Alexis Gatt
2004-12-08 18:53:24 UTC
Hi everyone,
I used the signal detection theory to analyse some of my experimental data.
Observers were asked to perform two distinct tasks, first to discriminate
between two stimuli, and then to identify the type of one of them. I have
performed all the calculations following the explanation given in the
excellent new version of Macmillan and Creelman's book, but as I am
unexperienced in this field, I am unsure on how to interpret the results. So
I would like to ask advices from some of you who are more experienced with
SDT.
The results for the discrimination experiment is d'=2.4. How would you
qualify such a performance? Would it be "reliable discrimination ability",
"good discrimination ability", "very good discrimination ability"? If that
helps, the corresponding p(c) is 0.61 for an unbiased observer. As I used a
same-different design, the equivalent p(c) for a simple yes-no design would
be 0.76.
For the identification experiment (which is a yes-no design in this case),
the d' falls to 1.01, and the p(c) is 0.57 is this case. Same question: how
would you interpret the results in this case?
Many thanks
Alexis
'`'`'`'`'`'`'`'`'`'`'`'`'`''`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`
sci.psychology.research is a moderated newsgroup.
Before submitting an article, please read the guidelines which are posted
here bimonthly or the charter on the web at http://psychcentral.com/spr/
Submissions are acknowledged automatically.
I used the signal detection theory to analyse some of my experimental data.
Observers were asked to perform two distinct tasks, first to discriminate
between two stimuli, and then to identify the type of one of them. I have
performed all the calculations following the explanation given in the
excellent new version of Macmillan and Creelman's book, but as I am
unexperienced in this field, I am unsure on how to interpret the results. So
I would like to ask advices from some of you who are more experienced with
SDT.
The results for the discrimination experiment is d'=2.4. How would you
qualify such a performance? Would it be "reliable discrimination ability",
"good discrimination ability", "very good discrimination ability"? If that
helps, the corresponding p(c) is 0.61 for an unbiased observer. As I used a
same-different design, the equivalent p(c) for a simple yes-no design would
be 0.76.
For the identification experiment (which is a yes-no design in this case),
the d' falls to 1.01, and the p(c) is 0.57 is this case. Same question: how
would you interpret the results in this case?
Many thanks
Alexis
'`'`'`'`'`'`'`'`'`'`'`'`'`''`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`
sci.psychology.research is a moderated newsgroup.
Before submitting an article, please read the guidelines which are posted
here bimonthly or the charter on the web at http://psychcentral.com/spr/
Submissions are acknowledged automatically.