Saturday, May 21, 2005

A big big sample

I began to look through my results of the statistical study of the Kassel Stuttering Therapy (KST). I need to submit the article at the beginning of June for the Oxford Dysfluency Conference . The paradoxical thing about conference submissions is that the researcher FIRST has to submit an abstract, which most often leads to them writing the abstract (i.e. the summary) before writing the article!

The KST sample is very good. All patients fill out questionnaires just before the therapy, just after the intensive 3-week phase, one year later, and three years later. In addition they have to do speech samples in four different speaking situations, reading, phone call, speaking to therapist, and interviewing a by-passer on the public's attitude on stuttering (!).

There have been 383 patients recorded over 7 years, I think, as of June 2003. This is a very large sample, and it's complete as all patients are in it. This dedication in time and effort (--> MONEY) on the part of the KST (led by Alex von Gudenberg , by the way the picture is an older and thinner one.. :-) is mostly to provide the German health insurers with evidence that the therapy is indeed beneficial, but also a genuine interest in going beyond the practice-based therapy.

In my article, I will only focus on the profile of the KST patients. Who are they? How much do they stutter? In which speaking situations do they stutter most? I have also looked at the outcome of the therapy, but this will be presented by Prof Euler at ODC.

Before discussing some results, I want to say a few words on therapy research. I have seen many studies, also at ODC 2001, on therapy outcome, and I think many are of little value. There are often serious issues, in my view: 1) the sample is often much too small (less than 30) 2) the sample is biased in some way e.g. only patients who have finished the therapy phase are included, 3) no long-term data, and 4) they look at statistical significance even though the important measure to look out for is effect size. Having said this, there are also researchers who do it well. But generally, they have a science background, and not a therapy background. Dont get me wrong I think therapists have an important contribution to make, but I, as a hard-core quantitative person, would prefer to see more first/second person experiences like case studies and discussion of them. Quantitative analysis is only useful for big and unbiased samples, and useless to get across complex issues/situations.

In my next post, I will talk about the sex ratio in the KST sample, and discuss why it's not balanced in PDS.

No comments: