https://orcid.org/0000-0001-6042-599X
Figures
Abstract
The present research built on the Self-Reported Delinquency interview and the Antisocial Behavior Scale to develop an updated brief instrument to measure antisocial behavior. College students (n = 3188, 67.75% women) from the USA, Argentina, the Netherlands and Spain completed an online survey. Analyses that combined approaches from the Classical Test Theory and Item Response Theory were conducted to select the items for the brief version. Findings suggested that a 13-item Brief Antisocial Behavior Scale (B-ABS) fulfilled the high-quality criteria: salient factor loadings, adequate discrimination, variability in response endorsement, adequate fit based on infit/outfit values, nondifferent item functioning across the four participating countries, and Cronbach’s alpha and ordinal omega coefficients higher than .70. The B-ABS scores generally significantly correlated with personality scores, mental health and marijuana outcomes, showing criterion-related validity evidence. Our overall findings suggest that B-ABS adequately assesses antisocial behavior in young adults from different countries/cultures.
Citation: Mezquita L, Bravo AJ, Pilatti A, Ortet G, Ibáñez MI, Cross-Cultural Addictions Study Team (2021) Preliminary validity and reliability evidence of the Brief Antisocial Behavior Scale (B-ABS) in young adults from four countries. PLoS ONE 16(2): e0247528. https://doi.org/10.1371/journal.pone.0247528
Editor: Eduardo Fonseca-Pedrero, University of La Rioja, SPAIN
Received: September 11, 2020; Accepted: February 9, 2021; Published: February 22, 2021
Copyright: © 2021 Mezquita et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data can be found at: https://osf.io/9a76c/.
Funding: Dr. Bravo was supported by a training grant (T32-AA018108) from the NIAAA during the duration of data collection for this project. Data collection across all sites was supported, in part, by grant T32-AA018108. NIAAA had no role in the study design, collection, analysis or interpretation of the data, writing the manuscript, or the decision to submit the paper for publication. Data collection in Spain and the publication fees was also supported, in part, by grants RTI2018-099800-B-I00 from the Spanish Ministry of Science, Innovation and Universities (MCIU), AICO/2019/197 from the Autonomous Government of Valencia, and UJI-A2017-18 and UJI-A2019-08 from the Universitat Jaume I. Data collection in Argentina was also supported by grants from the National Secretary of Science and Technology, (FONCYT, grant number #PICT 2015-849) and by grants from the Secretary of Science and Technology- National University of Córdoba (SECyT-UNC).
Competing interests: The authors have declared that no competing interests exist.
Introduction
Antisocial behavior is broadly defined as actions that violate societal norms and others’ personal or property rights [1]. Antisocial behavior begins in early adolescence and is related to higher risks of both criminal justice involvement and premature death [2, 3]. Antisocial behavior usually peaks in late adolescence or the beginning of adulthood showing a declining pattern after this life period [4]. In addition, some youths will show a persistent pattern of antisocial behavior throughout their lifetime [5, 6]. In fact in the USA, recent prevalence rates suggest that one in 4 US adults exhibit syndromal antisocial behavior, and antisocial syndromes are associated with other psychiatric disorders [e.g., bipolar I, posttraumatic stress disorder, borderline personality disorder, and schizotypal personality disorders; 7]. In young adults, antisocial behavior has been related to psychological distress (including symptoms of somatization, depression, anxiety, phobic anxiety, interpersonal sensitivity, obsessive-compulsivity, hostility, paranoid ideation, and psychoticism) [8], depression [9], alcohol use [10, 11], and illicit drugs use [12], among others. A recent meta-analysis has also showed that antisocial personality disorder has a prevalence of 3.05%. However, the majority of the studies included preceded from USA [13], suggesting the need of performing similar epidemiologic studies in other countries around the world.
To know whether a psychological phenomenon occurs across cultures is important for understanding the interplay between individuals and their cultural context [14, 15]. In this line of research, previous studies have suggested that rates of antisocial behavior and related symptoms appear to be more common in individualistic and industrialized cultures than in collectivistic cultures, that is in cultures that prioritize the needs of an individual over the needs of a group as a whole [15, 16]. However, other studies did not identify this pattern clearly, and suggest that income inequality could be responsible of higher antisocial behavior and violence by country [17, 18]. In any case, due to the lack of cross-national studies, further research is needed to better understand the role of the cultural context in the antisocial behavior. The scare number of these types of studies is in part due to the lack of equivalent measures to assess such behavior across different languages and nations.
History of the antisocial behavior scale
The Self-Reported Delinquency Interview [SRD; 19] was developed to assess antisocial behavior, understood as the transgression of laws against people and property, and as the violation of norms that imply common administrative sanctions in western societies. The original SRD contained 48 questions, administrated to boys (aged between 11–12 years) who were followed up 2 years later. The principal component analysis of the interview suggested the existence of a general factor of delinquency that underlay the structure of the questionnaire.
Based on Shapland’s [19] interview to assess antisocial behavior in boys aged 11–14 years, Pérez-Sánchez and Torrubia [20] translated and adapted items into Spanish, and also adapted the interview to be administrated as a self-report measure. First, they omitted 13 non-delinquent (e.g., “I have been fishing” or “I have been skating”) and added three mischievous (e.g., “I have dialed 999 just for joke) items from Shapland’s original version of 48 items. Second, they reworded three items that could reflect antisocial behavior in youngsters, but not in older people (e.g., “I have smoked cigarettes” to “I smoked cigarettes before the age of 16”). Finally, they added five items: “I have sprayed a message on a wall”, “I have used illegal gadgets to make a telephone call”, “I have broken the highway code (by failing to stop when lights are red or failing to respect other drivers’ right of way, etc.), “I have had a firearm in my possession for reasons other than for sport or for strict military use” and “I have obtained money through drug-trafficking”. The resulting Spanish SRD [S-SRD; 20] consists of a 37-item scale in which participants report whether they endorse (or not) each item (i.e., a yes/no scale) and a frequency score (i.e., with two response options: from 1 to 6 times, or 6 times or more). The authors showed that the sum of the total number of endorsed items correlated with sensation seeking in their sample of male and female students [20].
In a more recent study examining different etiological pathways to alcohol use and misuse, the Spanish SRD was revamped [21]. As the measure aimed to assess young adults in addition to adolescents, the name of the scale was modified to the Antisocial Behavior Scale (ABS), rather than using the term delinquency. In addition, when considering that the original scale was developed to assess adolescents aged from 11 to 14 years in order to assess more antinormative behaviors, the temporal specifications of some items changed from 16 or 18 years old to 14 years old (e.g., “I smoked cigarettes before the age of 14”). Furthermore, six items that were not considered antinormative in the Spanish socio-cultural context (e.g., “I have let off fireworks in the street”), or were old-fashioned (e.g., I have stolen school property worth more than about 100 pesetas) were deleted. Finally, four other items, including some related to the DSM and ICD criteria to diagnose Conduct Disorder, were included (e.g., I have mistreated animals). The response scale was also modified, and a 4-point scale (from never or almost never to very frequently) was employed rather than dichotomous items. The internal consistency of the scale was adequate (α = .88) and criterion-related validity evidence was shown. Specifically, the personality traits of (low) agreeableness and (low) conscientiousness were related to the antisocial behavior scores five years later, even when gender, age and previous alcohol use were controlled for, and antisocial behavior cross-sectionally predicted alcohol-related variables [21].
Although the scale has been shown sources of validity and reliability evidence, a 35-item scale may be too long for research endeavors in which assessment is time-limited or when the evaluation requires administering several tests/measures. The development of very brief measures is especially useful in cross-cultural research, in which the cost in terms of organization, coordination, and monetary price is very high. Therefore, there is a need, to obtain efficient instruments insofar as how much information they provide and the amount of time is needed to answer them; that is, one that is balanced in validity and utility terms [22].
Present study
A variety of measures assess antisocial behavior [e.g., the Antisocial Processes Screening Device; 23, Psychopathy Checklist-Revised; 24, A-D. Cuestionario de Conductas Antisociales-Delictivas; 25], but very few have been psychometrically tested to account for cultural differences. What is “normative behavior” varies across distinct cultures, nationalities and regions [see 26], and most psychological constructs are highly dependent on the cultural context in which tests are used. Therefore, a central aspect to test development is to determine whether measures operate in the same way across languages or groups of administration. Furthermore, comparisons are not valid until multicultural and multilanguage research confirms measurement equivalence [27]. The present study aimed to develop a brief version of the ABS (B-ABS) to assess antisocial behavior in two individualistic (USA and the Netherlands) and two collectivistic countries (Argentina and Spain). To accomplish this goal, a set of analyses based on Classical Test Theory (discrimination index and difficulty indices) and Item Response Theory (i.e., Rasch analysis, including differential item functioning analysis) were conducted. We expected that the resulting B-ABS shows evidence of reliability of its scores across countries (i.e., Cronbach’s alphas and Ordinal omega of .70 or higher). Additionally, associations between the B-ABS and different criterion-related variables (e.g., personality traits, mental health symptoms, and marijuana-related variables) were estimated to provide sources of criterion-related validity evidence of the scale. Specifically, we hypothesized that low agreeableness, low conscientiousness [28–30], impulsivity traits [31] and psychopathological symptoms would be related to the antisocial behavior [7–12]. We also expect higher correlations between the antisocial behavior and the drug use variables than with internalizing-related symptoms, as they are considered part of the same spectrum [32, 33]. Due to the lack of similar cross-national research, the study of the differences in the magnitude of these associations across countries are exploratory.
Method
Participants and procedures
The original sample included 3349 college students from four universities of USA (n = 1928, 66.89% females, Mean age = 19.95, SD = 4.16), one from Argentina (n = 375, 66.67% females, Mean age = 24.16, SD = 5.38), one from the Netherlands (n = 302, 73.51% females, Mean age = 20.87, SD = 2.96) and one from Spain (n = 754, 66.45% females, Mean age = 21.44, SD = 3.99), who completed an online survey study about personal mental health, personality traits and marijuana use behaviors [for more information, see 34]. For the USA sites, students were recruited from Psychology Department research pools and received research credit for their participation. In Argentina students were recruited through online social networks and email listings and entered into a raffle of seven prizes [one stay in a cottage located in a popular destination and six cash prizes (each of ≈US 36 at the time)]. In Spain, an email was sent to all the students at the university inviting them to participate in the research. Participants who completed the survey received €5 for their participation. In the Netherlands, students were recruited from the School of Social and Behavioral Sciences and received research credit. The study was approved by the institutional review board (or their international equivalent) at each participating university (USA single site IRB for all U.S. sites from the University of New Mexico [Approval Code: 112900–2]; the Tilburg University approval reference: EC-2018.03; Argentina and Spain resolution data: 12/13/2017 and 3/29/2018, respectively). All the participants completed the informed consent in the first page of the online survey.
For the present study, only the data of the students that completed the ABS were included. After excluding the individuals with more than 5% of missing values in the ABS (n = 63), and who did not disclose their gender (n = 13), the total analytic sample was composed of 3188 participants (USA: n = 1873, 67.49% females, mean age = 19.87, SD = 4.20); Argentina: n = 346, 65.61% females, mean age = 24.31, SD = 5.49); the Netherlands: n = 294, 74.49% females, mean age = 20.79, SD = 2.76); Spain, n = 675, 66.67% females, mean age = 21.41, SD = 3.95).
Measurement translation and adaptation
As the international guidelines for test adaptations suggested [35], an iterative procedure to develop the different versions of the questionnaires across languages was followed. In a previous project [see 36 for more information], two psychologists proficient in English and Spanish and with expertise in test adaptation and addictive behaviors translated the most updated version of the scale, the ABS [21], into English. Then a bilingual teacher unfamiliar with the inventory did a back translation. The analysis of the back translation indicated that the English version could be considered comparable to the original scale. Some minor changes were made in the Castilian Spanish version of the ABS to ensure that all the items were adequate for the Argentinian participants. The same procedure was followed in the project of which this study forms part to adapt the marijuana outcomes from English to Spanish [see 34 for more details]. To adapt the measure to Dutch, research team members, who are bicultural and proficient in English, Spanish and Dutch, with expertise in test adaptations, translated the original English/Spanish versions of the ABS, the personality questionnaires, the mental health measure and the marijuana outcomes questionnaires into Dutch (see below for a detailed description of the scales). Then the research team members compared the versions and, after a thorough discussion, composed a preliminary version of the instruments. Items from the English, Dutch and Spanish ABS versions are presented in S1 Table.
Measures
Antisocial behavior.
The ABS [21] includes 35 items that describe different antisocial behaviors (e.g. “I have used knives or sticks in fights” or “I have broken, ripped or damaged public properties”) on a 4-point response scale (1 = never or almost never, 2 = sometimes, 3 = frequently, 4 = very frequently or very often). The total score is obtained by summing the responses to all the items. A previous preliminary study showed that the scores for the English and Spanish ABS versions had good internal consistency. The different item functioning analyses showed that the items generally operate similarly across the USA, Argentina and Spain [36].
Personality traits.
Personality traits were assessed by the 50-item Big Five Personality Trait Short Questionnaire [BFPTSQ; 37] at the US sites, the Spanish version [38] at the sites in Spain and Argentina, and our own adaptation to Dutch in the Netherlands. The measure assessed the Five-Factor Model (FFM) broad personality domains on a 5-point Likert response scale (0 = strongly disagree, 4 = strongly agree): openness, extraversion, agreeableness, conscientiousness, and emotional stability. Items were averaged, thus scores ranged from 0 to 4 in the present study. Recent research has tested the measurement invariance across languages and cultures of the scale, and found that the BFPTSQ showed configural, metric and scalar invariance among Argentinian and Spanish youths and between the English and the Spanish versions of the questionnaire [39].
Short UPPS-P.
A 20-item English version of the UPPS-P Impulsive Behavior Scale [S UPPS-P; 40] was used to measure impulsivity facets of negative urgency, lack of perseverance, lack of premeditation, sensation seeking and positive urgency. The same 20 items were selected from the Spanish [41] and the Dutch UPPS-P [items were adapted in the current project; see 42] questionnaire to assess impulsivity among Argentinean, Spanish and Dutch youths. The scale response scale went from 1 to 4.
Mental health symptoms.
Past 2-week psychopathology was assessed by the 23-item DSM-5 Self-Rated Level 1 Cross-Cutting Symptoms Measure-Adult [43]. The Spanish version was administered to the Spanish-speaking students [44]. The measure was translated into Dutch for the Dutch-speaking students [see 45 for more datails]. Participants were asked, “In the past 2 weeks, how much (or how often) have you been bothered by the following problems?” and responded on a 5-point response scale (0 = none, not at all, 1 = slight or rare, less than a day or two; 2 = mild, several days; 3 = moderate, more than half the days, 4 = severe, nearly every day). A score of 2 or more in most domains, except substance use (score of 1 or more), suggests clinically-relevant mental health problems [46]. The measure has been validated with clinical [46] and college student [47] samples.
The prevalence rates of potential symptom presentation for the 13 domains (percentages were averaged for the domains with multiple items) of the DSM-5 Cross-Cutting Symptoms Measure in our analytic sample were as follows: depression (33.12%), anger (38.34%), mania (24.25%), anxiety (24.44%), somatic distress (18.31%), suicidal ideation (10.45%), psychosis (4.56%), sleep disturbance (29.75%), memory problems (16.58%), repetitive thoughts and behaviors (12.05%), dissociation (16.01%), and personality functioning (22.68%). For substance use, we present the rates for each specific substance: alcohol use (31.19%), tobacco (21.93%), and illicit drug use (18.35%).
Marijuana use.
The participants who reported using marijuana at least once in the previous month (i.e., last-month marijuana users) completed a visual guide showing different amounts of marijuana in grams. The same guide was used across countries. The Marijuana Use Grid [MUG; 48] was used to assess typical marijuana use. Participants reported the amount of grams they used in each 4-hour period on each typical week day (12p-4p on Monday, 4p-8p on Monday, etc.). All non-zero values, which reflect the number of time periods used in a typical week (possible range: 0–42), were added up to obtain an estimate of typical marijuana use frequency. By summing all the values, we obtained an estimate of the typical marijuana use quantity by reflecting the total number of grams used in a typical week. To address outliers, we winsorized quantity estimates > 3SDs above the mean.
Negative marijuana-related consequences.
Negative marijuana-related consequences were assessed using the B-MACQ [49]. Each item was scored dichotomously to reflect the presence/absence of the marijuana-related problem in the past month (0 = no, 1 = yes). The total score reflects the total number of consequences that an individual has experienced in the last 30 days. Recent research has tested the measurement invariance across languages and cultures of the scale, and found that a 20-item version showed configural and scalar invariance among the USA, Argentina, Uruguay and Spain. However, the questionnaire gave a poor fit in a sample of young adults from the Netherlands [34]. After considering these results, the 20-item version was herein used.
Statistical analysis
Missing data imputation.
From the 3188 participants, the 125 participants who present some missing data in the ABS came from USA (115 missing values from 65.555 possible values [35 items x 1873 participants], 0.18%) and from the Netherlands (10 missing values from 10.290 possible values [35 items x 294 participants], 0.10%). In Argentina and Spain, electronic prompts for each missing response were given and, therefore, there were no missing values. After checking that these missing values were at random (the maximum percentage of missing values in one item was 0.6%, which is lower than the recommended cut-off of 2%), we estimated the EFA using the full information maximum likelihood (FIML), while pairwise deletion was used in the classical test theory and item response theory analysis. Considering the low number of missing values, the mean score of the B-ABS and the rest of the scales included in the study were computed to provide evidences of criterion validity.
Exploratory factor analysis.
An exploratory factor analysis (EFA) was performed to: 1) explore the structure underlying the set of items; 2) select those items with the highest factor loadings for the brief version. By considering the response scale of the questionnaire, the items were analyzed as categorical variables and the WLSMV estimator was used. To evaluate the overall model fit, the model fit criteria suggested by Marsh, Hau and Wen [50] were used, including the Comparative Fit Index (CFI) >.90 (acceptable) > .95 (optimal), Tucker-Lewis Index (TLI) >.90 (acceptable) > .95 (optimal), Root Mean Square Error of Approximation (RMSEA) < .06, and Standardized Root Mean Square Residual (SRMR) < .08. These analyses were conducted with Mplus 7.4. [51].
Classical Test Theory and Item Response Theory analysis.
Following Meyer’s recommendations [52], we combined analyses and procedures based on the Classical Test Theory (discrimination index and difficulty indices) and the Item Response Theory (Rasch analyses, including differential item functioning) to identify the best items for the brief scale. We specifically examined the indices of discrimination for each item, which indicates the correlation between each item and the total test score (i.e., correlation item-total). By considering the number of points on the Likert scales, the discrimination index should lie between .43 and .83. Secondly, to reflect the item endorsability, we performed the item difficulty index (i.e., mean of the whole sample on each item). Third, we applied the Rasch Model [53], which is one particular IRT model, to evaluate person-item outfit and infit, and to determine whether items function in similar ways across countries (i.e., differential item functioning [DIF]). For the infit/outfit analyses, we used the unweighted mean square (UMS) and the weighted mean square (WMS) fit statistics, respectively. In both cases, values between .80 and 1.20 are recommended, while values between .50 and 1.5 are considered still productive for measurements [52]. Then we conducted a DIF analysis across each pair of countries (e.g., Spain and USA to explore the possible differences in the antisocial behavior estimates independently obtained for the college students from each country. We calculated the magnitude of the differences for each item between groups using the standardized P-DIF (sP-DIF), which is the result of dividing the standardized mean difference (SMD) by the item score range. An sP-DIF value below .05 indicates that items function similarly across college students from each country pair [52]. In other words, the absence of DIF suggests that the probability of each person endorsing every item is not affected by the country/culture of origin. To choose the final pool of items, we also followed some theoretical considerations (i.e., avoid item content and other associated constructs overlapping, such as drug use, or qualitative students’ and evaluators’ comments about understanding of item contents) apart from statistical considerations. These analyses were performed with the jMetrik software [52].
Internal consistency and sources of criterion-related validity evidence.
To examine the internal consistency of the brief version, we calculated Cronbach’s alpha (conducted with the SPSS 23 software) and the omega coefficient (conducted with Mplus 7.4). For both coefficients, values above .70 indicate adequate internal consistency. We conducted Pearson’s correlation analyses to evaluate the association between the total B-ABS score and personality traits, mental health outcomes and marijuana-related measures. These analyses were conducted in each country.
Results
Exploratory factor analysis
The EFA showed an optimal data fit for the one-factor model (χ2 (595) = 77574.50, CFI = .971, TLI = .969; RMSEA = .035, SRMR = .050). The inclusion of an additional factor did not significantly improve the model fit (ΔCFI = .007, ΔTLI = .006, ΔRMSEA = -.003), which supports the single factor structure. The factor loadings of the final factor solution were all salient (>.30) and significant at p < .05. These results are presented in Table 1.
Classical Test Theory and Item Response Theory analyses
Table 1 presents the items’ performance based on the discrimination index, level of item endorsement, infit/outfit indices and sP-DIF. The results about the category characteristic curves and the item information curves of the ABS are presented in S1 Fig. Based on the previously indicated inclusion criteria, we selected items with an adequate discrimination index (between .43 and .63) with adequate infit/outfit values (UMS and WMS values between .50 and 1.50), and that showed sP-DIF < .05 across countries. Items 1, 3, 5, 6, 7, 8, 10, 14, 17, 19, 21, 25, 29, 31 and 32 fulfilled these criteria. However, item 3 (“I smoked cigarettes before the age of 14“) was excluded to avoid content and measures of drug use overlapping. Item 32 (“I have used illegal gadgets to make a telephone call”) was excluded because its meaning may not be clear for participants (i.e., it is ambiguous if the item referred to robbing a mobile, breaking the phone booths, etc.). Therefore, the final list includes items 1, 5, 6, 7, 8, 10, 14, 17, 19, 21, 25, 29, and 31.
Internal consistency
Cronbach’s alphas for the whole sample (.890 [.884 .895]) and per country (USA: .910 [.904 .916]; Arg: .748 [.707 .785], Neth: .789 [.772 .806], Sp: .862 [.847 .877]) were significant at the 95% confidence interval (95%CI) and were higher than .70. Similar results were found with the ordinal omega at the 95%CI in the whole sample (.891 [.877 .904]) and per country (USA: .912 [.897 .923]; Arg: .751 [.672 .802], Neth: .817 [.681 .878], Sp: .863 [.827 .891]), which suggest adequate internal values consistency.
Sources of criterion-related validity evidence
The descriptive results (i.e., mean and standard deviation) of the final 13-item version (B-ABS) and all the criterion variables for each country are presented in Table 2. We found higher antisocial behavior in Argentina, followed by Spain, USA and the Netherlands. However, these differences were small in magnitude. A large difference appeared between the Netherlands and the other countries (the Netherlands > USA Argentina, Spain) on the mean scores of the DSM-V Level 1 Measure psychosis domain. Additional mean differences were found in agreeableness (Argentina > USA), somatic distress (USA, Argentina, Spain > the Netherlands), sleep disturbances (Spain > the Netherlands) and marijuana frequency used in a typical week in the last 30 days (USA, Argentina > the Netherlands).
Table 3 presents Pearson’s correlations between antisocial behavior and the criterion-related variables. Across countries, antisocial behavior was statistically significantly associated with low agreeableness, low conscientiousness, negative urgency, lack of premeditation, positive urgency, mania, somatic distress, suicidal ideation, psychosis, memory problems, repetitive thoughts, dissociation, personality functioning, tobacco use, illicit drug use, poor mental health, and negative marijuana-related consequences. In some, but not all the countries, antisocial behavior also showed significant associations with lack of perseverance (USA and Spain), sensation seeking (Argentina and Spain), depression (USA and the Netherlands), anger (USA and Argentina), anxiety (USA, Spain and the Netherlands), sleep disturbance (USA, Argentina and the Netherlands), alcohol use (USA, Spain and Argentina), frequency of marijuana use (USA>, Spain and Argentina) and quantity of marijuana use (USA, and Argentina).
The magnitude of the differences in the correlation coefficients across countries is also presented in Table 3. Across 186 possible comparisons, we found that the average difference in correlations was .072 (SD = .056). We considered differences < 1 SD to be small, between 1 SD and 2 SD were considered medium, between 2 SD and 3 SD were taken as being large, and differences above 3 SD were considered substantial. We found substantial differences between antisocial behavior and: a) sensation seeking between USA and Argentina, with a nonsignificant correlation in the USA (r = .026, p >.05) and a positive medium-size correlation in Argentina (r = .294, p < .01); b) alcohol use between the USA and the Netherlands, with a positive medium-size correlation in the USA (r = .299, p < .001) and a nonsignificant correlation in the Netherlands (r = .034, p >.05); c) marijuana quantity used in a typical week in the last 30 days between Argentina and Netherlands, with a positive medium-size correlation in Argentina (r = .247, p < .01) and a negative and nonsignificant correlation in the Netherlands (r = -.038, p >.05).
Discussion
The present research aimed to adapt the current version of the Spanish ABS scale to English and Dutch, and create a Brief version of the measure which could be particularly useful and necessary when administration time is limited. Following the international guidelines for test adaptations [35], after detecting the lack of measures to assess antisocial behavior in youths cross-nationally, an iterative adaptation of the ABS to the different languages was performed. Subsequently, confirmation guidelines were followed and evidence of the structure of the measure was provided, as the equivalence of the measure across countries, reliability of the scores and sources of criterion-validity evidences of the brief measure [35].
Regarding the structure of the ABS, we found evidence for the unidimensional structure underlying the set of items. To identify a brief set of items, without diminishing the validity and reliability properties, we followed a multi-prong approach. As previously recommended [52], we combined analyses from the Classical Test Theory and the Item Response Theory. The resulting brief scale (B-ABS) comprised 13 high-quality items that exhibited salient factor loadings, adequate discrimination, adequate values for infit/outfit indices, variability in its endorsability, and similar item functioning across different countries/cultures. The B-ABS showed adequate Cronbach’s alpha and ordinal omega indices in the four countries based on the standard cut-offs, which provide evidence of the reliability of the scores of the short version of the ABS.
Regarding sources of content validity evidence of the resulting B-ABS, it seems that the B-ABS is a more face valid measure of antisocial behavior understood as “actions that violate societal norms and others’ personal or property rights”. While the original SRD contains items like “I saw pornographic films before the age of fourteen” (item 2), “I smoked cigarettes before the age of fourteen” (item 3), “I drunk alcohol before the age of fourteen” (item 12), or “I have taken some type of drugs” (item 20), with a content that is difficult to justify as antisocial behavior, the B-ABS contains more pure antisocial behaviors and reflect the interpersonal component of the antisocial behavior construct. In addition, the B-ABS does not contain explicit personality content. Scales like the PCL-R [24], Antisocial Processes Screening Device [23], and other popular measures [e.g., Buss-Perry Aggression Questionnaire, 54] that assess antisocial behavior contain a good amount of trait content. Thus, if one is more interested in assessing antisocial behavior per se, the B-ABS has the advantage of not inflating effect sizes in particular research contexts due to criterion-predictor overlap.
When the antisocial behavior scores obtained in the B-ABS were compared across countries, we found higher antisocial behavior in Argentina, followed by Spain, USA and the Netherlands. These results contrast with the hypothesis than individualistic countries (e.g., USA and the Netherlands) have higher antisocial behavior than the collectivistic ones (e.g., Argentina and Spain) [15, 16]. In addition, it is partially supported that income inequality hypothesis could enhance antisocial behavior [17, 18], as the sample of youths from Netherlands, the country with lower income inequality [based on the Gini index of the World Bank estimate, 55], showed the lowest antisocial behavior rate while Argentina, the country with higher income inequality, showed the highest antisocial behavior score. However, it is important to note that the mean differences in the mean levels of antisocial behavior between the Netherlands and the rest of the countries were low, and that the sample size of the Netherlands was the smallest one in the study. Thus, additional studies should further explore if the differences in the income inequality are related or not with the higher/lower rates of antisocial behavior across countries.
Consistent with previous meta-analyses, the highest correlations between the B-ABS and personality traits were found with (low) agreeableness and (low) conscientiousness [28–30]. When we focused on specific impulsivity facets, we found that negative and positive urgency and premeditation were significantly related to antisocial behavior in all countries, while perseveration correlated with antisocial behavior in the USA and Spain, and sensation seeking in Argentina and Spain. Previous studies have also found that these five impulsivity facets are related to different aspects of antisocial behavior, such as aggression, rule-breaking behavior, theft and vandalism [56, 57], and that a genetic overlap occurs between these impulsivity facets and antisocial behavior [31]. The only substantial difference between the correlations of S UPPS-P and B-ABS was found with the sensation seeking subscale, which showed a nonsignificant correlation in the USA, and significant and positive medium-sized correlation in Argentina. Although the nonsignificant association in the USA sample was somewhat unexpected [31], there are reasons that might explain this result. For instance, sensation seeking displays an age-related pattern and peaks in teenage years and the early twenties [58, 59]. Notably, the difference in the mean age between the Argentinean (M = 24.31) and USA students (M = 19.97) was the biggest across all four countries.
Regarding the association of antisocial behavior with different mental health outcomes, we found that antisocial behavior correlated generally with externalizing (i.e., alcohol, tobacco and illicit drug use) and internalizing symptoms (i.e., depression, anxiety and somatic distress), which corroborates previous studies [60]. These associations were slightly higher with the externalizing than with the internalizing symptoms, suggesting that antisocial behavior and substance use are well conceptualized as part of the same spectrum [61]. One unexpected result was that alcohol use was not related to antisocial behavior in the Netherlands and, consequently, substantial differences in the magnitude of the correlations between antisocial behavior and alcohol use were found between USA (with positive medium-sized correlations) and the Netherlands. Future research is needed to corroborate these findings.
Previous studies have found that antisocial behavior is related to sleep problems [62], suicidal behavior [63], or depression symptoms [9]. The current research adds evidence to these findings by relating antisocial behavior with these domains and a wide variety of health outcomes in four different countries.
Antisocial behavior was also related to some marijuana outcomes in the four countries, and this result is consistent with studies reporting that individuals with conduct problems are at increased risk for substance (ab)use, including marijuana [64]. However, a substantial difference in the magnitude of the correlation of antisocial behavior with quantity of marijuana use appeared between Argentina and the Netherlands. Large differences in the magnitude of other correlations between antisocial behavior and negative marijuana consequences emerged between the Netherlands and the USA and Spain. However, these results should be taken with caution because the sample size of the Dutch marijuana users was small (n = 63).
Limitations
The present findings should be interpreted by considering a number of limitations. First, participants were recruited by different strategies (e.g., email listings, posting an invitation on online social networks, pool of students) across countries and different incentives for participation we also employed (e.g., raffles with prizes, research participation credit and money). These issues potentially limit generalizability to other youth samples. Second, we reported preliminary evidence of the structure and the invariance across countries of the resulting B-ABS. Thus, evidence of the adequacy of the structure of the B-ABS should be replicated in an independent cross-national sample performing Confirmatory Factor Analysis [65]. Third, our sample comprise a higher percentage of women than men. Considering than men usually report higher antisocial behavior than women [4], a higher representation of men in our sample would be advisable. Fourth, to avoid the burden produced by a long assessment protocol, only sex and age were assessed across countries. However, the inclusion of other demographic data could have provided a better description of the cross-national sample. Fifth, the sample size differed across sites, and was low in the case of the drug consumers of the Netherlands. Thus, we may have been underpowered to detect smaller differences in mean differences/correlations across cultures. Sixth, previous cross-national research has also demonstrated different levels of social desirability related to admitting antisocial behavior across countries [66]. Thus, objective information to assess antisocial behavior (e.g., criminal history) or infrequency and desirability scales could help to better reflect the similarities/differences in antisocial behavior among the samples of youths included in the present research. In any case, it is also remarkable that, although the higher/lower social desirability across countries could influence the prevalence of the reported behaviors, the self-reports have also demonstrated to be appropriate and valid to test theoretical correlates cross-nationally [66]. Seventh, although this brief measure provides an adequate assessment of antisocial behavior in college students from four countries, this version may not be appropriate for other college students or non-college students. Validation for Spanish-, English- and Dutch-speakers from other parts of the world will be an important next step. Finally, although we selected the best set of items following multiple quality criteria, we did not examine the association of this brief version with other (brief) measures of antisocial behavior.
Conclusion
The present study offers preliminary sources of reliability and validity evidence of the B-ABS to measure antisocial behavior. The results from both the Classical Test Theory and Item Response Theory offer encouraging evidence for the psychometric properties of a 13-item version of the measure and suggest that it can be used to assess antisocial behaviors in college students from different countries. Furthermore, cross-national comparisons of correlates of antisocial behavior offer avenues for future research to target the key risk factors associated with antisocial behavior. Overall, our findings suggest that the 13-item B-ABS scores represented a valid and reliable alternative to efficiently measure antisocial behaviors in young adults from different countries/cultures.
Supporting information
S1 Table. ABS instructions and items in English, Dutch and Spanish.
https://doi.org/10.1371/journal.pone.0247528.s001
(DOCX)
S1 Fig. The item response category characteristic curves and item information curves of the ABS.
The final 13 B-ABS items are marked in black squares.
https://doi.org/10.1371/journal.pone.0247528.s002
(DOCX)
Acknowledgments
This project was completed by the Cross-cultural Addictions Study Team (CAST), which includes the following investigators (in alphabetical order): Adrian J. Bravo, College of William & Mary, USA (Coordinating PI; ajbravo@wm.edu); James M. Henson, Old Dominion University, USA; Manuel I. Ibáñez, Universitat Jaume I de Castelló, Spain; Laura Mezquita, Universitat Jaume I de Castelló, Spain; Generós Ortet, Universitat Jaume I de Castelló, Spain; Matthew R. Pearson, University of New Mexico, USA; Angelina Pilatti, National University of Cordoba, Argentina; Mark A. Prince, Colorado State University, USA; Jennifer P. Read, University of Buffalo, USA; Hendrik G. Roozen, University of New Mexico, USA.
References
- 1.
World Health Organization. International classification of diseases for mortality and morbidity statistics (11th Revision). 2019.
- 2. Border R, Corley RP, Brown SA, Hewitt JK, Hopfer CJ, McWilliams SK, et al. Independent predictors of mortality in adolescents ascertained for conduct disorder and substance use problems, their siblings and community controls. Addiction. 2018;113: 2107–2115. pmid:30091161
- 3. Aalsma MC. Commentary on Border et al. (2018): The public health burden of conduct disorder, early mortality and criminal justice involvement. Addiction. 2018;113: 2116–2117. pmid:30215875
- 4. Duell N, Steinberg L, Icenogle G, Chein J, Chaudhary N, Di Giunta L, et al. Age patterns in risk taking across the world. J Youth Adolesc. 2018;47: 1052–1072. pmid:29047004
- 5. Moffitt TE. Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychol Rev. 1993;100: 674–701. pmid:8255953
- 6. Silberg JL, Rutter M, Tracey K, Maes HH, Eaves L. Etiological heterogeneity in the development of antisocial behavior: The Virginia Twin Study of Adolescent Behavioral Development and the Young Adult Follow-Up. Psychol Med. 2007;37: 1193–1202. pmid:17376258
- 7. Goldstein RB, Chou SP, Saha TD, Smith SM, Jung J, Zhang H, et al. The epidemiology of antisocial behavioral syndromes in adulthood: Results from the National Epidemiologic Survey on Alcohol and Related Conditions-III. J Clin Psychiatry. 2017;78: 90–98. pmid:27035627
- 8. Basto-Pereira M, Maia Â. Persistence in crime in young adults with a history of juvenile delinquency: The role of mental health and psychosocial problems. Int J Ment Health Addict. 2018;16: 496–506. http://dx.doi.org/10.1007/s11469-017-9847-7
- 9. Wiesner M, Kim HK, Capaldi DM. Developmental trajectories of offending: Validation and prediction to young adult alcohol use, drug use, and depressive symptoms. Dev Psychopathol. 2005;17: 251–270. pmid:15971769
- 10. Harford TC, Muthén BO. Adolescent and young adult antisocial behavior and adult alcohol use disorders: A fourteen-year prospective follow-up in a national survey. J Stud Alcohol. 2000;61: 524–528. pmid:10928722
- 11. Mason WA, Hitchings JE, McMahon RJ, Spoth RL. A test of three alternative hypotheses regarding the effects of early delinquency on adolescent psychosocial functioning and substance involvement. J Abnorm Child Psychol. 2007;35: 831–843. pmid:17534712
- 12. Wiesner M, Windle M. Young adult substance use and depression as a consequence of delinquency trajectories during middle adolescence. J Res Adolesc. 2006;16: 239–264. http://dx.doi.org/10.1111/j.1532-7795.2006.00131.x
- 13. Volkert J, Gablonski TC, Rabung S. Prevalence of personality disorders in the general adult population in Western countries: Systematic review and meta-analysis. Br J Psychiatry. 2018;213: 709–715. pmid:30261937
- 14. Basto-Pereira M, Queiroz-Garcia I, Maciel L, Leal I, Gouveia-Pereira M. An international study of pro/antisocial behavior in young adults. Cross-Cultural Res. 2020;54: 92–105. http://dx.doi.org/10.1177/1069397119850741
- 15. Thalmayer AG, Rossier J. Agreeableness, antagonism, and mental health across cultures. The handbook of antagonism: Conceptualizations, assessment, consequences, and treatment of the low end of agreeableness. 2019. pp. 97–111. http://dx.doi.org/10.1016/B978-0-12-814627-9.00007-4
- 16. Gawda B. Cross-cultural studies on the prevalence of personality disorders. Curr Issues Personal Psychol. 2018;6: 318–329. http://dx.doi.org/10.5114/cipp.2018.80200
- 17.
WHO. World report on violence and health: summary. Geneva: World Health Organization; 2002.
- 18.
OECD. Society at a Glance 2011—OECD Social Indicators. 2011. Available: http://www.oecd.org/social/societyataglance2011.htm
- 19. Shapland JM. Self-reported delinquency in boys aged 11 to 14. Br J Criminol. 1978;18: 255–266.
- 20. Pérez J, Torrubia R. Sensation seeking and antisocial behaviour in a student sample. Pers Individ Dif. 1985;6: 401–403. http://dx.doi.org/10.1016/0191-8869(85)90068-6
- 21. Mezquita L, Ibáñez MI, Moya J, Villa H, Ortet G. A longitudinal examination of different etiological pathways to alcohol use and misuse. Alcohol Clin Exp Res. 2014;38. pmid:24797208
- 22. Mullins-Sweatt SN, Widiger TA. Clinical utility and DSM-V. Psychol Assess. 2009;21: 302–312. pmid:19719343
- 23.
Frick PJ, Hare RD. Antisocial process screening device: APSD. Multi-Health Systems Toronto. Toronto: Multi-Health Systems; 2001. http://dx.doi.org/10.1093/ije/dyt021
- 24.
Hare RD. The Hare Psychopathy Checklist-Revised (PCL-R). Toronto: Multi-Health Systems; 2003.
- 25.
Seisdedos N. A-D. Cuestionario de Conductas Antisociales-Delictivas. Madrid: TEA Ediciones; 1995.
- 26. Graham J, Meindl P, Beall E, Johnson KM, Zhang L. Cultural differences in moral judgment and behavior, across and within societies. Curr Opin Psychol. 2016;8: 125–130. pmid:29506787
- 27.
International Test Commission. International guidelines for practitioner use of test revisions, obsolete tests, and test disposal. 2015.
- 28. Vize CE, Collison KL, Miller JD, Lynam DR. Using Bayesian methods to update and expand the meta-analytic evidence of the five-factor model’s relation to antisocial behavior. Clin Psychol Rev. 2019;67: 61–77. pmid:30292437
- 29. Vize CE, Miller JD, Lynam DR. FFM facets and their relations with different forms of antisocial behavior: An expanded meta-analysis. J Crim Justice. 2018;57: 67–75. http://dx.doi.org/10.1016/j.jcrimjus.2018.04.004
- 30. Jones SE, Miller JD, Lynam DR. Personality, antisocial behavior, and aggression: A meta-analytic review. J Crim Justice. 2011;39: 329–337. http://dx.doi.org/10.1016/j.jcrimjus.2011.03.004
- 31. Mann FD, Engelhardt L, Briley DA, Grotzinger AD, Patterson MW, Tackett JL, et al. Sensation seeking and impulsive traits as personality endophenotypes for antisocial behavior: Evidence from two independent samples. Pers Individ Dif. 2017;105: 30–39. pmid:28824215
- 32. Caspi A, Houts RM, Belsky DW, Goldman-Mellor SJ, Harrington H, Israel S, et al. The p factor: One general psychopathology factor in the structure of psychiatric disorders? Clin Psychol Sci. 2014;2: 119–137. pmid:25360393
- 33. Mezquita L, Ibáñez MI, Villa H, Fañanás L, Moya-Higueras J, Ortet G. Five-factor model and internalizing and externalizing syndromes: A 5-year prospective study. Pers Individ Dif. 2015;79: 98–103. http://dx.doi.org/10.1016/j.paid.2015.02.002
- 34. Bravo AJ, Pearson MR, Pilatti A, Mezquita L. Negative marijuana‐related consequences among college students in five countries: measurement invariance of the Brief Marijuana Consequences Questionnaire. Addiction. 2019;114: 1854–1865. pmid:31066108
- 35. Muñiz J, Elosua P, Hambleton RK. Directrices para la traducción y adaptación de los tests: segunda edición. Psicothema. 2013;25: 151–157. pmid:23628527
- 36. Mezquita L, Bravo AJ, Ortet G, Pilatti A, Pearson MR, Ibáñez MI. Cross-cultural examination of different personality pathways to alcohol use and misuse in emerging adulthood. Drug Alcohol Depend. 2018;192: 193–200. pmid:30268069
- 37. Morizot J. Construct validity of adolescents’ self-reported big five personality traits: importance of conceptual breadth and initial validation of a short measure. Assessment. 2014;21: 580–606. pmid:24619971
- 38. Ortet G, Martínez T, Mezquita L, Morizot J, Ibáñez MI. Big Five Personality Trait Short Questionnaire: Preliminary validation with Spanish adults. Span J Psychol. 2017;20: E7. pmid:28181474
- 39. Mezquita L, Bravo AJ, Morizot J, Pilatti A, Pearson MR, Ibáñez MI, et al. Cross-cultural examination of the Big Five Personality Trait Short Questionnaire: Measurement invariance testing and associations with mental health. Sudzina F, editor. PLoS One. 2019;14: e0226223. pmid:31846470
- 40.
Lynam DR. Development of a short form of the UPPS-P Impulsive Behavior Scale. Unpublished Technical Report; 2013.
- 41. Verdejo-García A, Lozano Ó, Moya M, Alcázar MÁ, Pérez-García M. Psychometric properties of a Spanish version of the UPPS–P impulsive behavior scale: Reliability, validity and association with trait and cognitive impulsivity. J Pers Assess. 2010;92: 70–77. pmid:20013458
- 42. Pilatti A, Bravo AJ, Prince MA, Pearson MR, Mezquita L, Pautassi RM, et al. Marijuana-related perceptions as mediators of the association between trait impulsivity and marijuana outcomes. Manuscr Submitt Publ. 2020.
- 43.
American Psychological Association (APA). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author; 2013.
- 44.
American Psychiatric Association. Manual diagnóstico y estadístico de los trastornos mentales. 5th ed. Madrid: Editorial Médica Panamericana; 2014.
- 45. Roozen HG, Bravo AJ, Pilatti A, Mezquita L, Vingerhoets A. Cross-Cultural Examination of the Community Reinforcement Approach Happiness Scale (CRA-HS): Testing measurement invariance in five countries. Curr Psychol. 2020. http://dx.doi.org/10.1007/s12144-020-00818-w
- 46. Narrow WE, Clarke DE, Kuramoto SJ, Kraemer HC, Kupfer DJ, Greiner L, et al. DSM-5 field trials in the United States and Canada, Part III: Development and reliability testing of a cross-cutting symptom assessment for DSM-5. Am J Psychiatry. 2013;170: 71–82. pmid:23111499
- 47. Bravo AJ, Villarosa-Hurlocker MC, Pearson MR, Protective Strategies Study Team. College student mental health: An evaluation of the DSM-5 Self-Rated Level 1 Cross-Cutting Symptom Measure. Psychol Assess. 2018;30: 1382–1389. pmid:30070557
- 48. Pearson MR, Marijuana Outcomes Study Team. Marijuana Use Grid: A brief, comprehensive measure of marijuana use. Manuscr Submitt Publ. 2020.
- 49. Simons JS, Dvorak RD, Merrill JE, Read JP. Dimensions and severity of marijuana consequences: Development and validation of the Marijuana Consequences Questionnaire (MACQ). Addict Behav. 2012;37: 613–621. pmid:22305645
- 50. Marsh HW, Hau KT, Wen Z. In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Structural Equation Modeling. 2004. pp. 320–341. http://dx.doi.org/10.1207/s15328007sem1103_2
- 51.
Muthén LK, Muthén BO. Mplus user’s guide. Eigth Edition. Muthén LK, Muthén BO, editors. Los Angeles, CA; 2018.
- 52.
Meyer JP. Applied Measurement with JMetrik. New York: Taylor & Francis; 2014.
- 53.
Rasch G. Probabilistic models for some intelligence and achievement tests. Copenhagen, Denmark: Denmarks Paedagogiske Institut; 1960. http://dx.doi.org/10.1016/S0019-9958(61)80061-2
- 54. Buss AH, Perry M. The Aggression Questionnaire. J Pers Soc Psychol. 1992;63: 452–459. pmid:1403624
- 55.
The World Bank. Gini index (World Bank estimate). Available: https://data.worldbank.org/indicator/SI.POV.GINI?type = shaded&view = map&year = 2016
- 56. Maneiro L, Gómez-Fraguela JA, Cutrín O, Romero E. Impulsivity traits as correlates of antisocial behaviour in adolescents. Pers Individ Dif. 2017;104: 417–422. http://dx.doi.org/10.1016/j.paid.2016.08.045
- 57. Carlson SR, Pritchard AA, Dominelli RM. Externalizing behavior, the UPPS-P Impulsive Behavior scale and reward and punishment sensitivity. Pers Individ Dif. 2013;54: 202–207. http://dx.doi.org/10.1016/j.paid.2012.08.039
- 58.
Zuckerman M. Sensation seeking and risky behavior. Washington, DC, US: American Psychological Association; 2007. http://dx.doi.org/10.1037/11555-002
- 59. Martins SS, Wall MM, Eisenberg R, Blanco C, Santaella J, Ramos-Olazagasti M, et al. Trajectories of sensation seeking among Puerto Rican children and youth. J Am Acad Child Adolesc Psychiatry. 2015;54: 1042–1050. pmid:26598479
- 60. Hyatt CS, Owens MM, Gray JC, Carter NT, MacKillop J, Sweet LH, et al. Personality traits share overlapping neuroanatomical correlates with internalizing and externalizing psychopathology. J Abnorm Psychol. 2019;128: 1–11. pmid:30589303
- 61. Forbes MK, Kotov R, Ruggero CJ, Watson D, Zimmerman M, Krueger RF. Delineating the joint hierarchical structure of clinical and personality disorders in an outpatient psychiatric sample. Compr Psychiatry. 2017;79: 19–30. pmid:28495022
- 62. Madrid-Valero JJ, Ordoñana JR, Klump KL, Burt SA. Children sleep and antisocial behavior: Differential association of sleep with aggression and rule-breaking. J Abnorm Child Psychol. 2019;47: 791–799. pmid:30280364
- 63. McCloskey MS, Ammerman BA. Suicidal behavior and aggression-related disorders. Curr Opin Psychol. 2018;22: 54–58. pmid:28829989
- 64. Tielbeek JJ, Vink JM, Polderman TJC, Popma A, Posthuma D, Verweij KJH. Genetic correlation of antisocial behaviour with alcohol, nicotine, and cannabis use. Drug Alcohol Depend. 2018;187: 296–299. pmid:29702338
- 65. Knekta E, Runyon C, Eddy S. One size doesn’t fit all: Using factor analysis to gather validity evidence when using surveys in your research. CBE Life Sci Educ. 2019;18: 1–17. pmid:30821600
- 66.
Enzmann D, Kivivuori J, Haen Marshall I, Steketee M, Hough M, Killias M. A global perspective on young people as offenders and victims. Cham: Springer International Publishing; 2018. http://dx.doi.org/10.1007/978-3-319-63233-9