Using the Internet as a Survey Method
The Internet has advantages and disadvantages like any other survey method. It is often used because of its low cost and its ability to reach specific audiences. British Airways posted a questionnaire on its website exclusively used by the airline's Executive Club members and 9000 people responded within 9 months. Reuters has also been active in online research. A questionnaire was posted on its website to research the potential demand for a new online news service for Irish expatriates. Traffic was driven to it by the registration of key words on search engines. For example, anyone searching for 'Irish News' would have been offered a link to the questionnaire site. The commercial think-tank Future Foundation used banner advertising on search engines Yahoo! and Excite to drive people to a questionnaire, which yielded 826 respondents. A problem when using e-mail to survey populations is the absence of accurate lists. Even when lists can be found, researchers need to treat them very carefully as 'spamming'-the sending of junk mail-is perceived as very offensive by most e-mail users. Nevertheless, the medium can offer tangible advantages over other survey methods particularly when surveying those groups who are accustomed to spending time online, most notably buyers of high-tech products.
Questionnaire design
To obtain a true response to a question, three conditions are necessary. First, respondents must understand the question; second, they must be able to provide the information; and, third, they must be willing to provide it. Figure 4.5 shows the three stages in the development of the questionnaire: planning, design and pilot.
The planning stage involves the types of decision discussed so far in this chapter. It provides a firm foundation for designing a questionnaire, which provides relevant information for the marketing problem that is being addressed.
The design stage deals with the actual construction of the survey instrument and involves a number of important decisions. The first relates to the ordering of topics. It is sensible to start with easy-to-answer questions, in order to relax the respondent, and leave sensitive questions until last. Effective questionnaires are well structured and have a logical flow. Second, the type of question needs to be decided. Generally, three types are used: dichotomous questions (allow two possible answers, such as 'Yes' / 'No'), multiple-choice questions, which allow more than two answers; and open questions, where the respondents answer by expressing their opinions.
Great care needs to be taken with both the wording and instructions used in the questionnaire and its layout. Questionnaire designers need to guard against asking ambiguous or leading questions, and using unfamiliar words (see Table 4.3). In terms of layout, the questionnaire should not appear cluttered and, where possible, answers and codes should each form a column so that they are easy to identify.
The use of 'scales' is very common in questionnaire design. For example, respondents are given lists of statements (e.g. 'My company's marketing information system allows me to make better decisions') followed by a choice of five positions on a scale ranging from 'strongly agree' to 'strongly disagree'. 'Probes' are used to explore or clarify what a respondent has said. Following a question about awareness of brand names, the exploratory probe 'Any others?' would seek to identify further names. Sometimes respondents use vague words or phrases like 'I like going on holiday because it is nice.' A clarifying probe such as, 'In what way is it nice?' would seek a more meaningful response. 'Prompts', on the other hand, aid responses to a question. For example, in an aided recall question a list of brand names would be provided for the respondent. Coding involves the assignment of numbers to specific responses in order to facilitate the analysis of the questionnaire later on.
Once the preliminary questionnaire has been designed it should be piloted with a representative subsample, to test for faults; this is known as the 'pilot stage'. Piloting tests the questionnaire design and helps to estimate costs. Face-to-face piloting, where respondents are asked to answer questions and comment on any problems concerning a questionnaire read out by an interviewer, is preferable to impersonal piloting where the questionnaire is given to respondents for self-completion and they are asked to write down any problems found. Once the pilot work proves satisfactory, the final questionnaire can be administered to the chosen sample.