CAPI and CATI as forms of electronic questionnaires

CAPI, CATI and all forms of electronic questionnaires have a number of advantages over paper questionnaires which have already been touched on in Chapter 2. Electronic questionnaires from all of the major software suppliers can:

• cope with complex routeing;
• rotate or randomize the order in which questions are asked;
• rotate or randomize the order in which responses are displayed;
• adapt questions depending on answers to previous questions;
• adapt response lists depending on answers to previous questions.

The programs generally offer a range of standardized formats that can be customized to the research organization’s conventions and layouts. This means, however, that many of the issues of layout are predetermined and thus taken out of the hands of the questionnaire writer. It also means that interviewers become used to a common format, which should reduce interviewer error.

The issues that remain are not dissimilar to those encountered with paper questionnaires, namely ensuring that all of a question and its responses appear on one page or screen so that the interviewers can read questions and response codes easily, and distinguish between questions to be asked and instructions to themselves.

However, electronic questionnaires should not just be seen as paper questionnaires transferred to screen. They offer many opportunities for questionnaire writers to be more creative in the way in which they ask questions, to ask more complex questions that do not appear to be so, and to use prompt material that would not otherwise be possible.

1. ‘Don’t know’ and ‘Not answered’ codes

CAPI and CATI questionnaires will tend to have ‘Don’t know’ or ‘Not answered’ codes for most questions. The interviewer may not be able to proceed to the following question without entering a response, and the respondent may refuse any answer other than a ‘Don’t know’ or refusal. Where the answer is used for quota purposes or the responses are to be used for routeing, these codes may be omitted. Even then, the question­naire writer should have a strategy for routeing the genuine ‘Don’t knows’ from such questions.

The inclusion of a ‘Not answered’ category on all questions as a matter of course is a question of individual preference, but the author’s view is that it is likely to lead to interviewers too readily accepting refusals and ambiguities in response, with a consequent increase in lost data.

2. Checking the questionnaire

The questionnaire layout should always be thoroughly checked from the standpoint of the interviewer, the coder, the data entry, the data processor and, if it is intended for self-completion, the respondent.

Checking for sense and usability will be repeated as part of the pilot survey (Chapter 10). Before the pilot survey is reached, though, the ques­tionnaire should be thoroughly proofread, and all interviewer and route- ing instructions double checked. Routeing instructions in electronic questionnaires should be checked and checked again.

Source: Brace Ian (2018), Questionnaire Design: How to Plan, Structure and Write Survey Material for Effective Market Research, Kogan Page; 4th edition.