Reliability Testing - split halves method

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Reliability Testing - split halves method
Message posted by Kimberly on June 9, 2000 at 12:00 AM (ET)

HELP!

I want to use the split half method to test for reliablity on a questionaire I used for a research project.

I have split the questionaires into two groups and found mean scores for each group.

I understand that I want to compare them..... But against what? How close should they be? How do I figure this out>

Any help would be appreciated.

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Reliability Testing - split halves method
Message posted by Phil on June 9, 2000 at 12:00 AM (ET)

My understanding of the split halves method is that it is for checking the internal reliability of "tests", not "questionnaires". You score each half of the test and then calculate the correlation coefficient between the first half and second half scores. The higher the score, the greater the reliability.

This method doesn't make sense for questionnaires, since they are not a "test".

For questionniares, you can test reliability with the Cronbach Alpha method. Refer to a Research methods book or look around the web.

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.