RESUMO
Relationship of different symptoms, personal history, some measures of Rorschach, neuroticism, ego strength score with prognosis of depression was studied. Out of 60 items studied in 60 cases, eleven were found to have statistically significant association with improvement. Multiple regression of percentage of improvement on 23 items (12 from personal and family history and 13 measurable items of Rorschach) was determined. To check the ability of this multiple regression to expl. in differences in responsiveness, predicted values were obtained for 24- cases (not included in the previous regression analysis). In most cases considerable differences was found in observed and predicted values. Then a regression coefficient using only 12 items (other than 13 Rorschach items) was determined similarly. Using this regression equation, predicted Values for percentage of improvement was obtained for the same 24 cases. There was insignificant association between observed and predicted Values. Thereafter only five measurable characteristics strongly related with percentage of improvement were taken and multiple regression on these items was determined and checked similarly on the same 24- cases. This also showed insignificant association between observed and predicted Values. Linear regression is unable to explain the difference in prognosis of population very similar to one on which it was developed. Inclusion of more history and social interaction factors might lead to more fruitful results.
