The short answer is It Depends. There are lots of factors involved in whether any adult learning method is effective. I've tried to cover the major points below.
There are many unknowns in the training effectiveness field. We can say that in any training situation, some people learn more than others, but the question is how to determine what affects whether the training helps. (And yes, I'm considering self-development seminars as a subset of training -- since that's the general academic term for adult learning in a non-academic setting.)
The first point, and I suspect what many people worry about in the self-improvement classes (as opposed to therapy) is that one-size fits all doesn't. The best trainers first perform a needs analysis to make sure that they know what people need to learn. See Assessing training needs: Critical levels of analysis. Training and development in organizations and A Proactive Model for Training Needs Analysis as sample articles. Or look at this Google Scholar search for more.)
If the people conducting a self-improvement seminar don't conduct a needs analysis, they're going to end up sticking to basic concepts that may or may not help the attendees. There are certain concepts that you can cover in any such seminar -- such as the fact that if you want things to change, you have to take action. If you keep doing the same thing, nothing will change. (Which, of course, is common sense. But there are psychological motivations that keep people from acting -- and good self-improvement seminars would help people identify some of those factors for themselves.)
Design of Class and Teacher Style
In addition, the design of the training and the effectiveness of delivery affect the results. See Effectiveness of training in organizations: A meta-analysis of design and evaluation features for a statistical analysis of design factors. For discussion of the affect of the trainer's style on results, see The relationship between verbal teacher immediacy behaviors and student learning or this Google scholar search.
In general, we can quantify the effectiveness of training by two means -- information retention and behavior change.
One site I found references an article that suggests that the initial training transfer rate is 62% (Saks, A. M., & Belcourt, M. (2006). An investigation of training activities and transfer of training in organizations. Human Resource Management, Winter 2006, Vol. 45, No. 4, Pp. 629648). In other words, immediately after training, attendees retain about 62% of the information presented in that training. The numbers drop over time, however. In addition, that rate is simply for information.
10 to 20% of training transfers to actual job performance. (Baldwin, T.T, Ford, J.K (1988), Transfer of training: a review and directions for future research, Personnel Psychology, Vol. 41 pp.63-105.)
Outside Factors: Attitudes, Context, etc.
So why so little change? There are a number of outside factors that can affect how well attendees at a training learn.
For example, the attitudes and motivation of the trainees can affect how well they integrate what they learn. (Trainee's Attributes and Attitudes: Neglected Influences on Training Effectiveness. Academy of Management Review. 1996) Also see The Influence of Trainee Attitudes on Training Effectiveness: Test of a Model and Toward an Integrative Theory of Training Motivation: A Meta-Analytic Path Analysis of 20 Years of Research.
Similarly, factors external to the training -- such as the personality of the trainees and the work environment of the people being trained can affect how well the training works (Training Effectiveness: Accounting for Individual Characteristics and the Work Environment).
Now, having said all that, there are some studies on particular types of personal development seminars. For example, EST was discussed in Observations on 67 patients who took Erhard Seminars Training, and in Psychiatric disturbances associated with Erhard Seminars Training: II. additional cases and theoretical considerations. You can always try to learn more about any particular system by doing similar research.