The "validity" and "reliability" refer to the way questions are written to obtain objective results from subjective perspectives (the people surveyed). They remain valid as long as they are applied in the original language to the validated test populations.

Changes in language or cultural understanding by either the questionnaire or the participants would require revalidation and tests for reliability.

For example, a questionnaire written in English and intended for participants within the Canadian culture with a certain level of education would be valid and reliable throughout Canada with that understanding of education and culture.

However, if it was translated into French, or used in England, or with a population with a different educational level, it would need to be re-validated to avoid errors in understanding and response.

This applies to studies conducted by interviewing people. If you mean a test protocol of physical observations (statistical analysis) then a validated protocol will remain valid for any similar test.

Steven Cooke, A questionnaire has been recognized internationally as a valid questionnaire. The questionnaire has also been tested for validity and reliability with the language in the country where the research will be carried out. However, because the test was carried out on a different population group than that which will be studied later. So the researchers re-tested the validity of the population group according to the research plan (n=20). If the results of the validity and reliability tests show that there are several invalid statement numbers, can the statements in the questionnaire still be used? because the questionnaire has been recognized internationally, even in the same population in other countries.

Wahyu Hidayati - no, the purpose of the test is to determine that validity, "recognized" previously internationally or not. It may be a good time to mention that in terms of the "international" acceptance. It doesn't mean that the questions need to be abandoned, but it DOES mean that they must be rephrased if you want to obtain valid and useful responses from your target population. Another approach would be to see if exclusion of the "problem" questions would seriously affect the overall survey. If not, you might just leave them out. It's much better to know that you need MORE data than to publish results based on data that is known to be suspect.

Thank you, Steven Cooke for your explanation.