The collected results will be treated statistically. The realization of student's T distribution has allowed us to analyse the significance of the difference in the results obtained by our sample of alcoholic subjects to the 2 terms ("when I drank" / usually"). SPSS ®, specialized software, will enable us to draw up a matrix of correlations of BravaisPearson. This matrix will serve as the basis for the calculation of the regressions of data obtained by the first mode ("when I drank") on the one hand, and secondly, by the second mode ("usually"). The results of the regression analysis step allowed us [21,22] to establish a model for each of the 2 terms. Each of the models obtained on the basis of this analysis will be tested by appropriate structural equations modelling software, AMOS®. According to literature [23,24] a model is consistent when it does not differ from the mathematical one (chi square nonsignificant); therefore some central indexes (global fit index (GFI) or central fit index (CFI) both closed to 1) are often used to complete the conditions of the consistence [17,[21][22][23][24][25][26].
Results
Global outcomesAccording to Table 1, it can be seen few stability in the selfperception of the persons among they have drank or when they don't.Average GSI indicate that alcoholic subjects in our sample to represent their psychological as difficulties being more serious when they think themselves to be under the influence of alcohol (1.22 vs. 0.95, p<.01). On the same way, the diversity of the perceptions of their problems changes whether they describe themselves "drunk" or "as usual" (53.83% vs. 46.93%; p<.05). Finally, their degree of discomfort (PST) is higher when they imagine to be under the influence of alcohol (1.96) when they imagine not be (1.73).With regard to the sub-scores obtained by our sample of alcoholic subjects to the nine psychological factors, they are also higher when the alcoholic subjects in our sample imagine be under the influence of alcohol when they imagine not be. We also note that only the features of anxiety, phobia and the "various" did not vary the set point. What would imply a degree of stability in the self-perceptions of their clinical features? Therefore, we can say that, descriptively, there are differences between the scores obtained by our sample when they imagine to be "in alcohol" and these same scores when they imagine to be alcoholfree. But what are the incidences of the different cognitions on the psychological suffering?Model of psychological difficulties among the condition "when I drank" and "usually"If one performs analyses of regression step by step from the score of the GSI, one realizes that two clinical features mainly underpin it: anxiety (β=92%) and depression (β=91%), and that the other traits are involved also in the composition of the score. This confirms, if it were needed, the factor nature of the SCL-90-R.However, if considering, by following the same statistical method, the score of anxiety, invariant of one condition to another, one realizes th...