Although one dominating idea is that addicts and non-addicts are characteristically different, possibly but not necessarily as a result of genetic disease, the prevailing view would probably be that defining the point at which excess becomes addiction is highly problematic (cf Vaillant. 1985). Whatever view one takes, it is clear that engagement in an addictive behaviour does not necessarily indicate addictive use, and that even experimentation with highly addictive substances does not necessarily predict addiction.

Acknowledging the fact that these behaviours are used in a normal fashion, a method is required for distinguishing between normal and abnormal usage. Psychiatric classification (e.g. DSM-IV) aims to do this. However, moving on from the more formal classification systems and methods of detecting problematic use, a conceptual framework that builds upon the use of addictive behaviours within a Non-treatment population is of value.

Research has very much concentrated on problematic use with only a few studies investigating addictive tendencies within a normal population, such as Von Knorring and Oreland’s study (1985) where it was found that smokers were more prone to the abuse of alcohol, glue, cannabis, amphetamines and morphine. As these behaviours are routinely used there is a possibility that even without pathological levels of use, it may still be possible to study orientations within a Non-treatment population, akin to those within the Treatment population. The results of such a study would be of considerable theoretical, and practical, interest.

6.2. General statistical information on the use of addictive substances in a non-Treatment population

Information on the nation’s involvement in the more notorious addictive substances is routinely collected, and looking at some basic statistics reveals that the involvement in addictive behaviours is quite substantial. In 1998 the British Crime Survey showed that 29% of those aged between 16 and 24 in England and Wales had used an illegal drug in the past year. The most commonly used drug was cannabis, which had been used by a third of young men and over a fifth of young women in the previous year. Although only 3% of young people reported using cocaine in the last year there had been a significant increase in the proportions using this drug since 1996.

The current Department of Health advice on alcohol is that consumption of between 3 and 4 units a day for men and 2 to 3 units for women is within “safe” limits. However, the proportion of men drinking over 21 units a week has remained broadly similar since 1988, at around 27%. The proportion of women drinking over 14 units has increased from 10 to 15 per cent between 1988 to 1999 (General Household Survey, 2000). These figures suggest that addictive levels of use are increasingly commonplace.

There is evidence which suggests that there may be similar relationships between addictive behaviours in Non-treatment group (i.e. non-addicts) as in those receiving treatment. In a study by Gilbert, Gilbert and Schultz, (1998) it was suggested that there is a high degree of similarity in withdrawal response across a number of addictive areas (alcohol, nicotine, caffeine, food and relationships) in both addicted and “non-addicted” populations. Similarities in withdrawal symptoms across substances suggested that individuals experiencing severe symptoms associated with withdrawal from one type of dependence were also likely to experience severe withdrawal symptoms from another. In addition to this, the observed co-morbidity between eating and drinking problems in clinical populations has been found to be present in non clinical samples of adolescent boys and girls (Krahn, Piper & King, 1996). Where commonalities such as these exist, there may be similar mechanisms e.g. of habit reinforcement, in these different populations. To search for mechanisms specific to the onset of addiction “proper” may be misguided. Personality, physiological or behavioural correlates of engagement in addictive behaviours may be equally evident in Non-treatment populations.

If there are similar patterns of usage in this Non-treatment population it would imply that “addicts” are not a discrete or especially distinctive population. Rather, they may be viewed as somewhat akin to neurotics, as those whose location at one extreme of a normal continuum, have experienced particular problems of adjustment to their environment. The orientations towards addictive behaviours may hence appear in the Treatment population in an amplified, exaggerated and problematic way. They are, nevertheless, orientations in terms of which the Non-treatment population may be characterised and assessed.

Alternatively, it is not hard to envisage that the normal results will fail to replicate the results with the Treatment group, and that the factors discovered in the Treatment group are not evident in those who claim not to have problems with addiction and who have not sought treatment. A Non-treatment group might be expected to develop preferences for different substances in an idiosyncratic way, according to personal inclination and learned experience.

Having found the addictive orientations in the Treatment group, the expectation that the findings of the Treatment group would be broadly replicated seemed the more appropriate expectation. The addictive orientations seem to reflect attitudes towards experience that are quite close to those of personality orientations discovered in conventional psychometric work in normal populations. This topic will be further explored in a later chapter.

For the moment, let us pose the question: Will the results from a Non-treatment group broadly replicate the factors found in the Treatment group? Do those who are not addicted express preferences for addictive substances and behaviours in the fashion of the orientations towards those substances and behaviours exhibited by the Treatment group? The answers to these questions will serve to direct the further study of the role of personality in the development of addictive behaviours.

The SPQ was administered to 199 male and 309 female participants making a total of 508. These participants were obtained by convenient opportunity, one group from a student population and another from the community. The only sampling criteria were that that could read the SPQ and were not receiving treatment for any type of addictive behaviour.

Approximately 200 participants were students at the University of Kent and were obtained through the Psychology Department’s Research participation scheme (RPS). People studying psychology are required, as part of their course requirement, to participate in a number of pieces of psychological research and are awarded “credits” for participation and it was from this population that this sample was drawn.

The remainder were gathered from outside of the University. This was conducted using a “pyramid gathering” technique where up to five questionnaires were sent to various contacts who were asked to complete one themselves and for the remainder to be given to friends or colleagues to complete. Confidentiality was insured as the participants were not required to enter their name, and envelopes were provided for each participant to place their questionnaires into.

The same analysis which was utilised with the Treatment population was conducted using data from a sample of 508 declared non-addicts. The sample included 199 males and 309 females. As this sample has a disproportionately high number of females, more precisely comparable results may be obtained when the analyses are completed by gender. The MANOVA/ANOVA tests later will test whether there is indeed a significant difference between genders and whether this factor needs to be taken into account. The descriptive statistics indicate that the mean values of each variable and the standard deviations are significantly smaller (in most of the cases less than half) than the values in the Treatment population (see Table 5.2. for details of Treatment group).

Table 6.1: Non-treatment population. Table of means and standard deviations of the SPQ items

Table 6.2: Non-treatment population. Correlation table for the 16 scales of the SPQ

Alcohol (AL), Shopping (SH), Food Bingeing (FB), Compulsive Helping Submissive (CHS), Nicotine (NI), Gambling (GA), Food Starving (FS), Compulsive Helping Dominant (CHD), Drugs (DR), Sex (S), Work (W), Relationships Dominant (RD), Caffeine (CA), Prescription Drugs (PD), Exercise (EX), Relationships Submissive (RS).

The correlation matrix for the sixteen scales is shown in table 6.2. It can be seen that the correlations are almost entirely positive, to an even greater extent than in the Treatment group. In the same way as in the Treatment sample, it seems that the matrix indicates that factor analysis will yield a solution which reflects the generally positive pattern of co variation. Where r is the estimated correlation and n is the number of observations it is confirmed that correlations as low as 0.09 are significant as the 5% level.

As with the Treatment sample, the preliminary Principal Components analysis Scree plot (figure 6.1) shows that there are possibly between 2 and 4 dimensions that explain the data well. In the present sample however, the largest eigenvalues of the correlation matrix, which represent the dimension that explains the most variation (the first PC) are much higher.

Figure 6.1: Non-treatment population: Scree plot of the sixteen scales from the SPQ

Table 6.3. Non-treatment population: Results from the principal components analysis

In the Treatment group factor analysis it was seen that a four factor model explained the data well. The task now is to investigate the factorial structure in the Non-treatment sample to see whether there are any similarities in factor structure.

Table 6.4. Non-treatment population: Results for the rotated factor loadings and their communality estimates.

Figure 6.2: Non-treatment population. Graph of the communality values

Again the graph of communality values shows that the model explains a large part of the variability for most of the variables. It is of note that the Compulsive helping variables are most, and that Alcohol and Gambling are least, explained, which is in direct correspondence with the 4 Factor Solution for the Treatment population (see figure 5.2 for graph of the communality variables for the Treatment population). It is to be expected that there are slight differences between the Treatment and Non-treatment groups with the communality values and rank orders, but basically the models seem to be consistent at this point.

Figure 6.3. Non-treatment population: Loadings for factor 1

Looking at figure 6.3. it can seen that the main constituents of Factor 1 are both the Food dimensions and Shopping, which corresponds exactly to the 3 variables with the highest correlation with Factor 1 for the Treatment population. So again this factor may represent “Self orientated nurturant behaviours”. However, using our 0.5 criterion would lead us also to include Relationship submissive for the Non-treatment group as opposed to Caffeine with the Treatment sample This identifies a slight differences between Non-treatment and Treatment groups. In addition we may also include Prescription drugs as this dimension’s factor loading is 0.498.

Figure 6.4. Non-treatment population: Factor loadings for factor 2

Looking at figure 6.4. it can be seen that the main constituents for Factor 2 are Compulsive Helping Submissive, Work, and Compulsive Helping Dominant. This is identical to Factor 2 for the Treatment group, and closely replicates the “Other-orientated nurturant” factor found.

Figure 6.5. Non-treatment population: Factor loadings for factor 3

Looking at figure 6.5 on the previous page it can be seen that the main constituents of factor 3 are Sex and Relationship Dominant. Gambling and Relationship Submissive are also worth considering as contributing a significant amount to this factor. Together these 4 variables contribute the most, which matches up with the main constituents for Factor 4 for the Treatment group. So essentially these Factors are measuring the same thing, but the constituents are weighted slightly differently in the two groups. The fact that it is the third most important Factor for the Non-treatment group and the fourth for Treatment group, is a trivial finding. The eigenvalues, which characterise the relative importance of a factor, and are an indication of the amount of variance explained, are of the same magnitude for the third and fourth factors for both Non-treatment and Treatment groups.

Figure 6.6. Non-treatment population: Factor loadings for factor 4

It can be seen from figure 6.6 that the main contributions are from Nicotine and Drugs, which, with the exception of Prescription drugs, matches up with Factor 3 for the Treatment group. The different role played by Prescription Drugs may be a result of the different perceptions and attitudes to Prescription Drugs held by the Treatment and the Non-treatment groups. Interestingly, Caffeine and Alcohol also contribute to this factor, although to a lesser extent. These are other addictive substances, which for the Treatment group did not feature consistently in this factor.

Even though there are differences between the samples there are clusters of behaviours which remain constant, especially the Food scales and Shopping, Drugs, Nicotine, Gambling and Sex. For the most part, the highest loading variables are consistently found in the equivalent factors in the two groups.

The Complete Linkage Algorithm for cluster analysis produced the dendrogram as below.

Figure 6.7: Non-treatment population. Cluster analysis of the SPQ scales.

If the cut off point is set at around 0.60 this partitions the variables into five main groupings:

Cluster 2: Compulsive Helping Submissive, Compulsive Helping Dominant, Work, Relationship Dominant, Relationship Submissive and Exercise.

The first cluster matches up to Factor 1 (“Self orientated nurturance”), the second cluster with Factor 2 (“Other orientated nurturance”). Cluster 3 and Cluster 4 combined correspond to the Factor 4 (“Sensation seeking hedonism”) and Cluster 5 corresponds to Factor 3 (“Power related hedonism”). Whilst these are quite good matches the cluster analysis is not so consistently confirmatory as with the Treatment group. This perhaps indicates some of the differences between the populations. In particular the “additional” factor of Caffeine and Prescription Drugs is an anomaly. Nonetheless, the overall pattern replicates the factor analysis closely in the major divide between the Hedonistic and Nurturant orientations, and in the further sub-division of each of those into two essentially similar clusters. Moreover, the “anomalous” pairing of Caffeine and Prescription Drugs has close links with the Hedonistic group.

The aim of this study was to investigate patterns of co-variation in a large variety of addictive behaviours in a Non-treatment population to see whether some form of orientation found in the Treatment group would be replicated.

On the basis of the scree plot and the principal components analysis a four factor solution was chosen as the best model to explain the data. Additional evidence for the appropriateness of this number of factors was given by the cluster analysis. The results indicate that the four factor model as found in the Treatment sample closely mirrors the “normal” pattern, but with interesting deviations which may throw light on the pathology of problematic indulgence in the Treatment group. These results have important implications for both the conceptualisation and measurement of addiction.

Characterising the similarities and differences between the Non-treatment and Treatment populations.

The similarity of factor structure in the two populations may indicate that the process. or motive to use is similar across the populations. If this is the case it may be said that there are naturally occurring orientations towards these clusters of addictive behaviours that when “used” or engaged in a non addictive way constitute a normal pattern of behaviour. This may perhaps be attributable to personality or temperament. Maybe then some additional factor or combination of factors, such as genetic predisposition, psycho social stressors, or developmental pressures, that then tips an individual user over from a normal pattern of involvement into the compulsive, addicted behavioural realm.

Though there are a number of very notable similarities across the populations there are, however, some interesting differences. One of these differences is in the position of prescription drugs in the factor structure. In the Treatment group that scale is linked to recreational drugs, whilst in the Non-treatment group it constitutes a separate cluster with caffeine, and is linked more directly to the Nurturant orientation. This exception to the rule may be symptomatic of the broader differences in style of use between the Treatment and Non-treatment populations. It is possible that in the Non-treatment population’s usage of prescription drugs focuses on pain killers, anti-anxieties and anti-depressants that are used for their prescribed or stated medicinal effects. As prescription drugs falls into the “Self orientated nurturant behaviours” in the Non-treatment group it is likely that this behaviour is used legally in a self soothing way, and not as a means to alter mood in the manner of recreational drugs.

In the Treatment population the use of prescription drugs may be more of a deliberate and aggressive nature, possibly involving the ingestion of more than the recommended dose, obtaining them illegally, or as a substitute for unobtainable recreational drugs. This difference in “style” is, of course, what characterises the difference between addictive and non-addictive use more generally. The different location of prescription drugs in the factor structure highlights the point.

It is of interest to note that Alcohol in the Non-treatment sample does not load on any of the factors. This is possibly to do with how acceptable and pervasive the use of alcohol is, and until usage becomes “alcoholic” or problematic it coexists alongside with many behaviours, without contributing conspicuously to any one factor.

Theoretical ideas which may shed light on the differing patterns of addictive involvement in Non-treatment and Treatment populations.

In the previous chapter Jacob’s (1989) idea of two factors having to be present in order for an addiction to develop was mentioned. The first necessary factor is thought to be an abnormal physiological resting state that is either excessively excited, or depressed. The second is an emotional or pathological state which stems from childhood. This helps to shed light on why similar patterns of orientations are found in Non-treatment and Treatment populations, as it could be argued that without pathological or extreme levels of the predisposition of hyper and hypo arousal what may be produced is a normal involvement or tendency to be more involved in certain addictive areas.

It is possible that people in general can be characterised as experiencing either of these arousal states but not in an extreme way. Contributing evidence for this stems from studies which have found that those who like to engage in extreme sports are more likely to be extroverted and have higher levels of external locus of control (Hughes & Coakley, 1991). It is thought that these characteristics contribute to the attraction of these high risk pursuits. If this is the case, these naturally occurring levels may then determine what types of activities are sought out. For instance for those who are already aroused to some degree, Gambling or Drug taking may be too stimulating, so more comforting behaviours which soothe the slightly elevated arousal state, may be a more natural preference. So a person with a hypertensive arousal state may find relief in substances or behaviours which have a resultant calming affect such as exercise with the resulting endorphins burning off excess energy or anxiety. A person with a depressed physiological arousal level may find relief in a stimulating and exciting activity such as gambling, thus temporarily eliminating their boredom and possible depression. In this way activities are determined by the naturally occurring arousal levels. With this goal achieved, the resting state is altered and a more comfortable homeostatic state is found.

The presence of an extreme level of arousal, whether over or under aroused, increases the likelihood of an addiction developing. According to the theory, an additional set of environmental conditions are required, leading to the development of a certain psychological “state”, before addictions are formed. It is this additional experiential factor, that affects the sense of selfhood, that maybe the necessary prerequisite that pushes this natural orientation from the normal range into the range of active addictive usage.

A further possibility is that the orientations can be understood in terms of Social learning theory, as mentioned in the previously. If this was the case then an inclination towards specific clusters of behaviours could be seen as a cultural and a behavioural phenomenon, with the orientations forming part of a learnt behavioural repertoire, influenced by parental figures, peers etc. In this case, it is feasible that the generalised patterns of use apparent in both of the populations, is a consequence of culture, personality and learning etc., but that in the Treatment population these naturally occurring inclinations are taken to an extreme. Hence the real difference is to do with degree of psychological or emotional significance and involvement which produces the person with characteristic addictive problems. Of course, the “physiological” and social learning theories may be seen as complementary to one another, in providing a basis for the orientations.

As differences were noted in chapter three regarding gender differences in various addictive behaviours, (e.g. the dominance of females suffering from eating disorders Andersen and Holman, 1997), in the next chapter differences between the two populations according to gender are explored.