Two shortcomings of attributional therapy and therapy-related research are pointed out. First, crucial comparisons between attributional inductions are not being made, so that the variables mediating change remain uncertain and, secondly, the full theory is not being used, thereby greatly limiting the possible change programmes.