The training analysis concept assumes that training analysts (TAs) can more effectively treat analyst-patients than non-TAs can. This study tests this assumption empirically by comparing satisfaction with analytic treatment by a TA with that by a non-TA in the same analyst-patients. Extensive literature critical of training analysis led us to hypothesize that analysis of analyst-patients by TAs would actually be less satisfactory than personal analysis by a non-TA. The validity of the analyst's questionnaire ratings of satisfaction was supported by independent ratings by 2 senior analysts of transcribed individual interviews of participants. It correlated significantly with participants’ questionnaire ratings of satisfaction. Theoretically, treatment by TAs should be more satisfactory than treatment by non-TAs. This study, however, found no significant difference between satisfaction with analytic treatment by a TA compared with treatment by a non-TA. This same lack of difference in satisfaction had been reported in a previous unrelated clinical interview study. In addition, there was no difference between TAs and non-TAs in the proportion of analysts who reached a mutually agreed termination and no difference in treatment duration. Because no study has reported that treatment by a TA is more satisfactory than treatment by a non-TA, the burden of proof falls on those psychoanalytic organizations who utilize a TA conception to demonstrate that treatment by a TA is more satisfactory.