The Applied Psychology of Addictive Orientations | Chapter 7

For many years most studies of drug and alcohol addiction used only male subjects, as it was assumed that these findings would also apply to women (Miller, 1997). However, more recent studies comparing women and men have found significant gender differences in epidemiology, clinical presentation, pharmacology and socio-cultural aspects of addiction (op cit.). It is, therefore, important in the present instance to investigate differences in addictive orientation according to gender, in both of the populations investigated so far. First a further look will be taken at the kind of gender differences which have been found previously in those with drug, alcohol and eating problems and in so doing expand on the differences that were noted in chapter three.

Research has shown that male and female substance abusers report different problem histories and pathways (Hodgins, El-Guebaly, Addington, 1997). For example in a study by Weiss, Martinez-Raga, Griffin, Greenfield & Hufford (1997), of cocaine treatments, males were reported as having an increased incidence of anti-social personality disorder whereas females had significantly more severe family and social problems.

Related findings were revealed in a study by Luthar, Cushing & Rounsaville, (1996) who looked at gender differences among opioid addicts. Here, results indicated that female addicts had significantly lower levels of conduct problems and higher levels of “internalising” problems during childhood, while males had significantly higher child conduct problems. From these findings it was suggested that opioid addiction for males may be more related to conduct problems in childhood and for females more related to the internalisation of problems during childhood.

Linking to this, it is well documented that, in general, boys and men are reported as exhibiting more delinquency than women and girls (Mears, Ploeger & Warr, 1998). In order to clarify this there has been a diverse range of explanations for the pervasive finding of greater risk taking among young men. For example the influence of social roles (Berger, 1989), the impact of societal structure and associated patterns of family interaction (Grasmick, Hagan, Blackwell, & Arneklev, 1996) have all been used to discuss these differences found between young men and women. As these are general, pervasive gender differences it is reasonable to propose that they may also impact the choice of addictive behaviour from initial experimentation through to pathological involvement.

There is evidence which corroborates this position, first, with regards to alcohol involvement. The results from a general population survey of over 2,500 men and women looking at factors associated with alcoholism indicated that being male, engaging in serious antisocial behaviour, and having a family history of problem drinking all increased the probability of alcoholism (Lewis & Bucholz, 1991). In conjunction to this in a survey of over 22,000 participants classified as current drinkers, it was found that the average estimated alcohol intake of males exceeded that of females by a factor of 2 (Dawson, 1996).

Turning now to eating disorders the most remarkable difference between men and women is that men are reported as representing only 10% of eating disorder cases (Andersen & Holman, 1997). However, it is thought that there is an increase in incidence in young adult males which is likely to continue (Hartley, 1996). In addition to this a piece of epidemiological research looking at newly diagnosed cases of anorexia and bulimia between 1988 and 1994 fully emphasises the differences between men and women. The relative risks of females to males were calculated as 40:1 for anorexia nervosa and 47:1 for bulimia nervosa (Turnbull, Ward, Treasure, & Jick 1996).

Looking at evidence from the Non-treatment population sheds light on the possible origin of such findings. In a study by Shapiro, Newcomb & Loeb (1997) it was shown that societal directives to be thin are perceived among children, and that the discontentment with body and behaviours, associated with eating disorders, begins before adolescence, and is most prevalent in females. It was concluded from this study that components that may lead to the development of eating disorders or dis-regulated or restrained eating in adolescents may be both internalised and expressed at a very early age (op cit.). It seems then that, early in socio-cultural development of children, gender seems to be an influential factor in the selection of ideal body size and in determining body satisfaction.

It may be the case that due to these influences females are in a way “prepared” or sensitised to be overly concerned with body image and are thus channelled into a range of sub-clinical eating and food related attitudes. In one study investigating the attitudes of young men and women (Tiggemann, Winefield, Winefield, & Goldney, 1994) this idea was partially supported, as it was shown that women tend to view themselves as more overweight than did men regardless of their true weight.

A further study that supports this position investigated the relationship between gender differences in eating patterns among college students (Hesse-Biber, 1989). Results indicated that a considerable number of college women but few men showed behavioural patterns associated with an eating disorder (anorexia or bulimia). These findings implied that the eating disturbances of these women may be eating problems which only partially resemble clinical eating disorders, as although the female subjects displayed the behavioural symptoms associated with anorexia and bulimia they exhibited few of the constellation of psychological traits associated with these disorders (op cit.).

The findings of this study indicated that even though there were many similarities between men and women, there were differences in factor structure. This was taken to indicate that there may be differences in the way that men and women use and are involved in addictive behaviour. The main difference was that Alcohol formed a clear component of the Hedonism factor in the female sample. In this study it was also indicated that women were as Hedonistic as men, but were more Nurturant than men regardless of diagnosis.

These previous findings have important implications for the comprehension and treatment of addicts, and given the evidence it is appropriate to consider what gender differences may be predicted.

First, it is predicted that there will be differences between the male and female Treatment populations’ factorial structure, perhaps with the most pertinent and notable differences observed between the Sensation Seeking Hedonism factor (drug) and the Self Orientated Nurturance factor (food). With males having perhaps a better defined Sensation Seeking Hedonism factor due to the general findings of greater risk taking among young men (Mears, Ploeger & Warr, 1998). So here there may be a difference in how the hedonistic behaviours cluster together, perhaps reflecting the link between problems of conduct in males and the internalisation of problems in females (Luthar, Cushing & Rounsaville, 1996). In support of this expectation Orford (1985) has stated that excessive gambling and sexuality, alcoholism and drug addiction have commonly been thought to be addictions mainly in the male domain. An additional difference may be the role of alcohol as in Stephenson et al’s (1995) study it was found that alcohol played a more “Hedonistic” role for women and a more “Nurturant” role for men, and this may be replicated in this study.

It seems from the literature that the dominance of eating disorders in females may lead to a stronger Self Orientated Nurturance factor, in comparison to the males. Especially with the relative risks of females to males being calculated as 40:1 for anorexia nervosa and 47:1 for bulimia nervosa (Turnbull et al, 1996).

Second, it is predicted that there will be differences between the male and female Non-treatment factorial structure, and, as in the Treatment population, with the most pertinent differences observed between the Sensation Seeking Hedonism factor (drugs) and the Self Orientated Nurturance factor (food). Perhaps with similar differences as found in the Treatment population being observed but to a less marked extent. For instance, concerns with weight and body size are better represented by normal females than males, and behaviours, associated with eating disorders, are more prevalent in normal females (Shapiro et al, 1997).

The analysis will be conducted in three parts. First some preliminary statistics are calculated for the Treatment sample to look at general differences in how males and females score on the individual scales. After this a factor analysis is conducted to see whether there are differences in “addictive orientation” in male and female Treatment populations.

Second, in the same way as for the Treatment sample, preliminary statistics are calculated for the Non-treatment sample to look at general differences in how males and females score on the individual scales. After this a factor analysis is conducted to investigate whether there are differences in addictive orientation in male and females in this population, and whether these patterns are similar to or different from the patterns found in the Treatment population.

Finally a Manova on the combined set of data is performed to see whether there are systematic differences between gender and Treatment/Non-treatment status.

First of all the means and standard deviations were calculated for males and females in the Treatment population, (table 7.1). Following this two sample t-tests were calculated to establish whether males and females score differently from one another on the individual scales of the SPQ.

Table 7.1: Treatment population. Means, medians, standard deviations and t-values for males and females

In general males have higher mean values for Gambling, Drugs, Sex and the “dominant” Relationship dimension. However when looking at the t-tests it can be seen that the difference between males and females on the Alcohol dimension doesn’t quite reach significance (p <0.06). The other behaviours, Gambling, Drugs, and Sex are shown to be highly significantly different from the females scores, in all cases p< 0.001 except for the Drug and the Relationship dominant dimension where the p value was found to be p < 0.01.

Females on the other hand score much more highly on the Food dimensions; Food bingeing (Females = 22.56, Males = 8.8, Shopping (Females = 17.05, Males = 9.44), the Compulsive Helping dimensions, and Relationships submissive. When looking at the t-tests it can be seen that females do score significantly higher than males with Shopping, both the Food dimensions and both the compulsive helping dimensions having a p value of <0.001 Females were also found to be significantly higher than males on the Caffeine and relationship submissive dimensions (p < 0.01).

No significant differences were found in the way that male and female addicts score on the following dimensions. Nicotine, Work, Prescription drugs and Exercise.

The factor analysis was repeated for men and women and the analysis suggested a four factor solution in both cases. The factor loadings and the communality values given in tables 7.2. and 7.3 show that these models explain most of the variability for the variables except, as seen before, the Gambling, Exercise and Alcohol scales. These findings support the findings from Stephenson et al’s (1995) study where the role of Alcohol was found to be different in men and women, a Hedonistic behaviour for woman and more of a Nurturant role for men.

Table 7.2: Treatment population. Rotated factor loadings for the female Treatment population

Table 7.3: Treatment population. Rotated factor loadings for the male Treatment population

As in the previous “Treatment” analysis (chapter five) the components of this factor for the male Treatment group are the behaviours which are associated with food intake and body image etc. these being both the Food dimensions, Shopping and Caffeine. This result matches perfectly with the cumulative analysis of the Treatment group. This however is not the case for the females as the first factor includes Drugs, Nicotine, Gambling and Alcohol, while the Food disorder dimensions are included in Factor 3.

This factor for both males and females comprises of both the Compulsive helping dimensions and Work.

The composition of this factor for the female sample can be seen as a cluster of addictive areas that are associated predominately with involvement or concern with eating regulation and body size/shape. For males this factor is made up of both the Relationship dimensions and Sex.

For the male sample this factor contains Prescription drugs, Drugs and Nicotine. For females the factor is made up of both the Relationship dimensions and Sex.

First of all the means and standard deviations were calculated for males and females, (table 7.4). Following this two sample t-tests were calculated to see whether males and females score significantly different from one another on the individual scales of the SPQ.

Table 7.4: Non-Treatment population. Means, medians, standard deviations and t-values for males and females

In general males have higher mean values for Nicotine, Work, Alcohol, Gambling, Drugs, Sex and Relationship dominant. When looking at the results of the t-tests however it can be seen that males are significantly higher than females on the following dimensions Alcohol, Gambling, Drugs, Sex (all at p <0.001) and Work (p< 0.05).

Females on the other hand have higher mean values for Shopping, Food bingeing, Food starving, both Compulsive helping dimensions and Relationship submissive. Again when looking at the significance of these differences Shopping (p<0.001) both the food dimensions (p<0.001) Compulsive helping submissive (p< 0.05) Compulsive helping dominant (p< 0.01) are found to be significantly different from the males scores. Even though females score more highly on Relationship submissive this failed to reach significance. The scores between males and females on the following dimensions were also found not be significantly different Nicotine, Relationships dominant, Caffeine, Prescription Drugs and Exercise.

Comparison of the gender differences found in the Non-treatment and Treatment population

If the Non-treatment means are compared to the Treatment means it can be seen that in general across the behaviours that the Treatment group achieve higher scores than the Non-treatment group. However, in particular Drugs and Prescription drugs are nearly triple in the Treatment population for both males and females.

Drugs Female Treatment = 16.7, Female Non-treatment = 3.1, Male Treatment = 21.8, Male Non-treatment = 8.1.

Prescription drugs Female Treatment = 12.6, Female Non-treatment = 3.0, Male Treatment = 11.8, Male Non-treatment = 2.7.

Substantial differences between the females in the two samples can be seen in the Food dimensions

Food bingeing Female Treatment = 22.5, Female Non-treatment = 14.2, Male Treatment = 8.8, Male Non-treatment = 8.1.

Food Starving Female Treatment = 17.9, Female Non-treatment = 9.4, Male Treatment = 7.5, Male Non-treatment = 5.5.

Of interest was the similarity in the location of the gender differences between the Treatment and Non-treatment populations. With males scoring higher than females on Alcohol, Drugs, Gambling, Relationship dominant and Sex. Females on the other hand scored higher than the males on Prescription drugs, Shopping, both Food dimensions, both the Compulsive helping dimensions and Relationship submissive.

As differences in factorial structure were found between men and women in the Treatment population it was thought that the factor structure may also vary in the Non-treatment population. So the factor analysis was repeated and the analysis again suggested a four factor solution in the case of both men and women. Having accepted a four factor model as the most suitable one for this data, the obtained results from the analysis of the male and female samples are now detailed and comparisons drawn with the previous analysis.

Table 7.5: Non-Treatment population. Rotated factor loadings for the female Non-treatment population

Table 7.6: Non-treatment population. Rotated factor loadings for the male Non-treatment population

If 0.5 is used as the threshold for the inclusion of each variable into a factor, a number of the variables do not reach this threshold and have relatively small loadings, such as Sex, Alcohol, Gambling, Caffeine and Prescription drugs in the female sample. However in this sample some of the groupings reflect the analysis from the Treatment sample. For example both the Food disorder dimensions appear with Exercise, but Shopping in this case is grouped with the Relationship dimensions.

Comparing the factors from the Treatment and Non-treatment populations, across gender, there is a correspondence between Factors 2 for the four groups (male - female and Non-treatment - Treatment), this being the presence of the Compulsive helping factors and Work. There is also a correspondence between Factor 4 of the Male Non-treatment and Female Non-treatment groups, with Nicotine and Drugs being present in both.

There is a correspondence in Factor 1 between the two male groups with both the Food dimensions and Shopping being present. In the male Treatment group this cluster was also joined by Caffeine and in the Non-treatment male group the cluster was joined by Prescription drugs. Again comparing the two male groups there is a high correspondence between Factors 3 with both the Relationship dimensions and Sex clustering together.

Table 7.4: Non-Treatment population. Means, medians, standard deviations and t-values for males and females

Comparison of the gender differences found in the Non-treatment and Treatment population

Drugs Female Treatment = 16.7, Female Non-treatment = 3.1, Male Treatment = 21.8, Male Non-treatment = 8.1.

Prescription drugs Female Treatment = 12.6, Female Non-treatment = 3.0, Male Treatment = 11.8, Male Non-treatment = 2.7.

Substantial differences between the females in the two samples can be seen in the Food dimensions

Food bingeing Female Treatment = 22.5, Female Non-treatment = 14.2, Male Treatment = 8.8, Male Non-treatment = 8.1.

Food Starving Female Treatment = 17.9, Female Non-treatment = 9.4, Male Treatment = 7.5, Male Non-treatment = 5.5.

Table 7.5: Non-Treatment population. Rotated factor loadings for the female Non-treatment population

Table 7.6: Non-treatment population. Rotated factor loadings for the male Non-treatment population

In order to obtain a more precise evaluation of the interaction between gender and Treatment/Non-treatment status, it was decided to test formally whether there is a difference between Treatment group (i.e. addicts) and gender, using multivariate analysis of variance. To perform this multivariate analysis of variance it was necessary to combine the two samples (Non-treatment and Treatment), fit a factor model to the combined data and then to formally examine the effects of treatment status and gender. In the combined data there were 1051 observations, each one classified as Male or Female and as Treatment or Non-treatment. Performing the Factor Analysis in the usual manner the following loadings were produced (see table 7.7. on the next page).

Table 7.7: Combined populations. Rotated factor loadings

It was anticipated that the same four factors would emerge as when the analysis was done separately. This is in fact the case with the factors main contributors listed in order below:

Factor 1: Both Food dimensions, Shopping and Caffeine.
Factor 2: Both Compulsive Helping dimensions and Work.
Factor 3: Drugs, Prescription drugs, Nicotine and Alcohol.
Factor 4: Both Relationships dimensions and Sex.

The four factors correspond to the factors which emerged before the merger the two populations. The first factor explains nearly 16% of the total variance, the second factor 14% the third, 12.6% and factor 4 explains 6%. Total =

This particular analysis can be considered as testing four different classifications: Male Treatment, Female Treatment, Male Non-treatment and Female Non-treatment. However, this is an unbalanced design as there are a different number of observations for each “treatment”, and so the Sums of Squares needed in the analysis are constructed in two different ways to ensure correct inference. The two methods used are:

Type 1 SS – lists the Sum of Squares in the order they were entered into the model

Type 2 SS – lists the Sum of Squares of each variable as if they were entered last into the model

Both these methods produced consistent answers and so no further reference is made to method of calculation. There are four “treatments” and so three contrasts are constructed to test the differences. The model we are testing is the two factorial experiment with interaction.

In the multivariate sense (testing all four factors together) the following test statistics were obtained:

Table 7.8. Wilk’s Lambda test results for significance of gender, Treatment/Non- treatment and interaction on factor scores.

These statistics indicate that all three terms are highly significant. Clearly we would expect GENDER and TREATMENT/NON-TREATMENT to affect the four factor scores, but what is not so expected is the significance of the interaction or GENDER*TREATMENT/NON-TREATMENT term. Without an interaction it would be expected that the differences between male and female Treatments be the same as the differences between male and female Non-treatment group, but the test suggests that this is not the case. This either means the difference between Men and Women is greater for the Treatment group than the Non-treatment group or vice-versa. To investigate this interesting interaction further, a univariate Anova was conducted on each factor separately. Only Factor 1 (F=26.66) proved to be highly significant but Factor 4 also produced a reasonably high F-statistic with a significance level of 11%. These interactions are best represented graphically as follows:

Figure 7.1: Graph illustrating the interaction of Factor 1 (Self Orientated Nurturance “food”)

If there was no interaction for Factor 1 then we would expect these lines to be parallel. However we can see from the graph how the two male scores do not differ, whereas the Treatment females show a hefty increase from the Non-treatment baseline. Non-treatment females, it should also be observed, score much more highly than the Non-treatment men. It appears that with pathological involvement the female addicts build upon an already high baseline level of involvement in this cluster of behaviours.

Figure 7.2: Graph illustrating the interaction of Factor 4 (Power related hedonism “Relationships”)

Here we can observe the same pattern, but in reverse. Treatment and Non-treatment females differ little on this factor, whereas men have a high baseline level, which sharply increases in the addicted population. Again, we may suggest that an existing baseline “problem”, in this case concerning the exploitative use of power, may become acutely problematic for that gender in particular.

The results of the other two factors can be presented just in terms of Type or Gender, as there is no significant added effect (interaction) See figures 7.3 and 7.4. on the next page.

Figure 7.3: Graph illustrating the interaction of Factor 2 (Other Orientated Nurturance “Compulsive helping”)

Figure 7.4: Graph illustrating the interaction of Factor 3 (Sensation seeking hedonism “Drugs”)

First it was established that, in general, there were differences in how Treatment males and females score on the individual scales of SPQ. Following on from this it was predicted that there would be differences in factorial structure between males and females and this was found to be the case, although not dramatically so. In general the same clusters of behaviours appeared across the male and female samples (e.g. Prescription drugs, Drugs and Nicotine appear in both samples for females in Factor 1 and for males in Factor 4).

However it was also predicted that the most notable differences would be between males and females in the Sensation Seeking Hedonism factor (drugs) and the Self Orientated Nurturance factor (food). Looking at the factorial breakdown, the main difference can be seen to be located in the Sensation Seeking Hedonism factor, as this factor appears to be “better” represented in females than it is in the male population. The factor seems stronger as it is comprised of Drugs, Prescription drugs, Nicotine, Gambling and Alcohol, thus containing more of what are generally perceived as hedonistic behaviours than does the male Sensation Seeking Hedonism factor. With the Self Orientated Nurturance factor (food) both the Food dimensions and Caffeine are present in the factorial structures for both the male and female populations. However this constellation of behaviours is joined by Exercise in the female sample, whilst it is joined by Shopping in the male sample, so comparing the male and female samples on this, it can be seen that they are actually quite similar.

First it was established that, in general, there were differences in how Non-treatment males and females score on the individual scales of SPQ. Following on from this it was predicted that there would be a difference in factorial structure between males and females in the Non-treatment population, and in the same way as for the Treatment population this prediction is only partially supported. Even though slight differences were found it was noted that in general the same clusters of behaviours appeared across the male and female samples (e.g. both forms of compulsive helping and work appear in both samples for males and females in Factor 3).

It was also predicted that the most notable differences would be between males and females in the Sensation Seeking Hedonism factor (drug) and the Self Orientated Nurturance factor (food). Looking at the factorial breakdown the main difference can be seen to be located in the Self Orientated Nurturance factor (food), as this factor contains additional behaviours in the male population, and so is comprised of both the Food dimensions (same as the females) Shopping and Prescription drugs. With the Sensation Seeking Hedonism factor (Drugs) this factor contains the same behaviours in both the populations namely Nicotine and Drugs.

In terms of correspondence between the Non-treatment and Treatment populations it was seen that even though differences were found, in general there was a high degree of similarity in terms of the factorial composition. The factor that was most similar was Factor 2 (Other Orientated Nurturance), but of course women score more highly than men on the factor.

Combining the data from Treatment and Non-treatment populations, it was found that differences in factor scores for gender depend on Treatment status. What was found was that the main effects of gender, and Treatment/Non-treatment status, and of the interaction were highly significant (p<.001). Without this interaction it would be expected that the differences between males and females from the Treatment group to be the same as the differences between the males and females from the Non-treatment group. This was not the case. In investigating the interaction further it was found that only the interaction for the Self Orientated Nurturance factor (food) proved to be highly significant. Looking at the interaction graphically it was seen that Treatment women score much higher than would be expected from the Non-treatment baseline score. A similar effect in reverse was noted with respect to Factor 4 (Exploitative Relationships), with male scores rising disproportionately in the Treatment group. It is suggested that a prevailing problem in each case (Food for women, and Power for men) is reaching pathological proportions predominantly in the respective gender group.

Though there are differences it must be pointed out that across the four samples (male Treatment, female Treatment, male Non-treatment, female Non-treatment) there are a number of stable dimensions which cluster together.

Substance factor: Drugs, Nicotine and Prescription drugs.
Relationship factor: Relationship dimensions, and Sex.
Food factor: Food bingeing, Food starving and Caffeine.
Helping factor: Compulsive Helping dimensions and Work.

Another way of looking at the similarities and the differences between these populations is to compare Non-treatment females with Treatment females and then to compare Non-treatment males with Treatment males. It can be seen below that even though there are differences in the factorial structures, there are a number of areas which cluster together across the samples.

Substance factor: Drugs, Nicotine and Alcohol (doesn’t reach threshold for inclusion but this behaviour came closest to contributing to this factor).

To make sense of the similarities and differences found between the populations each factor will be taken in turn and the Non-treatment factorial compositions will be examined to see whether a meaningful rationale for the various clusters of behaviours can be established. This is because these behaviours are used routinely by people who are not suffering from addiction, and it seems logical to propose that there may be a common drive or function that underlies the clustering of these areas.

This achieved, the Non-treatment structure may be seen to form a base line of addictive orientation, or in other words, areas which are commonly used together in a non-addicted fashion. If attention is paid to these “basic core structures”, this enables the characteristic orientation which occurs in the Non-treatment population to be seen as a foundation for subsequent addictive use.

There are a number of reasons why certain behaviours may occur together and why differences may be found between genders. For instance an individual’s gender and associated set of gender-role expectations may exert a dominating influence on one’s sense of self in terms of what are viewed as acceptable behaviours to be engaged in. The discussion will center on the possibility of the differing experiences of the two genders contributing to the orientation towards certain behaviours. Basically an argument will be made that connects males’ substance use with conduct disorder, delinquency and social roles, as their involvement in behaviours in Factor 3 (Substances) and Factor 4 (Relationships) are already primed. For females an argument will be made which links being female and a propensity to eating disordered behaviour, via a pre-existing orientation to the behaviours characteristic of Factor 1 (Food).

The core of this factor seems to originate with the association of Nicotine and Drugs, as it is these two areas which are found to co-occur in both the Non-treatment male and female populations. This association has been revealed in previous research. In a study by Von Knorring & Oreland (1985) it was found that regular smokers were found to be more prone to the abuse of alcohol, and to a variety of drugs such as glue, cannabis, amphetamines and morphine. In another study investigating the prevalence and use of alcohol, nicotine and other drugs in a family practice setting, a definite overlap between nicotine and other substance use was found (Mar, Johnson Pistorello, Rigmaiden & Veach, 1995). These previous studies in conjunction with the present results suggest that there may be a general orientation to these addictive substances. This may be taken to mean that if one is engaged in first, most likely nicotine, then this increases the likelihood of other substances being used.

However there are differences between males and females scoring on this dimension. For instance comparing the male and female Non-treatment Drug scores, males score significantly higher on average than females (Males = 8.1; Females score 3.1 (p <0.001). It is possible that these addictive behaviours are better represented in males as perhaps the uptake and involvement in these behaviours is in part determined by a difference in gender roles and socialisation, with these areas being more acceptable in males and less tolerated in females. Evidence which in part supports this position comes from research showing that men, as compared with women, are more likely to engage in risk behaviours (Byrnes, Miller, & Schafer, 1999).

Looking at further evidence to support this position it was mentioned in the introduction that it is well documented that, in general, boys and men are reported as exhibiting more delinquency (Mears, Ploeger, & Warr, 1998) and drug taking may be considered as delinquent behaviour. In a study investigating the factors associated with aggressive and delinquent behaviours in adolescents, it was found that adolescents who engage in these behaviours are more likely to use alcohol and other drugs, to engage in sexual risk behaviours, have problems with peer and family relationships and report more mental health symptoms (Durant, Knight, & Goodman, 1997).

Further supporting evidence for this position includes Gilligan’s (1982) suggestion that females in general are more socialised, in such a way that they become more subject to moral pressures. If this is the case then females will be more reluctant than males to engage in behaviours that would attract criticism and negative evaluations from others.

Another hypothesis proposes that gender differences in health behaviour are due primarily to two complementary aspects of gender roles. First, males’ greater risk taking and females’ greater health concerns (Waldron, 1997). This hypothesis proposes that males are socialised to take risks whereas females are socialised to be more cautious and more concerned with protecting health. Consequently males engage in more risky behaviour while females engage in more preventative health behaviour (op cit.).

Men are consistently higher on this dimension than are females (p<0.01) although the increase for females is proportionately higher than it is for men: Non-treatment male; 8.1. Treatment male; 21.8. Female Non-treatment; 3.1. Treatment female; 16.7.

There is an interesting difference between the populations in the role of Prescription drugs in this factor. What seems to occur is that when an individual moves into the realm of problematic usage, another behaviour which becomes associated is Prescription drugs, where again the difference between the male Non-treatment score and the Treatment score is great: Non-treatment male score; 2.7 and Treatment male score; 11.8. This leap in score in the Treatment group perhaps suggests that Prescription drugs are not typically used to alter mood and are not likely to be abused unless an individual is already in the realms of addictive use. This is perhaps why this behaviour does not load on Sensation Seeking Hedonism for the Non-treatment males. When drugs are used to excess it seems that Prescription drugs take a more important role, as this score is almost four times higher in the addicted population than in the Non-treatment population: (Male Treatment; 11.8. male Non-treatment; 2.7. Female Treatment; 12.6 Female Non-Treatment: 3.0).

What is particularly interesting with the female Treatment group is that there is a strong and well defined cluster of Hedonistic behaviours, which again seems to based on the core components of Nicotine and Drugs which appear in the Non-treatment female sample. Perhaps then this is not the most obvious route of excessive behaviour for females to be engaged in. However, when involved, it seems that their use escalates. This is apparent initially when looking at the composition of the Treatment population as there are many more males in treatment for drug and alcohol addiction.

Looking at the drug scores, the average female Non-treatment drug score is; 3.1 and the female Treatment score is; 16.7. This Treatment score is therefore five times greater than in the Non-treatment population (the difference in magnitude between the male Treatment and the male Non-treatment is under three). This may suggest that when a female becomes involved in such areas the involvement in some ways is more marked and extreme in conjunction with being more unusual. In addition to this it can be seen that, in general, the female’s scores on this factor are lower than the males (apart from Prescription drugs which is slightly higher).

Support for these ideas stems from Heimer (1996, p.42), as when commenting on delinquency in men and women, female delinquency was termed as different from male delinquency as it is, “Doubly deviant”. This is because it is inconsistent with societal gender role expectations as well as with laws and codes of conduct. This is because inappropriate role behaviour, which would include the behaviours on this factor may elicit more negative and perhaps more extreme societal judgement in females than in men. It would be interesting to look at these women as it may be found that their personality traits may have some features of male gender identity, which serves to orientate them towards this set of behaviours. This is, in part, because it has been suggested that females with higher levels of masculine personality traits are more likely to use alcohol or drugs (Lucke, 1998).

Across the four samples (male and female Treatment, and male and female Non-treatment) it was found that both the Compulsive Helping dimensions occur together with Work. It can be seen that there differences between males and females and between the two populations, though females tend to score higher than the males (i.e. Compulsive Helping submissive: Male Treatments score 21.0, Female Treatments 24.8, (p<0.001) Male Non-treatments 18.5 Female Non-treatment 20.3 (p< 0.05)). The presence of both Compulsive Helping and Work can possibly be explained by the fact that the two are conceptually quite close anyway, so it is perhaps unsurprising that these behaviours cluster together.

Thinking about stereotypical concepts used to describe women and men it may be expected that women would score more highly on these types of scales, as it may be argued that their socialisation promotes supportive and dependent roles. Looking at the basic compositions of social stereotypes, being feminine means being attuned to and responsive to the needs of others, more emotional and more nurturant (Williams, Satterwhite & Best, 1999). Independence, competence, assertiveness and aggression are typically used terms which describe the characteristics that are to do with masculinity (op cit.). Gilligan (1982) suggests that moral development in females is guided by the importance and centrality of human relationships and by a dominating obligation to care for others. This “other” orientated quality of female moral development contrasts sharply with the moral socialisation of males which is thought to be more associated with self interest (op cit.).

Some evidence which provides support for these ideas comes from a study by Eisenberg, Fabes & Shea (1989), who observed that from the age of about 11 or 12 girls are more concerned with others in relation to their moral reasoning than are boys.

The core of this factor seems to originate with the association of the Relationship dimensions, as it is these two areas which are found to co-occur in both the Non-treatment male and Non-treatment female populations.

In the same way as in the previous factors it may be expected that these areas are “better” represented or inflated in the Treatment populations, as these normal orientations are taken to extreme by addicts. Interestingly, Shopping appeared in the female Non-treatment and female Treatment factors, and Sex appeared in the male Non-treatment and Treatment factors, thus contributing to the interpretation of this factor as a concern with stereotypical role-related exercise of power.

With reference to the Relationship scales perhaps it is to be expected that these scores in the male population load with the Sex dimension, as stereotypically males are seen to be orientated toward hierarchy, mastery and control whereas women are orientated towards the maintenance of relationships, (Stiles, Lyall, Knight, Ickes, Waung, et al, 1997). Another relevant issue is the relationship between dominance and masculinity which is often cited (e.g. Kaufman, 1997 and Spence & Buckner, 1995) as this may, in part, contribute to the comprehension of this pattern of scoring.

In relation to these issues, power has been seen as an important factor in the study of romantic relationships since the 1950’s (Waller & Hill, 1951 in Browning, Kessler, Hatfield & Choo, 1999). The two major definitions of power are as follows: the ability to influence another person’s attitude or behaviour (Cromwell & Olsen, 1975; McCormick & Jessor, 1982 in Browning, Kessler, Hatfield & Choo, 1999) and the capacity to produce intended effects (Gray-Little & Burks, 1983 in Browning, Kessler, Hatfield & Choo, 1999). Both of these features are central to the Relationship dimensions.

In terms of the present results, controlling another person for personal advantage may be achieved in two main ways, namely in a submissive (tranquillising) and in a dominant (stimulating) form. It seems that when looking at the males association with the Sex dimension in conjunction with their Relationship scores indicating that they score higher than women on the Relationship dominant scale, this it may be argued represents a more characteristically male than female way of being which originates in “normal” behaviour. Support for this position can be found in a study by Felmless (1994) where the power balances in romantic relationships were investigated and less than half saw their relationships to be equal in the distribution of power, and men were over twice as likely as women to be viewed as the partners having more power.

Sex forming part of this factor for Non-treatment males maybe has something to do with the prevailing attitudes of acceptability, as even though double standards with regard to sexual behaviour has diminished since the 60’s (Lottes, 1993) they have by no means vanished. Oliver & Hyde’s (1993) meta-analysis of gender differences in sexuality found large gender differences in both sexual permissiveness and casual intercourse. For instance, incidence rates and approval of premarital sexual behaviour and of many sexual risk practices continue to be higher for men than for women (Oliver & Hyde, 1993), as do the rates of other risk behaviours such as delinquency and to a lesser extent, alcohol and drug use (e.g. Magurie & Pastore, 1997; Maxim & Keane, 1992). So what seems to be appearing is a more global acceptance of hedonistic or sensation-seeking behaviours in men. This concept of hedonism or sensation seeking is useful as it seems to be thematic in the male involvement in these areas and is related to many risk behaviours including sexual risk (Arnett, 1996).

Oliver & Hyde (1993) have emphasised that a variety of different perspectives such as socio-biological, social learning, social role and script theories all expect women to have more negative attitudes toward casual, premarital sex than do men. According to this line of reasoning the very low score that Non-treatment females achieved on Sex and that it failed to load on any of the factors in the Non-treatment sample seems to fit closely with previous work conducted in this field.

A study which provides evidence and supports this line of reasoning investigated the relationship between sexual behaviour and non-traditional gender role and masculine gender traits in young people. The findings indicated that women with higher levels of masculine personality traits and egalitarian gender role attitudes were more likely to have multiple partners use and alcohol or drugs (Nicholson, White & Duncan, 1999).

In conjunction to this young women are more likely than men to be committed to their relationships, to report being in love and to view love as important while men subscribe to more permissive sexual attitudes (Hendrick & Hendrik, 1995). This indicates that women’s motivations and involvement may be quite different than are men’s. In research looking at motivations for sex, young men are more likely to give reasons to do with physical pleasure, whereas young women emphasise emotional and relational reasons (e.g. Murstein & Tuerkheimer, 1998). Interestingly, in the female Treatment sample Sex was found to appear on the Power Related Hedonism factor. Perhaps what occurs here, in the same way as on the Sensation Seeking Hedonism factor, is when an individual gets into problematic behaviour the constraints of morality and expected social role are reduced and it is this which permits engagement in behaviours otherwise considered taboo or “off limits”.

In conjunction with Relationships the behaviour which also appeared in this factor for women in both samples was Shopping. It has already been mentioned that shopping is possibly a way of self-medicating depression (Faber & Christenson, 1996), and in comparison to other forms of mood altering is certainly a more socially acceptable method of “feeling better”. Looking at the significance of this association, previous research has indicated that self image, self-presentation and shopping behaviour is more closely linked for women than for men in both ordinary consumers (shoppers) and excessive shoppers (Dittmar & Drury, 2000). Importantly, controlling the household budget is a source of power and control in the family for many women, and spending itself may also induce the feeling of control and capability and authority.

The core of this factor appears to be the two Food dimensions, as it is these which appear in both the male and female Treatment and Non-treatment populations. As both are to do with food consumption it is likely that there is something to do with the regulation of food intake, perhaps related to over indulgence on some occasions and under eating by way of compensation on others. In the same way as with the previously mentioned factors, investigating these patterns in a Non-treatment population permits some assumptions to be made about the Treatment population.

Comparing the male and female Non-treatment Food bingeing and Food starving scores, males score on average 5.5 and females 9.4 for food starving (p<.001) and males score 8.1 and females 13.6 for food bingeing (p< .001). Looking at this difference between males and females, and at the interaction found, it is logical to deduce that this factor holds more significance and meaning for females and it is this which causes the higher scores in women. There�s one main question to be addressed when considering these results, why are Non-treatment women scoring highly on the Food dimensions?

One possible idea is that it is today’s cultural emphasis on preoccupation with food, body shape and dieting which is most commonly directed at women which contributes to these higher scores. In support of this, in one study investigating the content of men’s and women’s magazines it was found that women’s magazines contained on average 10.5 times as many advertisements and articles promoting weight loss than men’s magazines (Andersen & DiDomenico, 1992). Interestingly these present findings indicating more of a preoccupation with food and weight issues are well placed in relation to other research which has been conducted in this area.

It was mentioned in the introduction that it is well documented that women tend to view themselves as more overweight than do men in general (Tiggemann, Winefield, Winefield, & Goldney, 1994). It could be posited from this finding that this would increase the likelihood of women engaging in dieting type behaviours and exhibiting a greater tension around food. This may help to account for the higher score for Non-treatment women.

From the review in the introduction of this chapter, it was also seen that gender differences in body image and dieting appear before adolescence and that it is females who experience the most struggle with this area (Shapiro, Newcomb & Loeb, 1997). Further findings which support this account include a study by Rolls, Fedoroff, & Guthrie, (1991) who found that adult women in general experience more food related conflict than men in that they like fattening foods but perceive that they should not eat them. In conjunction to this it was also shown in this study that women are more dissatisfied with their body weight than men (op cit.). As there seems to be a dominance of these concerns in the Non-treatment female population it seems possible that socio-cultural influences in conjunction with psychological factors may be important in the development of eating disorders, which are much more prevalent in females.

If the above arguments are valid then again it would be expected that the SPQ scores would be higher in the Treatment population as they would be better represented, as these behaviours are being used in an excessive and perhaps pathological way. It can be seen from a comparison of the means that this seems to be the case, see table 7.9.

Table 7.9: Means for food starving and food bingeing in male and female Treatment and Non-treatment populations

The results, looking at possible interactions, revealed that there was an interaction effect on Factor 1. This indicated that Treatment females score much higher than would be expected on this factor, given the “baseline” figures from the Non-treatment population. This fits in with the previously mentioned evidence, as it seems that women in general are more concerned with the main areas contained in this factor.

The pertinent question here, which seems to be also reflected in the present results, is why is there such a gender discrepancy in the eating disorders? It can be seen that the food dimension scores for women in general are substantially higher than those for the males. In conjunction to this it was mentioned in the introduction to this chapter that men are reported as representing only 10% of eating disorder cases (Andersen & Holman, 1997). Much has been written on gender issues and eating disorders (e.g. Dolan & Gitzinger, 1994) and a brief view of some of these thoughts will be looked at.

On a socio-cultural level it is thought that the dominance of women in the eating disorders is maybe to do with the contrived and impossible demands in which women are placed. The link here is that the thin ideal for women espoused in the media is related to the high rates of eating disorders (Waller & Shaw, 1994). Possibly as a response to this, dieting has been characterised as normal eating for North American women and is thought to have been caused by a shift in the female beauty ideal toward a thin physique (Polivy & Herman, 1987).

It may be the case that women who score relatively highly on these dimensions are exhibiting some of the behaviours or symptoms of an eating disordered pattern. It was mentioned in the introduction that in one study it was found that in a female college sample behavioural symptoms associated with anorexia and bulimia were displayed but interestingly few of the constellation of the psychological traits were exhibited which are associated with these disorders (Hesse-Biber, 1989). What was concluded in this study was that, “The aetiology of eating problems may be partly related to women wanting to be thinner than is medically desirable and may represent a response of “normal” women to the new, more demanding cultural and super-cultural standards for thinness” (Hesse-Biber, 1989, p.71) This provides an interesting perspective on the present results.

A further interesting element from the results is the involvement of Exercise on this factor. Whereas Exercise failed to load on any of the factors in either of the male samples, in the female samples it loads strongly with the Self Orientated Nurturant factor appearing with Caffeine and both of the Food dimensions. It has already been said that this factor seems to be concerned with self control especially regarding body image, and if this is the case then Exercise fits well with this interpretation. This is because if compulsive exercise is used in conjunction with under eating and the use of caffeine (perhaps to moderate over eating), this constellation of behaviours will all contribute to weight regulation.

The presence of Exercise in the Food factor is likely to be related to the previously mentioned tension which exists between food and body image. A piece of research which relates to this is a study by McDonald & Thompson (1992), who investigated levels of eating disturbance, body dissatisfaction self esteem etc. of physically active males and females. The results indicated that for women the motivation for exercise was more often related to weight and muscle tone reasons than that of men. Interestingly for both genders, exercising for weight, muscle tone and attractiveness reasons was highly correlated with other measures of disturbance (op cit.). It can be seen that if only the means of the Exercise dimension are considered there is little variation across the samples, and it was found that there wasn’t a significant difference between the male and female Treatment population and the male and females Non-treatment population. It is only when this dimension is considered alongside other behaviours that its possible relevance is revealed.

Alcohol is a very commonly used substance due to its availability and socially accepted use. However, in general it has been seen from the previous literature that men consume more alcohol than women (e.g. Dawson, 1992). It may be, in part, for these reasons that alcohol doesn’t load highly on one particular factor in the Non-treatment and Treatment male populations. With females, however, alcohol appears alongside Nicotine and Drugs in the Treatment population and comes closest to loading on this factor for the Non-treatment population.

One of the main differences between the male and female samples is that Alcohol in the male sample doesn’t form a stable component of any of the factors, though it comes closest to contributing to Factor 2 which is a Nurturant factor, containing Compulsive Helping in both forms and Work. Here its role may reflect its use as a response to, or use in conjunction with, the commitment to working hard, looking after others (family responsibilities) essentially keeping on top of everyday obligations. It is conceivable that the effects of alcohol are sought as a socially acceptable release from the stress of keeping a “normal” life together and performing the role of the traditional male bread-winner. Though this idea is being stated in stereotypical terms it can be seen that with the consideration of the more nurturant behaviours being engaged in alongside alcohol use such as Work, the picture portrayed is potentially quite different. From the previous literature it may have been expected that alcohol would load most highly with the other forms of hedonistic behaviours as it seems to be most often reported alongside drug use and other anti-social activities.

It seems possible then that there may be a number of orientations towards alcoholism, as alcohol seems to play a part in a range of addictive activities. And so perhaps the consideration of the possibility of two broad categories would be useful, a Hedonistic alcoholic and a Nurturant alcoholic.

In the Non-treatment female sample Alcohol comes closest to appearing in the Sensation Seeking Hedonism dimension along with Nicotine and Drugs. In the female Treatment population, Alcohol contributes to the Sensation Seeking Hedonism factor appearing alongside Drugs, Nicotine, Prescription drugs and Gambling. As Alcohol appears with the other more aggressive and obvious substance based forms of mood altering in conjunction with Gambling, it is possible that this reflects the way that alcohol is used in a different capacity in comparison to its use in men. The majority of these areas involve deliberate and obvious acts of mood altering and it’s plausible that here alcohol is used more as an “upper” or a mood enhancer rather than for its tranquillising effects.

Again Heimer’s (1996) idea of double deviance is interesting here, as perhaps alcohol use is still not as acceptable in women’s behavioural repertoires as it is in men’s and therefore is seen as being associated more comfortably with the more extreme and deviant forms of mood alteration. In support of this the incidence rates and approval of delinquency and to a lesser extent, alcohol and drug use continue to be higher for men than for women (e.g. Magurie & Pastore, 1997; Maxim & Keane, 1992)

It seems from the results and discussion that part of the comprehension of addictive orientation will involve the careful consideration of gender. It seems that it is, in part, gender which orientates people towards the normal use of (and subsequent addictive use of) addictive substances and behaviours. Though it must be said that as both males and females are affected by addiction there may be a number of common factors which increase the likelihood of addictive use occurring. It seems that it may be gender which propels males and females into different courses of activity.

Theorists argue that focusing on lives in context is needed to understand differences in risk behaviours (Chesney-Lind, 1989). This is especially applicable in the field of addiction as it may be said that conventional ideas of addiction have been largely constructed to explain male addiction. It may be argued that consideration of and investigation into the complexity and richness of men’s and women’s lives will reveal important insights into the many questions surrounding gender and addiction. It may even be the case that separate theories are required to account for male and female “orientations” to addiction.

A requisite component of future research may be to attempt to contextualise the sociological and cultural variations in women’s and men’s lives, as it is important to place addiction within the context of contemporary society; to talk about both theory and life experiences of addicts, as it is likely that women and men may engage in the same behaviours for different reasons (Waldron, 1997). One implication is that this may cause differences in how to best approach and treat these problems.

In order to further explore these gender differences, it would be interesting to investigate the relevance of personality and associated psychiatric symptomatology in relation to the factors, especially with reference to the relevance of social and conduct disorder elements. In the next two chapters these areas will be explored.