- Email: [email protected]
Assessment of measures of impulsivity in healthy male volunteers
Pergamon
Person. individ. Di@T Vol. 19. No. 6, pp. 927-935. 1995 Copyright 01995 Elsevier Science Ltd 0191~8869(95)00107-7 Printed in Great Britain. All rights reserved 0191-8869/95 $9.50 + 0.00
ASSESSMENT OF MEASURES OF IMPULSIVITY IN HEALTHY MALE VOLUNTEERS Karin F. Helmers,‘,* Simon N. Young’ and Robert 0. Pihl’,’ Departments of ‘Psychiatry and *Psychology, McGill University, 1033 Pine Avenue West, Montreal, Quebec, Canada H3A I Al (Received 20 October 1994; received for publication 22 June 1995)
Summary-The current study evaluated the relationship between self-report and behavioral measures of impulsivity in 98 young adult male subjects. Principal components analysis of the self-report measures revealed four primary factors: the first demonstrated significant loadings with Eysenck’s Impulsivity and Nonplanning scales and all subscales of the Barratt Impulsivity Scale, the second with Eysenck’s Venturesome, Thrill and Adventure Seeking and Experience Seeking Scales, the third with the Boredom Susceptibility and Disinhibition subscales of Zuckerman’s Sensation Seeking Scale, and the fourth with Zuckerman’s State Sensation Seeking and the Kipnis Scale. The relationship between scores on these primary factors and performance on four behavioral measures of impulsivity, the Draw a line Slowly, Matching Familiar Figures, the Porteus Maze, and Gomo Go Discrimination tasks, was investigated. Only the fourth primary factor demonstrated a positive relationship, with errors of commission on the Go/No Go Task, and with the Q score in the Porteous Maze. The fourth factor was the only one that was independent of educational level, age and IQ. Of the behavioral tasks, the Go/No Go task is an operationalization of impulsivity as defined by Gray, Owen, Davis & Tsaltas (Biological Bases ofSensation Seeking, ImpuWiq, andilnrieg, 1983) and may not be subject to demand characteristics found in the other behavioral and paper and pencil measures of impulsivity. Increased understanding of impulsivity is needed, and the fourth factor we defined, and the Go/No Go task, may hold promise for future research on one type of impulsivity.
INTRODUCTION
Interest in the evaluation of impulsivity has increased due to the association of impulsivity with aggressive behaviors and because of the suggestion that specific neurotransmitter abnormalities may underlie impulsive behaviors. Impulsivity is central to a number of psychiatric disorders either as a principal diagnostic feature (i.e. histrionic, antisocial, borderline personality disorders) or as a feature commonly associated with a diagnostic category (i.e. conduct, narcissistic personality and compulsive personality disorders). Impulsivity may also be an important precipitant of aggression (Valzelli, 1984; Roy, Adinoff & Linnoila, 1988) and suicide (Roy, Virkkunen & Linnoila, 1990; Brown, Ebert, Goyer, Jimerson, Klein & Bunney, 1982), and has a genetic component (Roy, 1983; Plutchik & van Praag, 1986). It is believed that low levels of the central nervous system neurotransmitter serotonin (5hydroxytryptamine, 5-HT) may be related to impulsivity, which in turn may be the common link between behavioral acts such as suicide and violent assault, and the impulsive behavioral components of other psychiatric disorders (Soubrie, 1986; Linnoila, Virkkunen, Scheinin, Nuutila, Rimon & Goodwin, 1983; Lidberg, Tuck, Asberg, Scalia-Tomba & Bertilsson, 1985). However, very few clinical studies have measured impulsivity directly. Rather the role of impulsivity is typically inferred from behavioral psychopathology (i.e. suicide attempts, violent assaultive behavior). Major issues in the study of impulsivity are the definition of this personality construct and its measurement in controlled laboratory settings. The definition of impulsivity varies across studies, including the failure to evaluate a situation as risky or dangerous (Eysenck & McGurk, 1980), acting without thinking (Barratt & Patton, 1983), the inability to plan ahead (Barratt & Patton, 1983; Eysenck & Eysenck, 1977; Buss & Plomin, 1975), the tendency to respond quickly to stimuli rather than inhibiting responses (Barratt & Patton, 1983; Prior & Sanson, 1986; Buss & Plomin, 1975), and the failure to withhold a response that will lead to punishment or a deficit in passive avoidance learning (Gray, Owen, Davis & Tsaltas, 1983). In a clinical setting, impulsivity is generally defined by deviant behaviors in the life history of an individual. For example, Glueck and Glueck (1968) stated that impulsivity is identified by the repetition of the deviant behavior and not the uniqueness of deviant behavior. Similarly, Oas (1983) defined chronic impulsivity as a persistent tendency to emit *To whom all correspondence should be addressed. 921
928
Karin F. Helmers et al.
inappropriate or pathological behavior characterized by a lack of reflectivity and delay. The DSM-IV characterizes impulse-control disorders as the failure to resist an impulse, drive, or temptation to perform an act that is harmful to the person or to others (American Psychiatric Association, 1994). These latter definitions do not allow for the evaluation of impulsivity in normal control 5s. Further, the measurement of impulsivity should be separate from the clinical diagnostic categories. Within the laboratory, the assessment of impulsivity can be accomplished by several paper-and-pencil questionnaires: Eysenck’s impulsivity questionnaires (Eysenck & Eysenck, 1977; Eysenck & McGurk, 1980) the Kipnis questionnaire (Kipnis, 1971), and Barratt’s Impulsivity scale (Barrett & Patton, 1983). Impulsivity also appears to be a component of psychopathy (Hare, 1991) and sensation seeking (Zuckerman, 1979). However, the subscale of Thrill and Adventure Seeking obtained from the Sensation Seeking Scale appears to be associated with risk-taking behaviors that are often carefully planned to minimize physical danger as found in mountain climbers (Zuckerman, 1993). Thus the subscale of Thrill and Adventure Seeking may be conceptually different from other self-report measures of impulsivity. Behavioral measures of impulsivity have been constructed in order to measure different components of impulsivity. These include the inability to plan ahead (i.e. Porteus Maze, Trails A and B), tests measuring speed of response and number of errors (i.e. Matching Familiar Figures Test, reaction time tests), ability to inhibit responding or passive avoidance learning (i.e. Go/No Go discrimination tasks), and ability to inhibit motor control (i.e. “Draw a line slowly” or “Walk a line slowly”). The purpose of the present study was to examine the relationship between behavioral measures of impulsivity and first and second order factors derived from paper-and-pencil measures of impulsivity. We hypothesized that factors related to impulsivity would be associated with increased errors and speed of response, whereas factors associated with thrill and adventure seeking would be associated with decreased errors and speed of response.
METHOD Subjects
A total of 101 English speaking males were recruited via a newspaper advertisement in the metropolitan area of Montreal. Ss were between the ages of 18 and 35 yr, denied any current psychopathology or psychiatric history (e.g. affective disorders, assaultive violent behavior), and medical problems (e.g. cardiovascular disease, diabetes, gastrointestinal problems). Each S completed a battery of paper-and-pencil questionnaires and behavioral tasks. Assessment procedure Paper-and-pencil questionnaires. The following paper-and-pencil questionnaires were completed by Ss: Barratt Impulsivity Scale [BIS, version 11 (Barratt & Patton, 1983)] with subscales measuring Motor, Nonplanning and Attentional impulsivity obtained from 34 items scored as a 0 = ‘rarely’ to 3 = ‘almost always’. Several scales are derived from Eysenck’s questionnaires which include narrow Impulsivity (Eys-lmpulsivity) and Venturesomeness (Eys-Venturesome) (Eysenck & McGurk, 1980), risk-taking (Eys-Risktaking) and non-planning (Eys-Nonplanning) (Eysenck & Eysenck, 1977). The Kipnis scale ( 197 1) is derived from 56 autobiographical items relevant to activities during childhood and adulthood. The Sensation Seeking Scale (SSS) (Zuckerman, 1979) consists of the subscales entitled Boredom Susceptibility, Disinhibition, Experience Seeking and Thrill and Adventure Seeking. These latter subscales are measured each by 10 items in which Ss choose between two statements. State sensation seeking (States) was measured by 15 adjectives rated for the current time period and scored from 1 = ‘not at all’ to 5 = ‘very much’ (Zuckerman, 1979). The Hare psychopath checklist consists of 29 items scored on a five-point Likert scale from ‘Strongly disagree’ to ‘Strongly agree’ (Hare, 1991). Ss also completed the Beck Depression Inventory (Beck & Steer, 1987), a 21 item scale in which the S chooses one out of four statements which best described his mood during the past l-2 weeks. Furthermore, the Vocabulary subtest of the Wechsler Adult Intelligence Scale-Revised (WAIS-R) was administered. Raw scores were converted to scaled scores, and the intelligence quotient (IQ) equivalent score was determined by the appropriate age group. Previous research on the WAIS-R has
Assessment of impulsivity
929
indicated that the vocabulary subtest, in comparison to other subtests, demonstrates the greatest association with the full-scale IQ score (reviewed in Kaufman, 1990). Behavioral measures Matching familiarfigures test (MFFT). A series of 10 sets of figures in which the S is required to identify one of six very similar pictures to a ‘target’ picture (Kagan, 1966). The mean latency of response to the first choice (MFFTime) and the total number of errors (MFFTerr) were calculated. Porteus maze. The adult Vineland Revised version of the Porteus mazes (Porteus, 1965) were scored for impulsivity (Q score). Q scores are obtained by measuring number of times the pencil was lifted, touched the boundary, etc. Draw a line slowly (DALS). This test measures inhibition of motor activity (Rohrbeck & Twentyman, 1986). Ss draw a line from top to bottom of a 9” X 1” column without crossing the boundaries as slowly as possible (measured in msec). ‘Go/No Go ’ discrimination tusk. Ss learn by trial-and-error when to respond by pressing a button and when to withhold responding. The stimuli are eight two-digit numbers, repeated 10 times in different, randomized orders for a total of 80 trials. Four of the eight stimuli are designated as positive stimuli (S + ) and the other half as negative stimuli (S - ). Numbers range from 01 to 99, were randomly chosen, and were presented as white lettering against a blue computer screen. Stimuli remain on the screen for 3 set or until the S presses the button. The interstimulus interval is 1 sec. Correct responses are signalled with a high-pitched tone, the word ‘correct’ on the computer screen, and five cents is added to the running tally of the Ss’ earnings. Incorrect responses are signalled by a low pitched tone, ‘wrong’ on the computer screen, and the subtraction of five cents from the running tally of the Ss’ earnings. Four conditions were tested in randomized order. For each condition, errors of commission (failure to inhibit responses to negative stimuli = passive avoidance errors) and errors of omission (failure to respond to positive stimuli) were calculated. The four conditions were: a reward-punishment (RP) condition in which Ss start with fifty cents, responses to S + ‘s are rewarded, and S - ‘s are punished. Errors of omission (RP-omi) and commission (RP-corn) were calculated. A punishment-only condition (Pun), in which Ss begin with $2.00, have no opportunity to win more money, but lose money if they respond to S - ‘s or fail to respond to S + ‘s. Pun-omi and Pun-corn were calculated (errors of omission and commission during Punishment only condition). A reward-only condition (Rew) in which Ss start with no money, but could earn money by responding to S + ‘s and not responding to S - ‘s (Rew-omi and Rew-corn were calculated for errors of omission and commission). Finally, a punishment-reward (PR) condition in which subjects start with fifty cents and a non-response to S + is punished and a non-response to S - is rewarded (PR-omi and PR-corn were calculated for errors of omission and commission, respectively). Prior to each condition there is a pretreatment of 12 trials in which the S + :S - stimuli are presented in randomized order in a 2: 1 ratio.
STATISTICAL
ANALYSES
To obtain primary factor scores, a principal components analysis using an oblique rotation was conducted on all paper-and-pencil subscales related to impulsivity (Beck depression scores were excluded). To obtain a secondary factor, a principal components analysis using an oblique rotation was performed on the primary factor scores to obtain a General Impulsivity score for each S. Behavioral measures of impulsivity did not allow for a factor analysis because of low inter-correlations and low communalities (Stevens, 1992). To assess the association between demographic variables and impulsivity, Analyses of Variance (ANOVA) were conducted on categorical demographic variables and primary and second factor scores of impulsivity. Pearson product-moment correlations evaluated the relationships between age, IQ scores and obtained factors. The relationship between behavioral measures and primary and secondary factors were evaluated by Pearson product-moment correlations and partial correlations in which demographic variables (age and IQ) were controlled. One Ss was excluded from statistical analyses due to failure to understand the written instructions and two were excluded for Beck Depression scores greater than 20. These latter Ss were given
Karin F. Helmers et al.
930 Table
1. Correlation coefficients for factors obtained by principal component analysis of paper-and-pencil subscales of impulsivity
Impulsive Eys-lmpulsivity Nonplanning Motor Attentional Eys-nonplanning Hare Psychopathy Eys-venturesome Thrill and Adventure Seeking Experience Seeking Boredom Susceptibility Disinhibited Eys-risktaking Statess Kipnis
0.862* 0.849* 0.809* 0.708* 0.635* 0.618* 0.225 0.011 0.319 0.250 0.342 0.347 - 0.070 0.257
*Significant factor loadings, P < 0.00
I.
Adventurous
Enjoyment of Experiences
State Sensation Seeking
0.46 1 0.198 0.189 0.437 0.153 0.597* 0.319 0.272 0.261 0.854* 0.703* 0.579* 0.022 0.257
0.146 -0.106 0.236 - 0.103 0.045 0.419 0.376 0.289 - 0.220 -0.119 0.328 0.404 0.802* 0.659*
0.101 0.131 0.379 0.022 0.604* 0.449 0.846* 0.770* 0.743* 0.170 0.410 0.495 0.008 0.504
appropriate information about the treatment of depression and where to seek treatment. Two additional S did not complete the Go/No Go task due to previous experience with the task.
RESULTS First and second order factors of impulsivity
A principal component analysis using an oblique rotation (See Table 1) was conducted on all paper-and-pencil measures of impulsivity. Four factors were obtained with eigenvalues greater than 1.0 accounting for 68.5% of the variance (Factor 1 = 35.9%, Factor 2 = 15.5%, Factor 3 = 9.3% and Factor 4 = 7.7%). Each subscale loaded significantly onto a factor if the correlation coefficient was greater than 0.52 [two times the critical value for a correlation coefficient at a = 0.01 (Stevens, 1992)]. The first factor is named Impulsive as Eys-Impulsivity, Eys-Nonplanning, and all subscales of the BIS demonstrated significant loadings. The second factor is named Adventurous as the Eys-Venturesome, Thrill and Adventure Seeking and Experience Seeking loaded onto this factor. The third factor is named State Sensation Seeking as States and the Kipnis scale load significantly onto this factor. The fourth factor measures enjoyment of unpredictability (Boredom Susceptibility) and new experiences (Disinhibition) and is named Enjoyment of Experiences. It should be noted that both the Hare and Eys-nonplanning loaded significantly onto two factors. A second order factor analysis was performed on the four primary factors using an oblique rotation. One factor was obtained and is named Impulsive/Adventurous (Imp/Adv). Correlation coefficients between the Imp/Adv and the four primary factors were 0.67 for Impulsive, 0.69 for Adventurous, 0.71 for Enjoyment of Experiences, and 0.41 for State Sensation Seeking.
Association between demographic
variables and factor scores
Demographic characteristics of the Ss are presented in Table 2. Univariate ANOVA’s were conducted to assess the association between demographic variables and factor scores. Results indicated that educational status (High school or less, some college, finished college and some graduate courses) was significantly associated with Adventurous [F(3,94) = 4.2, P < 0.011, and the Imp/Adv [F(3,94) = 3.1, P < 0.051. Inspection of means indicates that Ss who had completed some graduate work report the least Imp/Adv or adventurous scores in comparison to other educational categories. Employment status was associated with Imp/Adv [F(2,95) = 3.5, P < 0.051 with working Ss (full- or part-time employment collapsed together) reporting the least Imp/Adv in comparison to students and unemployed Ss. Whether or not a S smoked was significantly associated with Adventurous [F( 1,96) = 10.9, P < 0.011 and the Imp/Adv [F(1,96) = 4.6, P < 0.051, with smokers reporting greater Imp/Adv or adventurous scores than non-smokers. Amount of alcohol consumed during a week by self-report (none, l-3,4-6 or greater than 6 drinks/week) was significantly associated with Impulsive
Assessment of impulsivity
931
Table 2. Associations between primary and secondary factors and demographic variables
Impulsive Education 5 High school (I 1) Some college (41) College (23) Some graduate (23) Employment Students (44) Employed (24) Unemployed (30) Smoke No (50) Yes (48) Alcohol/week None (21) l-3 (30) 4L6 (24) > 6 (22) Age (24.8 2 4.6 yr) IQ (104.3? 15.9)
Adventurous
Enjoyment of Experiences
State Sensation Seeking
+
Imp/Adv
0.34 0.16 -0.14 - 0.30
- 0.33 0.17 0.37 -0.51
0.35 0.09 - 0.02 - 0.30
0.26 0.04 0.00 - 0.20
-0.01 - 0.28 0.24
0.16 - 0.32 0.02 *
0.09 - 0.25 0.08
0.1 I - 0.36 0.12
- 0.05 0.05 *
- 0.31 0.32 *
- 0.03 0.03
-0.16 0.17 *
* 0.22 0.19 0.09 - 0.53 * 0.13 - 0.46 0.17 * - 0.21 0.22 *
- 0.15
- 0.35
~ 0.17 - 0.13 0.55 - 0.36** 0.0 I
- 0.18 0.07 0.56 - 0.24** 0.39**
- 0.34 ~ 0.07 0.02 0.46 - 0.26** 0.00
- 0.28 - 0.04 - 0.17 0.52 - 0.09 - 0.19
- 0.44 - 0.19 - 0.06 0.81 - 0.39** 0.17
Mm: Mean scores are presented for the various categories of education. employment, smoking and drinking. Numbers for age and IQ are Pearson product-moment correlations with primary and secondary factor\. *Indicates P < 0.05 for the demographic category and factor. **Pi 0.05.
[F(3,93) = 2.9, P < 0.051, Adventurous [F(3,93) = 3.7, P < 0.051, State Sensation Seeking [F(3,93) = 2.9, P < 0.051, and Imp/Adv [F(3,93) = 7.6, P < 0.011. Inspection of these means indicate that Ss reporting more than six drinks per week scored greater on these factors than Ss reporting less than six drinks per week. Finally, age of Ss was inversely correlated with all factors except State Sensation Seeking and IQ was positively associated with Adventurous (P’s < 0.05, See Table 2). Association between demographic variables and behavioral measures The association between demographic variables and behavioral measures of impulsivity were assessed with Univariate ANOVA’s, but no significant associations were found. Age was positively correlated with Q score, RP-corn, Pun-omi, Pun-corn, Rew-corn, and PR-omi (r’s ranging between 0.21 and 0.36). IQ was inversely associated with RP-omi, Pun-omi, Pun-corn, Rew-corn (r’s ranging between - 0.22 and - 0.32). Association between factor scores and behavioral measures Table 3(a) presents the correlation matrix for the associations between the factors and behavioral measures of impulsivity. Using a correlation of 0.21 as a cutoff for significance (PC 0.05), 10 associations are considered significant. The factor score for Adventurous is positively associated with DALS, and inversely associated with Pun-omi, Rew-corn, and PR-corn. The State Sensation Seeking factor score is positively associated with Q score and Rew-corn. The factor score for Impulsivity is negatively associated with RP-omi. The Enjoyment of Experiences factor is not significantly associated with any of the behavioral measures of impulsivity. The second order factor, Imp/Adv is inversely associated with RP-omi, Rew-corn and PR-corn. When IQ and age are partialled out in the above correlation matrix, six associations remain significant [see Table 3(b)]. Impulsive and RP-omi are inversely associated, Adventurous and PR-corn are inversely associated, State Sensation Seeking is positively associated with Q score and Rew-corn, and Imp/Adv is positively associated with Q score and negatively associated with PR-corn. DISCUSSION
The current study demonstrated significant associations among demographic variables and the factors Adventurous and Imp/Adv, but not behavioral measures of impulsivity. Those Ss who score high on the Imp/Adv factor tend to be younger, have completed high school or some college, are unemployed or students, smoke cigarettes and drink more than six drinks/week. Similarly, those Ss
Karin F. Helmers et al.
932 Table
3a. Pearson
product
moment
correlations between impulsivity
Enjoyment of Experiences
Impulsive DALS MPFTime MPFTelT Q score RP-omi RP-corn Pun-omi Pun-corn Rew-omi Rew-corn PR-omi PR-corn *Indicates
factors
-
0.22* - 0.06 -0.14 0.08 -0.18 0.04 -0.19 - 0.23* -0.18 - 0.32* - 0.01 - 0.33*
0.05 - 0.01
0.01 - 0.03 - 0.27* - 0.07 -0.13 -0.15 0.08 -0.13 - 0.05 - 0.07
0.11 0.09 - 0.06 0.01 - 0.18 0.06 - 0.1 I - 0.09 0.1 I - 0.20 0.00 -0.19
and
behavioral
measures
State Sensation Seeking
Imp/Adv
- 0.07 - 0.09 0.04 0.27’ 0.07 - 0.05 0.17 0.07 0.05 0.24* 0.02 - 0.08
0.10 - 0.01 - 0.07 0.10 - 0.25* 0.00 -0.14 -0.18 0.02 - 0.22* - 0.03 - 0.27*
of
r’s 2 0.21. P < 0.05
Table 3b. Partial correlations
between factors and behavioral age and IQ
measures of impulsivity
Impulsive
Adventurous
Enjoyment of Experiences
State Sensation Seeking
DALS MFFTime Mmerr Q score RP-omi RP-corn Pun-omi Pun-corn Rew-omi Rew-corn PR-omi PR-corn
- 0.09 0.06 0.03 0.06 - 0.23* 0.01 - 0.01 - 0.02 0.13 0.01 0.05 - 0.03
0.17 - 0.05 - 0.07 0.14 - 0.06 0.13 - 0.05 - 0.05 -0.11 - 0.16 0.09 - 0.26*
0.09 0.14 - 0.05 0.07 -0.15 0.1 1 - 0.03 0.00 0.14 -0.13 0.07 -0.17
~ 0.04 ~ 0.07 0.01 0.31’ 0.06
*Indicates
r’s 2 0.21, P-C
- 0.04 0. IX 0.06 0.04 0.24* 0.03 -0.11
controlling
for
ImplAdv 0.06 0.05 - 0.04 0.2 I * ~0.18 0.10 0.0 I -0.01 0.09 - 0.07 0.10 - 0.24*
0.05.
who score high on the Adventurous factor tend to be younger, have a greater IQ, have some college or a college degree, smoke and drink more than six drinks/week. Ss who scored high on the Impulsive and State Sensation Seeking factor also tended to drink more than six drinks/week. Of the primary factors, the more Adventurous Ss tended to report more risk-taking behaviors (e.g. smoking and alcohol use) in spite of having greater IQ’s than less Adventurous Ss. These data suggest that there may be two groups of normal control Ss who participate in risk-taking behaviors. The Adventurous males may participate in more risk-taking behaviors in the spirit of wanting to experience these activities and the Impulsive Ss may participate in risk-taking behaviors without thinking about the consequences. Previous studies have demonstrated low correlations between paper-and-pencil and behavioral measures of impulsivity (Kipnis, 1971; Paulsen & Johnson, 1980; Milich & Kramer, 1984). The present study replicates and extends previous studies in that primary and secondary factors obtained from self-report measures of impulsivity also demonstrate low correlations with behavioral measures of impulsivity. It was hypothesized that factors associated with impulsivity would demonstrate greater errors and faster response in the behavioral tasks, whereas factors associated with thrill and adventure seeking would demonstrate decreased errors and slower response (or a more careful response). Some support for this hypothesis was obtained as the Adventurous factor is associated with the ability to respond slowly and carefully as evidenced by the ability to draw a line slowly, and demonstrates fewer errors as indicated by decreased errors of commission in three of the four conditions of the Go/No Go task. However, IQ is positively associated with the Adventurous score and inversely associated with errors on the Go/No Go task. When IQ is controlled, only one of the four significant findings remained, decreased errors of commission on the Pun-Rew condition. These results are similar to previous studies in children in which the number of significant correlations among self-report and behavioral measures of impulsivity decreases dramatically when IQ and age are controlled (Milich
Assessment of impulsivity
933
& Kramer, 1984; Paulsen & Johnson, 1980). Thus IQ may be considered a competing hypothesis that might account for an individual behaving impulsively on these tasks. The primary factor Impulsive is inversely associated with errors of omission during the Rew-Pun condition even after controlling for age and IQ. These results are in contrast to the hypothesis that impulsive Ss will demonstrate greater errors. Previous studies have demonstrated that impulsive individuals (e.g. delinquents, substance abusers, psychopaths), in comparison to healthy controls, report greater scores on self-report measures of impulsivity (Saklofske & Eysenck, 1983; Eysenck & McGurk, 1980; McGown, 1988) and demonstrate greater errors on measures of behavioral impulsivity (Rohrbeck & Twentyman, 1986; Newman & Kosson, 1986; Newman, 1987; Iaboni, Douglas & Baker, 1995). It is not clear why the current study demonstrates that the Impulsive factor is associated with less errors. It is possible that this may have been a type I error or that self-report measures and behavioral tasks are measuring different constructs of impulsivity in individuals with a life history of behaving impulsively and in normal control Ss. The primary factor, State Sensation Seeking is associated with behavioral impulsivity (even after controlling for IQ and age) as observed by increased errors of commission during the Reward condition of the Go/No Go task and increased Q scores on the Porteus mazes. Thus Ss scoring high on the State Sensation Seeking factor endorse feeling elated, adventurous, imaginative, zany, enthusiastic, etc. and demonstrate behavioral impulsivity. This may be similar to the Carefree factor found in an item analysis of numerous impulsivity and sensation seeking questionnaires (Gerbing, Ahadi & Patton, 1987). Individuals who score high on the Carefree factor report being Happy-go-lucky which suggests feeling enthusiastic, zany, etc. It is of interest that this State Sensation Seeking factor is not confounded by the demographic variables of age, IQ and educational level. In healthy control Ss, it appears that the State Sensation Seeking factor may demonstrate greater associations with behavioral impulsivity measures than self-report measures of impulsivity. The Enjoyment of Experiences factor did not demonstrate any association with behavioral impulsivity measures. The second order factor, Impulsive/Adventurous, appears to be a mixture of both impulsive and adventurous characteristics in that Ss demonstrate less errors on the Go/No Go task, but increased Q scores. It is possible that in these Ss the monetary reward in the Go/No Go task gave them sufficient motivation to overcome their tendency to behave impulsively which was indicated by their performance in the Porteus Mazes. Paper-and-pencil and behavioral measures of impulsivity may be subject to various methodological biases which can alter the accuracy of scores. For example, paper-and-pencil questionnaires are subject to self-awareness and demand characteristics that may lead to inaccurate scores. Behavioral tasks are not dependent upon self-awareness, and the use of monetary rewards as in the Go/No Go task should override any demand characteristics. Thus the Go/No Go task may be without some of the limitations found in other behavioral and paper-and-pencil measures of impulsivity. Research using the Go/No Go task has increased during the past decade. Gray et al. (1983) has proposed that impulsive Ss are insensitive to stimuli associated with punishment, may display an excessive sensitivity to stimuli associated with reward, and display a deficit in passive avoidance learning. Passive avoidance learning is the ability to withhold a response that would have lead to punishment. Thus, impulsive Ss should exhibit greater errors of commission than nonimpulsive Ss. Previous work by Newman and colleagues has demonstrated that the Go/No Go task discriminates between impulsive and non-impulsive Ss. Specifically, in comparison to controls, incarcerated psychopaths, extraverts, and juvenile delinquents demonstrated greater errors of commission, but similar errors of omission (Newman, 1987; Newman & Kosson, 1986; Newman, Widom & Nathan, 1985; Patterson, Kosson & Newman, 1987). These differences were found only in the reward-punishment condition (Newman, Patterson, Howland & Nichols, 1990) in which Ss demonstrate a dominant response for reward (S + ) that makes it difficult to inhibit responding to punished stimuli (S - ). Thus, the Go/No Go task is an operationalization of impulsivity as theorized by Gray et al. (1983), and research has supported this operationalization in various clinical and healthy control 5s. While the Go/No Go task is an operationalization of impulsivity as theorized by Gray et al. (1983), some of the other behavioral tasks fit more closely other definitions of impulsivity. Thus, high scores on the Porteus maze may reflect the inability to plan ahead (Barratt & Patton, 1983; Buss & Plomin, 1975; Eysenck & Eysenck, 1977). Other definitions include the tendency to respond quickly to stimuli (Barratt & Patton, 1983; Buss & Plomin, 1975) and acting without thinking (Barratt & Patton, 1983)
934
Karin F. Helmers et al.
which may be operationalized by the MFFT, and in some ways fits the DALS. A fundamental difference between these latter definitions and that of Gray et al. (1983) concerns the concept of punishment. The idea that impulsivity is the failure to withhold a response that will lead to punishment is an appealing one from the clinical point of view, as it is the failure to control behaviors that lead to punishment that characterizes the clinically impulsive patient. In the current study, there were few associations between the factors and the Go/No Go task. Of the four primary factors, State Sensation Seeking appears to be most closely associated with behavioral impulsivity, whereas Impulsive and Adventurous appears to be inversely associated with behavioral impulsivity. These results are in contrast to previous studies in which the Go/No Go task discriminated between impulsive and nonimpulsive Ss (Newman, 1987; Patterson et al., 1987; Iaboni et al., 1995). These differences may be due to the populations sampled, comparing healthy control Ss with clinical Ss. It is possible that the impulsive behavior found in clinical populations is independent from and not an extreme form of impulsive behavior found in healthy control Ss. A complicating factor is that age and IQ appear to be associated with both self-report and behavioral measures of impulsivity. Further studies are needed to evaluate the Go/No Go task simultaneously in normal volunteers and clinical populations in order to determine if behaving impulsively is a continuum or if these groups of Ss demonstrate independent behaviors of impulsivity. It is possible that the State Sensation Seeking factor may differentiate normal volunteers who behave impulsively from clinical populations. In sum, the measurement of impulsivity in the laboratory is still in a preliminary stage. The present data suggests that a state measurement of sensation seeking may be the most closely associated with behaving impulsively in healthy control Ss. Of the other factors, Adventurous appears to be associated with less errors and greater motor inhibition, though IQ scores confound this relationship. Of the behavioral tasks examined, the Go/No Go task is appealing because it appears to be the operationalization of impulsivity as seen in the clinically impulsive patient. Further studies are needed to assess whether impulsivity in the clinical population is an extreme form of impulsivity observed in control Ss, or if impulsivity is independent in these two groups. Acknowledgemenrs-KFH is a recipient of a research fellowship from the Medical Research Council of Canada and a computer equipment grant from Marion Merrell Dow, Inc. This research was supported by a grant from the Medical Research Council of Canada. We gratefully acknowledge the assistance of Anna Guyetsky and Franceen Lenoff for their help in the data collection.
REFERENCES American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders: DSM-IV (4th edn). Washington, DC: American Psychiatric Association. Barratt, E. S. &Patton, J. H. (1983). Impulsivity: Cognitive, behavioral and psychophysiological correlates. In M. Zuckerman. (Ed.), Biological buses of sensation seeking, impulsivity, and anxiety (pp. 77-l 16). Hillsdale, NY: Lawrence Erlbaum. Beck, A. & Steer, R. (1987) The Beck depression inventory. New York: Harcourt Brace Jovanovic. Brown, G. L., Ebert, M. H., Goyer, P. F., Jimerson, D. C., Klein, W. J., Bunney, W. E., et al. (1982). Aggression, suicide, and serotonin: Relationships to CSF amine metabolites. American Journal of Psychiatry, 139, 74 l-746. Buss, A. & Plomin, R. (1975). A temperament theory of personality development, New York: Wiley. 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, 5768.
Eysenck, S. B. G. & McGurk, B. J. (1980). Impulsiveness and venturesomeness in a detention center population. Psychological Reports, 47, 1299-1306.
Gerbing, D. W., Ahadi, S. A. & Patton, J. H. (1987). Toward a conceptualization of impulsivity: Components across the behavioral and self-report domains. Multivariate Behavioral Research, 22, 357-379. Glueck, S. & Glueck, E. (1968) Delinquents and nondelinquents in perspective. Cambridge, MA: Harvard University Press. Gray, J. A., Owen, S., Davis, N. & Tsaltas, E. (1983). Psychological and physiological relations between anxiety and impulsivity. In M. Zuckerman (Ed.), Biological bases of sensation seeking, impulsivity, and nnxiety, Hillsdale, NY: Lawrence Erlbaum. Hare, R. D. (1991) The Hare psychopathy checklist-revised, Toronto, ON: Multi-Health systems. Iaboni, F., Douglas, V. I. & Baker, A. G. (1995). Effects of reward and response costs on inhibition in ADHD children. Journal of Abnormal Psychology, 104, 232-240.
Kagan, J. (1966). Reflection-Impulsivity: The generality and dynamics of conceptual tempo. Journal ofAbnormal Psychology, 71, 17-24. Kaufman, A. S. (1990) Assessing adolescent and adult intelligence, Boston, MA: Allyn and Bacon. Kipnis, D. (1971) Character structure and impulsiveness, New York: Academic Press.
Assessment
of impulsivity
935
Lidberg, L., Tuck, J. R., Asberg, M., Scaha-Tomba, G. P. & Bertilsson, L. (1985). Homicide, suicide and CSF SHIAA. Acra Psychiatrica Scandinavica,
71, 230-236.
Linnoila, M., Virkkunen, M., Scheinin, M., Nuutila, A., Rimon, R. & Goodwin, F. K. (1983). Low cerebrospinal fluid 5-hydroxindoleacetic acid concentration differentiates impulsive from nonimpulsive violent behavior. Life Sciences, 33, 2609-26 14.
McGown, W. G. (1988). Multi-impulsive personality disorder and multiple substance abuse: Evidence from members of self-help groups. British Journal of Addiction, 83, 43 l-432. Milich, R. & Kramer, J. (1984). Reflections on impulsivity: An empirical investigation of impulsivity as a construct. In K. Gadow (Ed.), Advances in learning and behavioral disabilities (Vol. 3, pp. 57-93). Greenwich, CT: JAI Press. Newman, J. P. (1987). Reaction to punishment in extraverts and psychopaths: Implications for the impulsive behavior of disinhibited individuals. Journal of Research in Personality, 21, 464480. Newman, J. P. & Kosson, D. S. (1986). Passive avoidance learning in psychopathic and nonpsychopathic offenders. Journal of Abnormal Psychology, 95, 252-256.
Newman, 3. P., Patterson, C. M., Howland, E. W. & Nichols, S. L. (1990). Passive avoidance in psychopaths: The effects of reward. Personality and Individual Differences, II, 1101-I 114. Newman, J. P., Widom, C. S. & Nathan, S. (1985). Passive avoidance in syndromes of disinhibition: Psychopathy and extraversion. Journal of Personality and Social Psychology, 48, 13 16-l 327. Oas, P. (I 983). Impulsive behavior and assessment of impulsivity with hospitalized adolescents, Psychological Reports, 53, 764-766.
Patterson, C. M., Kosson, D. S. &Newman, J. P. (1987). Reaction to punishment, reflectivity, and passive avoidance learning in extraverts. Journal of Personality and Social Psychology, 52, 565-575. Paulsen, K. &Johnson, M. (1980). Impulsivity: A multidimensional concept with developmental aspects. JournalofAbnormal Child Psychology, 8, 269-277.
of suicidality, aggressivity and impulsivity. Clinical Neuropharmacology, 9, 38&382. Porteus, S. D. (1965). Porteus maze tesrt Fifty years’ application, Palo Alto, CA: Pacific Books. Prior, M. & Sanson, A. (1986). Attention deficit disorder with hyperactivity: A critique. Journal of Child Psychology and Psychiatry, 27, 307-3 19. Plutchik, R. & van Praag, H. M. (1986). The measurement
Rohrbeck, C. A. & Twentyman, C. T. (1986). Multimodal assessment of impulsiveness in abusing, neglecting, and nonmaltreating mothers and their preschool children. Journal of Consulting and Clinical Psychology, 54, 23 l-236. Roy, A. (1983).Family history of suicide. Archives of General Psychiatry, 40, 97 l-974. Roy, A., Adinoff, B. & Linnoila, M. (1988). Acting out hostility in normal volunteers: Negative correlation with levels of SHIAA in cerebrospinal fluid. Psychiatry Research, 24, 187-194. Roy, A., Virkkunen, M. & Linnoila, M. (1990). Serotonin in suicide, violence, and alcoholism. In E. F. Coccaro & D. L. Murphy (Eds), Serotonin in major psychiatric disorders, Washington, DC: American Psychiatric Press. Saklofske, D. H. & Eysenck, S. B. G. (1983). Impulsiveness and venturesomeness in Canadian children. Psychological Reports, 52, 147-152.
Soubrie, P. ( 1986). Reconciling the role of central serotonin neurons in human and animal behavior. The Behavioral and Bruin Sciences.
9, 3 19-364.
Stevens, J. (I 992) Applied multivariate statisticsfor the social sciences, 2nd edn. Hillsdale, NJ: Lawrence Erlbaum Associates. Valzelli, L. ( 1984). Reflections on experimental and human pathology of aggression. Progress in Neuro-Psychopharmacological and Biological Psychiatry, 8, 3 I l-325.
Zuckerman, M. (1979).Sensation seeking: Beyond the optimal level of arousal Hillsdale, NJ: Lawrence Erlbaum. Zuckerman, M. (1993). Sensation seeking and impulsivity: A marriage of traits made in biology? In W. G. McCown, J. L. Johnson & M. B. Shure. (Eds), The impulsive clienf: Theory, research and freatment, Washington. DC: American Psychological Association.
Person. individ. Di@T Vol. 19. No. 6, pp. 927-935. 1995 Copyright 01995 Elsevier Science Ltd 0191~8869(95)00107-7 Printed in Great Britain. All rights reserved 0191-8869/95 $9.50 + 0.00
ASSESSMENT OF MEASURES OF IMPULSIVITY IN HEALTHY MALE VOLUNTEERS Karin F. Helmers,‘,* Simon N. Young’ and Robert 0. Pihl’,’ Departments of ‘Psychiatry and *Psychology, McGill University, 1033 Pine Avenue West, Montreal, Quebec, Canada H3A I Al (Received 20 October 1994; received for publication 22 June 1995)
Summary-The current study evaluated the relationship between self-report and behavioral measures of impulsivity in 98 young adult male subjects. Principal components analysis of the self-report measures revealed four primary factors: the first demonstrated significant loadings with Eysenck’s Impulsivity and Nonplanning scales and all subscales of the Barratt Impulsivity Scale, the second with Eysenck’s Venturesome, Thrill and Adventure Seeking and Experience Seeking Scales, the third with the Boredom Susceptibility and Disinhibition subscales of Zuckerman’s Sensation Seeking Scale, and the fourth with Zuckerman’s State Sensation Seeking and the Kipnis Scale. The relationship between scores on these primary factors and performance on four behavioral measures of impulsivity, the Draw a line Slowly, Matching Familiar Figures, the Porteus Maze, and Gomo Go Discrimination tasks, was investigated. Only the fourth primary factor demonstrated a positive relationship, with errors of commission on the Go/No Go Task, and with the Q score in the Porteous Maze. The fourth factor was the only one that was independent of educational level, age and IQ. Of the behavioral tasks, the Go/No Go task is an operationalization of impulsivity as defined by Gray, Owen, Davis & Tsaltas (Biological Bases ofSensation Seeking, ImpuWiq, andilnrieg, 1983) and may not be subject to demand characteristics found in the other behavioral and paper and pencil measures of impulsivity. Increased understanding of impulsivity is needed, and the fourth factor we defined, and the Go/No Go task, may hold promise for future research on one type of impulsivity.
INTRODUCTION
Interest in the evaluation of impulsivity has increased due to the association of impulsivity with aggressive behaviors and because of the suggestion that specific neurotransmitter abnormalities may underlie impulsive behaviors. Impulsivity is central to a number of psychiatric disorders either as a principal diagnostic feature (i.e. histrionic, antisocial, borderline personality disorders) or as a feature commonly associated with a diagnostic category (i.e. conduct, narcissistic personality and compulsive personality disorders). Impulsivity may also be an important precipitant of aggression (Valzelli, 1984; Roy, Adinoff & Linnoila, 1988) and suicide (Roy, Virkkunen & Linnoila, 1990; Brown, Ebert, Goyer, Jimerson, Klein & Bunney, 1982), and has a genetic component (Roy, 1983; Plutchik & van Praag, 1986). It is believed that low levels of the central nervous system neurotransmitter serotonin (5hydroxytryptamine, 5-HT) may be related to impulsivity, which in turn may be the common link between behavioral acts such as suicide and violent assault, and the impulsive behavioral components of other psychiatric disorders (Soubrie, 1986; Linnoila, Virkkunen, Scheinin, Nuutila, Rimon & Goodwin, 1983; Lidberg, Tuck, Asberg, Scalia-Tomba & Bertilsson, 1985). However, very few clinical studies have measured impulsivity directly. Rather the role of impulsivity is typically inferred from behavioral psychopathology (i.e. suicide attempts, violent assaultive behavior). Major issues in the study of impulsivity are the definition of this personality construct and its measurement in controlled laboratory settings. The definition of impulsivity varies across studies, including the failure to evaluate a situation as risky or dangerous (Eysenck & McGurk, 1980), acting without thinking (Barratt & Patton, 1983), the inability to plan ahead (Barratt & Patton, 1983; Eysenck & Eysenck, 1977; Buss & Plomin, 1975), the tendency to respond quickly to stimuli rather than inhibiting responses (Barratt & Patton, 1983; Prior & Sanson, 1986; Buss & Plomin, 1975), and the failure to withhold a response that will lead to punishment or a deficit in passive avoidance learning (Gray, Owen, Davis & Tsaltas, 1983). In a clinical setting, impulsivity is generally defined by deviant behaviors in the life history of an individual. For example, Glueck and Glueck (1968) stated that impulsivity is identified by the repetition of the deviant behavior and not the uniqueness of deviant behavior. Similarly, Oas (1983) defined chronic impulsivity as a persistent tendency to emit *To whom all correspondence should be addressed. 921
928
Karin F. Helmers et al.
inappropriate or pathological behavior characterized by a lack of reflectivity and delay. The DSM-IV characterizes impulse-control disorders as the failure to resist an impulse, drive, or temptation to perform an act that is harmful to the person or to others (American Psychiatric Association, 1994). These latter definitions do not allow for the evaluation of impulsivity in normal control 5s. Further, the measurement of impulsivity should be separate from the clinical diagnostic categories. Within the laboratory, the assessment of impulsivity can be accomplished by several paper-and-pencil questionnaires: Eysenck’s impulsivity questionnaires (Eysenck & Eysenck, 1977; Eysenck & McGurk, 1980) the Kipnis questionnaire (Kipnis, 1971), and Barratt’s Impulsivity scale (Barrett & Patton, 1983). Impulsivity also appears to be a component of psychopathy (Hare, 1991) and sensation seeking (Zuckerman, 1979). However, the subscale of Thrill and Adventure Seeking obtained from the Sensation Seeking Scale appears to be associated with risk-taking behaviors that are often carefully planned to minimize physical danger as found in mountain climbers (Zuckerman, 1993). Thus the subscale of Thrill and Adventure Seeking may be conceptually different from other self-report measures of impulsivity. Behavioral measures of impulsivity have been constructed in order to measure different components of impulsivity. These include the inability to plan ahead (i.e. Porteus Maze, Trails A and B), tests measuring speed of response and number of errors (i.e. Matching Familiar Figures Test, reaction time tests), ability to inhibit responding or passive avoidance learning (i.e. Go/No Go discrimination tasks), and ability to inhibit motor control (i.e. “Draw a line slowly” or “Walk a line slowly”). The purpose of the present study was to examine the relationship between behavioral measures of impulsivity and first and second order factors derived from paper-and-pencil measures of impulsivity. We hypothesized that factors related to impulsivity would be associated with increased errors and speed of response, whereas factors associated with thrill and adventure seeking would be associated with decreased errors and speed of response.
METHOD Subjects
A total of 101 English speaking males were recruited via a newspaper advertisement in the metropolitan area of Montreal. Ss were between the ages of 18 and 35 yr, denied any current psychopathology or psychiatric history (e.g. affective disorders, assaultive violent behavior), and medical problems (e.g. cardiovascular disease, diabetes, gastrointestinal problems). Each S completed a battery of paper-and-pencil questionnaires and behavioral tasks. Assessment procedure Paper-and-pencil questionnaires. The following paper-and-pencil questionnaires were completed by Ss: Barratt Impulsivity Scale [BIS, version 11 (Barratt & Patton, 1983)] with subscales measuring Motor, Nonplanning and Attentional impulsivity obtained from 34 items scored as a 0 = ‘rarely’ to 3 = ‘almost always’. Several scales are derived from Eysenck’s questionnaires which include narrow Impulsivity (Eys-lmpulsivity) and Venturesomeness (Eys-Venturesome) (Eysenck & McGurk, 1980), risk-taking (Eys-Risktaking) and non-planning (Eys-Nonplanning) (Eysenck & Eysenck, 1977). The Kipnis scale ( 197 1) is derived from 56 autobiographical items relevant to activities during childhood and adulthood. The Sensation Seeking Scale (SSS) (Zuckerman, 1979) consists of the subscales entitled Boredom Susceptibility, Disinhibition, Experience Seeking and Thrill and Adventure Seeking. These latter subscales are measured each by 10 items in which Ss choose between two statements. State sensation seeking (States) was measured by 15 adjectives rated for the current time period and scored from 1 = ‘not at all’ to 5 = ‘very much’ (Zuckerman, 1979). The Hare psychopath checklist consists of 29 items scored on a five-point Likert scale from ‘Strongly disagree’ to ‘Strongly agree’ (Hare, 1991). Ss also completed the Beck Depression Inventory (Beck & Steer, 1987), a 21 item scale in which the S chooses one out of four statements which best described his mood during the past l-2 weeks. Furthermore, the Vocabulary subtest of the Wechsler Adult Intelligence Scale-Revised (WAIS-R) was administered. Raw scores were converted to scaled scores, and the intelligence quotient (IQ) equivalent score was determined by the appropriate age group. Previous research on the WAIS-R has
Assessment of impulsivity
929
indicated that the vocabulary subtest, in comparison to other subtests, demonstrates the greatest association with the full-scale IQ score (reviewed in Kaufman, 1990). Behavioral measures Matching familiarfigures test (MFFT). A series of 10 sets of figures in which the S is required to identify one of six very similar pictures to a ‘target’ picture (Kagan, 1966). The mean latency of response to the first choice (MFFTime) and the total number of errors (MFFTerr) were calculated. Porteus maze. The adult Vineland Revised version of the Porteus mazes (Porteus, 1965) were scored for impulsivity (Q score). Q scores are obtained by measuring number of times the pencil was lifted, touched the boundary, etc. Draw a line slowly (DALS). This test measures inhibition of motor activity (Rohrbeck & Twentyman, 1986). Ss draw a line from top to bottom of a 9” X 1” column without crossing the boundaries as slowly as possible (measured in msec). ‘Go/No Go ’ discrimination tusk. Ss learn by trial-and-error when to respond by pressing a button and when to withhold responding. The stimuli are eight two-digit numbers, repeated 10 times in different, randomized orders for a total of 80 trials. Four of the eight stimuli are designated as positive stimuli (S + ) and the other half as negative stimuli (S - ). Numbers range from 01 to 99, were randomly chosen, and were presented as white lettering against a blue computer screen. Stimuli remain on the screen for 3 set or until the S presses the button. The interstimulus interval is 1 sec. Correct responses are signalled with a high-pitched tone, the word ‘correct’ on the computer screen, and five cents is added to the running tally of the Ss’ earnings. Incorrect responses are signalled by a low pitched tone, ‘wrong’ on the computer screen, and the subtraction of five cents from the running tally of the Ss’ earnings. Four conditions were tested in randomized order. For each condition, errors of commission (failure to inhibit responses to negative stimuli = passive avoidance errors) and errors of omission (failure to respond to positive stimuli) were calculated. The four conditions were: a reward-punishment (RP) condition in which Ss start with fifty cents, responses to S + ‘s are rewarded, and S - ‘s are punished. Errors of omission (RP-omi) and commission (RP-corn) were calculated. A punishment-only condition (Pun), in which Ss begin with $2.00, have no opportunity to win more money, but lose money if they respond to S - ‘s or fail to respond to S + ‘s. Pun-omi and Pun-corn were calculated (errors of omission and commission during Punishment only condition). A reward-only condition (Rew) in which Ss start with no money, but could earn money by responding to S + ‘s and not responding to S - ‘s (Rew-omi and Rew-corn were calculated for errors of omission and commission). Finally, a punishment-reward (PR) condition in which subjects start with fifty cents and a non-response to S + is punished and a non-response to S - is rewarded (PR-omi and PR-corn were calculated for errors of omission and commission, respectively). Prior to each condition there is a pretreatment of 12 trials in which the S + :S - stimuli are presented in randomized order in a 2: 1 ratio.
STATISTICAL
ANALYSES
To obtain primary factor scores, a principal components analysis using an oblique rotation was conducted on all paper-and-pencil subscales related to impulsivity (Beck depression scores were excluded). To obtain a secondary factor, a principal components analysis using an oblique rotation was performed on the primary factor scores to obtain a General Impulsivity score for each S. Behavioral measures of impulsivity did not allow for a factor analysis because of low inter-correlations and low communalities (Stevens, 1992). To assess the association between demographic variables and impulsivity, Analyses of Variance (ANOVA) were conducted on categorical demographic variables and primary and second factor scores of impulsivity. Pearson product-moment correlations evaluated the relationships between age, IQ scores and obtained factors. The relationship between behavioral measures and primary and secondary factors were evaluated by Pearson product-moment correlations and partial correlations in which demographic variables (age and IQ) were controlled. One Ss was excluded from statistical analyses due to failure to understand the written instructions and two were excluded for Beck Depression scores greater than 20. These latter Ss were given
Karin F. Helmers et al.
930 Table
1. Correlation coefficients for factors obtained by principal component analysis of paper-and-pencil subscales of impulsivity
Impulsive Eys-lmpulsivity Nonplanning Motor Attentional Eys-nonplanning Hare Psychopathy Eys-venturesome Thrill and Adventure Seeking Experience Seeking Boredom Susceptibility Disinhibited Eys-risktaking Statess Kipnis
0.862* 0.849* 0.809* 0.708* 0.635* 0.618* 0.225 0.011 0.319 0.250 0.342 0.347 - 0.070 0.257
*Significant factor loadings, P < 0.00
I.
Adventurous
Enjoyment of Experiences
State Sensation Seeking
0.46 1 0.198 0.189 0.437 0.153 0.597* 0.319 0.272 0.261 0.854* 0.703* 0.579* 0.022 0.257
0.146 -0.106 0.236 - 0.103 0.045 0.419 0.376 0.289 - 0.220 -0.119 0.328 0.404 0.802* 0.659*
0.101 0.131 0.379 0.022 0.604* 0.449 0.846* 0.770* 0.743* 0.170 0.410 0.495 0.008 0.504
appropriate information about the treatment of depression and where to seek treatment. Two additional S did not complete the Go/No Go task due to previous experience with the task.
RESULTS First and second order factors of impulsivity
A principal component analysis using an oblique rotation (See Table 1) was conducted on all paper-and-pencil measures of impulsivity. Four factors were obtained with eigenvalues greater than 1.0 accounting for 68.5% of the variance (Factor 1 = 35.9%, Factor 2 = 15.5%, Factor 3 = 9.3% and Factor 4 = 7.7%). Each subscale loaded significantly onto a factor if the correlation coefficient was greater than 0.52 [two times the critical value for a correlation coefficient at a = 0.01 (Stevens, 1992)]. The first factor is named Impulsive as Eys-Impulsivity, Eys-Nonplanning, and all subscales of the BIS demonstrated significant loadings. The second factor is named Adventurous as the Eys-Venturesome, Thrill and Adventure Seeking and Experience Seeking loaded onto this factor. The third factor is named State Sensation Seeking as States and the Kipnis scale load significantly onto this factor. The fourth factor measures enjoyment of unpredictability (Boredom Susceptibility) and new experiences (Disinhibition) and is named Enjoyment of Experiences. It should be noted that both the Hare and Eys-nonplanning loaded significantly onto two factors. A second order factor analysis was performed on the four primary factors using an oblique rotation. One factor was obtained and is named Impulsive/Adventurous (Imp/Adv). Correlation coefficients between the Imp/Adv and the four primary factors were 0.67 for Impulsive, 0.69 for Adventurous, 0.71 for Enjoyment of Experiences, and 0.41 for State Sensation Seeking.
Association between demographic
variables and factor scores
Demographic characteristics of the Ss are presented in Table 2. Univariate ANOVA’s were conducted to assess the association between demographic variables and factor scores. Results indicated that educational status (High school or less, some college, finished college and some graduate courses) was significantly associated with Adventurous [F(3,94) = 4.2, P < 0.011, and the Imp/Adv [F(3,94) = 3.1, P < 0.051. Inspection of means indicates that Ss who had completed some graduate work report the least Imp/Adv or adventurous scores in comparison to other educational categories. Employment status was associated with Imp/Adv [F(2,95) = 3.5, P < 0.051 with working Ss (full- or part-time employment collapsed together) reporting the least Imp/Adv in comparison to students and unemployed Ss. Whether or not a S smoked was significantly associated with Adventurous [F( 1,96) = 10.9, P < 0.011 and the Imp/Adv [F(1,96) = 4.6, P < 0.051, with smokers reporting greater Imp/Adv or adventurous scores than non-smokers. Amount of alcohol consumed during a week by self-report (none, l-3,4-6 or greater than 6 drinks/week) was significantly associated with Impulsive
Assessment of impulsivity
931
Table 2. Associations between primary and secondary factors and demographic variables
Impulsive Education 5 High school (I 1) Some college (41) College (23) Some graduate (23) Employment Students (44) Employed (24) Unemployed (30) Smoke No (50) Yes (48) Alcohol/week None (21) l-3 (30) 4L6 (24) > 6 (22) Age (24.8 2 4.6 yr) IQ (104.3? 15.9)
Adventurous
Enjoyment of Experiences
State Sensation Seeking
+
Imp/Adv
0.34 0.16 -0.14 - 0.30
- 0.33 0.17 0.37 -0.51
0.35 0.09 - 0.02 - 0.30
0.26 0.04 0.00 - 0.20
-0.01 - 0.28 0.24
0.16 - 0.32 0.02 *
0.09 - 0.25 0.08
0.1 I - 0.36 0.12
- 0.05 0.05 *
- 0.31 0.32 *
- 0.03 0.03
-0.16 0.17 *
* 0.22 0.19 0.09 - 0.53 * 0.13 - 0.46 0.17 * - 0.21 0.22 *
- 0.15
- 0.35
~ 0.17 - 0.13 0.55 - 0.36** 0.0 I
- 0.18 0.07 0.56 - 0.24** 0.39**
- 0.34 ~ 0.07 0.02 0.46 - 0.26** 0.00
- 0.28 - 0.04 - 0.17 0.52 - 0.09 - 0.19
- 0.44 - 0.19 - 0.06 0.81 - 0.39** 0.17
Mm: Mean scores are presented for the various categories of education. employment, smoking and drinking. Numbers for age and IQ are Pearson product-moment correlations with primary and secondary factor\. *Indicates P < 0.05 for the demographic category and factor. **Pi 0.05.
[F(3,93) = 2.9, P < 0.051, Adventurous [F(3,93) = 3.7, P < 0.051, State Sensation Seeking [F(3,93) = 2.9, P < 0.051, and Imp/Adv [F(3,93) = 7.6, P < 0.011. Inspection of these means indicate that Ss reporting more than six drinks per week scored greater on these factors than Ss reporting less than six drinks per week. Finally, age of Ss was inversely correlated with all factors except State Sensation Seeking and IQ was positively associated with Adventurous (P’s < 0.05, See Table 2). Association between demographic variables and behavioral measures The association between demographic variables and behavioral measures of impulsivity were assessed with Univariate ANOVA’s, but no significant associations were found. Age was positively correlated with Q score, RP-corn, Pun-omi, Pun-corn, Rew-corn, and PR-omi (r’s ranging between 0.21 and 0.36). IQ was inversely associated with RP-omi, Pun-omi, Pun-corn, Rew-corn (r’s ranging between - 0.22 and - 0.32). Association between factor scores and behavioral measures Table 3(a) presents the correlation matrix for the associations between the factors and behavioral measures of impulsivity. Using a correlation of 0.21 as a cutoff for significance (PC 0.05), 10 associations are considered significant. The factor score for Adventurous is positively associated with DALS, and inversely associated with Pun-omi, Rew-corn, and PR-corn. The State Sensation Seeking factor score is positively associated with Q score and Rew-corn. The factor score for Impulsivity is negatively associated with RP-omi. The Enjoyment of Experiences factor is not significantly associated with any of the behavioral measures of impulsivity. The second order factor, Imp/Adv is inversely associated with RP-omi, Rew-corn and PR-corn. When IQ and age are partialled out in the above correlation matrix, six associations remain significant [see Table 3(b)]. Impulsive and RP-omi are inversely associated, Adventurous and PR-corn are inversely associated, State Sensation Seeking is positively associated with Q score and Rew-corn, and Imp/Adv is positively associated with Q score and negatively associated with PR-corn. DISCUSSION
The current study demonstrated significant associations among demographic variables and the factors Adventurous and Imp/Adv, but not behavioral measures of impulsivity. Those Ss who score high on the Imp/Adv factor tend to be younger, have completed high school or some college, are unemployed or students, smoke cigarettes and drink more than six drinks/week. Similarly, those Ss
Karin F. Helmers et al.
932 Table
3a. Pearson
product
moment
correlations between impulsivity
Enjoyment of Experiences
Impulsive DALS MPFTime MPFTelT Q score RP-omi RP-corn Pun-omi Pun-corn Rew-omi Rew-corn PR-omi PR-corn *Indicates
factors
-
0.22* - 0.06 -0.14 0.08 -0.18 0.04 -0.19 - 0.23* -0.18 - 0.32* - 0.01 - 0.33*
0.05 - 0.01
0.01 - 0.03 - 0.27* - 0.07 -0.13 -0.15 0.08 -0.13 - 0.05 - 0.07
0.11 0.09 - 0.06 0.01 - 0.18 0.06 - 0.1 I - 0.09 0.1 I - 0.20 0.00 -0.19
and
behavioral
measures
State Sensation Seeking
Imp/Adv
- 0.07 - 0.09 0.04 0.27’ 0.07 - 0.05 0.17 0.07 0.05 0.24* 0.02 - 0.08
0.10 - 0.01 - 0.07 0.10 - 0.25* 0.00 -0.14 -0.18 0.02 - 0.22* - 0.03 - 0.27*
of
r’s 2 0.21. P < 0.05
Table 3b. Partial correlations
between factors and behavioral age and IQ
measures of impulsivity
Impulsive
Adventurous
Enjoyment of Experiences
State Sensation Seeking
DALS MFFTime Mmerr Q score RP-omi RP-corn Pun-omi Pun-corn Rew-omi Rew-corn PR-omi PR-corn
- 0.09 0.06 0.03 0.06 - 0.23* 0.01 - 0.01 - 0.02 0.13 0.01 0.05 - 0.03
0.17 - 0.05 - 0.07 0.14 - 0.06 0.13 - 0.05 - 0.05 -0.11 - 0.16 0.09 - 0.26*
0.09 0.14 - 0.05 0.07 -0.15 0.1 1 - 0.03 0.00 0.14 -0.13 0.07 -0.17
~ 0.04 ~ 0.07 0.01 0.31’ 0.06
*Indicates
r’s 2 0.21, P-C
- 0.04 0. IX 0.06 0.04 0.24* 0.03 -0.11
controlling
for
ImplAdv 0.06 0.05 - 0.04 0.2 I * ~0.18 0.10 0.0 I -0.01 0.09 - 0.07 0.10 - 0.24*
0.05.
who score high on the Adventurous factor tend to be younger, have a greater IQ, have some college or a college degree, smoke and drink more than six drinks/week. Ss who scored high on the Impulsive and State Sensation Seeking factor also tended to drink more than six drinks/week. Of the primary factors, the more Adventurous Ss tended to report more risk-taking behaviors (e.g. smoking and alcohol use) in spite of having greater IQ’s than less Adventurous Ss. These data suggest that there may be two groups of normal control Ss who participate in risk-taking behaviors. The Adventurous males may participate in more risk-taking behaviors in the spirit of wanting to experience these activities and the Impulsive Ss may participate in risk-taking behaviors without thinking about the consequences. Previous studies have demonstrated low correlations between paper-and-pencil and behavioral measures of impulsivity (Kipnis, 1971; Paulsen & Johnson, 1980; Milich & Kramer, 1984). The present study replicates and extends previous studies in that primary and secondary factors obtained from self-report measures of impulsivity also demonstrate low correlations with behavioral measures of impulsivity. It was hypothesized that factors associated with impulsivity would demonstrate greater errors and faster response in the behavioral tasks, whereas factors associated with thrill and adventure seeking would demonstrate decreased errors and slower response (or a more careful response). Some support for this hypothesis was obtained as the Adventurous factor is associated with the ability to respond slowly and carefully as evidenced by the ability to draw a line slowly, and demonstrates fewer errors as indicated by decreased errors of commission in three of the four conditions of the Go/No Go task. However, IQ is positively associated with the Adventurous score and inversely associated with errors on the Go/No Go task. When IQ is controlled, only one of the four significant findings remained, decreased errors of commission on the Pun-Rew condition. These results are similar to previous studies in children in which the number of significant correlations among self-report and behavioral measures of impulsivity decreases dramatically when IQ and age are controlled (Milich
Assessment of impulsivity
933
& Kramer, 1984; Paulsen & Johnson, 1980). Thus IQ may be considered a competing hypothesis that might account for an individual behaving impulsively on these tasks. The primary factor Impulsive is inversely associated with errors of omission during the Rew-Pun condition even after controlling for age and IQ. These results are in contrast to the hypothesis that impulsive Ss will demonstrate greater errors. Previous studies have demonstrated that impulsive individuals (e.g. delinquents, substance abusers, psychopaths), in comparison to healthy controls, report greater scores on self-report measures of impulsivity (Saklofske & Eysenck, 1983; Eysenck & McGurk, 1980; McGown, 1988) and demonstrate greater errors on measures of behavioral impulsivity (Rohrbeck & Twentyman, 1986; Newman & Kosson, 1986; Newman, 1987; Iaboni, Douglas & Baker, 1995). It is not clear why the current study demonstrates that the Impulsive factor is associated with less errors. It is possible that this may have been a type I error or that self-report measures and behavioral tasks are measuring different constructs of impulsivity in individuals with a life history of behaving impulsively and in normal control Ss. The primary factor, State Sensation Seeking is associated with behavioral impulsivity (even after controlling for IQ and age) as observed by increased errors of commission during the Reward condition of the Go/No Go task and increased Q scores on the Porteus mazes. Thus Ss scoring high on the State Sensation Seeking factor endorse feeling elated, adventurous, imaginative, zany, enthusiastic, etc. and demonstrate behavioral impulsivity. This may be similar to the Carefree factor found in an item analysis of numerous impulsivity and sensation seeking questionnaires (Gerbing, Ahadi & Patton, 1987). Individuals who score high on the Carefree factor report being Happy-go-lucky which suggests feeling enthusiastic, zany, etc. It is of interest that this State Sensation Seeking factor is not confounded by the demographic variables of age, IQ and educational level. In healthy control Ss, it appears that the State Sensation Seeking factor may demonstrate greater associations with behavioral impulsivity measures than self-report measures of impulsivity. The Enjoyment of Experiences factor did not demonstrate any association with behavioral impulsivity measures. The second order factor, Impulsive/Adventurous, appears to be a mixture of both impulsive and adventurous characteristics in that Ss demonstrate less errors on the Go/No Go task, but increased Q scores. It is possible that in these Ss the monetary reward in the Go/No Go task gave them sufficient motivation to overcome their tendency to behave impulsively which was indicated by their performance in the Porteus Mazes. Paper-and-pencil and behavioral measures of impulsivity may be subject to various methodological biases which can alter the accuracy of scores. For example, paper-and-pencil questionnaires are subject to self-awareness and demand characteristics that may lead to inaccurate scores. Behavioral tasks are not dependent upon self-awareness, and the use of monetary rewards as in the Go/No Go task should override any demand characteristics. Thus the Go/No Go task may be without some of the limitations found in other behavioral and paper-and-pencil measures of impulsivity. Research using the Go/No Go task has increased during the past decade. Gray et al. (1983) has proposed that impulsive Ss are insensitive to stimuli associated with punishment, may display an excessive sensitivity to stimuli associated with reward, and display a deficit in passive avoidance learning. Passive avoidance learning is the ability to withhold a response that would have lead to punishment. Thus, impulsive Ss should exhibit greater errors of commission than nonimpulsive Ss. Previous work by Newman and colleagues has demonstrated that the Go/No Go task discriminates between impulsive and non-impulsive Ss. Specifically, in comparison to controls, incarcerated psychopaths, extraverts, and juvenile delinquents demonstrated greater errors of commission, but similar errors of omission (Newman, 1987; Newman & Kosson, 1986; Newman, Widom & Nathan, 1985; Patterson, Kosson & Newman, 1987). These differences were found only in the reward-punishment condition (Newman, Patterson, Howland & Nichols, 1990) in which Ss demonstrate a dominant response for reward (S + ) that makes it difficult to inhibit responding to punished stimuli (S - ). Thus, the Go/No Go task is an operationalization of impulsivity as theorized by Gray et al. (1983), and research has supported this operationalization in various clinical and healthy control 5s. While the Go/No Go task is an operationalization of impulsivity as theorized by Gray et al. (1983), some of the other behavioral tasks fit more closely other definitions of impulsivity. Thus, high scores on the Porteus maze may reflect the inability to plan ahead (Barratt & Patton, 1983; Buss & Plomin, 1975; Eysenck & Eysenck, 1977). Other definitions include the tendency to respond quickly to stimuli (Barratt & Patton, 1983; Buss & Plomin, 1975) and acting without thinking (Barratt & Patton, 1983)
934
Karin F. Helmers et al.
which may be operationalized by the MFFT, and in some ways fits the DALS. A fundamental difference between these latter definitions and that of Gray et al. (1983) concerns the concept of punishment. The idea that impulsivity is the failure to withhold a response that will lead to punishment is an appealing one from the clinical point of view, as it is the failure to control behaviors that lead to punishment that characterizes the clinically impulsive patient. In the current study, there were few associations between the factors and the Go/No Go task. Of the four primary factors, State Sensation Seeking appears to be most closely associated with behavioral impulsivity, whereas Impulsive and Adventurous appears to be inversely associated with behavioral impulsivity. These results are in contrast to previous studies in which the Go/No Go task discriminated between impulsive and nonimpulsive Ss (Newman, 1987; Patterson et al., 1987; Iaboni et al., 1995). These differences may be due to the populations sampled, comparing healthy control Ss with clinical Ss. It is possible that the impulsive behavior found in clinical populations is independent from and not an extreme form of impulsive behavior found in healthy control Ss. A complicating factor is that age and IQ appear to be associated with both self-report and behavioral measures of impulsivity. Further studies are needed to evaluate the Go/No Go task simultaneously in normal volunteers and clinical populations in order to determine if behaving impulsively is a continuum or if these groups of Ss demonstrate independent behaviors of impulsivity. It is possible that the State Sensation Seeking factor may differentiate normal volunteers who behave impulsively from clinical populations. In sum, the measurement of impulsivity in the laboratory is still in a preliminary stage. The present data suggests that a state measurement of sensation seeking may be the most closely associated with behaving impulsively in healthy control Ss. Of the other factors, Adventurous appears to be associated with less errors and greater motor inhibition, though IQ scores confound this relationship. Of the behavioral tasks examined, the Go/No Go task is appealing because it appears to be the operationalization of impulsivity as seen in the clinically impulsive patient. Further studies are needed to assess whether impulsivity in the clinical population is an extreme form of impulsivity observed in control Ss, or if impulsivity is independent in these two groups. Acknowledgemenrs-KFH is a recipient of a research fellowship from the Medical Research Council of Canada and a computer equipment grant from Marion Merrell Dow, Inc. This research was supported by a grant from the Medical Research Council of Canada. We gratefully acknowledge the assistance of Anna Guyetsky and Franceen Lenoff for their help in the data collection.
REFERENCES American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders: DSM-IV (4th edn). Washington, DC: American Psychiatric Association. Barratt, E. S. &Patton, J. H. (1983). Impulsivity: Cognitive, behavioral and psychophysiological correlates. In M. Zuckerman. (Ed.), Biological buses of sensation seeking, impulsivity, and anxiety (pp. 77-l 16). Hillsdale, NY: Lawrence Erlbaum. Beck, A. & Steer, R. (1987) The Beck depression inventory. New York: Harcourt Brace Jovanovic. Brown, G. L., Ebert, M. H., Goyer, P. F., Jimerson, D. C., Klein, W. J., Bunney, W. E., et al. (1982). Aggression, suicide, and serotonin: Relationships to CSF amine metabolites. American Journal of Psychiatry, 139, 74 l-746. Buss, A. & Plomin, R. (1975). A temperament theory of personality development, New York: Wiley. 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, 5768.
Eysenck, S. B. G. & McGurk, B. J. (1980). Impulsiveness and venturesomeness in a detention center population. Psychological Reports, 47, 1299-1306.
Gerbing, D. W., Ahadi, S. A. & Patton, J. H. (1987). Toward a conceptualization of impulsivity: Components across the behavioral and self-report domains. Multivariate Behavioral Research, 22, 357-379. Glueck, S. & Glueck, E. (1968) Delinquents and nondelinquents in perspective. Cambridge, MA: Harvard University Press. Gray, J. A., Owen, S., Davis, N. & Tsaltas, E. (1983). Psychological and physiological relations between anxiety and impulsivity. In M. Zuckerman (Ed.), Biological bases of sensation seeking, impulsivity, and nnxiety, Hillsdale, NY: Lawrence Erlbaum. Hare, R. D. (1991) The Hare psychopathy checklist-revised, Toronto, ON: Multi-Health systems. Iaboni, F., Douglas, V. I. & Baker, A. G. (1995). Effects of reward and response costs on inhibition in ADHD children. Journal of Abnormal Psychology, 104, 232-240.
Kagan, J. (1966). Reflection-Impulsivity: The generality and dynamics of conceptual tempo. Journal ofAbnormal Psychology, 71, 17-24. Kaufman, A. S. (1990) Assessing adolescent and adult intelligence, Boston, MA: Allyn and Bacon. Kipnis, D. (1971) Character structure and impulsiveness, New York: Academic Press.
Assessment
of impulsivity
935
Lidberg, L., Tuck, J. R., Asberg, M., Scaha-Tomba, G. P. & Bertilsson, L. (1985). Homicide, suicide and CSF SHIAA. Acra Psychiatrica Scandinavica,
71, 230-236.
Linnoila, M., Virkkunen, M., Scheinin, M., Nuutila, A., Rimon, R. & Goodwin, F. K. (1983). Low cerebrospinal fluid 5-hydroxindoleacetic acid concentration differentiates impulsive from nonimpulsive violent behavior. Life Sciences, 33, 2609-26 14.
McGown, W. G. (1988). Multi-impulsive personality disorder and multiple substance abuse: Evidence from members of self-help groups. British Journal of Addiction, 83, 43 l-432. Milich, R. & Kramer, J. (1984). Reflections on impulsivity: An empirical investigation of impulsivity as a construct. In K. Gadow (Ed.), Advances in learning and behavioral disabilities (Vol. 3, pp. 57-93). Greenwich, CT: JAI Press. Newman, J. P. (1987). Reaction to punishment in extraverts and psychopaths: Implications for the impulsive behavior of disinhibited individuals. Journal of Research in Personality, 21, 464480. Newman, J. P. & Kosson, D. S. (1986). Passive avoidance learning in psychopathic and nonpsychopathic offenders. Journal of Abnormal Psychology, 95, 252-256.
Newman, 3. P., Patterson, C. M., Howland, E. W. & Nichols, S. L. (1990). Passive avoidance in psychopaths: The effects of reward. Personality and Individual Differences, II, 1101-I 114. Newman, J. P., Widom, C. S. & Nathan, S. (1985). Passive avoidance in syndromes of disinhibition: Psychopathy and extraversion. Journal of Personality and Social Psychology, 48, 13 16-l 327. Oas, P. (I 983). Impulsive behavior and assessment of impulsivity with hospitalized adolescents, Psychological Reports, 53, 764-766.
Patterson, C. M., Kosson, D. S. &Newman, J. P. (1987). Reaction to punishment, reflectivity, and passive avoidance learning in extraverts. Journal of Personality and Social Psychology, 52, 565-575. Paulsen, K. &Johnson, M. (1980). Impulsivity: A multidimensional concept with developmental aspects. JournalofAbnormal Child Psychology, 8, 269-277.
of suicidality, aggressivity and impulsivity. Clinical Neuropharmacology, 9, 38&382. Porteus, S. D. (1965). Porteus maze tesrt Fifty years’ application, Palo Alto, CA: Pacific Books. Prior, M. & Sanson, A. (1986). Attention deficit disorder with hyperactivity: A critique. Journal of Child Psychology and Psychiatry, 27, 307-3 19. Plutchik, R. & van Praag, H. M. (1986). The measurement
Rohrbeck, C. A. & Twentyman, C. T. (1986). Multimodal assessment of impulsiveness in abusing, neglecting, and nonmaltreating mothers and their preschool children. Journal of Consulting and Clinical Psychology, 54, 23 l-236. Roy, A. (1983).Family history of suicide. Archives of General Psychiatry, 40, 97 l-974. Roy, A., Adinoff, B. & Linnoila, M. (1988). Acting out hostility in normal volunteers: Negative correlation with levels of SHIAA in cerebrospinal fluid. Psychiatry Research, 24, 187-194. Roy, A., Virkkunen, M. & Linnoila, M. (1990). Serotonin in suicide, violence, and alcoholism. In E. F. Coccaro & D. L. Murphy (Eds), Serotonin in major psychiatric disorders, Washington, DC: American Psychiatric Press. Saklofske, D. H. & Eysenck, S. B. G. (1983). Impulsiveness and venturesomeness in Canadian children. Psychological Reports, 52, 147-152.
Soubrie, P. ( 1986). Reconciling the role of central serotonin neurons in human and animal behavior. The Behavioral and Bruin Sciences.
9, 3 19-364.
Stevens, J. (I 992) Applied multivariate statisticsfor the social sciences, 2nd edn. Hillsdale, NJ: Lawrence Erlbaum Associates. Valzelli, L. ( 1984). Reflections on experimental and human pathology of aggression. Progress in Neuro-Psychopharmacological and Biological Psychiatry, 8, 3 I l-325.
Zuckerman, M. (1979).Sensation seeking: Beyond the optimal level of arousal Hillsdale, NJ: Lawrence Erlbaum. Zuckerman, M. (1993). Sensation seeking and impulsivity: A marriage of traits made in biology? In W. G. McCown, J. L. Johnson & M. B. Shure. (Eds), The impulsive clienf: Theory, research and freatment, Washington. DC: American Psychological Association.