Keith and everyone else,

It has taken a while to digest all your comments here.  I'm also looking forward to looking up your article.  I'll use your comments as a springboard to launch another discussion question to the list - after I read your article I'll probably have more specific questions regarding your post.

I thought the first discussion on the use of experiential exercises in FAP was pretty interesting, although I think there are a lot more opinions out there.  It turns out it is hard to get consistent activity going on a list serve...

2)  How important is it to nail down functional assessment in FAP?  Put another way, what is an acceptable CRB?  Some examples:  No one will argue that CRBs identified by Glenn's FIAT system are the wrong thing to do?  But are there other "right" things?   What about CRBs not so identified?  Are you doing FAP if you really haven't specified what you're targeting, but you "know it when you see it"?  It is so easy to do treatment without a clear conceptualization, simply going after what comes up in the relationship.  Does this meet our standards for FAP (we all do it, except maybe Glenn)? In FECT, cognitive distortions, as long as the content of the distortion had to do with the therapy or the therapist, were considered CRBs.  Is this FAP?  (I'm purposefully pushing buttons on that last one.)

General thoughts about the relationship between functional assessment and FAP?

I ask this because I know it is important to Reno FAP; indeed it is important to behavior analysis in general.  It has been growing in importance in how I conceive it, but it is nowhere to be found in Kohlenberg and Tsai (1991), and in fact is not even in much of what Bill or other Reno FAPpers have written about FAP (although Bill has written extensively on functional assessment, he just hasn't tied it to FAP in these writings, as far as I know).

Discuss?

Jonathan



keithsonn wrote:

hi glenn,

I know it's been a while since you entered your thoughts on the FIAT
and reliability here. I didn't know about this forum until recently.
We met briefly at the FAP workshop in Seattle. I shared my excitement
about your work on establishing tools for assessing CRBs, by using a
standard language.

I agree that interpersonal repertoires comprise important clinical
issues. At the workshop, I referred you to my article -- Sonnanburg,
K. (1996), Meaningful Measurement in Psychotherapy, /Psychotherapy,
Special Issue: Psychotherapy Outcomes/, 33(2-Sum), 160-169 -- in hope
that it might provide a wider framework for yoour continued research
into these matters (e.g., adding indices of personal distress,
impaired performances, and health risks). Hope for a common language
is the central theme of that paper.

Although clients may evince idiographic profiles of operants and
respondents, the relationships between responses and their
contingencies are invariant. We do not need to prove the validity of
the methodology of FAP, but rather of the results (i.e., do the
changes effected matter?...what really has improved?). The neutral
technology of change cannot indicate its own targets.

You say that unique target behaviors make comparisons across clients
difficult, and so reliability measures will suffer. The demonstration
of reliability only establishes that measurements made can be
attributed to non-error variance (i.e., to valid variance in the
variables of interest). It is because target behaviors are unique that
we must first know that what we measure has valid meaning in the
clinical context.

Reliability coefficients correlate separate measurement events.
Measuring water at the boiling point, we expect replicable results;
measuring lake temperature in summer and then in winter, we expect
change. Validity coefficients, of course, are correlations between
measurements and some criterion to be predicted. What do we know about
changes expected in clients' CRBs? I believe that the primary question
is this? does the organism's repertoire enable flourishing in the
inhabited environmental niches (and by corollary, what does it mean to
flourish)?

There seems to be some consensus in the FAP community re: the implicit
assumptions that clinicians make about behaviors that are desirable or
undesirable. Apparently, face validity is a given. I believe your work
focuses on content validity (delineating a representative sample of
clinically relevant responses). The idiographic aspects of the FAP
approach should lend themselves well to concurrent validity (both
within and between subjects). There appears to be no "substantive",
"structural", nor "external" construct validity studies. After these
matters are settled, issues of generalizability can be explored.

You say that the FIAT-Q is internally consistent and has
convergent/divergent validity, but it is normed as nomothetic (moving
away from idiographic assessment). Norms are, by their nature,
nomothetic. Yet the roles that each normed behavior plays in a
client's life may still be idiographic. In my referenced article, one
suggestion I made was to use Q-sorts as a way to capture idiographic
measures that could then be subjected to various validity and
reliability tests.

You pointed out that reliability issues may result from flaws in the
insturments tested, but also from the metrics used to asses the
instrument. The basic forms of reliability are: stimulus consistency,
test-retest reliability, and inter-rater reliability. You report that
the FIAT-Q has internal consistency. Jonathan has complained about the
test-retest reliability of client reports. It sounds like measures of
inter-rater reliability are underway. Again, I don't believe that the
metrics for reliability are problematic; I would first suggest a focus
on issues of validity. This is what I appreciate about the work you
have done thus far.

I hope you find some of these remarks helpful. I would like to review
any materials on the FIAT or the FAPRS content that you are willing to
share. Thank you.

Keith Sonnanburg, Ph.D.
Seattle, WA
keson@...



--- In functionalanalyticpsychotherapy@yahoogroups.com,
"drglenncallaghan" <drglenncallaghan@y...> wrote:
>
> Hi all,
> Sorry I have been absent. My wife and baby-to-be are fine, but baby
> is impatient like her Daddy and is trying to come out to this world
> a little early. That has kept me very busy.
> I read the posts on the FIAT and FAPRS. Here's what I can say about
> the FIAT.
>
> I have worked on this idiographic system for a very long time, and
> it has real advantages and some problems. I am happy to provide it
> to those who request if from me specifically. I don't want it on the
> website yet. I have had the great fortune of woring with Jonathan
> Kanter on the FAPRS and FIAT, and I look forward to years of
> collaboration with him and with all FAP-minded people out there. I
> think what we are trying to do is unique and very important to the
> field of behavioral therapies, if not psychotherapy.
>
> Herein is the problem; what we are doing is unique and actually
> requires a very thoughtful approach to the question of assessment
> and measurement. The issue of parametrics is a tough one, and I have
> been fortunate in the kind of job I have that I can explore these
> issues and still have a job to go to without having to solve them
> and publish quickly.
>
> The FIAT is, at risk of tooting my own horn,
> the most innovative assessment idea I have worked on. It seeks to
> create unique tailored individual assments using a common language
> for problems but attempts to retain the idiographic analysis of each
> person.
>
> In other words, the FIAT is attempting to capture what is important
> about
> behavior analysis and behavioral assessment, more specifically
> funcitonal analysis and provide a common language to do this. At
> least, I am trying to do this by starting with FAP. I have already
> started using it for interpersonal problems assessed and treated
> bheaivorally and congitive behaviorally (you heard it here, I am
> still involved with the C part of CBT) including depression and
> PTSD.
>
> Now the big problem we all know about (as does Haynes, not
> necessarily Hayes, and others) is that funcitonal and idiographic
> analyses creat apples and oranges comparison problems. This means
> that standard approaches to psychometrics such as basic measures of
> reliability will struggle.
>
> Here is the point I want to make though: This is a problem with the
> metrics we imported into psychotherapy research. Reliability MUST be
> addressed differently for this kind of work. The failure of the FIAT
> to get reliability at this stage is possibly the insturment, but
> also the metric we use to assess the instrument. In no way am I
> saying the instrument is unflawed. It is being revised as each
> researcher tries to use it.
>
> I am saying that our measures of reliability are flawed. I spent 3
> years creating a questionnaire version of the FIAT, we call (not so
> cleverly) the FIAT-Q. This instrument has great psychometrics with
> respect to internal consistency and convergent and disciminant
> validity. I began to write it up and realized the FIAT-Q actually
> gets me and the users further away from the purpose of the FIAT,
> namely idiographic assessment. Instead, the insturment is attempting
> to be normed, and is, in essence, nomothetic. I only use the Q as a
> set of quesitons to get me to the FIAT and the ongoing assessment of
> my clients.
>
> OK, so that was a lot. To summarize:
> Idiographic hard
> FIAT good, but challenges exist
> Standard, nomothetic assessments not bad, but not good for behavior
> analysts
> FIAT Q good by normal standards, bad for us
>
> Now, I should say, "any thoughts?"
>
> Glenn

-- Jonathan W. Kanter, Ph.D.
Assistant Professor and Clinic Coordinator
Department of Psychology
University of Wisconsin-Milwaukee
P.O. Box 413
Milwaukee, WI 53201
Office: Garland Hall 238C
(414) 229-3834