Research Methodology 4.1 Introduction
4.4 Research Method used in this Study
This study adopts a survey approach. Survey research is a frequently used mode of observation in the social sciences. Surveys can be used for descriptive, explanatory and exploratory purposes. Survey research is described as “probably the best method
are an effective vehicle for measuring peoples’ attitudes, beliefs, opinions, motivations, and behaviour. Survey research is widely regarded as being inherently quantitative and positivistic (De Vaus 2002).
The term ‘survey’ is defined by Bryman (2004, p. 43) as a:
Cross sectional design in relation to which data are collected predominantly by self- completion questionnaire or by structured interview on more than one case and at a single point in time in order to collect a body of quantitative or quantifiable data in connection with two or more variables which are then examined to detect patterns of relationship between variables.
The main advantage of the survey method is that it rarely requires a study of the whole population. Random statistical samples are commonly drawn from a population and inferences can then be made from the sample responses to the population. Therefore, surveys provide an efficient and economical means of gathering and analysing data about people’s attitudes and orientations. The contribution of survey research to the methodology of the social sciences is emphasised by Kerlinger (1986, p. 381) when he describes its most important contributions as:
…rigourous sampling procedures, the overall design and the implementation of the design of studies, the unambiguous definition and specification of the research problem, and the analysis and interpretation of data.
Weaknesses of survey research may commonly involve: a lack of depth in the survey data; large investments of time and resources; sampling errors; inadequate measurement that may produce unreliable and invalid results; and, non-response bias (Kerlinger 1986). For the current study, the treatments of these issues are dealt within this chapter.
Surveys can be broadly classified into three categories: mail and electronic questionnaire, telephone interview, and personal interview; they are the most common method of collecting survey data (De Vaus 2002). Each survey method has its place in social research. After careful consideration of all the advantages and disadvantages that each method offers, a well balanced choice needs to be made so as to meet the requirements of the research.
4.4.1.1 Mail Questionnaire
Mail questionnaires are posted out to respondents who then answer questions by completing the questionnaire personally and returning it by mail. Mail questionnaires can be administered quickly and usually at a relatively lower cost than the other two survey methods. This advantage is enhanced when the researcher is dealing with large geographically dispersed samples, however, the questionnaires may take several weeks before they are completed and returned. Another noted advantage is that there is an absence of interviewer impact on answers provided by the respondent (David & Sutton 2004). The respondent is in no way influenced by an interviewer when completing the questionnaire. Studies have found that the presence of any interviewer at all can produce biased responses (Groves, Fowler, Couper, Lepkowski, Singer & Tourangeau 2004). Furthermore, as there is no interviewer present, there is no variability in the order and way the questions are asked. Respondents also decide the time and speed at which they respond to the questionnaire. This convenience may further assist the response rate.
One of the limitations of the mail questionnaire method is the response rate. Generally mail questionnaires will typically result in low response rates. It is suggested that the main disadvantage of the mail survey is often the low response rate obtained (David & Sutton 2004). Czaja and Blair (2005) caution that mail surveys are prone to response bias because of low response rates, as a result, non-response may bias the research results. Further, errors may arise as the researcher has no control over the process as they do not get the opportunity to prompt or probe. Partially answered
being unable to collect additional data and the fatigue factor experienced by respondents answering long questionnaires (Bryman 2004).
4.4.1.2 Electronic Questionnaire
A recent variation of the mail-out questionnaire is the electronic mail questionnaire. This can involve distributing the questionnaire as an attachment forming part of an electronic mail (email). With this method the survey is formatted with either word processor or with web-formatting language and attached to an email. The questionnaire is answered and returned by email by the respondent. A more recent alternative is to send an email with a hyperlink to a website containing the questionnaire. This type of electronic survey is often described as a web-based survey.
Many studies have compared response rates on mail-out and web-based surveys, and the results are mixed. A meta-analyses of thirty-nine study results published in the last decade shows that mail surveys have higher response rates than web-surveys in general (Shih & Fan 2008). Kaplowitz, Hadlock and Levine (2004) argue that web- surveys can achieve similar response rates to a mailed questionnaire. Whereas, Cobanoglu, Warde and Moreo (2001) found, that when respondents were given choice between completing a mail or web-based survey, the web-based option produced a higher response rate. Porter (2004, p. 9) suggests that there …“is difficulty in drawing conclusions about response rates from such studies”.
One of the major sources of error in any survey is non-response (Archer 2007). Errors relating to non-response result when not all of the potential respondents complete the survey lead to non-response bias. Crawford, Couper and Lamias (2001) argue that non-response represents the biggest challenge for web-based surveys. It is suggested that an acceptable response rate for a mail questionnaire is above 50 percent (Mangione 1995). However, response rates between five and ten percent are common (Alreck & Settle 1985). Schonlau, Fricker and Elliot (2001) conducted a review of
2008; Solomon 2001; Cook, Heath & Thompson 2000).
Most of the advantages and limitations of mail-out questionnaires also apply to electronic questionnaires. However, electronic surveys can be administered at a greater speed via the internet than mail-out questionnaires and may save time for the researcher (Wright 2005). When completed and submitted, the responses to the questionnaire are received immediately, either through a return email or a dedicated data base. Furthermore, responses are also likely to be received much earlier as they are returned electronically. Kwak and Radler (2002) found that a web-based survey had significantly shorter turn-around time compared to a mail survey. Web-based surveys are now relatively inexpensive and achieve satisfying response rates compared to the classical mail-out questionnaire (Ganassali 2008). The use of electronic questionnaires eliminates the need for paper and other costs such as printing, postage and data entry (Llieva, Baron & Healey 2002).
4.4.1.3 Telephone Interview
Telephone interviews involve making telephone contact with a selected sample and asking questions over the telephone. These interviews are usually administered from a central location as conducting the interviews from a centralised location allows for quality control and greater supervision of interviewers. When compared to personal interviews this method is cheaper and also quicker to administer. The cost saving arises from the fact that interviewers do not have to travel and thus the avoidance of related costs. This method provides the opportunity to interview a geographically dispersed sample more quickly. Moreover, better responses to embarrassing or personal questions may be achieved without face-to-face contact (McMurray, Spoonley, Pearson, Pace & Scott 2004). However, it is not possible to interview people that are not contactable by telephone and further difficulty may also be encountered when interviewing the hearing impaired. Frey (2004) claims that telephone interviews are less sustainable in time compared to personal interviews. He also claims that telephone interviews achieve slightly lower response rates when
therefore respond to facial or other signs of confusion or unease when asking a question (Bryman 2004).
4.4.1.4 Personal Interview
The personal interview method involves the collection of data from a respondent often using a face-to-face interview with a trained interviewer. Because of the face-to-face contact the quality of the responses in interviews is superior to mail-out questionnaires because of the depth and quality of data that can be collected. This approach lends itself more to open-ended questions where more detailed information can be gathered. The interviewer has the ability to provide explanations to questions where necessary and to probe further if they deem it important. This approach also provides the interviewer with the opportunity to use visual aids to assist with responses. The personal interview is usually the most costly survey method, but it is likely to produce a better response rate compared to other survey methods as respondents are less likely to terminate a face-to-face interview. Disadvantages of this method include a lack of anonymity which may cause some problems when dealing with certain confidential or sensitive questioning. Furthermore, an interviewer may introduce biases when questioning and subsequently influence responses.
4.4.2 Strengths and Weaknesses of Survey Research
Surveys, as with other data collection methods, have both strengths and weaknesses. These need to be kept in mind in determining whether a survey is appropriate for the research objectives. Drawing partly on Babbie (2008), Bryman (2004) and DeVaus (2002), an overview was developed to summarise the strengths and weaknesses of survey research. This overview is presented in Table 4.2.
Large samples are feasible Time consuming Useful in describing characteristics of a
large population
Can have low response rates Can be administered from a remote
location
The participants may have problems recalling information
Reliability is strong Respondents may not answertruthfully
on controversial or threatening questions
Can be inexpensive The design is generally inflexible
Data are collected in a standardised format
Is comparatively weak on validity