The Applied Psychology of Addictive Orientations | Chapter 9

In this chapter pertinent information from the previous review chapter will be summarised in order to introduce results linking the four factor solutions from previous chapters with personality characteristics. First, inherent problems with the attempt to link personality and addictive orientation will be highlighted. Following this, the possibility of the addictive orientations having a relationship to physiological correlates will be discussed, then the possible role of Anti-social personality disorder will be addressed. Finally, drawing upon the previous chapter, hypotheses will be generated which will attempt to predict which personality variables may be associated with the four factors.

A number of inherent problems detract from a simple cause and effect relationship between personality and addiction. The first problem involves the lack of predictive studies that link the existence of personality as assessed prior to the onset of addiction. In the present study the problem of causality cannot be addressed. Nonetheless, investigating patterns of personality in relation to the found addictive orientation is still a valid pursuit, and the studies which have been reviewed in the previous chapter will contribute to a base line level of expectation.

Second, great variability in personality characteristics has been found across the different forms of addictive behaviour. Looking at the predictive studies in conjunction with the correlational studies for different addictive areas it was noted in the previous chapter that a wide range of personality variables have been implicated. Summarising some conceptual similarities from these areas will help to contribute to the predictions in the present study.

An area of particular interest from the review is the possibility of a relationship between a physiological model of addiction (Eysenck, 1997) and Khantzian’s (1985) idea of self-medication. Both theoretical positions argue that addictive behaviour serves a useful function for the individual, and the nature of the functions they fulfil is related to the personality profile of the addict. Eysenck (op cit) proposed that those who are neurotic tend to be highly reactive to environmental and psychological stimuli. This it is thought results in individuals who are typically easily startled and shaken. In contrast, he proposed that individuals who are extrovert are not as reactive to stimuli, and that this characteristic requires extroverts to create their own excitement. Khantzian (1985) argued that addicts use their addiction as a form of self-medication to combat and cope with their psychological problems and that the drug of choice is not random but rather specific to their psychopathologies.

These analogous views may be taken to suggest that the baseline activation level of individuals who are orientated towards the more sensation-seeking hedonistic behaviours (contained in the Power related hedonism factor and the Sensation seeking hedonism factors) are low, and occurring in extraverted individuals. If this is the case then, possibly in conjunction with some form of psychological problem, this low baseline activity serves to orient them towards certain behaviours in order to satisfy or self medicate an internal state.

By contrast those orientated toward the Nurturant behaviours (Self orientated and Other orientated nurturance), maybe be more likely better characterised by neuroticism, as their baseline level of activation requires behaviours which soothe an already reactive and aroused state. So in this case this level of activation, possibly in conjunction with some form of distress or psychological problem, channels them towards these more soothing behavioural constellations.

In the review of the literature on alcohol’s effects on emotional states (Sher, 1987) and other factors (Meusler et al, 1998) there seems to be plenty evidence which suggests that alcohol and drugs alleviate a variety of different problems. This rather contradicts findings reported in previous chapters that alcohol didn’t contribute strongly to the drug-dominated factor, but contributed more weakly to a number of factors. In relation to personality it seems unlikely that alcoholism and alcohol use is straightforwardly related to personality. Evidence which supports this position includes the number of attempts which have been made to sub-type alcoholism (e.g. Cloninger, 1987, Babor et al, 1992); it seems that there are a number of different types of alcoholics, possibly reflecting different personality orientations.

Anti-social personality disorder has been a popular avenue of investigation, and much research has shown that this personality disorder is strongly related to both alcohol and drug addiction (Regier et al, 1990). Of particular interest are the personality dimensions which have been found to be linked with Anti-social personality disorder. These include neuroticism, impulsivity and extraversion (Sher & Trull, 1994). The unifying themes across these conditions include deficits in inhibition and excesses in rule breaking or norm violating behaviour. Individuals with Anti-social personality disorder have also been shown to exhibit deficits in delay of gratification, to engage in frequent sensation seeking behaviour and to be vulnerable to boredom (Quaym, 1965). Interestingly, these dimensions have also attracted attention in the alcohol literature. In the same way as with Anti-social personality disorder, impulsivity/disinhibition appears to be most relevant to alcoholism (see section on alcohol later in this chapter).

Interestingly, it has been shown that childhood conduct disorder is often a precursor of later Anti-social personality disorder (Muesler et al, 1998). As this evidence is prospective this gives additional weight to the argument that Anti-social personality disorder may be a contributing factor for addiction and most importantly that it is a factor which may precede addiction. In relation to the present study, the personality characteristics which are associated with Anti-social personality disorder suggest that the extraversion and related anti-social characteristics will be related to the hedonistic factors.

As this was a clinical sample, and Neuroticism is a personality dimension which has been found to be associated with many addictive behaviors, it was predicted that Neuroticism would be positively associated with all of the factors. In particular, it was expected that there would be a particularly strong association with the Self orientated nurturance factor (the “eating disorder” factor), in line with results from previous research.

In the review one of the main findings suggested that personality variables of Neuroticism and Psychoticism are positively related to drug abuse and across a large number of studies these findings are quite robust (Eysenck & Eysenck, 1991). Substance dependence has also been found to be associated with elevated levels of negative emotionality and depressed levels of constraint relative to controls (Krueger, Caspi, Moffitt, Silva & Mcgee, 1996 in McGue, Slutske & Iacono, 1999). In the previous chapter it was mentioned that the link between Anti-social personality disorder and drug and alcohol addiction has been a popular avenue of investigation, as much research has shown that this personality disorder is strongly related to both these addictions (Regier et al, 1990). For instance a study comparing those with a substance use disorder with those without, found that rates of conduct disorder, Anti-social personality disorder and nicotine dependence were consistently higher among those having a substance use disorder than among those not (McGue, Slutke & Iacono, 1999). This particular study seems to relate to one of the findings from the factor analyses, this being the strong association of nicotine with drugs.

In the previous chapter Sensation seeking was found to be another widely studied personality correlate of drug use disorders, with the finding that indicators of this personality characteristic tend to be associated with drug use disorder status (McGue, Slutske & Iacono 1999). In conjunction to this Gambling has also been described as a form of risk taking or sensation seeking behaviour (see chapter two). Pathological gamblers are known to be high sensation seekers, scoring significantly higher than social gamblers on the Disinhibition, Boredom Susceptibility, and Experience Seeking sub scales of Zuckerman’s Sensation Seeking Scales (Dickerson, Hincy, & Faber, 1987). What is also pertinent to this area is that males are more likely to be gamblers and sensation seekers than females (Spunt, Lesieur, Liberty & Hunt, 1996).

Interestingly work has been conducted where links have been made between sensation seeking behaviour, gambling and drinking (e.g. Zuckerman’s 1994, in Gupta & Derevensky, 1998). These findings from Zuckerman (op cit) suggest that individuals with higher sensation seeking levels more frequently engage in gambling activities and are more likely to engage in high risk activities such as drinking and reckless driving.

If these findings are taken together what may be suggested is that personality variables which may be related to Anti-social personality disorder and conduct disorder, notably and especially Conscientiousness, and also Agreeableness, may be related to the Sensation Seeking Hedonism factor. In the present study this leads to the prediction that the personality variables Agreeableness and especially Conscientiousness will be negatively related to the Sensation seeking Hedonism factor.

As more males than females are diagnosed with Anti-social personality disorder it was thought that the representation of personality traits in the male and female samples may differ. Ball (1995) in his study found that cocaine misusing women scored higher than men on Impulsivity/sensation seeking, and Neuroticism/anxiety dimensions. This, he suggested, indicated greater sociopathic traits in addicted women. In sum, looking at the evidence it seems both that men are more likely on the whole to be diagnosed with drug and alcohol addiction, and that Anti social personality disorder is related to these diagnoses. However, the women who are involved in these behaviors seem to exhibit a more extreme manifestation of the personality characteristics. It would seem then that differences in personality according to gender would be expected.

So, considering material from the previous chapter and looking at this selection of representative studies for this orientation it was expected not only that the personality dimensions of Agreeableness and Conscientiousness would be negatively associated with Sensation seeking hedonism for both males and females but also that conscientiousness may be more relevant for men. In addition to this it was also predicted that the presence of Neuroticism would be associated with this factor and this association may be more important for women.

Negative conscientiousness (especially for men)
Negative agreeableness
Neuroticism (especially for women)

There is ongoing discussion around the inclusion of relationship addiction being a bona fide addictive behaviour to include under the term addiction. As a result of this, research investigating the personality characteristics of this area can be seen as in part exploratory. It can be suggested that the use of relationships in an addictive fashion is predominantly to do with control, manipulation and the use of others in order to achieve satisfaction for the self. This being the case, Dominance and negative Agreeableness seemed to be appropriate predictions. In a similar fashion as sex addiction seems also to be to do with power and the use of others for personal gain it seems possible that dominance may be the main predictor for this factor. As this factor is closely related to the previously discussed Hedonism factor, it was thought that negative conscientiousness may also be associated.

As the majority of addiction research has concentrated on alcohol and drug abuse the available research for this area is a little thin on the ground. Cermark (1986) has outlined a number of characteristics typically associated with “co-dependency”, in particular, anxiety in periods of intimacy or separation, neurotic symptoms and taking responsibility of meeting other people’s needs at the expense of their own. More recently co-dependency has been found to be related to self defeating personality characteristics (Wells, Glickauf-Hughes & Bruss, 1998).

It can reasonably be suggested that this constellation of behaviours involves gaining satisfaction through the pursuit of attending to other’s needs. As this factor contains Work in addition to the Compulsive helping dimensions it seemed logical that a positive involvement in Agreeableness and Conscientiousness would be expected. However, given the association of obsessive compulsive traits (Seybold, Salomone, 1994), perfectionism (Spence & Robbins, 1992) and Type A personality (coronary prone) (Robinson, 1996 & 1999), the prediction of Neuroticism being associated with this factor was also made. In addition to this it was predicted that there may also be an association with negative Dominance. It is unlikely that individuals who compulsively help would exhibit dominant personality characteristics, as this would contradict their drive to please others at their own expense. This would also be contrary to the findings that co-dependency has been found to be related to self defeating personality characteristics (Wells, Glickauf-Hughes & Bruss, 1998).

In the previous analysis looking at gender differences in the factorial structure it was seen that its composition was the same for both males and females. However, instead of predicting no gender differences for the personality configuration for this factor it was recognised that this factor is better represented by females, and that the expectations would be correspondingly greater for women. This is because the compulsive helping behaviours embrace strong care taking and nurturing of others which reflects a more typically stereotypical female behaviour.

Agreeableness (especially for women)
Conscientiousness (especially for women)
Neuroticism
Negative dominance

Evidence which suggests that personal ineffectiveness and low self-esteem (Mayhew & Edelmann, 1989) combined with perfectionism and obsessive compulsive traits (Fahy, 1991) may be principal personality correlates of eating disorders seems to relate quite well to the more global personality constructs being used in this study. It seemed appropriate to predict Conscientiousness and Neuroticism as being relevant to this set of behaviours as these resonate with the idea of a sense of mastery and need for control, arguably features of the constellation of eating disorders.

Further evidence for these predictions comes from research findings from exercise addiction, as here again a poorly established sense of self and the use of exercise or diet as a means of gaining a sense of control has been found (Yates, 1987) In a similar vein it is thought that the prime motive for obligatory exercisers is self regulation and control of athletic performance, physical appearance or both (op cit). Kagan and Squires (1985) also found that compulsive exercisers are very likely to score towards the extreme end of scales measuring behaviours such as eating control, body dissatisfaction, low self-esteem and self-efficacy. It may also be suggested that low self-esteem, characteristic of eating disorders, reflects a position of negative dominance. Finally, low self-esteem may also predispose those high on this factor towards agreeableness.

As eating disorders are still predominately female disorders it was predicted that there may be differences in personality structure for males and females, as it was thought that the contribution of personality may have something to do with being female. If this addictive orientation is somehow essentially connected to being female, it was predicted that for males that a negative score on dominance may be especially important. As in general it seems that eating disorders are well characterised by the Neuroticism dimension, it was predicted that Neuroticism may play a more important role for females.

Neuroticism (especially for women)
Conscientiousness
Agreeableness
Negative dominance (especially for men)

Personality was assessed by self-report and this data was collected at the same time as the SPQ. The same clinical sample as used in the previous clinical study was used, and again the sample was split into males and females. This was done as in addition to global predictions, differences were predicted for the male and female. The sample in this study is reduced to approximately 350 from the original 543 patients as the personality scores were not available for the entire sample. This was due to a change in questionnaire use at the treatment centre.

The Five factor measure employed in the study examines what are now commonly regarded as the five principal dimensions of personality initially developed by factor-analysis (Cattell, 1946). This line of investigation has been continued and further refined by a number of investigators with recent work being completed by Peabody and Goldberg (1989). The Five factor model (FFM) dimensions are Dominance and Agreeableness, Neuroticism, Conscientiousness and Openness to experience, and has been well developed empirically (Costa & McCrae, 1992). The Five factor model is also thought to be well suited for exploratory research in relatively uncharted domains for which specific hypotheses regarding individual differences variables are lacking (Trapnell & Wiggins, 1990). It has been found that much of the variance of these dimensions is heritable (Heath, Neale, Kessler, Eaves & Kendler, 1992) and therefore can be seen not just as summaries of personality but as genotypic tendencies of individuals to think, act and feel in consistent ways.

Personality questionnaire: The Interpersonal Adjective Scales-Big Five. (IASR-B5)

The IASR-B5 (Trapnell and Wiggins, 1990) is a 124 item adjectival rating questionnaire which combines the circumplex and the Five factor models of personality. This questionnaire was developed in response to the growing recognition that the Dominance and Nurturance dimensions of the circumplex model correspond closely to the Extraversion and Agreeableness dimensions of the Five factor model. So the questionnaire measures the interpersonal circumplex dimensions of Dominance and Nurturance as well as the dimensions of the remaining dimension of the Five-factor, these being Neuroticism, Conscientiousness and Openness to experience. The Five factors may be briefly described as follows.

Openness to experience: intellectual curiosity, awareness of inner feelings, need of variety in actions.

In terms of the validity of the measure, the five scale measure, “Was found to have excellent structure on the item level, internally consistent scales, and promising convergent and discriminant properties” (Trapnell & Wiggins, 1990 p. 781). (Full questionnaire in Appendix 3).Instructions were given and respondents were asked to rate the self-descriptive accuracy of each single adjective on an 8 point Likert scale ranging from extremely inaccurate to extremely accurate.

The major study questions were addressed with regression analysis, with the factor score measures being derived from the previous clinical study.

What was of interest was the investigation of the relationship between the four factors from the previous study, and measures of personality, using the IASR-B5 (Trapnell & Wiggins, 1990). Initially the question to be answered was whether the factor scores could be “explained” by the personality measures, and if so, whether there was a statistical model for this relationship, and what was its level of predictive power.

By design the factors are independent and so a linear model was fitted to each of the 4 factors separately with the 5 personality measures as the regressors. The method of fitting the linear model was that of Least Squares, which produces the best linear unbiased estimators for the regression parameters. The normality assumption was checked through diagnostic plots, which indicated that the procedure was appropriate.

When assessing the fit of a linear regression model one of the most popular statistics to use is the R2 statistic. This measures the proportion of variation explained by the model. This is a general measure of fit for the whole model as it measures the multiple correlation between the dependent variable y and the explanatory variables.

What was of interest was which of the personality measures are useful in explaining and predicting the factor scores. A priori it can be assumed that it is unlikely that all the personality measures are needed to explain the separate factor scores and so some sort of model selection has to take place. There is a number of approaches to use for selecting which explanatory variable to include in the linear model. In essence these methods reduce to comparing the variation explained by competing models. For this particular case where there is a small number of possible regressors the most suitable method is step-wise selection. Stepwise selection works by choosing whether to include or remove variables on the basis of an F-test. The F-test is obtained through comparing two models, one with and one without the variable being tested. The proportion of additional variation explained by including the variable, as measured through the sum of squares, is then tested to see whether it is significant. If it is, then the model with the extra term is kept; if not significant then the model without the extra term is kept.

Basically the F-test compares a model with (p – 1) parameters with a model with p terms. So the F-statistic has 1 and (n – p – 1) degrees of freedom, which for the large number of observations and small number of parameters has approximately the same critical values as z2, the square of the standardised normal

Table 9.1. and 9.2. presents results of all the regression analyses and these are summarised in table 9.3. on page 204.

Table 9.1. Regression analyses of personality predicting factor scores in males.

Male: Factor 1 Self-orientated nurturance.
(Food bingeing, Caffeine, Food starving and Shopping)

Male: Factor 2 Other Orientated Nurturance.
(Compulsive helping submissive, Work, and Compulsive helping dominant)

Male Factor 4 Sensation seeking hedonism.
(Prescription drugs, Drugs and Nicotine)

Table 9.2. Regression analyses of personality predicting factor scores in females.

Female: Factor 1 Sensation seeking hedonism.
(Drugs, Prescription drugs, Nicotine, Gambling and Alcohol)

Female: Factor 2. Other orientated nurturance
(Compulsive helping in both forms and Work)

Female: Factor 3 Self orientated nurturance.
(Food starving, Food bingeing, Exercise, Caffeine)

Female: Factor 4 Power related hedonism.
(Relationships in both forms and Shopping)

Table 9.3. Summary table: Regression analyses of personality predicting factor scores in males and females.

Male factors	Male personality variables and R²	Female factors	Female personality variables and R²
Factor 1 (Both food dimensions, Caffeine, and Shopping)	Dominance (-ve) Agreeableness (-ve) Neuroticism 14.49%	Factor 3 (both food dimensions, caffeine and exercise)	Neuroticism Openness to experience (-ve) 16.66%
Factor 2 (Both Compulsive helping dimensions and Work)	Neuroticism Agreeableness Conscientiousness 9.87%	Factor 2 (both forms of compulsive helping and work)	Agreeableness Conscientiousness 8.39%
Factor 3 (Both Relationship dimensions and Sex)	Neuroticism Agreeableness (-ve) Dominance 22.7%	Factor 4 (Both relationship dimensions and shopping)	Neuroticism Conscientiousness (-ve) Dominance 27.01%
Factor 4 (Prescription drugs, Drugs and Nicotine)	Conscientiousness (-ve) 15.02%	Factor 1 (Prescription drugs, drugs, nicotine, gambling and alcohol)	Conscientiousness (-ve) 6.37%

It can be seen that not all the personality measures seem to influence the factor scores. In particular the Openness to experience doesn’t seem to effect any of the models except for the female Self orientated nurturance factor (food). The R2 values are not exceptionally high, but the included variables are statistically significant at the levels indicated. It is also clear that there is a correspondence between the male and female models.

Regression analyses comparing the factors and the associated personality variables

The first prediction was that Neuroticism would be associated with all of the factors, and that there would be a particularly strong association with the Self orientated nurturance factor (eating disorders). This prediction was partially supported as Neuroticism was found to be present in most of the models, the exceptions being: the male Sensation seeking factor (drug) and for females the Sensation seeking hedonism factor (drug) and Other orientated nurturance (compulsive helping) factors.

However, it was found that Neuroticism though present in both the male and female Self orientated nurturance factor had a particularly strong association with the female population. This association was significant at the 0.1% level and in conjunction with a low score on Openness accounted for 17% of the variance. This appearance of Openness to Experience is particularly interesting, and merits further comment (see section on Self orientated nurturance).

Out of the five personality dimensions a further pivotal finding is the presence or absence of Conscientiousness; with Conscientiousness being positively associated with the “Other Orientated Nurturance” factor and negatively with both the male and female Sensation seeking hedonism factors and the Power related hedonism factor in the female sample.

Examination of the regression analyses primarily reveals that the main predictions are confirmed. In the next section each addictive orientation is taken in turn and examined in order to look at the contribution and relevance of the personality variables for both males and females.

Looking at the Sensation seeking factors for both males and females (Factor 3 for males and Factor 1 for females) and the Power related hedonism factor for females (Factor 4) it can be seen that a negative score on conscientiousness is associated with all three of these factors, thus suggesting that, in general, this personality variable is important in hedonistic behaviours.

One way of interpreting this is that there is a possibility that negative Conscientiousness in some way channels individuals into the use of these clusters of addictive behaviours. Individuals with personality characteristics which contribute to negative conscientiousness may be at increased risk of drug abuse because they are less likely to accept, and less fearful of, the consequences of failing to follow cultural norms governing drug use or concerning sexual conduct. Drug use and abuse, the manipulation of others for personal gain and sexual permissiveness tend to be discouraged by conventional cultural standards and are not considered as acceptable ways of behaving. However, as individuals who exhibit these characteristics can be characterised by the unwillingness or inability to inhibit behavioural impulses, a lack of caution, and a failure to endorse conventional moral standards, it is possible that it is these qualities which increases the likelihood of the hedonistic behaviours being attractive.

It was predicted that in general that the contributing variables for this addictive orientation would be Neuroticism (+ve), Agreeableness (-ve) and Conscientiousness (-ve). In the event, the results indicated that a negative score on Conscientiousness was the only personality variable which predicted these factor scores for both males and females. It was also predicted that Neuroticism may play a more important role for females and that low Conscientiousness would play a more important role for males. For the male sample low Conscientiousness explained 15% of the variance, and for females it explained 6%. So it can be seen that the prediction of low Conscientiousness playing a more important role for males was supported, whereas the prediction of Neuroticism playing a more important role for females was not supported.

It must be noted that there was a substantial difference in how much variance for this addictive orientation was explained by the conscientiousness variable (6% for females and 15% for males). This may be explained perhaps simplistically through differences in gender expectations, in that it is more acceptable for males to be delinquent and to be engaged in these behaviours. However, it must be said that as the percentage of variance explained for both males and females is low it reveals that there are other factors which may provide further explanation for the orientation.

However, a number of previously mentioned studies relate to the finding of negative Conscientiousness. In Krueger, Caspi, Moffitt, Silva & McGee’s study (1996 in McCue, Slutske & Iacono, 1999) it was found that substance dependence was associated with depressed levels of constraint relative to controls. This seems to resonate with negative Conscientiousness in the present study, showing an inability to control actions seems to well illustrate the idea of negative Conscientiousness. A further correspondence can be found in the study by McGue, Slutke & Iacono (1999) where rates of conduct disorder and Anti-social personality disorder were found to be associated with those with a substance use disorder. Again it would be expected that those with conduct disorder and/or Anti-social personality disorder low Conscientiousness would be associated, as this forms an intrinsic part of the characterisation of these disorders.

Dominance (+ve), Neuroticism (+ve), Conscientiousness (-ve), and Agreeableness (-ve), were predicted to explain this factor. These predictions were partly supported, as for both males and females the common personality variables that were found to contribute to this factor in males and females were Dominance (+ve) and Neuroticism (+ve). For both males and females Neuroticism was found to be the best predictor in the models, so the tendency to experience negative effect seems too important in the comprehension of this orientation. However, differences were found between males and females. For instance for males negative Agreeableness played a role for males and not for females. For females negative Conscientiousness contributed to this factor’s prediction and this was not the case for the males. For the females low Agreeableness did not form part of the factor’s explanation.

It can be said that the personality variables were found to be particularly relevant for the both the male and female Power-related hedonism factors, as a substantial amount of the variance - males 23% and females 27% - were explained by the personality variables.

It is difficult to place these findings in the context of previous research as little empirical work has been conducted in the area. Though through this type of analysis the generation of ideas may begin. So, it is possible that for males the orientation towards, and the subsequent use, of Relationships and Sex has something to do with a disregard for other’s feelings as a negative score on Agreeableness perhaps indicates a selfish attitude (i.e. lack of interest in offering emotional support). This feature in addition to being assertive etc. (Dominance) seems to correspond with the underlying significance of this factor, this being the manipulation of others for the individual’s own gain.

For females it seems that if personality in any way drives these behaviors that maybe the key to involvement with this orientation has more to do with the tendency towards being undisciplined (-ve Conscientiousness) in conjunction with being assertive (Dominance).

Co-dependency (compulsive helping) has been found to be related to self defeating personality characteristics (Wells, Glickauf-Hughes & Bruss, 1998), and Cermark, (1986) has outlined a number of characteristics such as anxiety in periods of intimacy or separation, neurotic symptoms and taking responsibility of meeting other people’s needs at the expense of their own.

Conscientiousness and Agreeableness were predicted as being associated with this factor and this prediction has been supported, as both of these personality characteristics appear in both the male and female models. It was also predicted that Neuroticism would contribute to this factor’s explanation. For males this prediction was supported and Neuroticism was found to be the most important variable for this factor’s explanation whereas for females Neuroticism did not feature in the model and the most important personality variable was found to be Agreeableness. Interestingly across all the factors for both males and females positive Conscientiousness only featured in the “Other Orientated Nurturance”. This possibly suggesting that this dimension may help in terms of distinguishing the factors.

It was also predicted that negative Dominance and Neuroticism would contribute to this factor’s explanation. The contribution of negative dominance was not indicated in either the male or female models. However, even though a number of the predictions were supported for this factor at a highly significant level, the percentage of variance which was explained amounted to 8% for the females and 10% for the males, which is quite low. Looking at gender differences it was predicted that for females Conscientiousness and Agreeableness would be particularly important for this factor, however as these variables also appeared in the male model this indicates that the hypothesis was not supported.

With the presence of Agreeableness and Conscientiousness this seems to elaborate, to a certain degree, the meaningfulness of this factor. As the incorporation of personality variables to this constellation of behaviours further illuminates its significance, as the dimensions fit well with descriptions of co-dependency. For example one can see that the drive to meet other people’s needs may be well served by having an agreeable and conscientious disposition. The picture portrayed is that of an individual who is eager to please others through being helpful and supportive and willing to take on responsibility.

Neuroticism, Conscientiousness, Agreeableness and negative Dominance were predicted as being able to explain, in part, this factor and these predictions were only partially supported. It was hypothesised that the best predictor for females for the “Self orientated nurturance” factor (the food factor) would be high Neuroticism, for men it was predicted that it would be (negative) Dominance, and both of these predictions were supported. Conscientiousness was found not to contribute to either populations.

Negative Openness to Experience contributed to the variance explained in the female model, and it is noteworthy that this personality variable did not contribute to any other models. It has been mentioned that eating disorders have traditionally been seen as a predominately female problem. Perhaps the contribution of negative Openness reflects a withdrawal from curiosity and creativity as a consequence of the introspective nature of obsessional thinking and preoccupation associated with eating disorders.

In conjunction to this it has previously been mentioned that it has been proposed that problems are generally internalised in females (Luthar, Cushing & Rounsaville, 1996). Perhaps this internalisation of problems, in part, has a degree of correspondence with females being less involved in the outward looking and expansive.

It was hypothesised that Agreeableness would play a role in this factor, however it was found to be relevant only for the male sample and in a negative fashion i.e. negative score on Agreeableness contributing to the factor’s explanation. Dominance was also found to contribute negatively to the explanation of the males’ factor score and not the females’. In conjunction to this it was also found that Neuroticism contributed to the male model, and so the hypothesis for Neuroticism was supported.

Taking these results and bearing in mind the evidence which indicates that male eating disorder cases are on the increase, what may be indicated is that where there is a combination of males adopting stereotypically female attitudes towards body image and control, and exhibiting personality features which may be judged as more feminine (e.g. negative dominance) this serves to orientate them into the Self orientated nurturance arena.

Taking the above factors and considering them in the light of Eysenck’s (1997) physiological resource model it can be seen quite clearly that the Hedonistic factors with their associated personality variables fit quite succinctly into a pattern of extravert behaviour. If this is the case they may be understood in Eysenck’s (op cit) physiological terms, in other words it is possible that these individuals may be characterised by being less reactive to stimuli and thus in turn needing larger amounts of stimulation to feel satisfied. Although this general Eysenckian level of explanation is supported, the results suggest that it is possible to find personality differences in various groupings of the addictive areas, indicating a further elaboration of personality characteristics serving to orientate individuals to specific addictive areas.

The Nurturant factors also fit closely with Eysenck’s (1997) explanation of Neuroticism. One can see that with the attraction to compulsive helping, work and eating disorders may be more comfortable for those who are perhaps already emotionally aroused and therefore highly reactive. Again, though differences have been found within this group of Nurturant behaviours, it may be the case that the neurotic and extraversion substrates do serve as fundamental building blocks which orientates people into certain directions i.e. outward extravert expression or more neurotic introvert expression. It is possible that it is a later stage of development perhaps influenced through learning or availability that a branching off occurs which leads to specific behaviors or combinations to be used.

A similar pattern can be discerned in Khantzian’s self-medication model (1985), in that the more nurturant behaviours may be related to the desire to satisfy a problem-driven life which occurs in conjunction with a personality characterized by a drive to be conscientious, agreeable and socially acceptable. This is clearly a different pattern from the drug orientated factor which seems to be related to a very different characterological structure. Perhaps the commonality behind the factors is some form of dissatisfaction with the self which becomes manifested in different forms due to the nature of the personality.

It was said earlier that research on Anti-social personality disorder and drug addiction (Regier et al, 1990) has suggested that these disorders tend to co-occur, and the results from the drug factor for males and females relates to this previous finding for the following reasons. The most consistently reported personality dimension of Anti-social personality disorder is impulsivity/disinhibition (Sher & Trull, 1994) together with deficits in delay of gratification etc. (Quay, 1965). It is clear that this constellation of variables fits well with a negative score on Conscientiousness which was found to be the best predictor for the Self orientated hedonism in the present study. It is possible that it is this factor which may be most associated with Anti-social personality disorder, and this would be an interesting further investigation.

Alcohol features only in the female sample contributing a small degree to the “Sensation seeking hedonism” factor, but overall it doesn’t feature highly as it is only a weak contributing addictive dimension for the female sample and in the male sample alcohol doesn’t actually load on any of the factors. It is possible that this occurred because alcohol consumption can be seen as a universal behaviour and therefore is commonly used in conjunction with many different combinations of addictive outlets.

What may be of interest here is previous work looking at the possibility of the presence of more than one type of alcoholic orientation. This may help shed light on the confusing picture that Miller (1976) humorously illustrates in the next sentence. In his review of the personality of alcoholics he states that, “One could conclude from this research that the average alcoholic is a passive, overactive, inhibited, acting-out, withdrawn, gregarious psychopath with a conscience, defending against poor defences as a result of excessive and insufficient mothering” (1976, p.657). This is obviously a tongue in cheek comment aimed at the smorgasbord of descriptors that have been found to be associated with alcoholism. It can be said that from clinical observation that often alcoholics present in many different ways for example suffering from additional problems of drug abuse etc. or involvement with workaholism and compulsive helping. Here looking at dimensions of addiction and the relevance of personality what may be revealed are certain patterns of personality which may be involved in these dimensions.

The idea that there may be two types of alcoholics (Cloninger 1987 & Babor et al, 1992) is of relevance here. The first type is characterised by a relatively late age of alcoholism onset, a low familial loading, and an absence of other anti-sociality, whereas the second is characterised by a relatively early age of alcoholism onset, a high familial loading, and the existence of other anti-social behaviour. Of particular interest in relation to this study is the idea that personality factors are differentially associated with the two forms of alcoholism. The first type of alcoholism is thought to be specifically associated with relatively high levels of negative emotionality, whereas the second type is hypothesised to be associated with relatively high levels of behavioural dis-inhibition (Sher & Trull, 1994).

These descriptors seem to lend themselves to a Hedonistic and Nurturant interpretation of the differing patterns of alcoholism. In the present study it can be seen that even though alcohol did not feature highly in contributing to a factor in the male clinical sample, it came closest to contributing to the Other orientated nurturance factor (compulsive helping) with a loading of .36. This may be seen to fit with the type one alcoholic especially when considering the absence of other anti-sociality which has been associated with this type of alcoholism. In terms of the reported high levels of negative emotionality this factor was best predicted by positive Agreeableness, but this was in conjunction with the presence of Neuroticism. So it could be the case that the combination of activities i.e. working hard and drinking to excess are behaviours used to cover extreme feelings of negative emotionality concern or depression. Bearing in mind the terms mentioned by Miller (1976) it is interesting to note that Nurturant alcoholics may be characterised by a personality which is passive inhibited and withdrawn whereas the hedonistic alcoholic may be overactive, acting out and gregarious.

By way of comparison in the female sample it was found that alcohol made a contribution to the “Sensation seeking hedonism” appearing alongside drugs prescription drugs, nicotine and gambling. This seems to fit with the type two characterisation of alcoholism with its associated existence of other “anti-social” behaviour as drug taking can be considered to be anti-social. Here the only personality variable to contribute to this factor was negative Conscientiousness which has been previously linked to anti-sociality (Lewis & Bucholz, 1991).

Of relevance here are the results from a general population survey of over 2,500 men and women which indicated that serious anti-social behaviour (both conduct disorder and anti-social personality disorder), gender and a family history of problem drinking were all significantly associated with alcoholism. With regards to gender, being male increased the probability of alcoholism (Lewis & Bucholz, 1991). Interestingly it was also indicated that anti-social behaviour was a more powerful risk factor for women than for men. Is this because this behaviour is not typically female, and when engaged in indicates that social mores have been violated? Maybe, but despite this close association with alcoholism, having either anti-social personality or a positive family history of problem drinking identified only 53% of male alcoholics and 41% of female alcoholics. This indicates the importance of further influential factors.

In a study by (McGue, Slutske & Iacono, 1999) support was found for the idea that constraint and negative emotionality mark distinct pathways to alcoholism, but the proposition that these two pathways are primarily distinguished clinically by different levels of general antisocial behaviour received only inconsistent support. It was suggested however that co-morbid drug abuse might be the critical clinical factor differentiating the two types of alcoholism (op cit). Again the pattern here tends to lend itself to a Hedonistic and Nurturant orientation, drug involvement perhaps escalating the level of anti-social behaviour and thus distinguishing two types of alcoholism.

The major limitation of the study is that its design was cross sectional rather than longitudinal, and as a consequence it is not possible to determine whether the personality differences which were observed predated the onset or developed as a consequence of addictive behaviour.

Another limitation is that personality data was not collected from non-addicted participants. This was because the non-addict data for the factor analytic studies was collected over a number of other related investigations and was not standardised for present purposes. If it was found that correlations remained constant across a range of factors and personality in a normal population then this would strengthen the argument for the role of basic personality dimensions in addiction.

Another criticism involves the method of personality assessment. In general, participants found the questionnaire difficult and awkward to complete, finding it repetitive and some of the terms difficult to understand. This may have caused an under-estimation of the association with personality. In addition to this there may be personality characteristics which are associated with these different orientations to addiction that were not measured by this instrument.

Unfortunately much empirical and theoretical work simply inventories the personality characteristics of various addictive groups observed on various traits instead of presenting explanatory models of the disorders, and the interplay between personality and other etiologically relevant variables (Sher & Trull, 1994). In this study the aim was not to establish relationships between diagnosis or individual behaviours and personality, but to look at the possibility of underlying dimensions of addiction having concomitant personality features. There is inherent sense in this approach and given the extent of cross addiction it seems appropriate to look at the possibility of underlying dimensions. It is perhaps a valid point to make that orientation instead of diagnosis is more scientifically suited in the characterising instead of classifying of addicts as one is possibly more likely to get consistent associations with personality variables. In addition to this, the incorporation of personality dimensions in relation to these underlying dimensions may help researchers and practitioners think more clearly about if and how personality contributes in the development and maintenance of the constellation of behaviours used.

Overall, these results provide evidence for the utility of personality in the understanding of addictive orientations. It also extends previous research in the elaboration and characterisation of “addictive orientations” (Stephenson et al, 1995). Consequently one implication from this and earlier research is that the construct of personality should be considered to be of importance in the conceptualisation of addiction. In general, correlations were found between the personality variables and the factors in the directions predicted, and regression analyses indicated that personality made an important contribution to our understanding of the factors in all cases.

This study has helped to explain why there isn’t a single addictive personality, as it seems that from the Five factor model the dimensions seem to contribute to different clusters of addictive behaviour. It has also become evident that most likely culturally determined gender differences exist in the salience both of particular orientations and associated personality dimensions. There may be many motives and layers of explanation behind why compensation, solace or excitement may be desired (in a normal and a pathological way), though it seems that personality may be one of the influential elements which dictate the kind of substance or addictive behaviour that an individual may choose.

However, regarding personality as a causative factor is problematic and that consideration cannot be overlooked. But, considerable data do suggest that personality traits similar to those assessed in this study predispose the later development of substance use problems, deviant and other risky behaviours (Tarter, 1986). However, if personality does not cause repeated use of substances the question is what does, the answer, as with so many questions in psychology, is probably multi-factorial. The list of causative factors is likely to include environmental pressures, learning, and genetic influences amongst others. So it seems that until more evidence is available the claim that addicts are associated with any personality type must be made with caution.

Factors	Beta	t	p	Model R²
Dominance	-0.18	-2.85	< 0.01
Agreeableness	-0.15	-2.22	< 0.05
Neuroticism	0.15	2.08	< 0.05	14.49%

Factors	Beta	t	p	Model R²
Neuroticism	0.17	2.44	< 0.05
Agreeableness	0.17	2.41	< 0.05
Conscientiousness	0.12	2.05	< 0.05	9.87%

Factors	Beta	t	p	Model R²
Neuroticism	0.28	4.24	< 0.001
Agreeableness	-0.24	-3.74	< 0.001
Dominance	0.21	3.51	< 0.001	22.7%

Factors	Beta	t	p	Model R²
Conscientiousness	-0.25	-5.17	< 0.001	15.02%

Factors	Beta	t	p	Model R²
Conscientiousness	-0.18	-3.16	< 0.01	6.37%