Methods and Models for Surveying Using the Internet

Introduction – Trend In Data Collection

Conducting a statistical survey is complex, costly and lengthy process. Traditionally, social surveys were done on paper. It needed a lot of data to get the best results which were leading to substantial consumer budget only for data collection. Traditionally, surveys were based on three main methods of data collection: face to face interviews, telephone interviews and mail interviews. Each of these methods has been revolutionized by the widespread adoption of information technologies.

The transition from traditional methods computer to the computer assisted one took place in two stages (Tourangeau, 2004).

The development of computers has led to the development of two methods: Computer Assisted Telephone Interviewing (CATI) and Computer Assisted Personal Interviewing (CAPI) - sounding method using electronic devices). CATI method became popular in the 75s and CAPI began to take shape in the early 80s with the advent of portable computers (laptops). Those technologies have paved the way for the emergence of other methods: surveys via e-mail on the Internet (EMS - email survey), or using interactive voice response (Interactive Voice Response - IVR – a computer plays a pre-recorded questions by phone). The emergence of these computer-assisted methods not only transformed the classic trio (direct survey, by telephone or mail), but also created a series of hybrid approach of the respondents (Tourangeau, 2004).

Once portable computers began to be more easily manipulated and provided greater autonomy, it began a change of the traditional method of direct interview. The method CAPI gradually began to replace method PAPI at the end of 80s and throughout the 90s. Although this change was more visible to respondents (because they use computers instead of paper), the transition from PAPI to CAPI does not seem to have had a major effect on costs, response rates, sampling or responses of respondents like in the case of the methods who are using the phone (according to research conducted by Baker, Bradburn, & Johnson, 1995; Tourangeau, Rasinski, Jobe, Smith, & Pratt, 1997).

Mail surveys were the least affected by the first phase of technological change. The vast majority of large surveys (the scale censuses) also benefited from technical developments, but not by changing the procedures of collection answers but what looked centralizing data by introducing the technology of scanning and optical character recognition of the paper.

The second stage of computerization of the data collection process has made a more radical change by removing the interviewer and the administration of questionnaires by computer. This change had a much greater impact on costs and survey results.

As shown in the literature, one of the disadvantages of direct interview is even the interviewer. A large number of cases show that respondents may be reluctant to disclose information for an interview. Several national surveys collect information on illicit drug use, risky sexual behaviors, diseases and other potentially embarrassing topics and respondents' reluctance to be honest about such subjects in face to face interviews is a clear limitation of this method of data collection (De Leeuw, Hox, & Kef, 2003). This problem also affects the telephone interviews. Computer assisted surveys have overcome these problems and use complex questionnaires without the help of human operators, which is a great advantage when studying sensitive topics.