For each resident there
are multiple incidents of ‘noncompliance’ over time. Would the
reason for noncompliance change from time to time? What are the higher
level reasons for the noncompliance? Would the relationship with the
staff person make a difference in the compliance? Would the presence of
other residents make a difference in the compliance? Will the resident
become more compliant with time at the center? What would be the
explanation for the change?
The question I was raising was a
general one. For the above research, descriptive statistics might be all
that is needed.
I agree that descriptive statistics can be a first step in any kind of
behavioral research. But that’s only the beginning. The next thing you
have to do is look at the scatter in the data. If it is so large that you
wouldn’t like to base any important decisions (about an individual) on
it, you have to ask why there is so much variability. Probably the most
common answer you will find is that your hypothesis about what is going
on is wrong. Another is that there is some important variable factor that
you don’t know about. Both of these answers require that you re-analyze
the situation to see how you can reduce the variability. That means
switching to a different model or revising the model you have. You
continue this process of revising and testing until the probability of a
correct prediction for an individual is as high as you can get it. Then,
if that’s not high enough to satisfy your ethical or practical standards,
you decide against taking the implied action.
Unfortunately, psychologists and others do not like to change their
models. If they do a study of the effects of anxiety on performance, and
the relationship is highly variable, they increase the population size
until they get statistical significance, which can always be done given a
large enough budget and freedom from time constraints. They do not ask if
“anxiety” is a useful concept, or if the performance they are
measuring is something people actually do. They assume that what they see
happening is real, not imaginary, and that the relationships they notice
are the ones that are actually important.
I am really confused about what role statistics can play in PCT research.
It sounds as though descriptive statistics is OK but that inferenctial statistics is not OK.
I thought that multiple case studies was the way to go, but now I am not so sure.
It is my understanding that the goal of PCT research is to discover what variables of experience a person is controlling by his/her actions in a particular situation.
For example, at the Children’s Center where I work, the children will often refuse to follow the directions of staff and staff will write up ‘a critcial incident’ describing the incident. I would like to study this from a PCT viewpoint.
It is this sort of ‘noncompliant’ behavior that has resulted in their parents, or caretakers, saying that the children were ‘unmanageable’ and which led the parents to say that the children could not longer stay in the home.
What role can statistics play in the study? This is a ‘treatment center’ and the willingness of the children to cooperate with adult requests are a common treatment goal.
I am really confused about what role statistics can play in PCT research. It is my understanding >that the goal of PCT research is to discover what variables of experience a person is controlling by >his/her actions in a particular situation.
For example, at the Children’s Center where I work, the children will often refuse to follow the directions of staff and staff will write up ‘a critcial incident’ describing the incident. I would like to study this from a PCT viewpoint. It is this sort of ‘noncompliant’ behavior that has resulted in their parents, or caretakers, saying that the children were ‘unmanageable’ and which led the parents to say that the children could not longer stay in the home.
What role can statistics play in the study? This is a ‘treatment center’ and the willingness of the children to cooperate with adult requests are a common treatment goal.
This took me back to Ed Ford’s work with kids who would get similar descriptions, asking them,
“What do you want?” I wonder if before deciding on what statistical test you wanted to use, it would be useful first of all, simply to tabulate the array of answers you would get to that question.
Now that I think about it I wonder if Ed ever did that while developing his method. In any case I think that kind of data gathering might go on (and eventually be useful in research) while the treatment program continues to work in whatever way you have developed.
A brief MOL Therapy kind of session (less than 10 minutes) after an incident might help find out ‘what he/she wanted’ when the resident didn’t follow the staff’s request.
For each resident there are mulitple incidents of ‘noncompliance’ over time. Would the reason for noncompliance change from time to time? What are the higher level reasons for the noncompliance? Would the relationship with the staff person make a difference in the compliance? Would the presence of other residents make a difference in the compliance? Will the resident become more compliant with time at the center? What would be the explanation for the change?
The question I was raising was a general one. For the above research, descriptive statistics might be all that is needed.
David
···
----- Original Message -----
From:
Robertson Richard
To: CSGNET@LISTSERV.ILLINOIS.EDU
Sent: Friday, December 26, 2008 1:24 PM
Subject: Re: Re.: Statistics in Psychology
[From Dick Robertson(2008.12.26.1216CDT0]
[Goldstein (2008.12.25.15:48 EDT)]
I am really confused about what role statistics can play in PCT research.
It is my understanding >that the goal of PCT research is to discover what variables of experience a person is controlling by >his/her actions in a particular situation.
For example, at the Children’s Center where I work, the children will often refuse to follow the directions of staff and staff will write up ‘a critcial incident’ describing the incident. I would like to study this from a PCT viewpoint.
It is this sort of ‘noncompliant’ behavior that has resulted in their parents, or caretakers, saying that the children were ‘unmanageable’ and which led the parents to say that the children could not longer stay in the home.
What role can statistics play in the study? This is a ‘treatment center’ and the willingness of the children to cooperate with adult requests are a common treatment goal.
This took me back to Ed Ford’s work with kids who would get similar descriptions, asking them,
“What do you want?” I wonder if before deciding on what statistical test you wanted to use, it would be useful first of all, simply to tabulate the array of answers you would get to that question.
Now that I think about it I wonder if Ed ever did that while developing his method. In any case I think that kind of data gathering might go on (and eventually be useful in research) while the treatment program continues to work in whatever way you have developed.