One frequent method used to diffuse sensitive items is to make the questionnaire anonymous. For example, in a student questionnaire that asked the learners to evaluate their language teacher and the course (Clément, Dôrnyei, & Noels, 1994), using similar items to the ones employed in the Gliksman et al. (1982) study just mentioned, we felt it unlikely that the 16/17-year-old teenagers in the sample were going to agree to give us honest answers without being assured about the anonymity of the questionnaires. Following the same reasoning -and particularly when legal considerations, such as local research regulations, also necessitate it - researchers often feel 'forced' to make the survey anonymous. The main argument to support this practice is that anonymous respondents are likely to give answers that are less self-protective and presumably more accurate than respondents who believe they can be identified (Kearney, Hopkins, Mauss and Weisheit, 1984). Anonymity, however, raises two issues:
• Opinions differ widely as to whether respondent anonymity actually fulfills its purpose in encouraging honesty and willingness to disclose. As Aiken (1997) summarizes, most adults will probably give the same answers to questionnaire items whether or not their responses are anonymous. For example, Sudman and Bradburn (1983) report on a large-scale postal survey of college graduates, in which the researchers placed the mailing label (which naturally contained the respondent's name) on the back cover of the questionnaires and sent these out in window envelopes. Out of the 40,000 recipients, only five objected to this procedure and scratched out their names. On the other hand, in situations when an honest answer might cause embarrassment or pose actual threat to the respondent, anonymity does obviously matter. Thus, the question to consider is whether our questionnaires really falls into this category.
• Anonymity may not serve the purpose of the investigation. More often than not the researcher would like to link the data from the questionnaires to data coming from other sources; for example, motivational data obtained by questionnaires is often correlated to achievement scores coming from end-of-term course grades or proficiency tests. Without any identity marking on the questionnaires, we simply cannot link someone's scores in the two dataseis. Similarly, if we are conducting a longitudinal investigation we would not be able to follow a person's development if all the answers gathered from the multiple subjects at a time were anonymous.
In sum, sensitive items and anonymity are a serious issue that needs to be considered right from the beginning. In Section 3.4.3, I will present some approaches that have been successfully used in the past to reconcile confidentiality with the need for identification for research purposes.
Was this article helpful?
You won’t want to miss this. Get Paid Taking Surveys In Your Spare Time. I know what you’re thinking, ‘Oh great, another get rich quick scheme’. WRONG! This is your guide to making money from home by participating in paid surveys on the internet.