1. Questionnaire-Part 2

2. Translating a questionnaire Quality of the obtained data increases if the questionnaire is presented in the respondents’ own mother tongue The translated version of a questionnaire is usually the only version that the respondents see. (You don’t need to provide them with the English version) For that reason, the quality of translation is very important.

3. Translation can be done by the researcher himself/herself However, if it is administered without further study on it, there may be some problems in some items or the whole scale In that case, the researchers cannot do anything else, but exclude the items in the questionnaire

4. So, while translating there are mainly two things we should keep in mind: the need to produce a close translation of the original text so that we can claim that the two versions are equivalent, the need to produce natural-sounding texts in the target language, similar to the words people would actually say.

5. How can we do this? a) After the initial translation is completed, consult external reviewers b) recruit an independent translator to backtranslate the target language version into the source language. back-translation: involves an independent translator turning the L2 version of the questionnaire back into the source language and then comparing the two texts: If the back- translated version corresponds with the source language version, this is an indication that both instruments are asking the same questions, which attests to the accuracy of the translation

6. Piloting the Questionnaire In questionnaires so much depends on the actual wording of the items (even minor differences can change the response pattern) Thus, “field testing,” that is, piloting the questionnaire at various stages of its development on a sample of people who are similar to the target sample is important These help the researcher to collect feedback and make alterations and fine-tune the final version of the questionnaire.

7. Item Analysis Item analysis can be conducted at two different points in the survey process: After the final piloting stage: in this case, the results are used to fine-tune and finalize the questionnaire. After the administration of the final questionnaire: such a “post hoc analysis” is useful to screen out any items that have not worked properly.

8. The procedures in both cases are similar and usually involve checking three aspects of the response pattern: 1. Missing responses and possible signs that the instructions were not understood correctly If some items are left out by several respondents, that should serve as an indication that something is not right: Perhaps the item is too difficult, too ambiguous, or too sensitive; Perhaps its location in the questionnaire is such that it is easily overlooked.

10. Two main ways of dealing with missing data: Listwise deletion: one missing value deletes a whole case from all the analyses even if some of the available data could be used for certain calculations. Pairwise deletion: temporary deletion of a case from the analysis only when specific statistics are computed that would involve the particular missing value.

11. 2. The range of the responses elicited by each item. We should avoid including items that are endorsed by almost everyone or by almost no one, because they are difficult if not impossible to process statistically (since statistical procedures require a certain amount of variation in the scores) Although the lack of variation may well be the true state of affairs in the group, it may be useful in many cases to increase item variation by adding additional response categories or rewording the question.

12. 3. The internal consistency of multi-item scales. Questionnaires should contain multi-item scales, rather than single items, to focus on any particular content domain Multi-item scales are effective if the items within a scale work together in a homogeneous manner; that is, if they measure the same target area In psychometric terms this means that each item on a scale should correlate with the other items and with the total scale score, which has been referred “Internal Consistency” Following this principle, a simple way of selecting items for a scale is to compute correlation coefficients for each potential item with the total scale score and to retain the items with the highest correlations.

13. A word of caution: Before we discard an item on the basis of the item analysis, we should first consider how the particular item fits in with the overall content area of the whole scale Automatic exclusion of an item suggested by the computer may lead to narrowing down the scope of the content area too much If a problem item represents an important dimension of the targeted domain, we should try and alter its wording or replace it with an alternative item rather than simply delete it.