- Email: [email protected]
The development of Addiction-Prone Personality traits in biological and adoptive families
Personality and Individual Differences 82 (2015) 107–113
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
The development of Addiction-Prone Personality traits in biological and adoptive families Nozomi Franco Cea a,b,⇑, Gordon E. Barnes a,b,c a
School of Child and Youth Care at University of Victoria, PO Box 1700, STN CSC, Victoria, BC V8W 2Y2, Canada Centre for Addiction Research of British Columbia at University of Victoria, Canada c Centre for Youth and Society at University of Victoria, Canada b
a r t i c l e
i n f o
Article history: Received 4 November 2014 Received in revised form 2 February 2015 Accepted 27 February 2015 Available online 22 March 2015 Keywords: Addiction-Prone Personality Development Family environment Young adult Adoptive Longitudinal
a b s t r a c t This project investigated the predictors of Addiction-Prone Personality (APP) scores in youth and young adults from biological (N = 328, 53% female) and adoptive (N = 77, 53% female) families. The development of offspring’s APP traits was examined from three different angles: (1) patterns in biological and adoptive families, (2) offspring’s vs. parent’s perceptions of familial environment, and (3) different points in the life span. The offspring’s APP scores were found to be significantly predicted by parents’ APP scores in both biological and adoptive families. Parents’ APP scores and offspring’s gender consistently showed significant direct influences on offspring’s APP scores in biological families. The familial care factor (maternal and paternal care, family cohesion, and family adaptability) was found to be the consistent significant predictor of offspring’s APP scores in adoptive families even when offspring became older. These results are consistent in showing that the social environment plays an important role in the development of Addiction-Prone Personality traits. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction The abuse and misuse of a wide variety of substances and the consequences of such use have been a major societal problem all over the world (Kendler et al., 2012). Various studies have found significant associations between personality and substance use behaviors (e.g., Hicks, Durbin, Blonigen, Iacono, & McGue, 2012; Malmberg et al., 2012). The role of personality in the use of different substances (i.e., alcohol, tobacco, marijuana, etc.) as well as in polysubstance use has been well-established (e.g., Lackner, Unterrainer, & Neubauer, 2013). In recent years, more studies on personality and substance use have been conducted in the general population of both adults and youth, with several of these studies utilizing longitudinal research methodology. By doing so, researchers intended to investigate not only the personality of those individuals who are addicted to substances, but also how personality predisposes certain individuals to future substance use (e.g., Anderson, Barnes, & Murray, 2011; Krank et al., 2011). Over the last two decades, certain dimensions of personality underlying under-controlled or disinhibited behavior (i.e., impulsivity and
⇑ Corresponding author at: School of Child and Youth Care, University of Victoria, PO Box 1700, STN CSC, Victoria, BC V8W 2Y2, Canada. Tel.: +1 250 721 7979. E-mail addresses: [email protected] (N. Franco Cea), [email protected] (G.E. Barnes). http://dx.doi.org/10.1016/j.paid.2015.02.035 0191-8869/Ó 2015 Elsevier Ltd. All rights reserved.
sensation seeking) have been identified as correlates of substance use among other forms of externalizing behaviors (Quinn & Harden, 2013). In their extensive review of personality development, Caspi, Roberts, and Shiner (2005) point out that among the copious research studies on personality development and the influence of personality on various outcomes, only a very few studies have contemplated the influences of parental personality (e.g., Prinzie et al., 2012; Schofield et al., 2012). This is a curious oversight because parental personality could shape individual’s behavior directly via modeling effects or indirectly through the influence of personality on parenting behaviors (Caspi et al., 2005). Schofield et al. (2012) investigated whether a parent’s positive personality characteristics (i.e., the alpha-linked personality – high conscientiousness, high agreeableness, and low neuroticism) predict similar adolescent personality traits over time along with the role played by positive parenting in this process. Schofield et al. (2012) found that; (1) parents’ personality positively predicted observed parenting; (2) higher levels of parental alpha-linked traits were associated with higher levels of adolescent alpha-linked traits; and (3) positive parenting positively predicted adolescent alphalinked personality traits. These findings suggest that both parents’ personality and the quality of parenting behaviors may play an important role in personality development during adolescence (Schofield et al., 2012).
108
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
Finally, there is a fair amount of literature on associations between family socialization and broad personality traits (e.g., five-factor personality; Saucier, Wilson, & Warka, 2007), and between family socialization and the development of substance use behaviors (Barnes, Murray, Patton, Bentler, & Anderson, 2000). While many studies have found associations between personality traits, family socialization and substance use, only a handful of those focus on the development of specific personality traits such as substance-use-proneness. The nature of these personality traits – origin, lifetime change, association with other factors – continues to be a topic of debate. Therefore, we need more studies on the development of such specific personality traits and their associations with genetic and environmental factors in the lifespan. 1.1. Addiction-Prone Personality Previous studies have shown that personality characteristics play an important role in predicting substance use behaviors and patterns (Barnes et al., 2000). The Addiction-Prone Personality (APP) scale is a new measure that has been designed to predict underlying vulnerability to substance use problems (Anderson, Barnes, Patton, & Perkins, 1999; Anderson et al., 2011; Barnes et al., 2000). The APP scale was originally developed by utilizing data from a large general population survey (Winnipeg Health and Drinking Survey; Barnes et al. 2000). Personality items that were linked with both a family history of alcoholism and a lifetime diagnosis of alcoholism were selected from a large battery of personality tests (see Anderson et al., 2011 for the APP-21 items). The content of this scale suggests that individuals who score high on this measure are characterized by high novelty seeking and low self-regulation. Earlier studies with the APP scale found that this test is very good for discriminating drug addicts from non-addicts, and predicting the severity of addiction and likelihood of remission during recovery (Barnes et al., 2000). The APP scale is also a useful instrument for predicting alcohol and other substance misuse across both gender and age cohorts (Anderson et al., 1999). The APP scale was found to be significantly (p < .001) correlated with three of the Five-Factor Personality Scales – high APP scores are correlated with high Neuroticism, low Agreeableness, and low Conscientiousness (Barnes et al., 2000). The associations between parents’ and offspring’s APP were examined in one of the early studies of APP (Vancouver Family Survey (VFS); Barnes et al., 2000). The term offspring is used to refer to children of both biological and adoptive parents in this report. It found that offspring’s APP traits were significantly associated with their parents’ APP in biological families. While these correlations were not significant in the smaller sample of adoptive families, the order of magnitude of the effects observed were roughly the same. The same study has also shown that a nurturing family environment was significantly associated with lower APP scores in offspring (Barnes et al., 2000). Results support the possible role of the social environment on the development of APP. To this day, there is no study examining the APP scale with family environment or with other types of addictive behaviors such as problem gambling, shopping, or eating disorder available. 1.2. The current study This project was a secondary data analysis of the longitudinal data set (VFS), and investigated the predictors of the AddictionProne Personality (APP) scores in a community sample consisting of biological and adoptive families. The cross-sectional analysis of the first wave of VFS found; (1) the APP scores are significantly correlated with substance use (alcohol, tobacco, marijuana and other illicit drugs) in both parents and offspring; (2) males score higher
on the APP scale; (3) the offspring’s APP scores are higher than parents’ scores; however, within the offspring sample, the APP scores are not significantly different by age; and (4) socioeconomic status variables (parents’ education, income and occupation) are not strong predictors of offspring’s APP scores. In the current analysis, we particularly wanted to see whether or not the parents’ APP scores and family socialization factors could predict the development of offspring’s APP over time. Offspring’s APP was assessed at two different points in time (7 years apart). The hypothesized theoretical model to be tested is presented in Fig. 1. The theoretical model was tested to answer three research questions: 1. Do the parents’ APP scores and family socialization factors predict the offspring’s APP scores and does the association pattern differ in biological and adoptive samples? 2. Do offspring’s and parents’ perceived family socialization factors show different association patterns with offspring’s APP scores? 3. Does the effect of the family environment on offspring’s APP scores diminish over time and does this diminished effect differ in biological and adoptive samples? 2. Methods 2.1. Participant selection The Vancouver Family Survey (VFS) was conducted as a two wave longitudinal survey in the Greater Vancouver area. The descriptions of the recruitment of participants and the original design of VFS are available elsewhere (Anderson et al., 1999; Barnes et al., 2000). At time 1 (1995–1996), extensive questionnaires were administered to mothers, fathers and youngest offspring in the 14–25 age range living at home from 473 biological and 128 adoptive families for a total of 601 families. Participating families received $50.00 per family. Offspring in the adoptive families had to be adopted before the age of five, and most adoptions occurred early in the first year of life. The sample excluded adoptive families where there existed any biological relationship with one of the parents. Beginning in 2003, follow-up data at time 2 were collected for the offspring sample of participants in the VFS (now ages 21–33). Potential participants were offered $25.00 as an incentive to participate. A total of 215 females and 190 males (n = 405) were re-interviewed. This represented 67% of the participants originally tested at time 1. The refusal rate for this project was 18% with the remaining 15% of the sample lost for other reasons including death or failure to locate the individual (Barnes, Anderson, & Jansson, 2008). The final sample was comprised of 328 offspring from biological families and 77 offspring from adoptive families. Data for time 1 participants who did not participate at time 2 were excluded from this analysis. 2.2. Measures 2.2.1. Demographic domain Demographic questions that were selected as being important in the current investigation as control variables were offspring’s gender (0 = female and 1 = male) and adoption status. 2.2.2. Addiction-Prone Personality (APP) The APP-21 measure was used to assess personality vulnerability to substance use. Each family member completed the APP-21 measure at time 1. Seven years later, 405 offspring completed the same APP measure at time 2. Recent study (Anderson et al., 2011) has shown that the APP-21 scale is a reliable measure, both
109
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
in terms of internal consistency, test–retest reliability and predictive validity for predicting new cases of substance dependence. This scale also showed good discriminant validity in relation to several of the major personality inventories (Anderson et al., 2011). In the VFS, internal consistencies for the APP-21 scale were .67 (time 1) and .71 (time 2) for offspring, .64 for mothers, and .67 for fathers. 2.2.3. Familial care factor Perceptions of the familial care factors were assessed by two different instruments for measuring family systems and parental care at time 1. The family system was assessed by asking each family member to complete the Family Adaptability and Cohesion Evaluation Scales II (FACES II; Olson & Tiesel, 1991). FACES II contains two scales designed to measure adaptability and cohesion in each family. In the VFS, internal consistencies for the adaptability scale were .79 for offspring, .74 for mothers, and .72 for fathers. For the cohesion scale, Cronbach’s alphas were .88 for offspring, .86 for mothers, and .86 for fathers (Barnes, Patton, & Marshall, 1997). Parental care was assessed by having each family member fill out the Parker Parental Bonding Instrument (PBI) which contains 12 items assessing parental care (Parker, Tupling, & Brown, 1979). In the VFS, offspring filled out the instrument twice, once for each parent, and parents were asked to report on their own parenting over the first 15 years of the offspring’s life. In the offspring’s report, reliabilities of the care scale were very high at .90 for mother care and .91 for father care. In each of the parent’s self-reports Cronbach’s alphas were .73 for mother care and .78 for father care (Barnes et al., 1997). 2.3. Data analyses The first step was to examine the mean differences of APP scores between time 1 and time 2 by conducting the dependent T-test. Next, correlations were computed between the parents’ APP scores and offspring’s APP scores in both the biological and adoptive samples. The parents’ mid-point APP scores were also computed and correlated with offspring scores. The parents’ midpoint scores were used in earlier adoption studies and found to be slightly better predictors of offspring’s personality scores than individual parent scores (e.g., Scarr, Webber, Weinberg, & Wittig, 1981). Latent variables for familial care were developed using three variables (family cohesion, family adaptability and parental care) in both the parent perception model and the offspring perception model. Overall structural equation models combined biological and adoptive family samples were then developed by
Offspring’s Gender
Table 1 APP correlations between parents and offspring in biological and adoptive families. Offspring’s APP Biological sample (n = 328)
Mother’s APP Father’s APP Parents Mid-point APP * ** ***
Family Cohesion Family Adaptability Father Care Mother Care
Time 1
Time 2
Time 1
Time 2
.18** .19** .24***
.16** .16** .20***
.28* .19 .30**
.19 .11 .20
p < .05. p < .01. p < .001.
utilizing EQS 6.1 (Bentler, 2004) to examine the association between the parents’ APP, offspring’s APP and familial care factors as shown in Fig. 1. The Lagrange Multiplier test and Wald tests as well as the Comparative Fit Index (CFI) and the Root Mean Square Error of Approximation (RMSEA) were used to guide modifications to the model until an acceptable level of fit was achieved. The final stage of analysis involved developing the multiple group constraints models to compare the biological and adoptive sample groups. 3. Results 3.1. Comparisons between time 1 and time 2 APP scores The result of the dependent T-test was t(404) = 8.171, p < .000. Due to the means of the two APP scores and the direction of the t-value, we can conclude that there was a statistically significant decline of APP scores over time from 10.15 ± 3.57 to 8.74 ± 3.77 (p < .000); a decrease of 1.42 ± 3.49. 3.2. Bi-variate analyses Correlations between the parents’ and offspring’s APP scores at time 1 and time 2 are presented in Table 1. The trend of correlations is very similar to the one found in the earlier study (Barnes et al., 2000). The correlation coefficients between the parents’ mid-point and offspring’s APP at time 2 in biological and adoptive families are exactly same; however, due to the sample size, results show lower statistical significance in the adoptive sample. Furthermore, both biological and adopted offspring data show significant correlations between time 1 and time 2 APP scores (r = .70, r = .49 respectively). This implies that time 1 and time 2 APP scores should be analyzed in separate models and compared; otherwise, for time 2 APP, influences of other variables would be diminished
Adoption Status
Parents’ APP
Offspring’s APP Time 1
Offspring’s Perception
Adoptive sample (n = 77)
Familial Care Parents’ Perceptions Mother Father Family Cohesion Family Cohesion Family Adaptability Family Adaptability Father Care Mother Care
Fig. 1. The development of APP traits conceptual model.
Offspring’s APP Time 2
110
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
Gender
Parents’ Mid-APP A; -.04 B; -.18*
A; .10 B; .21*
.23*
Cohesion
A; -.32* B; -.00
.84 .72
Adaptability
Familial Care
.64
Offspring’s APP T1 -.44*
Mother Care .85
.68
.99
Father Care
²= 53; DF = 29; CFI = 0.96; RMSEA = 0.06
Fig. 2. Offspring’s perception time 1 model.
Gender
Parents’ Mid-APP A; -.06 B; -.18*
A; .08 B; .34*
.19*
Cohesion
A; -.33* B; -.00
.84 .73
Adaptability
Familial Care
.63
Offspring’s APP T2
A; -.42* B; -.16*
Mother Care
.91
.69
.99
Father Care
²= 37; DF = 28; CFI = 0.99; RMSEA = 0.04
Fig. 3. Offspring’s perception time 2 model.
Parents’ Mid-APP
Gender
Cohesion Fa .96
Cohesion Mo
.56
-.20*
.61
Familial Care
Adaptability Fa Adaptability Mo Care Fa
.20*
.20*
.28 .49
Offspring’s APP T1 A; -.43* B; -.10 .94
.98
.25
Care Mo
²= 78; DF = 56; CFI = 0.97; RMSEA = 0.04
Fig. 4. Parents’ perception time 1 model.
by this highly correlated time 1 APP. Based on these findings, the final version of structural equation model analysis was determined to consist of four separate multiple group constrained models. 3.3. Structural equation models The results of the final structural equation models (constrained models) are shown in Figs. 2–5. In these final models, the constrained paths are presented as solid lines, while dotted lines represent paths that are needed to be free from constraints. In the final offspring’s perception models (see Figs. 2 and 3); the models fit the data well with CFI of .96 for time 1 and .99 for time 2. The parents’ perception models (see Figs. 4 and 5) also fit the data well in both the time 1 and time 2 (CFI = .97 and .97 respectively). The variable characteristics for all variables used in these models are summarized in Table 2. 3.3.1. Biological and adoptive families In the constrained models, all measurement paths and many structural paths were constrained to be equal with several
exceptions. The structural path from the familial care factor to the offspring’s APP scores showed a clear difference between the biological and the adoptive samples. It was the significant predictive path in all four models for adoptive samples; however, for biological samples, this path was significant only in the offspring models. Male gender directly predicts offspring’s higher APP scores in the biological samples, whereas, it affects APP scores through the lower familial care latent variable in the adoptive samples. Lastly, the parents’ APP scores predict family cohesion only in the biological offspring samples.
3.3.2. Offspring’s and parents’ perceptions Both offspring’s and parents’ perceptions of familial care fit well within the models and show similar structural paths. One difference is that the parents’ APP scores did not predict the offspring’s perceived familial care factor, whereas it was a consistent significant predictor in the parents’ perception models. In the offspring models, male gender strongly predicted lower scores of the familial care factor in the adoptive samples; however, gender did not
111
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
Parents’ Mid-APP
Gender
Cohesion Fa .97
Cohesion Mo
.56
-.20*
.60
Familial Care
Adaptability Fa Adaptability Mo
.27 .49
Care Fa
.31*
.15*
Offspring’s APP T2 A; -.32* B; -.11 .92
.98
.25
Care Mo
²= 78; DF = 56; CFI = 0.97; RMSEA = 0.04
Fig. 5. Parents’ perception time 2 model.
Table 2 Characteristics of variables used in the final models.
Time 1 Offspring’s age Parents’ Mid APP Offspring’s APP Father’s perspective Cohesion Adaptability Care Mother’s perspective Cohesion Adaptability Care Offspring’s perspective Cohesion Adaptability Paternal care Maternal care Time 2 Offspring’s age Offspring’s APP
Mean
Range
SD
17.86 4.93 10.15
14–25 0–13.50 0–19.89
2.54 2.30 3.57
4.68 5.10 27.73
1–8 2–8 11–36
1.54 1.41 5.29
4.95 5.24 31.22
2–8 1–8 16–36
1.51 1.38 4.23
3.82 4.19 24.42 28.26
1–8 1–8 0–36 6–36
1.63 1.75 7.54 6.62
25.84 8.74
21–33 0–18.00
2.59 3.77
predict the familial care factor in the parents’ models even for the adoptive samples. 3.3.3. Effect of family environment over time Despite the significant difference between time 1 and time 2 APP scores found by the dependent T-test, the models did not show much difference between the two times in both offspring and parents models. One difference is the path from the familial care factor to the offspring’s APP. The coefficient of this path declined over time in the biological samples. However, this effect remained high in the adoptive samples, requiring the removal of this constraint. For the biological samples, gender was the consistent significant predictor of offspring’s APP scores in both offspring and parents models. 4. Discussion This project intended to investigate the predictors of AddictionProne Personality (APP) traits in a general population sample. We conducted secondary data analyses of the VFS longitudinal data set to study three research questions. First, we examined whether the parents’ APP scores and the family socialization factors could predict the development of offspring’s APP. The results are consistent with existing studies. Both the current study and the study by Schofield et al. (2012) found a significant influence of parental socialization on offspring’s
personality as well as a significant association between parents’ and offspring’s personality traits in older offspring. Moreover, we were able to control for genetic influences on APP development by including both biological and adoptive families. In the biological sample, the direct pathway between the parents’ APP and offspring’s APP includes both biological and family socialization components. In the adoptive sample, this pathway includes only a socialization component. The finding that these pathways are similar in strength in our biological and adoptive samples suggests that genetic components contributing to the pathway between the offspring’s and the parents’ APP scores are not very strong. At the same time, the finding that these pathways are consistently significant in all four models, suggests that there might be effects of both environment and gene–environment interaction on the development of this trait in offspring. Next, we compared parents’ and offspring’s perception models and found that all models were similar at both times. This finding confirms that the effects of family socialization on offspring outcomes hold up across different reporting sources. However, the associations between parents’ APP and perceived familial factors are significant only in the parents’ perception models. There is still limited research on and thus little is known about how parent personality traits might affect parenting and family environment directly and/or indirectly, particularly when offspring are in adolescence and emerging adulthood (Oppenheimer, Hankin, Jenness, Young, & Smolen, 2013). Therefore, more studies of parent personality are urgently needed. Lastly, we examined whether the effect of the family environment diminished over time. We found that this effect diminished in the biological sample; whereas it remained high in the adoptive sample. It suggests the possibility that adopted offspring are more sensitive and susceptible to lasting family environment effects. Furthermore, The APP scores in the current study showed a significant decline when offspring got older which is consistent with existing studies of impulsivity and sensation seeking (e.g., Quinn & Harden, 2013). However, the APP scores of our adopted offspring sample did not decline as much (t = 2.25, p < .05) as biological offspring did (t = 7.96, p < .000). Therefore, there is the possibility that the family environment keeps predicting offspring’s APP score over time because of this unchanged or less changed APP scores in adopted offspring. 5. Limitations The current study was not without limitations. First, we examined only the total scores of the APP-21 measure which may contain several facets of the Addiction-Prone Personality. Anderson (2003) found that the APP is multi-faceted and identified three facets – impulsivity/recklessness, sensation-seeking and negative
112
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
view of self. At the same time, studies by Quinn and Harden (2013) and by Hicks et al. (2013) found that each facet or subscale of personality can have a different trajectory of development over time (e.g., impulsivity, sensation seeking, boldness, etc.). Therefore, by analyzing three facets of the APP, future studies will provide a more nuanced description of associations between parents’ APP, environment and offspring’s APP over time. Second, in the VFS, we had only two waves of data. To maximise the value of the VFS data, and to focus more on development, it will be meaningful to examine change/discrepancy of APP scores between two waves as a dependent variable in future studies. Third, since the current study utilized self-report measure, we need to consider possibilities of biased data due to the response distortion and social desirability (Viswesvaran & Ones, 1999). Therefore, our findings need to be interpreted with caution. Moreover, it should be acknowledged that there are possible multi-directional impacts and relations between offspring’s personality, family environment and offspring’s substance use. Research in developmental and personality psychology over the last several decades has increasingly emphasized the individuals’ role in influencing his/her own environment including the rearing environment and selection of social environment (Hicks et al., 2013; Oppenheimer et al., 2013). The theory of genotype-environment effects hypothesizes that individuals are exposed to, seek out, select and create their own environments (Scarr & McCartney, 1983). In other words, offspring’s personality traits and behaviors (substance use) can affect his/her environment. At the same time, substance use studies suggest that there are possible bi-directional relations between personality development and substance use (Hicks et al., 2012; Malmberg et al., 2012; Quinn and Harden, 2013). They point to the possibilities that the onset and trajectory of offspring’s substance use impact on his/her personality development as well as the impact of personality on substance use. By considering those possibilities, future studies should include changes in environment, behaviors (substance use) and personality in their investigation of APP development. 6. Conclusion The results are consistent in showing that the social environment plays an important role in the development of AddictionProne Personality traits. High familial care is protective against the development of Addiction-Prone Personality characteristics. Moreover, it remains protective in the adoptive families even when offspring became older. These findings will be useful for training programs for parents as well as for professionals working with children, youth, and families. Such programs should pay close attention to the promotion of caring family environments. Parents should be encouraged to be confidently involved in their offspring’s development even as offspring get older. Furthermore, based on the findings about the potential sensitivity and susceptibility of adopted offspring towards family environment and its long-term influence, it is crucial that all personnel who work with or live with adopted or foster children and youth are aware of the direct and indirect impact of the whole family environment. Future studies should continue to examine the development of APP traits in different age groups, the interplay between gene and environment on substance use, and multi-directional effects between these factors. Acknowledgements The first wave of data collection of Vancouver Family Survey (VFS) used in this study was funded by Health Canada through the National Health Research and Development Program
(NHRDP). The second wave of VFS was funded by a Community Alliance for Health Research grant from the CIHR and a New Emerging Team grant (#RA79917) funded by CIHR (Institutes of Neurosciences, Mental Health and Addiction, Institute of Aboriginal People’s Health, Institute of Human Development, Child and Youth Health), the Canadian Centre on Substance Abuse, and Canadian Tobacco Control Initiative. Resources in support of this study were also provided by the Centre for Addiction Research of British Columbia.
References Anderson, R. E. (2003). Investigating a quantitative measure of addiction-prone personality. ProQuest, UMI Dissertations Publishing. Anderson, R. E., Barnes, G. E., & Murray, R. P. (2011). Psychometric properties and long-term predictive validity of the addiction-prone personality (APP) scale. Personality and Individual Differences, 50(5), 651–656. Anderson, R. E., Barnes, G. E., Patton, D., & Perkins, T. M. (1999). Personality in the development of substance abuse. In I. Mervielde, I. J. Deary, F. De Fruyt, & F. Ostendorf (Eds.). Personal psychology in Europe (Vol. 7, pp. 141–158). The Netherlands: Tilburg University Press. Barnes, G. E., Patton, D., & Marshall, S. K. (1997). Family Environments and Substance Use. Final Report Submitted to NHRDP. Ottawa, Canada. Barnes, G. E., Anderson, R. E., & Jansson, M. (2008). Polysubstance use and mental health symptoms. Paper presented at the 34th alcohol epidemiology symposium of Kettil Bruun Society, Victoria, Canada. Barnes, G. E., Murray, R. P., Patton, D., Bentler, P., & Anderson, R. E. (2000). The addiction-prone personality. New York: Kluwer Academic/Plenum Publishers. Bentler, P. M. (2004). EQS 6.1 for Windows. Encino, CA: Multivariate Software Inc.. Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality development: Stability and change. Annual Review of Psychology, 56(1), 453–484. http://dx.doi.org/ 10.1146/annurev.psych.55.090902.141913. Hicks, B. M., Durbin, C. E., Blonigen, D. M., Iacono, W. G., & McGue, M. (2012). Relationship between personality change and the onset and course of alcohol dependence in young adulthood. Addiction, 107(3), 540–548. Hicks, B. M., Johnson, W., Durbin, C. E., Blonigen, D. M., Iacono, W. G., & McGue, M. (2013). Gene–environment correlation in the development of adolescent substance abuse: Selection effects of child personality and mediation via contextual risk factors. Development and Psychopathology, 25(1), 119–132. http://dx.doi.org/10.1017/S0954579412000946. Kendler, K. S., Sundquist, K., Ohlsson, H., Palmér, K., Maes, H., Winkleby, M. A., et al. (2012). Genetic and familial environmental influences on the risk for drug abuse: A national Swedish adoption study. Archives of General Psychiatry, 69(7), 690–697. http://dx.doi.org/10.1001/archgenpsychiatry.2011.2112. Krank, M., Stewart, S. H., O’Connor, R., Woicik, P. B., Wall, A., & Conrod, P. J. (2011). Structural, concurrent, and predictive validity of the substance use risk profile scale in early adolescence. Addictive Behaviors, 36(1), 37–46. http://dx.doi.org/ 10.1016/j.addbeh.2010.08.010. Lackner, N., Unterrainer, H., & Neubauer, A. C. (2013). Differences in big five personality traits between alcohol and polydrug abusers: Implications for treatment in the therapeutic community. International Journal of Mental Health and Addiction, 11(6), 682–692. http://dx.doi.org/10.1007/s11469-013-9445-2. Malmberg, M., Kleinjan, M., Vermulst, A. A., Overbeek, G., Monshouwer, K., Lammers, J., et al. (2012). Do substance use risk personality dimensions predict the onset of substance use in early adolescence? A variable- and personcentered approach. Journal of Youth and Adolescence, 41(11), 1512–1525. http:// dx.doi.org/10.1007/s10964-012-9775-6. Olson, D. H., & Tiesel, J. (1991). Faces II: Linear scoring and interpretation. St. Paul, MN: University of Minnesota. Oppenheimer, C., Hankin, B., Jenness, J., Young, J., & Smolen, A. (2013). Observed positive parenting behaviors and youth genotype: Evidence for gene– environment correlations and moderation by parent personality traits. Development and Psychopathology, 25(1), 175–191. http://dx.doi.org/10.1017/ S0954579412000983. Parker, G., Tupling, H., & Brown, L. B. (1979). A parental bonding instrument. British Journal of Medical Psychology, 52(1), 1–10. Prinzie, P., Dekovic, M., van den Akker, A., de Haan, A., Stoltz, S., & Hendriks, A. (2012). Fathers’ personality and its interaction with children’s personality as predictors of perceived parenting behavior six years later. Personality and Individual Differences, 52(2), 183–189. http://dx.doi.org/10.1016/ j.paid.2011.10.012. Quinn, P., & Harden, K. (2013). Differential changes in impulsivity and sensation seeking and the escalation of substance use from adolescence to early adulthood. Development and Psychopathology, 25(1), 223–239. http:// dx.doi.org/10.1017/S0954579412000284. Saucier, G., Wilson, K. R., & Warka, J. (2007). The structure of retrospective accounts of family environments: Related to the structure of personality attributes. Journal of Personality Assessment, 88(3), 295–308. http://dx.doi.org/10.1080/ 00223890701317012. Scarr, S., & McCartney, K. (1983). How people make their own environments: A theory of genotype ? environment effects. Blackwell Publishing. http://dx.doi.org/ 10.2307/1129703.
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113 Scarr, S., Webber, P. L., Weinberg, R. A., & Wittig, M. A. (1981). Personality resemblance among adolescents and their parents in biologically related and adoptive families. Journal of Personality and Social Psychology, 40(5), 885–898. http://dx.doi.org/10.1037/0022-3514.40.5.885. Schofield, T. J., Conger, R. D., Donnellan, M. B., Jochem, R., Widaman, K. F., & Conger, K. J. (2012). Parent personality and positive parenting as predictors of positive
113
adolescent personality development over time. Merrill-Palmer Quarterly, 58(2), 255–283. http://dx.doi.org/10.1353/mpq.2012.0008. Viswesvaran, C., & Ones, D. S. (1999). Meta-analysis of fakability estimates: Implications for personality measurement. Educational and Psychological Measurement, 59, 197–210.
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
The development of Addiction-Prone Personality traits in biological and adoptive families Nozomi Franco Cea a,b,⇑, Gordon E. Barnes a,b,c a
School of Child and Youth Care at University of Victoria, PO Box 1700, STN CSC, Victoria, BC V8W 2Y2, Canada Centre for Addiction Research of British Columbia at University of Victoria, Canada c Centre for Youth and Society at University of Victoria, Canada b
a r t i c l e
i n f o
Article history: Received 4 November 2014 Received in revised form 2 February 2015 Accepted 27 February 2015 Available online 22 March 2015 Keywords: Addiction-Prone Personality Development Family environment Young adult Adoptive Longitudinal
a b s t r a c t This project investigated the predictors of Addiction-Prone Personality (APP) scores in youth and young adults from biological (N = 328, 53% female) and adoptive (N = 77, 53% female) families. The development of offspring’s APP traits was examined from three different angles: (1) patterns in biological and adoptive families, (2) offspring’s vs. parent’s perceptions of familial environment, and (3) different points in the life span. The offspring’s APP scores were found to be significantly predicted by parents’ APP scores in both biological and adoptive families. Parents’ APP scores and offspring’s gender consistently showed significant direct influences on offspring’s APP scores in biological families. The familial care factor (maternal and paternal care, family cohesion, and family adaptability) was found to be the consistent significant predictor of offspring’s APP scores in adoptive families even when offspring became older. These results are consistent in showing that the social environment plays an important role in the development of Addiction-Prone Personality traits. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction The abuse and misuse of a wide variety of substances and the consequences of such use have been a major societal problem all over the world (Kendler et al., 2012). Various studies have found significant associations between personality and substance use behaviors (e.g., Hicks, Durbin, Blonigen, Iacono, & McGue, 2012; Malmberg et al., 2012). The role of personality in the use of different substances (i.e., alcohol, tobacco, marijuana, etc.) as well as in polysubstance use has been well-established (e.g., Lackner, Unterrainer, & Neubauer, 2013). In recent years, more studies on personality and substance use have been conducted in the general population of both adults and youth, with several of these studies utilizing longitudinal research methodology. By doing so, researchers intended to investigate not only the personality of those individuals who are addicted to substances, but also how personality predisposes certain individuals to future substance use (e.g., Anderson, Barnes, & Murray, 2011; Krank et al., 2011). Over the last two decades, certain dimensions of personality underlying under-controlled or disinhibited behavior (i.e., impulsivity and
⇑ Corresponding author at: School of Child and Youth Care, University of Victoria, PO Box 1700, STN CSC, Victoria, BC V8W 2Y2, Canada. Tel.: +1 250 721 7979. E-mail addresses: [email protected] (N. Franco Cea), [email protected] (G.E. Barnes). http://dx.doi.org/10.1016/j.paid.2015.02.035 0191-8869/Ó 2015 Elsevier Ltd. All rights reserved.
sensation seeking) have been identified as correlates of substance use among other forms of externalizing behaviors (Quinn & Harden, 2013). In their extensive review of personality development, Caspi, Roberts, and Shiner (2005) point out that among the copious research studies on personality development and the influence of personality on various outcomes, only a very few studies have contemplated the influences of parental personality (e.g., Prinzie et al., 2012; Schofield et al., 2012). This is a curious oversight because parental personality could shape individual’s behavior directly via modeling effects or indirectly through the influence of personality on parenting behaviors (Caspi et al., 2005). Schofield et al. (2012) investigated whether a parent’s positive personality characteristics (i.e., the alpha-linked personality – high conscientiousness, high agreeableness, and low neuroticism) predict similar adolescent personality traits over time along with the role played by positive parenting in this process. Schofield et al. (2012) found that; (1) parents’ personality positively predicted observed parenting; (2) higher levels of parental alpha-linked traits were associated with higher levels of adolescent alpha-linked traits; and (3) positive parenting positively predicted adolescent alphalinked personality traits. These findings suggest that both parents’ personality and the quality of parenting behaviors may play an important role in personality development during adolescence (Schofield et al., 2012).
108
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
Finally, there is a fair amount of literature on associations between family socialization and broad personality traits (e.g., five-factor personality; Saucier, Wilson, & Warka, 2007), and between family socialization and the development of substance use behaviors (Barnes, Murray, Patton, Bentler, & Anderson, 2000). While many studies have found associations between personality traits, family socialization and substance use, only a handful of those focus on the development of specific personality traits such as substance-use-proneness. The nature of these personality traits – origin, lifetime change, association with other factors – continues to be a topic of debate. Therefore, we need more studies on the development of such specific personality traits and their associations with genetic and environmental factors in the lifespan. 1.1. Addiction-Prone Personality Previous studies have shown that personality characteristics play an important role in predicting substance use behaviors and patterns (Barnes et al., 2000). The Addiction-Prone Personality (APP) scale is a new measure that has been designed to predict underlying vulnerability to substance use problems (Anderson, Barnes, Patton, & Perkins, 1999; Anderson et al., 2011; Barnes et al., 2000). The APP scale was originally developed by utilizing data from a large general population survey (Winnipeg Health and Drinking Survey; Barnes et al. 2000). Personality items that were linked with both a family history of alcoholism and a lifetime diagnosis of alcoholism were selected from a large battery of personality tests (see Anderson et al., 2011 for the APP-21 items). The content of this scale suggests that individuals who score high on this measure are characterized by high novelty seeking and low self-regulation. Earlier studies with the APP scale found that this test is very good for discriminating drug addicts from non-addicts, and predicting the severity of addiction and likelihood of remission during recovery (Barnes et al., 2000). The APP scale is also a useful instrument for predicting alcohol and other substance misuse across both gender and age cohorts (Anderson et al., 1999). The APP scale was found to be significantly (p < .001) correlated with three of the Five-Factor Personality Scales – high APP scores are correlated with high Neuroticism, low Agreeableness, and low Conscientiousness (Barnes et al., 2000). The associations between parents’ and offspring’s APP were examined in one of the early studies of APP (Vancouver Family Survey (VFS); Barnes et al., 2000). The term offspring is used to refer to children of both biological and adoptive parents in this report. It found that offspring’s APP traits were significantly associated with their parents’ APP in biological families. While these correlations were not significant in the smaller sample of adoptive families, the order of magnitude of the effects observed were roughly the same. The same study has also shown that a nurturing family environment was significantly associated with lower APP scores in offspring (Barnes et al., 2000). Results support the possible role of the social environment on the development of APP. To this day, there is no study examining the APP scale with family environment or with other types of addictive behaviors such as problem gambling, shopping, or eating disorder available. 1.2. The current study This project was a secondary data analysis of the longitudinal data set (VFS), and investigated the predictors of the AddictionProne Personality (APP) scores in a community sample consisting of biological and adoptive families. The cross-sectional analysis of the first wave of VFS found; (1) the APP scores are significantly correlated with substance use (alcohol, tobacco, marijuana and other illicit drugs) in both parents and offspring; (2) males score higher
on the APP scale; (3) the offspring’s APP scores are higher than parents’ scores; however, within the offspring sample, the APP scores are not significantly different by age; and (4) socioeconomic status variables (parents’ education, income and occupation) are not strong predictors of offspring’s APP scores. In the current analysis, we particularly wanted to see whether or not the parents’ APP scores and family socialization factors could predict the development of offspring’s APP over time. Offspring’s APP was assessed at two different points in time (7 years apart). The hypothesized theoretical model to be tested is presented in Fig. 1. The theoretical model was tested to answer three research questions: 1. Do the parents’ APP scores and family socialization factors predict the offspring’s APP scores and does the association pattern differ in biological and adoptive samples? 2. Do offspring’s and parents’ perceived family socialization factors show different association patterns with offspring’s APP scores? 3. Does the effect of the family environment on offspring’s APP scores diminish over time and does this diminished effect differ in biological and adoptive samples? 2. Methods 2.1. Participant selection The Vancouver Family Survey (VFS) was conducted as a two wave longitudinal survey in the Greater Vancouver area. The descriptions of the recruitment of participants and the original design of VFS are available elsewhere (Anderson et al., 1999; Barnes et al., 2000). At time 1 (1995–1996), extensive questionnaires were administered to mothers, fathers and youngest offspring in the 14–25 age range living at home from 473 biological and 128 adoptive families for a total of 601 families. Participating families received $50.00 per family. Offspring in the adoptive families had to be adopted before the age of five, and most adoptions occurred early in the first year of life. The sample excluded adoptive families where there existed any biological relationship with one of the parents. Beginning in 2003, follow-up data at time 2 were collected for the offspring sample of participants in the VFS (now ages 21–33). Potential participants were offered $25.00 as an incentive to participate. A total of 215 females and 190 males (n = 405) were re-interviewed. This represented 67% of the participants originally tested at time 1. The refusal rate for this project was 18% with the remaining 15% of the sample lost for other reasons including death or failure to locate the individual (Barnes, Anderson, & Jansson, 2008). The final sample was comprised of 328 offspring from biological families and 77 offspring from adoptive families. Data for time 1 participants who did not participate at time 2 were excluded from this analysis. 2.2. Measures 2.2.1. Demographic domain Demographic questions that were selected as being important in the current investigation as control variables were offspring’s gender (0 = female and 1 = male) and adoption status. 2.2.2. Addiction-Prone Personality (APP) The APP-21 measure was used to assess personality vulnerability to substance use. Each family member completed the APP-21 measure at time 1. Seven years later, 405 offspring completed the same APP measure at time 2. Recent study (Anderson et al., 2011) has shown that the APP-21 scale is a reliable measure, both
109
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
in terms of internal consistency, test–retest reliability and predictive validity for predicting new cases of substance dependence. This scale also showed good discriminant validity in relation to several of the major personality inventories (Anderson et al., 2011). In the VFS, internal consistencies for the APP-21 scale were .67 (time 1) and .71 (time 2) for offspring, .64 for mothers, and .67 for fathers. 2.2.3. Familial care factor Perceptions of the familial care factors were assessed by two different instruments for measuring family systems and parental care at time 1. The family system was assessed by asking each family member to complete the Family Adaptability and Cohesion Evaluation Scales II (FACES II; Olson & Tiesel, 1991). FACES II contains two scales designed to measure adaptability and cohesion in each family. In the VFS, internal consistencies for the adaptability scale were .79 for offspring, .74 for mothers, and .72 for fathers. For the cohesion scale, Cronbach’s alphas were .88 for offspring, .86 for mothers, and .86 for fathers (Barnes, Patton, & Marshall, 1997). Parental care was assessed by having each family member fill out the Parker Parental Bonding Instrument (PBI) which contains 12 items assessing parental care (Parker, Tupling, & Brown, 1979). In the VFS, offspring filled out the instrument twice, once for each parent, and parents were asked to report on their own parenting over the first 15 years of the offspring’s life. In the offspring’s report, reliabilities of the care scale were very high at .90 for mother care and .91 for father care. In each of the parent’s self-reports Cronbach’s alphas were .73 for mother care and .78 for father care (Barnes et al., 1997). 2.3. Data analyses The first step was to examine the mean differences of APP scores between time 1 and time 2 by conducting the dependent T-test. Next, correlations were computed between the parents’ APP scores and offspring’s APP scores in both the biological and adoptive samples. The parents’ mid-point APP scores were also computed and correlated with offspring scores. The parents’ midpoint scores were used in earlier adoption studies and found to be slightly better predictors of offspring’s personality scores than individual parent scores (e.g., Scarr, Webber, Weinberg, & Wittig, 1981). Latent variables for familial care were developed using three variables (family cohesion, family adaptability and parental care) in both the parent perception model and the offspring perception model. Overall structural equation models combined biological and adoptive family samples were then developed by
Offspring’s Gender
Table 1 APP correlations between parents and offspring in biological and adoptive families. Offspring’s APP Biological sample (n = 328)
Mother’s APP Father’s APP Parents Mid-point APP * ** ***
Family Cohesion Family Adaptability Father Care Mother Care
Time 1
Time 2
Time 1
Time 2
.18** .19** .24***
.16** .16** .20***
.28* .19 .30**
.19 .11 .20
p < .05. p < .01. p < .001.
utilizing EQS 6.1 (Bentler, 2004) to examine the association between the parents’ APP, offspring’s APP and familial care factors as shown in Fig. 1. The Lagrange Multiplier test and Wald tests as well as the Comparative Fit Index (CFI) and the Root Mean Square Error of Approximation (RMSEA) were used to guide modifications to the model until an acceptable level of fit was achieved. The final stage of analysis involved developing the multiple group constraints models to compare the biological and adoptive sample groups. 3. Results 3.1. Comparisons between time 1 and time 2 APP scores The result of the dependent T-test was t(404) = 8.171, p < .000. Due to the means of the two APP scores and the direction of the t-value, we can conclude that there was a statistically significant decline of APP scores over time from 10.15 ± 3.57 to 8.74 ± 3.77 (p < .000); a decrease of 1.42 ± 3.49. 3.2. Bi-variate analyses Correlations between the parents’ and offspring’s APP scores at time 1 and time 2 are presented in Table 1. The trend of correlations is very similar to the one found in the earlier study (Barnes et al., 2000). The correlation coefficients between the parents’ mid-point and offspring’s APP at time 2 in biological and adoptive families are exactly same; however, due to the sample size, results show lower statistical significance in the adoptive sample. Furthermore, both biological and adopted offspring data show significant correlations between time 1 and time 2 APP scores (r = .70, r = .49 respectively). This implies that time 1 and time 2 APP scores should be analyzed in separate models and compared; otherwise, for time 2 APP, influences of other variables would be diminished
Adoption Status
Parents’ APP
Offspring’s APP Time 1
Offspring’s Perception
Adoptive sample (n = 77)
Familial Care Parents’ Perceptions Mother Father Family Cohesion Family Cohesion Family Adaptability Family Adaptability Father Care Mother Care
Fig. 1. The development of APP traits conceptual model.
Offspring’s APP Time 2
110
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
Gender
Parents’ Mid-APP A; -.04 B; -.18*
A; .10 B; .21*
.23*
Cohesion
A; -.32* B; -.00
.84 .72
Adaptability
Familial Care
.64
Offspring’s APP T1 -.44*
Mother Care .85
.68
.99
Father Care
²= 53; DF = 29; CFI = 0.96; RMSEA = 0.06
Fig. 2. Offspring’s perception time 1 model.
Gender
Parents’ Mid-APP A; -.06 B; -.18*
A; .08 B; .34*
.19*
Cohesion
A; -.33* B; -.00
.84 .73
Adaptability
Familial Care
.63
Offspring’s APP T2
A; -.42* B; -.16*
Mother Care
.91
.69
.99
Father Care
²= 37; DF = 28; CFI = 0.99; RMSEA = 0.04
Fig. 3. Offspring’s perception time 2 model.
Parents’ Mid-APP
Gender
Cohesion Fa .96
Cohesion Mo
.56
-.20*
.61
Familial Care
Adaptability Fa Adaptability Mo Care Fa
.20*
.20*
.28 .49
Offspring’s APP T1 A; -.43* B; -.10 .94
.98
.25
Care Mo
²= 78; DF = 56; CFI = 0.97; RMSEA = 0.04
Fig. 4. Parents’ perception time 1 model.
by this highly correlated time 1 APP. Based on these findings, the final version of structural equation model analysis was determined to consist of four separate multiple group constrained models. 3.3. Structural equation models The results of the final structural equation models (constrained models) are shown in Figs. 2–5. In these final models, the constrained paths are presented as solid lines, while dotted lines represent paths that are needed to be free from constraints. In the final offspring’s perception models (see Figs. 2 and 3); the models fit the data well with CFI of .96 for time 1 and .99 for time 2. The parents’ perception models (see Figs. 4 and 5) also fit the data well in both the time 1 and time 2 (CFI = .97 and .97 respectively). The variable characteristics for all variables used in these models are summarized in Table 2. 3.3.1. Biological and adoptive families In the constrained models, all measurement paths and many structural paths were constrained to be equal with several
exceptions. The structural path from the familial care factor to the offspring’s APP scores showed a clear difference between the biological and the adoptive samples. It was the significant predictive path in all four models for adoptive samples; however, for biological samples, this path was significant only in the offspring models. Male gender directly predicts offspring’s higher APP scores in the biological samples, whereas, it affects APP scores through the lower familial care latent variable in the adoptive samples. Lastly, the parents’ APP scores predict family cohesion only in the biological offspring samples.
3.3.2. Offspring’s and parents’ perceptions Both offspring’s and parents’ perceptions of familial care fit well within the models and show similar structural paths. One difference is that the parents’ APP scores did not predict the offspring’s perceived familial care factor, whereas it was a consistent significant predictor in the parents’ perception models. In the offspring models, male gender strongly predicted lower scores of the familial care factor in the adoptive samples; however, gender did not
111
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
Parents’ Mid-APP
Gender
Cohesion Fa .97
Cohesion Mo
.56
-.20*
.60
Familial Care
Adaptability Fa Adaptability Mo
.27 .49
Care Fa
.31*
.15*
Offspring’s APP T2 A; -.32* B; -.11 .92
.98
.25
Care Mo
²= 78; DF = 56; CFI = 0.97; RMSEA = 0.04
Fig. 5. Parents’ perception time 2 model.
Table 2 Characteristics of variables used in the final models.
Time 1 Offspring’s age Parents’ Mid APP Offspring’s APP Father’s perspective Cohesion Adaptability Care Mother’s perspective Cohesion Adaptability Care Offspring’s perspective Cohesion Adaptability Paternal care Maternal care Time 2 Offspring’s age Offspring’s APP
Mean
Range
SD
17.86 4.93 10.15
14–25 0–13.50 0–19.89
2.54 2.30 3.57
4.68 5.10 27.73
1–8 2–8 11–36
1.54 1.41 5.29
4.95 5.24 31.22
2–8 1–8 16–36
1.51 1.38 4.23
3.82 4.19 24.42 28.26
1–8 1–8 0–36 6–36
1.63 1.75 7.54 6.62
25.84 8.74
21–33 0–18.00
2.59 3.77
predict the familial care factor in the parents’ models even for the adoptive samples. 3.3.3. Effect of family environment over time Despite the significant difference between time 1 and time 2 APP scores found by the dependent T-test, the models did not show much difference between the two times in both offspring and parents models. One difference is the path from the familial care factor to the offspring’s APP. The coefficient of this path declined over time in the biological samples. However, this effect remained high in the adoptive samples, requiring the removal of this constraint. For the biological samples, gender was the consistent significant predictor of offspring’s APP scores in both offspring and parents models. 4. Discussion This project intended to investigate the predictors of AddictionProne Personality (APP) traits in a general population sample. We conducted secondary data analyses of the VFS longitudinal data set to study three research questions. First, we examined whether the parents’ APP scores and the family socialization factors could predict the development of offspring’s APP. The results are consistent with existing studies. Both the current study and the study by Schofield et al. (2012) found a significant influence of parental socialization on offspring’s
personality as well as a significant association between parents’ and offspring’s personality traits in older offspring. Moreover, we were able to control for genetic influences on APP development by including both biological and adoptive families. In the biological sample, the direct pathway between the parents’ APP and offspring’s APP includes both biological and family socialization components. In the adoptive sample, this pathway includes only a socialization component. The finding that these pathways are similar in strength in our biological and adoptive samples suggests that genetic components contributing to the pathway between the offspring’s and the parents’ APP scores are not very strong. At the same time, the finding that these pathways are consistently significant in all four models, suggests that there might be effects of both environment and gene–environment interaction on the development of this trait in offspring. Next, we compared parents’ and offspring’s perception models and found that all models were similar at both times. This finding confirms that the effects of family socialization on offspring outcomes hold up across different reporting sources. However, the associations between parents’ APP and perceived familial factors are significant only in the parents’ perception models. There is still limited research on and thus little is known about how parent personality traits might affect parenting and family environment directly and/or indirectly, particularly when offspring are in adolescence and emerging adulthood (Oppenheimer, Hankin, Jenness, Young, & Smolen, 2013). Therefore, more studies of parent personality are urgently needed. Lastly, we examined whether the effect of the family environment diminished over time. We found that this effect diminished in the biological sample; whereas it remained high in the adoptive sample. It suggests the possibility that adopted offspring are more sensitive and susceptible to lasting family environment effects. Furthermore, The APP scores in the current study showed a significant decline when offspring got older which is consistent with existing studies of impulsivity and sensation seeking (e.g., Quinn & Harden, 2013). However, the APP scores of our adopted offspring sample did not decline as much (t = 2.25, p < .05) as biological offspring did (t = 7.96, p < .000). Therefore, there is the possibility that the family environment keeps predicting offspring’s APP score over time because of this unchanged or less changed APP scores in adopted offspring. 5. Limitations The current study was not without limitations. First, we examined only the total scores of the APP-21 measure which may contain several facets of the Addiction-Prone Personality. Anderson (2003) found that the APP is multi-faceted and identified three facets – impulsivity/recklessness, sensation-seeking and negative
112
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113
view of self. At the same time, studies by Quinn and Harden (2013) and by Hicks et al. (2013) found that each facet or subscale of personality can have a different trajectory of development over time (e.g., impulsivity, sensation seeking, boldness, etc.). Therefore, by analyzing three facets of the APP, future studies will provide a more nuanced description of associations between parents’ APP, environment and offspring’s APP over time. Second, in the VFS, we had only two waves of data. To maximise the value of the VFS data, and to focus more on development, it will be meaningful to examine change/discrepancy of APP scores between two waves as a dependent variable in future studies. Third, since the current study utilized self-report measure, we need to consider possibilities of biased data due to the response distortion and social desirability (Viswesvaran & Ones, 1999). Therefore, our findings need to be interpreted with caution. Moreover, it should be acknowledged that there are possible multi-directional impacts and relations between offspring’s personality, family environment and offspring’s substance use. Research in developmental and personality psychology over the last several decades has increasingly emphasized the individuals’ role in influencing his/her own environment including the rearing environment and selection of social environment (Hicks et al., 2013; Oppenheimer et al., 2013). The theory of genotype-environment effects hypothesizes that individuals are exposed to, seek out, select and create their own environments (Scarr & McCartney, 1983). In other words, offspring’s personality traits and behaviors (substance use) can affect his/her environment. At the same time, substance use studies suggest that there are possible bi-directional relations between personality development and substance use (Hicks et al., 2012; Malmberg et al., 2012; Quinn and Harden, 2013). They point to the possibilities that the onset and trajectory of offspring’s substance use impact on his/her personality development as well as the impact of personality on substance use. By considering those possibilities, future studies should include changes in environment, behaviors (substance use) and personality in their investigation of APP development. 6. Conclusion The results are consistent in showing that the social environment plays an important role in the development of AddictionProne Personality traits. High familial care is protective against the development of Addiction-Prone Personality characteristics. Moreover, it remains protective in the adoptive families even when offspring became older. These findings will be useful for training programs for parents as well as for professionals working with children, youth, and families. Such programs should pay close attention to the promotion of caring family environments. Parents should be encouraged to be confidently involved in their offspring’s development even as offspring get older. Furthermore, based on the findings about the potential sensitivity and susceptibility of adopted offspring towards family environment and its long-term influence, it is crucial that all personnel who work with or live with adopted or foster children and youth are aware of the direct and indirect impact of the whole family environment. Future studies should continue to examine the development of APP traits in different age groups, the interplay between gene and environment on substance use, and multi-directional effects between these factors. Acknowledgements The first wave of data collection of Vancouver Family Survey (VFS) used in this study was funded by Health Canada through the National Health Research and Development Program
(NHRDP). The second wave of VFS was funded by a Community Alliance for Health Research grant from the CIHR and a New Emerging Team grant (#RA79917) funded by CIHR (Institutes of Neurosciences, Mental Health and Addiction, Institute of Aboriginal People’s Health, Institute of Human Development, Child and Youth Health), the Canadian Centre on Substance Abuse, and Canadian Tobacco Control Initiative. Resources in support of this study were also provided by the Centre for Addiction Research of British Columbia.
References Anderson, R. E. (2003). Investigating a quantitative measure of addiction-prone personality. ProQuest, UMI Dissertations Publishing. Anderson, R. E., Barnes, G. E., & Murray, R. P. (2011). Psychometric properties and long-term predictive validity of the addiction-prone personality (APP) scale. Personality and Individual Differences, 50(5), 651–656. Anderson, R. E., Barnes, G. E., Patton, D., & Perkins, T. M. (1999). Personality in the development of substance abuse. In I. Mervielde, I. J. Deary, F. De Fruyt, & F. Ostendorf (Eds.). Personal psychology in Europe (Vol. 7, pp. 141–158). The Netherlands: Tilburg University Press. Barnes, G. E., Patton, D., & Marshall, S. K. (1997). Family Environments and Substance Use. Final Report Submitted to NHRDP. Ottawa, Canada. Barnes, G. E., Anderson, R. E., & Jansson, M. (2008). Polysubstance use and mental health symptoms. Paper presented at the 34th alcohol epidemiology symposium of Kettil Bruun Society, Victoria, Canada. Barnes, G. E., Murray, R. P., Patton, D., Bentler, P., & Anderson, R. E. (2000). The addiction-prone personality. New York: Kluwer Academic/Plenum Publishers. Bentler, P. M. (2004). EQS 6.1 for Windows. Encino, CA: Multivariate Software Inc.. Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality development: Stability and change. Annual Review of Psychology, 56(1), 453–484. http://dx.doi.org/ 10.1146/annurev.psych.55.090902.141913. Hicks, B. M., Durbin, C. E., Blonigen, D. M., Iacono, W. G., & McGue, M. (2012). Relationship between personality change and the onset and course of alcohol dependence in young adulthood. Addiction, 107(3), 540–548. Hicks, B. M., Johnson, W., Durbin, C. E., Blonigen, D. M., Iacono, W. G., & McGue, M. (2013). Gene–environment correlation in the development of adolescent substance abuse: Selection effects of child personality and mediation via contextual risk factors. Development and Psychopathology, 25(1), 119–132. http://dx.doi.org/10.1017/S0954579412000946. Kendler, K. S., Sundquist, K., Ohlsson, H., Palmér, K., Maes, H., Winkleby, M. A., et al. (2012). Genetic and familial environmental influences on the risk for drug abuse: A national Swedish adoption study. Archives of General Psychiatry, 69(7), 690–697. http://dx.doi.org/10.1001/archgenpsychiatry.2011.2112. Krank, M., Stewart, S. H., O’Connor, R., Woicik, P. B., Wall, A., & Conrod, P. J. (2011). Structural, concurrent, and predictive validity of the substance use risk profile scale in early adolescence. Addictive Behaviors, 36(1), 37–46. http://dx.doi.org/ 10.1016/j.addbeh.2010.08.010. Lackner, N., Unterrainer, H., & Neubauer, A. C. (2013). Differences in big five personality traits between alcohol and polydrug abusers: Implications for treatment in the therapeutic community. International Journal of Mental Health and Addiction, 11(6), 682–692. http://dx.doi.org/10.1007/s11469-013-9445-2. Malmberg, M., Kleinjan, M., Vermulst, A. A., Overbeek, G., Monshouwer, K., Lammers, J., et al. (2012). Do substance use risk personality dimensions predict the onset of substance use in early adolescence? A variable- and personcentered approach. Journal of Youth and Adolescence, 41(11), 1512–1525. http:// dx.doi.org/10.1007/s10964-012-9775-6. Olson, D. H., & Tiesel, J. (1991). Faces II: Linear scoring and interpretation. St. Paul, MN: University of Minnesota. Oppenheimer, C., Hankin, B., Jenness, J., Young, J., & Smolen, A. (2013). Observed positive parenting behaviors and youth genotype: Evidence for gene– environment correlations and moderation by parent personality traits. Development and Psychopathology, 25(1), 175–191. http://dx.doi.org/10.1017/ S0954579412000983. Parker, G., Tupling, H., & Brown, L. B. (1979). A parental bonding instrument. British Journal of Medical Psychology, 52(1), 1–10. Prinzie, P., Dekovic, M., van den Akker, A., de Haan, A., Stoltz, S., & Hendriks, A. (2012). Fathers’ personality and its interaction with children’s personality as predictors of perceived parenting behavior six years later. Personality and Individual Differences, 52(2), 183–189. http://dx.doi.org/10.1016/ j.paid.2011.10.012. Quinn, P., & Harden, K. (2013). Differential changes in impulsivity and sensation seeking and the escalation of substance use from adolescence to early adulthood. Development and Psychopathology, 25(1), 223–239. http:// dx.doi.org/10.1017/S0954579412000284. Saucier, G., Wilson, K. R., & Warka, J. (2007). The structure of retrospective accounts of family environments: Related to the structure of personality attributes. Journal of Personality Assessment, 88(3), 295–308. http://dx.doi.org/10.1080/ 00223890701317012. Scarr, S., & McCartney, K. (1983). How people make their own environments: A theory of genotype ? environment effects. Blackwell Publishing. http://dx.doi.org/ 10.2307/1129703.
N. Franco Cea, G.E. Barnes / Personality and Individual Differences 82 (2015) 107–113 Scarr, S., Webber, P. L., Weinberg, R. A., & Wittig, M. A. (1981). Personality resemblance among adolescents and their parents in biologically related and adoptive families. Journal of Personality and Social Psychology, 40(5), 885–898. http://dx.doi.org/10.1037/0022-3514.40.5.885. Schofield, T. J., Conger, R. D., Donnellan, M. B., Jochem, R., Widaman, K. F., & Conger, K. J. (2012). Parent personality and positive parenting as predictors of positive
113
adolescent personality development over time. Merrill-Palmer Quarterly, 58(2), 255–283. http://dx.doi.org/10.1353/mpq.2012.0008. Viswesvaran, C., & Ones, D. S. (1999). Meta-analysis of fakability estimates: Implications for personality measurement. Educational and Psychological Measurement, 59, 197–210.