Re: Psychotherapy

From: Harnad, Stevan (harnad@soton.ac.uk)
Date: Fri Jun 16 1995 - 20:58:45 BST


> From: "Maguire, Nick" <NICK92@psy.soton.ac.uk>
> Date: Mon, 5 Jun 1995 12:21:15 GMT
>
> You criticised me for not questioning data and theories that concur
> with my own agenda. In fact I do question them, but probably attach
> less significance to negative results that I read than positive ones.
> I feel that you (and everybody else in the scientific community) are
> probably just as guilty of this form of backing up one's own agenda.

I agree that everyone inadvertently does this to some degree (but not
everyone does to the same degree, though, don't you agree?).

> In the lectures you talked about, explained and emphasised hermeneutics
> at the same time as talking about psychotherapy. It was not difficult
> to pick up your own personal view

In every lecture, on every topic, I tagged my own view with "but that's
just my view, and there are others, and this course is about debates:
have a look at the rest of the views:...

> (unless it was a double bluff and you
> actually do find some merit in the concept of psychotherapy - which I
> somehow doubt);

I actually do, but I am troubled by how difficult it is to find what
really works, unlike in other branches of medicine.

> by explaining psychotherapy in terms of hermeneutics in
> such an emphatic fashion, the two are linked in our minds.

The hermeneutic thread ran through all of it -- from parapsychology and
hypnosis, through psychotherapy, sociobiology, and consciousness to
theory of mind. It's there, because psychology cannot shake it off. It
is what makes psychology psychology rather than something else.

> Those of us
> who go on to read Prioleau et al in depth realise that in fact this
> empirical evidence that you cite is not particularly convincing,
> especially if the commentaries are taken into consideration (more about
> this later),

That's right. But neither is the alternative convincing. The subject is
controversial, and unresolved. That's why it was in the course...

> but those who choose to follow up other topics and leave
> PT are left with the opinion that "it's all hermeneutics because Stevan
> said so". I am not saying that we are all mindless and cannot form
> opinions, but the information on which many formed those opinions was
> one sided. Even if the references were followed up (which, for
> practical reasons - i.e. time - was not possible for all of the
> treatments) it is only the commentaries that provide the alternative,
> pro-psychotherapy view. Hermeneutics provides ONE discourse; another is
> that PT does actually work.

I agree that I have to figure out a better way of ensuring that everyone
does at least some reading on each topic. In future years this will be
fixed.

As for hermeneutics being ONE discourse: The "discourse" view of things
IS hermeneutics. The alternative (realism) is that it's not just a bunch
of discourses. That there are facts of the matter, about which
discourses can be right or wrong...

> Prioleau et al attempt to provide empirical evidence indicating that PT
> is no better than placebo by examining a subset of the Smith et al
> (1980) meta-analysis. Their results are far from conclusive, and again
> seem to be a matter of interpretation.

I agree. They are not conclusive in either direction. But what is
conclusive is that other branches of medicine are not inconcluive in
this way...

> My major criticism is the
> averaging out of effect sizes from different therapies, a point that is
> made best by Andrews in the commentaries. Even the authors themselves
> quote figures of .74 for the effect size of verbal psychotherapies
> (p.304), but average them with an effect size of .35 for developmental
> therapies. This was justified by pointing out that these comparisons
> were made on different studies, and that such a comparison "may not be
> legitimate". They don't actually explain why.

Averaging effects is the whole point of metanalysis. If you did it
across a lot of branches of medicine, you'd get a nig positive effect,
not a very weak one that specialists cavilled was due to mixing apples
and oranges. That is because most of medicine really works.

> This issue is, I (and many others) believe is at the heart of the
> problem; different therapies with different clients produce varying
> results, and the averaging out of such varying methods of therapy is
> misleading.

Would you say the same of medicines? If you average together the effects
of a bunch of medicines (tested, validated medicines), prescribed for
the right illness, they will all contibute to a huge meta-analytic
effect. If they did not, we couldn't say that it was the medicines did
indeed work, but they were just not being well-matched to the patients:
Until they had been well-matched to the patients and their effects
shown, we could not conclude they worked at all. And that's the point of
a meta-analysis.

> It is hypothesised by Dahl that different people respond
> differentially to different PTs. Future research needs to control for
> this.

Not "control" for it, but sort it out to show that is indeed true,
rather than special pleading in the absence of reliable overall results,
as in the "anomalies" research...

> A brief BIDS search for the latest meta-analyses did not locate
> any recent evidence that this has been done.

I just did a search and found 60 new metanalyses since Prioleau et al's.
in '83. See:
ftp://cogsci.ecs.soton.ac.uk/pub/harnad/Bibliographies/psychotherapy.metaanalysis

> On reflection I suppose that my criticisms pertain to anyone who
> publicises positions and theories and who also has an agenda (including
> me). I was considering my message last night, and I think that I found
> a flaw in my own argument. Do you think that psychotherapy does work,
> because of the patient and therapist reinforcing each other in a
> hermeneutic fashion, but this nature leaves it impervious to empirical
> investigation; or is your criticism more global, in that the
> hermeneutic nature of PT is no better than placebo (as in Prioleau et
> al)?

It could be hermeneutic and real -- that is, whether you are happy with
life could be a matter of interpretation. More likely, though, for
serious problems, it's not, in which case if psychotherapy is, it won't
work.

Chrs, Stevan
mailto:harnad@soton.ac.uk



This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:17 GMT