Dear collegues,
I performed a cross-cultural study using two questionnaires (66 and 12 items). The original version of these questionnaires was used in the first country, and these tools were also translated into the second country´s language to be administered here. It was the first time the translated version was used for research purposes. The number of participants in the first country was 216, in the second 265. Is it required to perform confirmatory factor analyisis (for the two lignuistical versions separatedly), or is it enough to report internal consistency coefficients in this particular publication? What I am actually supposed to do to fulfill the required standards of reporting psychometric properties of translated questionnaires?
Thank you for your suggestions in advance.
It would be advisable to do one for each country and then a general one with multi-group invariance to evaluate the differences.
Ivana Mašková : I've seen numerous publications that collected data from various countries (with questionnaires being translated and back-translated). They report the psychometric properties for the whole dataset, not by country. I would recommend that you test for measurement equivalence. See below a paper that might help:
"Measurement Equivalence in Cross-National Research", Eldad Davidov, Bart Meuleman, Jan Cieciuch, Peter Schmidt, Jaak Billiet Annual Review of Sociology 2014 40:1, 55-75
The psychometrics properties are multiple and in multiples steps. It's not possible to identify all properties with only one test. What do you want made specificly ? The properties for a translate test is fucntion of the properties of the original test.
If you want, writ me privately.
CFA of just new version and u may have to do cross language validation,readability.
CFA is not suffisant. Before CFA, EFA is very very important and after Rasch model or IRT analysis is essential.
If you want discuss with me, only contact me at [email protected]
Hello Ivanna!
When you translate a test into another language you need to hold an entire validation of the new version of the scale. This means to ensure the validity and reliability of the scale. Whit the alpha you only asses internal consistency, and with CFA only which and how many factors are you assessing.
As some persons told you, you need to asses things like external validation (comparing results with an already validated scale), time stability or, a thing that I recommend to much, make an Item Response Theory analysis.
If you need anything, I'm glad to help.
Hi Ivana
I apologies for being such a pedant but I must remind you that it's the scores from a scale, and not the scale itself, that have psychometric properties. It is good to get into the habit of referring to properties of the scores, as when you write up results for publication this is something that a reviewer is likely to look out for.
VBW
Mark
If I understood correctly, you want to make sure of the validity of the translation. So, I suggest you make the comparison between the two tests: Test and retest on the same sample. The time between the two tests is 14 to 21 days.
- Badges
- Science topic
- Similar topics
- Psychology
- Psychometrics