Dear,
I built a survey in English and I have translated it to Arabic. How can I find the correlation between the English version and the Arabic version?
I looked on Youtube, but I could not find an answer, so please share any helpful guideline you know.
Thank you so much.
Hello Omar! you will have to collect responses in both languages and then analyze the correlation of the responses for each question in English and Arabic (if your samples in each language are supposed to answer in a similar way, you should then have a high correlation). You should also check that the survey maintains its reliability after translating it. One way to go is to calculate the internal consistency of both instruments and then compare their respective alphas using a Feldt’s (1969) test for equality of two independent alpha coefficients. I hope this helps, good luck.
Omar:
Another thing you can do is test for measurement invariance. To do this you need two samples, one that took your survey in Arabic and another in English. For each scale in your survey, you can do one or more invariance tests to show that the scales are equivalent. This can be done using IRT or SEM methods. The following paper has a section on measurement invariance, and provides citations to a number of sources.
Spector, P. E., Liu, C., and Sanchez, J. I. (2015). Methodological and substantive issues in conducting multinational and cross-cultural research. Annual Review of Organizational Psychology and Organizational Behavior, 2, 101-131.
Santiago's first suggestion is sound - indeed, it's the only way to know how strongly they correlate. Naturally you also want to establish things like the reliability and distribution of scores (or measurement invariance) on the Arabic translation, but no statistic based solely on one version or the other will answer your main question.
I believe that "best practice" among scale translators is to first translate the measure (as you have done), then have someone else (someone who does not know the original) do a "back-translation" into the original language. The original and the back-translation are then compared, and corrections are made for any items that have become garbled in the process. If there are many corrections, the process may need to be repeated.
- Badges
- Science topic
- Similar topics
- Psychology
- Psychometrics