For my Master's thesis, I need to measure participants' state anxiety before the presented stimuli (synonyms of the same word) and immediately after. There will be very little time, less than 5 minutes between the measures. Because of the short time frame, I would rather not use the same self-report measure both times. However, I need to be able to compare the two so they have to be comparable measures or different versions of the same measure. This is not my primary DV, but one that I believe will be a moderating factor. Does anyone have any suggestions? I could really use your help. Thank you.
Dear Kari,
The Spielberger State Anxiety Scale is widely used. It has 10 items. You could take your full sample size and randomly split it into two sets of items. Itemsubset 1 and Itemsubset 2. You would also create two random samples of subjects from your total sample size, Sample 1 and Sample 2. You could then give Sample 1, itemsubset 1 at pre-test and itemsubset 2 at post-test and reverse this for Sample 2. You should have decent intercorrelations of the items in both Itemsubset 1 and 2, but the alphas may not be great because of only 5 items in each set. I would just calculate the coeff alphas using a canned program and then ask for the mean intercorrelation of the items in each case. There should be good test-retest reliability on the parallel forms of the Spielberger that you have created and the test-retest correlation should be less biased because you will not be using the same items. However, if some of the great people on here know of a state anxiety measure (say the SCL or symptom checklist) that has parallel forms for anxiety, than that might be easier.
Good luck.Add your answer
You can use the tests like LIEBOWITZ SOCIAL ANXIETY SCALE but you can also do much complicated tests. like measure/monitoring of the heart beating , hormons or voice.
Thanks Aleksander. Unfortunately we do not have access to the equipment needed to do the biological measures, but I will check out the scale you suggested.
Dear Kari,
The Spielberger State Anxiety Scale is widely used. It has 10 items. You could take your full sample size and randomly split it into two sets of items. Itemsubset 1 and Itemsubset 2. You would also create two random samples of subjects from your total sample size, Sample 1 and Sample 2. You could then give Sample 1, itemsubset 1 at pre-test and itemsubset 2 at post-test and reverse this for Sample 2. You should have decent intercorrelations of the items in both Itemsubset 1 and 2, but the alphas may not be great because of only 5 items in each set. I would just calculate the coeff alphas using a canned program and then ask for the mean intercorrelation of the items in each case. There should be good test-retest reliability on the parallel forms of the Spielberger that you have created and the test-retest correlation should be less biased because you will not be using the same items. However, if some of the great people on here know of a state anxiety measure (say the SCL or symptom checklist) that has parallel forms for anxiety, than that might be easier.
Good luck.Add your answer
I agree, too with the STAI. It has good properties. My feeling is that it should work as well in repeated administration as only once. Subjects think about each question and answer from their own experience, so familiarity with the questions should not cause a problem.
- Badges
- Science topic
- Similar topics
- Psychology
- Psychometrics