I am planning to do test retest reliability for survey, I want to know process, how to choose my sample and sample size, what is the duration between two tests? how to get the same sample again?
Thanks in advance
1. Enroll at least 30 participants (more is better) to fill your questionnaire at baseline.
2. Specifically identify at least 30 of the the individuals above to fill the same questionnaire again in 2 weeks (follow up). [1 week interval is also acceptable].
3. Find correlation between the two sets (baseline and follow up) using Pearson Correlation Coefficient or Intraclass Correlation.
4. Correlation of >0.7 (70%) and above indicates good retest reliability.
Good luck!
Both answers are sufficient if you a sample size of 70 it is much better and conclusive in case of t-test you will use z-score in place of t-score.
Should I tell them from the beginning that I will call you again after two weeks to fill the same survey again? And give them the same number in two different set two correlate between their answers? I mean how can I know individual X is same in the two sets? Thanks all for your responding
Hi Louai,
If you can follow up a good number of them then no need to tell them. Informing them beforehand might introduce element of recall bias--which is exactly what retest reliability seeks to avoid.
Like you said, you can give study numbers (written on a nicely designed card) and on the questionnaire to your participants. Tell them to kindly keep it safe as you might be interested in additional study [do not mention same survey] in a few days or weeks. Inquire about the study number when you administer the follow up.
Hope this helps.
Thanks for your answer..can you suggest me an article or reference for more information or explai in more details? Thanks again
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size