to the fact that in the two-component model multicollinearity prevented one independent latent variable from contributing to the prediction. The equality of the pure-impulsivity model and the one-component model in predicting reasoning suggested that neuroticism did not provide a unique contribution to the prediction. Moreover, this parsimonious model showed an excellent degree of fit although no attempt was made to improve the model-data fit by adapting the model to characteristics of the data. The correlations observed for the MMPI Impulsivity Scale, the PRF Impulsivity Scale and the FPI Neuroticism Scale suggest a moderate relationship with reasoning, which can be interpreted as an influence which originates in the personality trait. It should be regarded as a sort of secondary influence since the reasoning ability itself is presumably not impaired. Instead, impulsive people are prevented from performing according to their potential in that at times the concentration on a difficult problem is interrupted. However, the disadvantageous influence on reasoning seems to be restricted. The sizes of the correlations roughly correspond to the sizes of the correlations observed between reasoning and mental speed. Moreover, the failure of the other personality scales to lead to substantial correlations with reasoning needs to be addressed. Extraversion was found to correlate with intelligence (Ackerman & Heggestad, 1997) so that there was reason to expect a substantial correlation. Surgency which was reported to correlate considerably with impulsivity (Stumpf et al., 1985) led to a marginal correlation only. Furthermore, achievement orientation did not correlate as expected because of the ambitious attitude assumed as a favorable precondition in reasoning. Although these are three cases in which the expectation was not supported, only one of the cases is a failure to replicate previous results. This one failure is probably due to the low effect size which includes a low degree of replicability.
Acknowledgements The author is grateful to Claudia Ehret and Andreas Hansen for their support in collecting the data for this study. The author would like to thank S. B. G. Eysenck and two anonymous reviewers for the useful suggestions which they provided. References Ackerman, P. L., & Heggestad, E. D. (1997). Intelligence, personality and interests: evidence for overlapping traits. Psychological Bulletin, 121, 219–245. Allport, G. W. (1961). Pattern and growth in personality. New York: Holt, Rinehart, & Winston. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: APA. Austin, E. J., Hofer, S. M., Deary, I. J., & Eber, H. W. (2000). Interactions between intelligence and personality: results from two large samples. Personality and Individual Differences, 29, 405–427. Brebner, J., & Stough, C. (1995). Theoretical and empirical relationships between personality and intelligence. In D. H. Saklofske, & M. Zeidner (Eds.), International handbook of personality and intelligence (pp. 321–347). New York: Plenum press.
1042
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
Buss, A. H., & Plomin, R. (1975). A temperament theory of personality development. New York: Interscience. Buss, A. H., & Plomin, R. (1984). Temperament: Early developing personality traits. Hillsdale, NJ: Erlbaum. Cattell, R. B., Eber, H. W., & Tatsuoka, M. M. (1970). Handbook for the Sixteen Personality Factor Questionnaire (16PF). Champaign: Institute for Personality and Ability Testing (IPAT). Cloninger, C. R., Svrakic, D. M., & Przybeck, T. R. (1993). A psychobiological model of temperament and character. Archives of General Psychology, 50, 975–1057. Costa, P. T. Jr., & McCrae, R. R. (1992). Revised NEO personality inventory manual. Odessa, FL: Psychological Assessment Resources. Depue, R. A., & Collins, P. F. (1999). Neurobiology of the structure of personality: Dopamine, facilitation of incentive motivation, and extraversion. Behavioral and Brain Sciences, 22, 491–569. Eysenck, H. J., Barrett, P., Wilson, G., & Jackson, C. (1992). Primary trait measurement of the 21 trait components of the P-E-N system. European Journal of Psychological Assessment, 8, 109–117. Eysenck, H. J., & Eysenck, M. W. (1985). Personality and individual differences. New York: Plenum Press. Eysenck, S. B. G., Daum, I., Schugens, M. M., & Diehl, J. M. (1990). A cross-cutural study of impulsiveness, venturesomeness and empathy: Germany and England. Zeitschrift fu¨r Differentielle und Diagnostische Psychologie, 11, 209–213. Eysenck, S. B. G., & Eysenck, H. J. (1977). The place of impulsiveness in a dimensional system of personality description. British Journal of Social and Clinical Psychology, 16, 57–68. Eysenck, S. B. G., Pearson, P. R., Easting, G., & Allsopp, J. F. (1985). Age norms for impulsiveness, venturesomeness, and empathy in adults. Personality and Individual Differences, 6, 613–619. Fahrenberg, J., Hampel, R., & Selg, H. (1994). Das Freiburger Perso¨nlichkeitsinventar FPI. 6. Aufl. Go¨ttingen: Hogrefe. Furnham, A., Forde, L., & Cotter, T. (1998). Personality and intelligence. Personality and Individual Differences, 24, 187–192. Gough, H. G. (1957). California psychological inventory manual. Palo Alto: Consulting Psychologists Press. Gupta, U. (1988). Effects of impulsivity and caffeine on human cognitive performance. Pharmacopsychologeia, 1, 33–41. Hathaway, S. R., & McKinley, J. C. (1951). Minnesota Multiphasic Personality Inventory (MMPI) (Rev. Ed.). New York: Psychological Corporation. Horn, W. (1983). Leistungs-Pru¨f-System. Go¨ttingen: Hogrefe. Humphreys, M. S., & Revelle, W. (1984). Personality, motivation, and performance: a theory of the relationship between individual differences and information processing. Psychological Review, 91, 153–184. Jackson, D. N. (1974). Manual for the Personality Research Form (2nd ed.). Goshen: Research Psychologists Press. Jo¨reskog, K., & So¨rbom, D. (1996). LISREL 8: User’s reference guide. Chicago, IL: Scientific Software International. Kagan, J. (1965). Impulsive and reflective children: Significance of conceptual tempo. In J. Krumboltz (Ed.), Learning and the educational process (pp. 133–161). Chicago: Rand McNaslly. Kagan, J., Moss, H. A., & Sigel, I. E. (1963). Psychological significance of style of conceptualisation. In J. C. Wright, & J. Kagan (Eds.), Basic cognitive processes in children. Monographs of the society for research in child development (pp. 73–112. Matthews, G. (1987). Personality and multidimensional arousal: a study of two dimensions of extraversion. Personality and Individual Differences, 8, 9–16. Messer, S. B. (1976). Reflection-impulsivity: a review. Psychological Bulletin, 83, 1026–1052. Pedersen, N. L., Plomin, R., McClearn, G. E., & Friberg, L. (1988). Neuroticism, extraversion, and related traits in adult twins reared apart and reared together. Journal of Personality and Social Psychology, 55, 950–957. Phillips, L. H., & Rabbitt, P. M. A. (1995). Impulsivity and speed-accuracy strategies in intelligence test performance. Intelligence, 21, 13–29. Revelle, W., Humphreys, M. S., Simon, L., & Gilliland, K. (1980). The interactive effect of personality, time of day, and caffeine: a test of the arousal model. Journal of Experimental Psychology: General, 109, 1–31. Salthouse, T. A. (1993). Influence of working memory on adult age differences in matrix reasoning. British Journal of Psychology, 84, 171–199. Salthouse, T. A. (1996). The processing-speed theory of adult age differences in cognition. Psychological Review, 103, 403–428. Schneewind, K. A., Schro¨der, G., & Cattell, R. B. (1994). Der 16PF. 2. Aufl. Go¨ttingen: Hogrefe.
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1043
Schweizer, K. (2000). Cognitive mechanisms as sources of success and failure in intelligence testing. Psychologische Beitro¨ge, 42, 190–200. Schweizer, K. (2001). On the role of mechanisms when the processing complexity is high. European Psychologist, 6, 133–143. Snyderman, M., & Rothman, S. (1987). Survey of expert opinion on intelligence and attitude testing. American Psychologist, 42, 137–144. Spreen (1963). MMPI Saarbru¨cken. Bern: Huber. Stott, D. H. (1985). Learning style or ‘‘intelligence’’?. School Psychology International, 6, 167–174. Stumpf, H., Angleitner, A., Wieck, T., Jackson, D. N., & Beloch-Till, H. (1985). Deutsche Personality Research Form. Go¨ttingen: Hogrefe. Tellegen, A. (1982). Multidimensional Personality Questionnaire manual. Minneapolis, MN: University of Minnesota Press. Whiteside, S. P., & Lynam, D. R. (2001). The five factor model of impulsivity: using a structural model of personality to understand impulsivity. Personality and Individual Differences, 30, 669–689.
Does impulsivity influence performance in reasoning? Karl Schweizer* Abt. Methodenlehre, Evaluation und Forschungsmethoden, Psychologisches Institut, Johann Wolfgang Goethe-Universita¨t, Mertonstr. 17, 60054 Frankfurt a. M., Germany Received 19 July 2001; received in revised form 30 October 2001; accepted 30 November 2001
Abstract The influence of impulsivity on reasoning was investigated in a sample of 108 high school and university students. Impulsivity was represented by the PRF Impulsivity Scale [Jackson, D. N. (1974). Manual for the Personality Research Form (2nd ed.). Goshen: Research Psychologists Press.], the MMPI Impulsivity Scale [Gough, H. G. (1957). California psychological inventory manual. Palo Alto: Consulting Psychologists Press.] and the FPI Impulsiveness Scale [Fahrenberg, J., Hampel, R., & Selg, H. (1994). Das Freiburger Perso¨nlichkeitsinventar FPI. 6. Aufl. Go¨ttingen: Hogrefe.]. Reasoning was measured by the Figural Reasoning Scale (Horn, W., 1983) and the Numeric/Alphabetical Reasoning Scale [Horn, W. (1983). Leistungs-Pru¨fSystem. Go¨ttingen: Hogrefe.]. Furthermore, measures of neuroticism, extraversion, achievement orientation and surgency were applied. Substantial negative correlations of reasoning with impulsivity and neuroticism were observed. Structural equation modeling was applied to predict reasoning by means of an impulsivity composite. A path coefficient of 0.33 was observed. This result suggests the interpretation that a high degree of impulsivity impairs performance in completing reasoning tasks. # 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction Impulsivity is a trait which is ascribed to persons a major characteristic of which is that their behavior is to a considerable degree guided by impulses. Persons high in impulsivity prefer the immediate over the well-reasoned response. In such persons new stimuli are rather likely to interrupt the execution of well-conceived behavioral plans which are unrelated to the new stimuli. Although impulsive behavior seems to be easily identifiable, impulsivity is described rather differently. In their attempt to clarify the meaning of impulsivity Depue and Collins (1999) found that it is characterized as including various lower-order traits such as sensation seeking, novelty seeking, boldness, adventuresomeness, boredom susceptibility, unreliability, and unorderliness. Furthermore, in the Diagnostic and Statistical Manual for the Mental Disorders (DSM-IV; * Corresponding author. Tel.: +49-69-798-22894; fax: +49-69-798-23847. E-mail address: [email protected] (K. Schweizer). 0191-8869/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0191-8869(01)00209-4
1032
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
American Psychiatric Association, 1994) impulsivity is associated with various disorders such as impulse-control disorders, antisocial personality disorder, mania, and dementia. A look at the various models of the structure of personality can help to obtain a better image of impulsivity. The 1977 version of Eysenck and Eysenck’s (1977) three factor theory of personality includes four specific dimensions of impulsivity: narrow impulsiveness, risk-taking, non-planning, and liveliness. The first of these dimensions was shown to be associated with neuroticism and the other ones with extraversion. Irrespective of these dimensions, impulsivity led to a robust factor in an attempt to construct a single scale (S. B. G. Eysenck, Pearson, Easting, & Allsopp, 1985) which even showed cross-cultural consistency (S. B. G. Eysenck, Daum, Schugens, & Diehl, 1990). In Cloninger’s model of personality which includes four temperament dimensions impulsivity is an aspect of novelty seeking (Cloninger, Svrakic, & Przybeck, 1993). According to Cloninger impulsivity is the biology-based tendency to respond fast to novel stimuli. The three factor model of personality by Tellegen (1982) also includes the aspect of impulsivity. This model is constituted by the factors positive emotionality, negative emotionality and constraint. A low degree of constraint is closely associated with a high degree of impulsivity. Finally, the Five Factor Model of Personality is to be considered. Costa and McCrae (1992) note that the neuroticism and conscientiousness factors include impulsivity facets. The various representations of impulsivity led Whiteside and Lynam (2001) to investigate the structure of impulsivity. They applied a large number of impulsivity scales to a sample of over 400 young adults and identified four impulsivity factors by means of exploratory factor analysis. They labeled these factors urgency, (lack of) premeditation, (lack of) perseverance, and sensation seeking. Most of the intercorrelations between the corresponding scores were low to moderate indicating a considerable degree of heterogeneity among the facets of impulsivity. Thus, the conceptual variety associated with the various models of personality finds its explanation. Impulsivity is also regarded as a cognitive style (Kagan, 1965; Kagan, Moss, & Sigel, 1963) showing two extremes one of which is denoted impulsive style and the other one reflective style. The impulsive style is characterized by fast decisions of which the certainty of being correct is low whereas the reflective style denotes the preference for late decisions associated with a high degree of certainty. The cognitive style has been assessed by means of tests which require the detection of differences between similar figures, and it has served well in describing children’s behavior (Messer, 1976). The investigation of impulsive behavior in children is especially attractive because the children’s behavior reflects social norms to a lower degree than the adults’ behavior. Furthermore, it needs to be emphasized that impulsivity is considered as a temperament (Buss & Plomin, 1975; Cloninger et al., 1993). A temperament is defined as a characteristic style of behaving, which is hereditary (Allport, 1961). However, evidence concerning heredity of impulsivity is not unequivocal. Pedersen, Plomin, McClearn, and Friberg (1988) provided positive evidence whereas Buss and Plomin (1984) did not. Additional support for the assumption of heredity results from the observation of impulsive behavior in children since children’s behavior is less determined by experiences than the adults’ behavior. 1.1. Possible consequences of impulsivity for reasoning Since impulsivity tends to interrupt the execution of well-conceived behavioral plans, a disadvantageous influence on performance can be expected when the demands are complex.
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1033
Reasoning is a prominent instance of a complex mental activity. Therefore, a disadvantageous influence of impulsivity on performance in reasoning tasks and related tasks can be hypothesized. Impulsive behavior can cause inappropriate responding in completing reasoning tasks, i.e. responding exceptionally fast and without a sufficient degree of certainty. Accordingly, impulsivity was supposed to favor the selection of cognitive associations which are easily available over other associations (Stott, 1985) and also the neglect of available information (Brebner & Stough, 1995). Such behavioral effects should be especially disadvantageous in reasoning since reasoning usually requires the execution of many cognitive operations which are to be performed according to a specific algorithm. The exclusive concentration on the task, combined with the suppression of other behavioral impulses, is advantageous for the successful completion of a reasoning task if it can be sustained for a sufficiently long period of time. The processes which are associated with the completion of reasoning tasks interact with each other and, therefore, are described as mechanisms (Schweizer, 2000, 2001). The successful completion of a reasoning task is likely when all the operations of the algorithm are performed appropriately and within the time limit (Salthouse, 1993, 1996). Deviations from the algorithm reduce the likelihood of the successful completion of the task in impulsive persons and, thus, contributes to the other sources of possible failure. Apparently, in persons showing a high degree of impulsivity the aforementioned reasons provide a disadvantage. 1.2. Evidence concerning the assumed influence of impulsivity on reasoning Despite the reasonableness of all these considerations, the correlations reported in a study by Furnham, Forde, and Cotter (1998) were only of small to moderate size. In this study impulsivity was represented by the corresponding scale of the Eysenck Personality Profiles (Eysenck, Barrett, Wilson, & Jackson, 1992). These findings are contrasted by the results of a study by Phillips and Rabbitt (1995) who reported insignificant results only. In this study a behavioral measure was applied for the assessment of impulsivity instead of a personality measure. Consequently, the difference may be due to the different types of measure. The personality measure is based on longterm experiences whereas the behavioral measure reflects momentary behavior only. Impulsivity has not been given much consideration in the investigation of the relationship between personality and intelligence. However, since it has been regarded as a component of neuroticism (e.g. Eysenck et al., 1992) and reasoning is considered as a component of intelligence (see Snyderman & Rothman, 1987) there is the possibility to deduce a hypothesis concerning the relationship between impulsivity and reasoning from the findings of studies investigating the relationship between neuroticism and intelligence. The available evidence suggests the existence of a small to moderate negative relationship between neuroticism and intelligence. In a meta-analytic study by Ackerman and Heggestad (1997) weak but significant negative relationships between sales interpretable as components of neuroticism and various measures of intelligence are reported. Further indirect evidence is provided by a more recent study which reports the data of 30,000 persons (Austin, Hofer, Deary, & Eber, 2000). In this study neuroticism is shown to influence the relationship between different ability measures. There are also some studies relating impulsivity, intelligence and arousal level (e.g. Gupta, 1988; Humphreys & Revelle, 1984; Matthews, 1987; Revelle, Humphreys, Simon, & Gilliland, 1980). The authors of these studies expect an influence of the arousal level on the relationship between
1034
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
impulsivity and intelligence. These studies even provide evidence that caffeine administration and time of day have an effect on the relationship although the results are not consistent. Despite these findings, Brebner and Stough (1995) still regard these studies as insufficient for unravelling the theoretical connections between impulsivity, intelligence and arousal level. 1.3. Aims of the study The investigation of the relationship between impulsivity and reasoning in a more comprehensive way than in the previous studies was the main aim of this study. It was expected that impulsivity and reasoning correlate substantially. Three different scales which were expected to measure impulsivity were selected for the new study. The set of these scales could be assumed to provide a better representation of impulsivity than the individual scales included in other studies. However, these scales were not expected to represent all the factors of the Whiteside and Lynam (2001) model of impulsivity since only a subset of factors could be regarded as important with respect to reasoning. Furthermore, extraversion and neuroticism which were assumed to include facets of impulsivity were considered. Two further personality scales were added in order to have a more complete description of personality.
2. Method 2.1. Participants The sample of this study included 108 participants (82 women and 26 men). The mean age was 22.78 years with a standard deviation of 4.81. The lower limit for participation was 17 years of age. The participants were high school and university students. 2.2. Measures 2.2.1. Reasoning scales Reasoning was assessed by means of the two reasoning scales included in the Leistungs-Pru¨fSystem (LPS; Horn, 1983). The first scale required figural reasoning and the second scale numeric/alphabetical reasoning. Accordingly, the first scale is denoted Figural Reasoning Scale and the second one Numeric/Alphabetical Reasoning Scale. In completing the items of these scales the individuals had to identify the inherent logic of a series of figures, numbers or letters and apply this logic inductively. Each series was constructed systematically with one exception which had to be identified (e.g. aa aa aa a aa). The correlation between the two reasoning scales was 0.62 in the sample of this study. For some of the statistical investigations reported in the following sections the scores obtained by means of the two scales were combined in order achieve an overall reasoning score. 2.2.2. Impulsivity scales Impulsivity was measured by three scales: the impulsivity Scale of the Personality Research Form (PRF; Jackson, 1974; German version: Stumpf, Angleitner, Wieck, Jackson, & Beloch-Till,
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1035
1985), the impulsivity Scale (Gough, 1957) which was based on the items of the Minnesota Multiphasic Personality Inventory (MMPI; Hathaway & McKinley, 1951; German version: Spreen, 1963), and the impulsiveness scale of the Freiburger Perso¨nlichkeitsinventar (FPI; Fahrenberg, Hampel, & Selg, 1994). These scales included 16 (PRF), 20 (MMPI) and 12 (FPI) items. These scales are denoted PRF Impulsivity scale, MMPI Impulsivity Scale and FPI Impulsiveness Scale respectively. All the items included a dichotomous response format. 2.2.3. Neuroticism scale The emotionality scale of the Freiburger Perso¨nlichkeitsinventar (Fahrenberg et al., 1994) was selected for this purpose since this scale was originally constructed to represent the neuroticism dimension of the Eysencks’ PEN system (Eysenck & Eysenck, 1985). It is addressed FPI Neuroticism Scale. 2.2.4. Other personality scales Three further personality scales were selected for this study: the surgency scale of the 16 Personality Factors questionnaire (16PF; Cattell, Eber, & Tatsuoka, 1970; German version: Schneewind, Schro¨der, & Cattell, 1994), the extraversion and achievement orientation scales of the Freiburger Perso¨nlichkeitsinventar (Fahrenberg et al., 1994). The surgency scale was selected because of a large correlation ( 0.41) which was reported between this scale and the PRF impulsivity scale (Stumpf et al., 1985). The extraversion scale was considered because it was found to correlate with intelligence in other studies (Ackerman & Heggestad, 1997) and the achievement orientation scale because of its content which suggested a positive relationship with reasoning. Furthermore, the scales allowed us to obtain a more comprehensive description of personality. These scales are denoted 16PF Surgency Scale, FPI Extraversion Scale and FPI Achievement Orientation Scale. 2.3. Procedure The reasoning tests and personality scales were applied in the following order: Figural Reasoning Scale, Numeric/Alphabetical Reasoning Scale, PRF Impulsivity Scale, 16PF Surgency Scale, FPI scales, MMPI Impulsivity Scale. The items of the FPI scales were included into one item list. The participants were tested in small groups.
3. Results The means and standard deviations of the raw scores of the scales to measure reasoning and personality traits are included in Table 1. The means varied between 6.44 and 64.10 and the standard deviations between 2.51 and 7.44. The differences between the means reflected the differences between the numbers of items included in the individual scales. The reasoning scores were quite large. This was obvious from the upper limits of the scales, which were 40 in both cases. However, these large scores were not unusual since the participants showed a high educational level. The other statistics did not indicate that the sample differed considerably from the population.
1036
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
3.1. Effects of sex and age The reasoning and personality scores were investigated for effects due to sex or age by means of computing Pearson correlations. A correlation of 0.22 (P<0.05) was observed between sex and the Numeric/Alphabetical Reasoning Scale. This meant that being male was associated with a slightly higher average reasoning scores than being female. All the other correlations with the reasoning scales were close to zero. Neither one of the impulsivity scales was correlated with sex or age, nor the FPI Neuroticism Scale. The 16PF Surgency Scale was correlated with age (r= 0.24, P<0.05) whereas all the other personality scales were not related to sex or age. These results did not give reason to suspect that effects due to either sex or age would influence the relationships between the reasoning and trait scores. 3.2. The relationships of reasoning and personality traits The Pearson correlations between the reasoning scales and the trait scales are presented in Table 2.
Table 1 Means and standard deviations for the raw scores of the Reasoning and Personality Scales Scale
Mean
SD
Reasoning (overall) Figural Reasoning Scale Numeric/Alphabetical Reasoning Scale PRF Impulsivity Scale MMPI Impulsivity Scale FPI Impulsiveness Scale FPI Neuroticism Scale 16PF Surgency Scale FPI Extraversion Scale FPI Achievement Orientation Scale
64.10 31.59 32.51 9.34 8.67 6.86 7.16 10.93 6.74 6.44
7.44 4.35 3.92 3.05 3.79 3.26 3.45 3.53 3.44 2.51
Figural reasoning
Numeric/alphabetical reasoning
Table 2 Pearson correlations between Reasoning Scales and Personality Scales Personality Scales PRF Impulsivity Scale MMPI Impulsivity Scale FPI Impulsiveness Scale FPI Neuroticism Scale 16PF Surgency Scale FPI Extraversion Scale FPI Achievement Orientation Scale * P< 0.05 ** P< 0.01
Reasoning 0.215* 0.321** 0.132 0.300** 0.021 0.045 0.105
0.199* 0.283** 0.134 0.289** 0.011 0.025 0.087
0.187 0.295** 0.103 0.248** 0.053 0.058 0.103
1037
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
All the correlations of the reasoning and impulsivity scales were negative. Five of the nine correlations concerning this relationship reached the level of significance. All the correlations of the MMPI Impulsivity Scale and reasoning were significant. The PRF Impulsivity Scale showed two substantial correlations with reasoning. In contrast, neither one of the three correlations between the FPI Impulsiveness Scale and reasoning were significant. Furthermore, the FPI Neuroticism scale correlated substantially with all the reasoning scales. In contrast, neither one of the further personality scales, 16PF Surgency Scale, FPI Extraversion Scale, and FPI Achievement Orientation Scale, led to a substantial correlation with reasoning. All these further correlations were close to zero. 3.3. The relationships among the neuroticism and impulsivity scales The heterogeneity of the results observed for the impulsivity scales suggested the investigation of the relationships among the impulsivity scales. The FPI Neuroticism Scale was additionally considered since neuroticism was assumed to include facets of impulsivity as a component. The intercorrelations among these scales are presented in Table 3. It is obvious that the intercorrelations varied considerably. The lowest correlation was 0.145 between the PRF Impulsivity Scale and the FPI Impulsiveness Scale and the highest correlation 0.618 between the MMPI Impulsivity Scale and the FPI Neuroticism Scale. All the correlations were positive; but only five of the six correlations reached the level of significance. The correlation matrix suggested interesting rankings of relationships among the impulsivity scales: the PRF Impulsivity Scale showed a moderate relationship with the MMPI Impulsivity Scale and proved to be independent of the FPI Impulsiveness Scale. The MMPI Impulsivity Scale was correlated with both the PRF Impulsivity Scale and the FPI Impulsiveness Scale, and the correlations were of approximately equal size. The correlation of the FPI Impulsiveness Scale with the MMPI Impulsivity Scale was high, and the correlation with the PRF Impulsivity Scale insignificant. Furthermore, the correlations of the FPI Neuroticism Scale and the impulsivity scales differed considerably. Whereas the correlations with the MMPI Impulsivity Scale and the FPI Impulsiveness Scale were high, the correlation with the PRF Impulsivity Scale was small. The correlation of the FPI Neuroticism Scale and the MMPI Impulsivity Scale differed from the correlation of the FPI Neuroticism Scale and the PRF Impulsivity Scale [t(97)=4.45, P<0.01] and the correlation of the FPI Neuroticism Scale and the FPI Impulsiveness Scale from the correlation of the FPI Neuroticism Scale and the PRF Impulsivity Scale [t(97)=3.51, P<0.01). Moreover, the correlation Table 3 Intercorrelations and factor loadings of the Impulsivity and Neuroticism Scales Scale
(PRF)
Correlation Impulsivity (MMPI)
PRF Impulsivity Scale MMPI Impulsivity Scale FPI Impulsiveness Scale FPI Neuroticism Scale
1.000 0.338** 0.145 0.209*
1.000 0.411** 0.618**
* P< 0.05 ** P< 0.01
(FPI)
1.000 0.583**
Neuroticism
Factor loading
1.000
0.311 0.727 0.616 0.869
1038
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
of the MMPI Impulsivity Scale and the FPI Impulsiveness Scale differed from the correlation of the FPI Impulsiveness Scale and the PRF Impulsivity Scale [t(97)=2.50, P<0.02). Obviously, the three impulsivity scales did not show the high degree of homogeneity, that would be observable if conceptual similarities among the three scales were high. Instead, the large differences between the correlations suggested that the scales must really represent different facets of impulsivity. Furthermore, factor analysis according to the common factor model was computed in order to learn more about the structure of the scales. The PRF Impulsivity Scale, the MMPI Impulsivity Scale, the FPI Impulsiveness Scale and the FPI Neuroticism Scale were included in this analysis. The largest eigenvalue was 2.213. There was no further eigenvalue larger than one so that no more than one component should be extracted from the data. The factor loadings are included in the last column of Table 3. They varied between 0.311 and 0.869. Apparently, the FPI Neuroticism Scale was most characteristic of this factor and the PRF Impulsivity Scale least characteristic. These loadings suggested homogeneity among the FPI Neuroticism Scale, the MMPI Impulsivity Scale and the FPI Impulsiveness Scale. In contrast, the PRF Impulsivity Scale seemed to share only a moderate amount of variance with the other scales. One can speculate whether the inclusion of a further indicator which would be similar to the PRF Impulsivity Scale would have increased the number of factors to two.1 The heterogeneity of the correlations with reasoning, the heterogeneity of the intercorrelations among the impulsivity scales and the results of factor analysis gave a reason to investigate the contents of the scale items for differences between the scales. This investigation of the items of the PRF Impulsivity Scale revealed descriptions of persons who were spontaneous and gave way to impulses and, additionally, descriptions of the contrary of well-controlled behavior. However, the items of this scale gave hardly any reason to suspect the presence of corresponding emotional or motivational states. In contrast, the investigation of the items of the MMPI Impulsivity Scale revealed the almost complete absence of descriptions of impulsive behavior. However, the contents of several items suggested the presence of emotional and motivational states which could be regarded as a precondition of impulsive behavior. These were states which could be assumed to prevent or disrupt the long-term concentration on a specific task. Some items suggested a high degree of similarity between this scale and the factors of premeditation and urgency identified by Whiteside and Lynam (2001). The investigation of the FPI Impulsiveness Scale showed the prevalence of another type of content. The items of this scale included emotional words related to annoyance or aggressiveness. Furthermore, behavior which seemed to be related to the states of annoyance or aggressiveness was described. This behavior was presented as the result of unfavorable stimuli. In sum, the investigation of the contents of the items revealed differences which supported the hypothesis that the three impulsivity scales represented different facets of impulsivity. As a result of this investigation, it was concluded that conceptual heterogeneity of the scales are the sources of the differences between the correlations. 3.4. Models for the prediction of reasoning Finally, structural equation modeling was applied in order to investigate the prediction of reasoning by means of the MMPI Impulsivity Scale, the PRF Impulsivity Scale, the FPI Impulsiveness 1
An anonymous reviewer contributed this interpretation of the result of factor analysis.
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1039
Scale and the FPI Neuroticism Scale. The FPI Neuroticism Scale was excluded in a later step because this scale could be expected to represent other characteristics of personality additionally. Three models were constructed. The goodness-of-fit and the parameters of these models were estimated by means of LISREL (Jo¨reskog & So¨rbom, 1996). The first model was the two-component model. This model was constructed because the results of factor analysis could tentatively be interpreted as suggesting two factors corresponding to two factors obtained by Whiteside and Lynam (2001): the MMPI Impulsivity Scale, the FPI Impulsiveness Scale and the FPI Neuroticism Scale on the one hand and the PRF Impulsivity Scale on the other hand.1 Accordingly, this model included two independent latent variables. The first one was represented by the MMPI Impulsivity Scale, the FPI Impulsiveness Scale and the FPI Neuroticism Scale and the second one by the PRF Impulsivity Scale. The Figural Reasoning Scale and the Numeric/Alphabetical Reasoning Scale served as manifest variables of the dependent latent variable which was denoted reasoning. The investigation of the goodness-of-fit of this model revealed a Chi-square of 10.13 (df=8), a RMSEA of 0.059, a GFI of 0.97, an AGFI of 0.91, a NFI of 0.94, a NNFI of 0.97 and a CFI of 0.98. All these measures of goodness-of-fit indicated an acceptable or good degree of agreement between data and model. The first path coefficient relating the first latent variable to reasoning was 0.29 (t= 2.15, P<0.05) indicating a moderate degree of influence of impulsivity on reasoning. In contrast, the second path coefficient relating the second latent variable to reasoning was 0.12 (t= 1.24, n.s.). This result suggested the lack of such a link. It was interesting to find that the value of the first path coefficient was lower than the value of the correlation between the MMPI Impulsivity Scale and reasoning. The second model was the one-component model. It was constructed to include one independent and one dependent latent variable. The MMPI Impulsivity Scale, the PRF Impulsivity Scale, the FPI Impulsiveness Scale and the FPI Neuroticism Scale were selected as manifest variables of the independent latent variable which was denoted impulsivity. Again, the Figural Reasoning Scale and the Numeric/Alphabetical Reasoning Scale served as manifest variables of the dependent latent variable. The investigation of the goodness-of-fit of this model revealed a Chi-square of 12.88 (df=9), a RMSEA of 0.063, a GFI of 0.96, an AGFI of 0.91, a NFI of 0.93, a NNFI of 0.97 and a CFI of 0.98. All these measures of goodness-of-fit indicated an acceptable or good degree of agreement between data and model. A substantial path coefficient of 0.33 (t= 3.25, P<0.01) relating impulsivity to reasoning was observed. Since the two models were nested, it was possible to identify the best model by means of the Chi-square difference. The two models differed by a Chi-square difference of 1.95 (df=1). Since the Chi-square table did not indicate that the difference reached the level of significance, preference had to be given to the more parsimonious model. The one-component model was the more parsimonious model since it included only one independent latent variable instead of two independent latent variables. However, since this model was heterogeneous insofar as it included one neuroticism scale beside several impulsivity scales, a third model called pure-impulsivity model was constructed. This model included the MMPI Impulsivity Scale, the PRF Impulsivity Scale and the FPI Impulsiveness Scale as independent latent variables and the Figural Reasoning Scale and the Numeric/ Alphabetical Reasoning Scale as dependent latent variables. A graphical representation of this model, which includes the estimates, is given in Fig. 1.
1040
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
Fig. 1. Pure-impulsivity model for the prediction of reasoning.
A Chi-square of 3.98 (df=5), a RMSEA of 0.000, a GFI of 0.99, an AGFI of 0.96, a NFI of 0.96, a NNFI of 1.02 and a CFI of 1.00 were obtained. An excellent degree of agreement between data and model was indicated by all these measures of goodness-of-fit. The path coefficient relating impulsivity to reasoning was 0.33 (t= 2.89, P<0.01). This coefficient suggested a moderate degree of influence of impulsivity on reasoning. Since this model showed the same degree of efficiency in predicting reasoning as the one-component model but was more homogeneous, it had to be selected as the most appropriate model.
4. Discussion The results of this study support the assumption that impulsivity influences reasoning. Negative correlations of small to moderate size between scales measuring impulsivity and reasoning were observed. This finding is in agreement with results obtained in investigating related concepts. Furthermore, it was interesting to find considerable differences between the sizes of the correlations for the three impulsivity scales. Differences between the contents of the items of the three scales are presumably responsible for the differences in correlation. The existence of such differences is even suggested by the findings of the Whiteside and Lynam (2001) study. The highest correlations were found for the MMPI Impulsivity Scale some items of which indicated the presence of emotional and motivational states related to impulsivity or its opposite. The other scales emphasized either the presence of instances of impulsive behavior or the dominance of annoyance and aggressiveness. The importance of the MMPI Impulsivity Scale may be due to several reasons. Firstly, the trait which is measured by this scale is not obvious to the respondent since hardly any one of the items describes an instance of impulsive behavior. This characteristic prevents respondents from describing themselves as impulsive intentionally. This may occasionally happen because impulsive behavior is regarded as desirable in some social groups. Secondly, items selected in giving preference to observations over theoretical considerations in the construction of a scale may integrate aspects of impulsivity, which are not apparent, but are as important as the most obvious instances of impulsive behavior.
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1041
Furthermore, it was interesting to find that the pure-impulsivity model and the one-component model provided better predictions of reasoning than the two-component model as was obvious from the path coefficients. This was presumably due to the fact that in the two-component model multicollinearity prevented one independent latent variable from contributing to the prediction. The equality of the pure-impulsivity model and the one-component model in predicting reasoning suggested that neuroticism did not provide a unique contribution to the prediction. Moreover, this parsimonious model showed an excellent degree of fit although no attempt was made to improve the model-data fit by adapting the model to characteristics of the data. The correlations observed for the MMPI Impulsivity Scale, the PRF Impulsivity Scale and the FPI Neuroticism Scale suggest a moderate relationship with reasoning, which can be interpreted as an influence which originates in the personality trait. It should be regarded as a sort of secondary influence since the reasoning ability itself is presumably not impaired. Instead, impulsive people are prevented from performing according to their potential in that at times the concentration on a difficult problem is interrupted. However, the disadvantageous influence on reasoning seems to be restricted. The sizes of the correlations roughly correspond to the sizes of the correlations observed between reasoning and mental speed. Moreover, the failure of the other personality scales to lead to substantial correlations with reasoning needs to be addressed. Extraversion was found to correlate with intelligence (Ackerman & Heggestad, 1997) so that there was reason to expect a substantial correlation. Surgency which was reported to correlate considerably with impulsivity (Stumpf et al., 1985) led to a marginal correlation only. Furthermore, achievement orientation did not correlate as expected because of the ambitious attitude assumed as a favorable precondition in reasoning. Although these are three cases in which the expectation was not supported, only one of the cases is a failure to replicate previous results. This one failure is probably due to the low effect size which includes a low degree of replicability.
Acknowledgements The author is grateful to Claudia Ehret and Andreas Hansen for their support in collecting the data for this study. The author would like to thank S. B. G. Eysenck and two anonymous reviewers for the useful suggestions which they provided. References Ackerman, P. L., & Heggestad, E. D. (1997). Intelligence, personality and interests: evidence for overlapping traits. Psychological Bulletin, 121, 219–245. Allport, G. W. (1961). Pattern and growth in personality. New York: Holt, Rinehart, & Winston. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: APA. Austin, E. J., Hofer, S. M., Deary, I. J., & Eber, H. W. (2000). Interactions between intelligence and personality: results from two large samples. Personality and Individual Differences, 29, 405–427. Brebner, J., & Stough, C. (1995). Theoretical and empirical relationships between personality and intelligence. In D. H. Saklofske, & M. Zeidner (Eds.), International handbook of personality and intelligence (pp. 321–347). New York: Plenum press.
1042
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
Buss, A. H., & Plomin, R. (1975). A temperament theory of personality development. New York: Interscience. Buss, A. H., & Plomin, R. (1984). Temperament: Early developing personality traits. Hillsdale, NJ: Erlbaum. Cattell, R. B., Eber, H. W., & Tatsuoka, M. M. (1970). Handbook for the Sixteen Personality Factor Questionnaire (16PF). Champaign: Institute for Personality and Ability Testing (IPAT). Cloninger, C. R., Svrakic, D. M., & Przybeck, T. R. (1993). A psychobiological model of temperament and character. Archives of General Psychology, 50, 975–1057. Costa, P. T. Jr., & McCrae, R. R. (1992). Revised NEO personality inventory manual. Odessa, FL: Psychological Assessment Resources. Depue, R. A., & Collins, P. F. (1999). Neurobiology of the structure of personality: Dopamine, facilitation of incentive motivation, and extraversion. Behavioral and Brain Sciences, 22, 491–569. Eysenck, H. J., Barrett, P., Wilson, G., & Jackson, C. (1992). Primary trait measurement of the 21 trait components of the P-E-N system. European Journal of Psychological Assessment, 8, 109–117. Eysenck, H. J., & Eysenck, M. W. (1985). Personality and individual differences. New York: Plenum Press. Eysenck, S. B. G., Daum, I., Schugens, M. M., & Diehl, J. M. (1990). A cross-cutural study of impulsiveness, venturesomeness and empathy: Germany and England. Zeitschrift fu¨r Differentielle und Diagnostische Psychologie, 11, 209–213. Eysenck, S. B. G., & Eysenck, H. J. (1977). The place of impulsiveness in a dimensional system of personality description. British Journal of Social and Clinical Psychology, 16, 57–68. Eysenck, S. B. G., Pearson, P. R., Easting, G., & Allsopp, J. F. (1985). Age norms for impulsiveness, venturesomeness, and empathy in adults. Personality and Individual Differences, 6, 613–619. Fahrenberg, J., Hampel, R., & Selg, H. (1994). Das Freiburger Perso¨nlichkeitsinventar FPI. 6. Aufl. Go¨ttingen: Hogrefe. Furnham, A., Forde, L., & Cotter, T. (1998). Personality and intelligence. Personality and Individual Differences, 24, 187–192. Gough, H. G. (1957). California psychological inventory manual. Palo Alto: Consulting Psychologists Press. Gupta, U. (1988). Effects of impulsivity and caffeine on human cognitive performance. Pharmacopsychologeia, 1, 33–41. Hathaway, S. R., & McKinley, J. C. (1951). Minnesota Multiphasic Personality Inventory (MMPI) (Rev. Ed.). New York: Psychological Corporation. Horn, W. (1983). Leistungs-Pru¨f-System. Go¨ttingen: Hogrefe. Humphreys, M. S., & Revelle, W. (1984). Personality, motivation, and performance: a theory of the relationship between individual differences and information processing. Psychological Review, 91, 153–184. Jackson, D. N. (1974). Manual for the Personality Research Form (2nd ed.). Goshen: Research Psychologists Press. Jo¨reskog, K., & So¨rbom, D. (1996). LISREL 8: User’s reference guide. Chicago, IL: Scientific Software International. Kagan, J. (1965). Impulsive and reflective children: Significance of conceptual tempo. In J. Krumboltz (Ed.), Learning and the educational process (pp. 133–161). Chicago: Rand McNaslly. Kagan, J., Moss, H. A., & Sigel, I. E. (1963). Psychological significance of style of conceptualisation. In J. C. Wright, & J. Kagan (Eds.), Basic cognitive processes in children. Monographs of the society for research in child development (pp. 73–112. Matthews, G. (1987). Personality and multidimensional arousal: a study of two dimensions of extraversion. Personality and Individual Differences, 8, 9–16. Messer, S. B. (1976). Reflection-impulsivity: a review. Psychological Bulletin, 83, 1026–1052. Pedersen, N. L., Plomin, R., McClearn, G. E., & Friberg, L. (1988). Neuroticism, extraversion, and related traits in adult twins reared apart and reared together. Journal of Personality and Social Psychology, 55, 950–957. Phillips, L. H., & Rabbitt, P. M. A. (1995). Impulsivity and speed-accuracy strategies in intelligence test performance. Intelligence, 21, 13–29. Revelle, W., Humphreys, M. S., Simon, L., & Gilliland, K. (1980). The interactive effect of personality, time of day, and caffeine: a test of the arousal model. Journal of Experimental Psychology: General, 109, 1–31. Salthouse, T. A. (1993). Influence of working memory on adult age differences in matrix reasoning. British Journal of Psychology, 84, 171–199. Salthouse, T. A. (1996). The processing-speed theory of adult age differences in cognition. Psychological Review, 103, 403–428. Schneewind, K. A., Schro¨der, G., & Cattell, R. B. (1994). Der 16PF. 2. Aufl. Go¨ttingen: Hogrefe.
K. Schweizer / Personality and Individual Differences 33 (2002) 1031–1043
1043
Schweizer, K. (2000). Cognitive mechanisms as sources of success and failure in intelligence testing. Psychologische Beitro¨ge, 42, 190–200. Schweizer, K. (2001). On the role of mechanisms when the processing complexity is high. European Psychologist, 6, 133–143. Snyderman, M., & Rothman, S. (1987). Survey of expert opinion on intelligence and attitude testing. American Psychologist, 42, 137–144. Spreen (1963). MMPI Saarbru¨cken. Bern: Huber. Stott, D. H. (1985). Learning style or ‘‘intelligence’’?. School Psychology International, 6, 167–174. Stumpf, H., Angleitner, A., Wieck, T., Jackson, D. N., & Beloch-Till, H. (1985). Deutsche Personality Research Form. Go¨ttingen: Hogrefe. Tellegen, A. (1982). Multidimensional Personality Questionnaire manual. Minneapolis, MN: University of Minnesota Press. Whiteside, S. P., & Lynam, D. R. (2001). The five factor model of impulsivity: using a structural model of personality to understand impulsivity. Personality and Individual Differences, 30, 669–689.