A REVIEW OF SALIENT DIMENSIONS OF ON-LINE MARKET RESEARCH: IS ON-LINE RESEARCH BETTER?

V. Aslihan Nasir Management Information Systems Department Bogazici University, Turkey

ABSTRACT

Information technology, particularly the Internet, provides alternative ways of conducting marketing research. In this paper, the role of the Internet in different kinds of research and the dimensions of research design and methodology, including sampling, measurement and data analysis, are discussed. The newly emerging literature on the benefits and limitations of on-line marketing research is presented. It is concluded that on-line market research and traditional market research should act hand in hand.

INTRODUCTION Since the early 1990s, the Internet has drawn the attention of the media, academics and the business world. As of February 2003, it was estimated that over a half billion people worldwide had access to the Internet (Nielsen-NetRatings, 2003). According to Nielsen-NetRatings, the US accounts for 29% of global Internet access, followed by Europe with 23%, and Asia-Pacific with 13%. The decline in the cost of Internet connection due to the competition between Internet service suppliers (ISPs), the decreased prices of personal computers -even notebooks- and enhanced software programs are among the major drivers that lead to the increased popularity of the Internet.

1

Many businesses are being transformed by the advent of electronic business (ebusiness); likewise, market research is being transformed by on-line research. Since the Internet has the potential of being a revolutionary way of collecting data without geographic restrictions and high costs, it is time to use computers and networks to change research processes, and to promote consumer and market research. It is accepted that successful marketing strategies in the future will rely on the information a company has been able collect about its customer segments (Glazer, 1991). The more information a company has about its customers, the greater the value that it can deliver to them. Since most customer information is collected via the company’s marketing research department, the value that a company can deliver to its customers will depend largely on the strength of its marketing research and data collection efforts. This is the fact that lies behind the increased importance of on-line research. As Internet usage around the world steadily increases, the researchers’ investigation of computer-mediated communications as a device for conducting research and collecting consumer data also increases. According to Johnson (2001), information technology leads to new methods of studying the digitized consumer. Miller and Dickson (2001, p. 139) define on-line research as “research conducted over the Internet, including mail surveys, web-browser-based surveys and concepts, on-line interviews and focus groups.” In other words, both qualitative and quantitative research designs can be conducted on-line. However, research designs that are conducted in on-line environments may lead to some drawbacks as well as benefits that are not common in traditional research methods. As the Internet, e-mail and other related technologies have diffused into a large population, researchers have begun to explore the possibility of using those technologies as a tool for marketing and consumer research. Some researchers have investigated the use of the Internet and e-mail as new survey medium (Dibb et al., 2001; Ilieva et al., 2002; Mehta

2

and Sivadas, 1995), while others have explored computer mediated interviewing and online focus groups (Curasi, 2001; Montoya-Weiss et al., 1998). Different qualitative research methodologies have emerged, such as Netnography (ethnography on the Internet), which adopts ethnographic research techniques to study the cultures and communities that are emerging through computer-mediated communications (Kozinets, 2002). Moreover, Rae (1999) examines web-based experimental designs and discusses the problems encountered. Furthermore, researchers have recently begun to conduct studies which compare traditional marketing survey tools (e.g., fax, mail and telephone) with on-line survey tools (e.g., e-mail surveys, web-base surveys) (Cobanoglu et al., 2001; Mehta and Sivadas, 1995). In their study, Schaefer and Dillman (1998) report on the development and testing of alternative sets of procedures for conducting e-mail surveys that build on knowledge of how to improve response to mail surveys. However, while some academicians see on-line research as a major change (Johnson, 2001), others have skeptical thoughts about this new research medium (Miller, 2001; Taylor, 2000). Researchers have begun recently to analyze on-line research method as a whole and to provide marketers with information that can assist in the design and implementation of computer-mediated research. As mentioned earlier, numerous researchers have recognized both the benefits and drawbacks that on-line research provides over traditional research. Whether it is a qualitative research design (e.g., focus groups, in-depth interviews) or a quantitative research design (e.g., surveys and experimental designs), all on-line research methods have some benefits and limitations. In the following parts, the benefits and limitations of on-line qualitative research are presented; then in the second part, benefits and limitations of on-line quantitative research are given. Finally, a general discussion and conclusion about the topic is presented.

3

THE BENEFITS AND LIMITATIONS OF ON-LINE QUALITATIVE RESEARCH On-line qualitative research includes on-line focus groups, in-depth interviews, bulletin boards and list-serve discussions. Chat or on-line focus groups offer real-time interaction between the participants and the researcher, and are therefore the most popular option for on-line qualitative research. On the other hand, bulletin boards offer the option of presenting well-crafted questions and well thought out answers. In-depth interviewing and focus groups are administered to better understand the experiences, opinions and interpretations of the characteristics of a phenomenon (Curasi, 2001). Since the Internet is a visual medium, it can be used for concept, advertising and package testing (Taylor, 2000). In this part, the features, disadvantages and advantages of on-line qualitative research of several research dimensions, such as sampling issues, data collection issues, data analysis and interpretation, reliability and validity, will be discussed. Some academicians assert that as on-line research is restricted to those who have access to the Internet it can not be used to represent the views of the general public (Dommeyer and Moriarty, 2000; Taylor, 2000). Moreover, according to Dommeyer and Moriarty (2000), Internet access is skewed to those people who have the following characteristics: high education, high income, residence in urban areas, and are younger in age. Therefore, it is particularly short of people who did not finish school, people with lower incomes and people over 65 (Taylor, 2000). On the other hand, Johnson (2001) supports on-line research and questions traditional research in terms of sample representatives. According to Johnson (2001), on-line research puts an end to the use of students as research respondents and thus increases validity. Another answer to the question of whether Internet users are a representative sample of the adult population comes from Miller and Dickson (2001). The authors offer demographic and propensity weighting as a method purported to make the people on-line look like the general

4

population. It is clear that there is no consensus among academicians and researchers about the representativeness of on-line research. At this point of the discussion, it will be beneficial to look at the recent studies about the demographic structure of Internet users. In the early days of e-commerce, on-line shoppers were generally wealthier, younger, better educated, more affluent and more techsavvy than the average off-line shopper, and they were predominantly male (Totty, 2003). However, according to the “Consumer Technographics” study of Forrester Research company (July 25 2003, p. 1), this trend begins to change, “as online consumers became the majority, their average age increased, their education level decreased, and their attitude toward technology relaxed considerably. ” Table 1 shows this change for U.S. consumers. “On-line consumers themselves didn’t change, they just came to resemble the overall U.S. population as they absorbed more and more average people into their ranks” (Forrester Research, Technographics Research, July 25 2003, p. 1).

Table 1 – Basic On-line U.S. Consumer Statistics: 1998 to 2003 1998

1999

2000

2001

2002

2003

100.0

101.0

102.1

103.2

104.3

105.5

Connected to Internet

25%

33%

43%

57%

61%

64%

Profile of on-line households

1998

1999

2000

2001

2002

2003

Average age

40.5

41.0

41.5

44.9

45.0

46.6

Male

57%

56%

53%

49%

50%

49%

College degree

49%

46%

44%

42%

42%

41%

Total U.S. Households (million)

Mean HH income (US $) 60,624 59,489 62,520 62,095 62,244 64,063 Agree with the statement: “Technology is important to me” Hours on-line per week for personal reasons Average years on-line

47%

43%

41%

33%

31%

30%

N/A

7.2

9.9

9.5

9.0

8.5

1.6

2.2

2.5

3.1

3.5

4.5

Have broadband at home

1%

2%

6%

10%

17%

23%

Source: Forrester Research, Inc. Consumer Technographics 1998-2003. Available [online] at www.forrester.com/Research

5

The reliability and validity of on-line data has also been questioned and found inferior to that of traditional research methods (Miller, 2001). Triangulation, the combination of multiple methods in a single project to increase the credibility of the findings, can be achieved by on-line focus groups and interviewing methods. “On-line interviewing can be effectively triangulated, or combined with other data collection methods to strengthen the credibility and conformability of the findings” (Curasi, 2001, p. 372). As an additional data collection method, on-line interviewing can be designed into a project to help assess its predictive, convergent and/or discriminant validity (Curasi, 2001). It has been argued that on-line interviews may be less valid than traditional face-to-face/in-person methods because on-line samples can be less representative, moderators have less control, and focus group and interview processes are less standardised. But the validity of on-line research improves when repeated measures are used to track customer preferences over time or when the focus is upon target segments (Miller and Dickson, 2001). It should be also noted that the interpretation of on-line data needs more caution due to the risk of sampling errors and non-sampling errors (Miller, 2001). In addition, special care must be given to evaluate the quality and ensure the validity of data collected in an on-line qualitative study. One of the most common concerns about online research is participant honesty, which causes inaccurate response (Montoya-Weiss et al., 1998). However, even in face-to-face settings, there is no guarantee of participant honesty. Additionally, because they can participate at their own leisure, on-line respondents actually tend to be very cooperative and honest. Certainly, establishing the researcher’s support for respondent confidentiality and honest research will help respondents feel more comfortable sharing valuable opinions and information (CASRO, 2000). There are also new technologies that can help to ensure that researchers are dealing with the correct on-line respondent when conducting research on-line.

6

Another point is that the low response rates and low level of voluntarism in on-line interviews and focus groups can be eliminated by simply having personal contact, either in-person or on the phone, in order to invite individuals to participate in on-line research (Curasi, 2001). Higher informant commitment or motivation is needed, since participants are required to exert more effort to complete an on-line interview. The more motivated the respondent, the better the interview data will be (Curasi, 2001). Moreover, cash and noncash incentives easily can be given during traditional research methods, however, it is impossible to include tangible incentives into computer-mediated environments (Cobanoglu et al., 2001). Furthermore, it is known that voluntarism (non-refusal, agreeing to participate in a focus group) is highly influenced by incentives provided to respondents (Miller, 2001). In addition to the limitations cited above, on-line qualitative research also has some benefits. For instance, on-line research has the advantage of recruitment speed. The speed of on-line research can create first-mover advantages, providing substantial economic returns and economic rents (CASRO, 2000). The elimination of travel costs and time required away from the office, respondents’ participation from the comfort of their own homes or offices are among the benefits of on-line research; in other words, when time, financial restraints and geographical boundaries limit an investigation, on-line qualitative research becomes crucial (CASRO, 2000). Clerical support and data entry are other costs that are not associated with on-line research (Ilieva et al., 2002). Another benefit is that open-ended questions result in more detailed, richer and longer replies than traditional face-to-face interviews and focus groups. Since on-line interviewers are not present to provide focus and direction, the length and amount of information is more dependent on the informants (Curasi, 2001).

7

Participants generally feel more comfortable and worry less about what others think of them when discussing controversial or sensitive issues (i.e. drug addiction, AIDS, sexual harassment, ovarian cancer), and therefore feel more comfortable expressing strong and negative feelings toward other participants when using the electronic approach. This is particularly attributed to the “anonymity” feature of on-line environments (Montoya-Weiss et al., 1998). In addition, the interviewer has less control over the success of the interview. In other words, on-line qualitative research is less susceptible to response bias (Miller and Dickson, 2001). Organizations have markets consisting of consumers with highly diverse opinions and preferences. Therefore, in order to gain a broader and more in-depth understanding of the target segment, market researchers must bring multiple individuals with various interests and perspectives together. Computer and communication technology can enable the conduct of focus groups composed of diverse individuals (Montoya-Weiss et al., 1998). Virtual groups do not allow for tactile touch/feel experiences or capturing the facial expressions (consternation, excitement, interest) of the respondents as they are exposed to ideas and concepts. However, the depth of transcripts that can be derived from just an hour of discussion is powerful — often 20-30 pages of verbatim responses, available immediately (CASRO, 2000). In other words, non-verbal paralinguistic cues can not be used online. However, there have been attempts to replace physical expressive functions (e.g., a smile) with special symbols and conventions in printed text in order to convey emotion and transmit subtle meanings. For example, “emoticons” (certain keys, which, when typed in combination, look like facial expressions) are used to exhibit emotion, while typing in all capital letters represents shouting (Montoya-Weiss et al., 1998). Furthermore, by using computer-assisted thematic content analysis to examine the transcripts, it is

8

observed that people use a higher proportion of strong words (both negative and positive) when working on-line (Miller, 2001). Follow-up probes can take place during the entire research process, so that the researcher is not limited to the probes that are remembered during the face-to-face interviews (Curasi, 2001). Moreover, respondents who participate in on-line research may allow for easier member checks, a technique of testing hypotheses, data, preliminary categories and interpretations with informants. Furthermore, it is easy to send informants a draft of the evolving manuscript for their comments (Curasi, 2001). On-line informants are able to read and reread their responses, making editorial revisions prior to returning their responses, but since on-line informants may exercise more caution over what is said, their responses may contain less spontaneity (Curasi, 2001). Another advantage of on-line qualitative research is that there is no need for transcription of the interview with an interview conducted on-line because the completed interview usually can be printed or entered into qualitative software program and soon be ready for analysis (Miller and Dickson, 2001). Unlike traditional focus groups, complete transcripts of the session are available minutes after the conclusion of the group. A complete qualitative report is then written by professional analysts and completed in less than a week. Finally, Nancarrow et al. (2001, p. 138) mention that the decision of whether or not to use Internet based research will depend on: intended use (idea generation on a subject, diagnosis, evaluation etc.), the topic (sensitivity), the characteristics of the population (eliteracy, geography, reticence etc.), the use of stimuli, special research techniques that might be needed, the importance of reading body language and tone of voice, the skills of the moderator, and budget and timing.

9

Up to now, the pitfalls and advantages of on-line qualitative research have been reviewed by referring to the literature on this topic. The discussion above highlights the challenges facing the research conducted in on-line environments. It is clear that information technology presents new opportunities and challenges, but this does not mean that on-line focus groups and interviews will or should totally replace traditional face-toface focus groups. Rather, it will be more appropriate when on-line focus groups and interviews are used as a part of a bundle of market research techniques. On-line research is more than just a new way to gather information, it represents a technological and cultural change. In the next part, the benefits and limitations of on-line quantitative research will be presented.

THE BENEFITS AND LIMITATIONS OF ON-LINE QUANTITATIVE RESEARCH On-line data collection methods increased significantly during the late 1990s. According to Ilieva et al. (2002), this was preceded by 1) a growing number of Internet and e-mail users, and 2) various computer-assisted data collection techniques such as CAPI (computer-assisted personal interviews) and CATI (computer-assisted telephone interviews). The on-line research debate about quantitative techniques mostly concentrates on on-line surveys (Ilieva et al. 2002; Sheehan and McMillan, 1999; Schaefer and Dilmann, 1998; Mehta and Sivadas, 1995; Oppermann, 1995). “Basically, there are three methods for conducting online surveys: e-mail surveys, HTML forms (HTML is an Internet programming language), and downloadable interactive survey applications” (Bowers, 1999). Investigation into the validity of on-line data collection has been grounded mainly in comparisons between on-line surveys and mail surveys (Cobanoglu et al., 2001; Miller et

10

al., 2001; Schaefer and Dilmann, 1998; Sheehan and McMillan, 1999). The analysis of the past research on the use of the Internet as a survey methodology has shown that there is a wide variation in response rates and the speed of response rates (Oppermann, 1995; Sheehan and McMillan, 1999). A consistent finding of the studies that compare the response speeds of surveys delivered via e-mail and postal mail is that e-mail responses are returned much more quickly than postal mail responses (Mehta and Sivadas, 1995; Oppermann, 1995; Schaefer and Dilmann, 1998). Cobanoglu et al. (2001) compare mail, fax and web-based surveys in terms of response speed and find that the fastest method is fax surveys followed by webbased surveys. Another widely studied issue about on-line surveys is the response rate. Although there is consistency about the response speed of e-mails, there is a lack of consensus among academicians about the superiority of response rates of e-mail surveys. Mehta and Sivadas (1995) and Schaefer and Dilmann (1998) find no significant differences in response rates between the two modes. There are other studies which find that e-mail and web-based surveys have significant advantages in terms of response rates (Cobanoglu et al., 2001; Oppermann, 1995). Moreover, Sheehan and McMillan (1999) discover that issue salience appears to have a positive effect on response rates for e-mail surveys and also prenotification appears to increase response speeds. Even though there is a broad body of literature about the comparison of the response rate and response speed of mail and email/web-based surveys, the quality of response of e-mail surveys has not often been analyzed by previous researchers. But those who have analyzed it find a high response quality. It has been shown to produce fewer item omissions (Schaefer and Dilmann, 1998) and a greater response to open-ended questions (Mehta and Sivadas, 1995; Schaefer and Dilmann, 1998). Therefore the hypothesis that e-mail and web-based surveys provide more complete and detailed information is supported by those studies.

11

Another crucial contribution about on-line surveys comes from Dommeyer and Moriarty (2000). The authors compare an embedded e-mail survey with an attached e-mail survey and find that the embedded survey yielded a significantly higher response rate than the attached survey; but no significant difference was found between the two methods on response speed, the number of item omissions or response bias. Although the embedded mail survey is the simplest method, it limits the researcher’s ability to formulate sophisticated questionnaires; that is, it is restricted to flat text format, and coloured or altered size, and graphic images and sound can not be used (Bowers, 1999). Another way of conducting an on-line survey is to send the questionnaire as an attachment in an e-mail (Dommeyer and Moriarty, 2000). Attached mail surveys can be given a format and appearance that will be inviting and pleasing to the respondent. One of the benefits of on-line quantitative research is that variable costs for data collection are much lower for on-line research. In addition, the transmission costs for online research are also very low. Furthermore, the speed of data collection is its significant advantage because of instant access to a wide audience, irrespective of its geographical location (Ilieva et al., 2002). If accurate response speeds are required for an international study, all countries must receive the questionnaire at the same local time (Miller, 2001). Also, weekends should be avoided for questionnaires targeted at the workplace, but may be preferred for those targeted at consumers in their own homes. Another advantage of on-line surveys is that it is more convenient to handle e-mail surveys. E-mail files can be saved into the outgoing mailbox and sent on to a large number of addresses simply by entering the respondent’s address (Dibb et al., 2001). However, postal mail surveys take many hours to prepare. Copying, labeling, folding, stuffing the envelopes and preparing the return envelopes takes considerable labor and financial resources (Cobanoglu et al., 2001). Luckily, once the survey is prepared, the e-mail

12

questionnaire can be sent to the all respondents with a click of the mouse. Furthermore, the sender is immediately notified of incorrect addresses and responses are returned with a notification of time and date (Dibb et al., 2001). Due to the speed at which a large number of respondents can be recruited over the Internet and the relatively low cost of doing so, models were estimated with large sample sizes, enabling meaningful comparisons across consumer subgroups (Miller and Dickson, 2001). Distributed survey applications and multimedia-capable respondent workstations provide a rich environment for consumer research, including conjoint studies, concept testing, virtual shopping experiments, and advertising-copy testing for video and Web pages (Miller and Dickson, 2001). Moreover, respondent queries about completing the questionnaire can be handled quickly (Dibb et al., 2001). Another major gain of on-line research is that the data are directly loaded in the data analysis software, thus saving the time and resources associated with the data entry process (i.e., labor and time costs) (Ilieva et al., 2002). The possibilities of data input errors can be reduced and time saved by linking respondents’ replies directly to a spreadsheet or other data capture software. In many research contexts, on-line surveys may be more efficient and more effective than traditional telephone and mail surveys. Hard-to-reach respondents are often more accessible using the Internet, for example, for busy executives or professionals (Miller and Dickson, 2001). E-mail questionnaires are currently relatively unusual and may therefore arouse respondents’ interest and attention. In addition, e-mail respondents tend to be less inhibited in their answers, writing more open-ended questions (Dibb et al., 2001). Furthermore, more extensive open-ended comments were obtained from on-line panelists than from mall-intercept panelists. Parallel testing shows that panels of on-line respondents were more willing than mall-intercept respondents to express negative feelings about new

13

products (Miller and Dickson, 2001). More urgency may be attached to an e-mail message and it is less likely to be perceived as junk mail than a mailed questionnaire (Dibb et al., 2001). In addition, e-mails are perceived as an environmentally friendly means of communication (Dibb et al., 2001). On the other hand, on-line quantitative research also receives criticism. Many on-line samples and panels are subject to criticism as biased or non-representative. Sample bias may arise from the low sample coverage. Research suggests that those most likely to answer an e-mail survey are well educated, younger, upper/middle income individuals interested in technology (Dibb et al., 2001). On-line surveys present special problems. It is difficult to specify a sampling frame when there is no centralized listing of Internet users, and when the users are both mobile and global (Miller and Dickson, 2001). However, it is argued that with adequate sample sizes, on-line data can be weighted to be representative by demographic variables such as age, sex, education, occupation and region (Taylor, 2000). Low sample coverage is a particular problem for on-line consumer surveys because many households have no chance of being included in on-line samples; moreover, those targeted need to be sufficiently confident to use this medium (Miller, 2001). Therefore if the use of on-line samples is desired to make inferences about the general population then it must be recognized that at the present time fewer households have access to the Internet. To overcome the sample coverage problem, Schaefer and Dilmann (1998) suggest a multimode strategy of data collection, i.e. approaching them via e-mail and by post/mail. This study argues that different modes of on-line and postal surveys complement each other and there are increased advantages in them being used together rather than applied separately. At this point in the discussion, it is also recommended to remember the changing nature of the on-line consumers’ demographic structure, which is presented in Table 1.

14

On-line surveys are not based on probability sampling, but on “volunteer” sampling or “convenience” sampling (Taylor, 2000). Furthermore, non-deliverable surveys in Webbased and e-mail based surveys are significantly more than those for postal mail and fax methods (Cobanoglu et al., 2001). The main difference for this reason is that people change their e-mail addresses and Internet service providers (ISPs) much more frequently than their postal addresses or fax numbers. Moreover, there is no central listing of e-mail addresses; even if such a listing existed, sending e-mails to potential respondents of on-line surveys who have not agreed beforehand to participate in on-line research would be like “spamming” and would be unacceptable to many people (Miller, 2001; Bowers, 1999). According to Bowers (1999, p. 46), “the most important rule for recruiting this new medium is to get respondent permission first.” Respondents who receive a large quantity of e-mail may respond negatively to being surveyed by this route and may discard the e-mail as junk mail (Dibb et al., 2001). E-mail makes it difficult for respondents to remain anonymous and this may affect willingness to respond. Furthermore sending unsolicited email to potential research participants is inappropriate, and is illegal in some jurisdictions (Miller, 2001). Because on-line samples can be less representative, questionnaires are selfadministered or contain fewer items, and survey processes are less standardized, on-line surveys may be less valid than traditional surveys; but the validity of on-line surveys improves when repeated measures are used to track customer preferences over time (Miller and Dickson, 2001). The reliability of a scale or measure is the degree to which it yields consistent measurements and one way to assess reliability is to use test-retest designs: measuring respondent attributes twice using the same scale (Churchill, 1999). The correlation between test and retest scores provides an estimate of reliability. Another way to assess reliability is to administer parallel forms of a scale. “The correlation between

15

scores on the parallel forms provides an estimate of reliability. Reliability can also be estimated from a single administration of a multi-item scale: Do respondents with the same scale scores have similar patterns of item responses? If they do, then the scale is reliable or internally consistent” (Miller and Dickson, 2001, p. 148). However, the recent research suggests that there are high levels of test-retest reliability for both on-line and paper-and pencil methods (Miller et al., 2001). In other words, under favorable conditions, on-line and paper-and-pencil tasks are highly reliable and highly comparable. Targeted respondents may try to access an on-line survey more than once or give access instructions to family members, friends, or colleagues (Miller, 2001). However, the new technology can eliminate this obstacle by IP numbers (Internet protocol numbers, a kind of identity for computers). Another effective way is “to generate a unique Internet survey for each invitee. This can be done by using a random generator to append a small tack-on address to the main survey URL (web address)…In conjunction with cookies and other technologies, this method allows respondents to click on a hot link and go to the survey site and either start to the survey from the beginning or pick up where he or she left off. As the respondent attempts to enter the survey, the pre-determined address code is compared to a database. If the survey code is valid and has not been submitted previously, the respondent is admitted to take the survey” (Bowers, 1999, p. 47). In addition, not everyone checks e-mail on a daily basis and so, unless the date of reading the e-mail is checked against the questionnaire return date, the average response speed may be underestimated (Ilieva et al., 2002). On-line survey transmission times vary by season, day of week, and time of day (low in summer, high on Monday) (Miller, 2001). Therefore, an e-mail survey should be left open for 1 to 10 days, whereas HTML form surveys have to be displayed 3 to 15 days

16

(Bowers, 1999). Furthermore, the reasonable length for a survey is no longer than 20 minutes; longer surveys require greater incentives (Bowers, 1999). Moreover, the lack of anonymity may deter some people from responding, and also an e-mail survey is very easy to ignore and discard (Dommeyer and Moriarty, 2000). It is more difficult to include tangible incentives and this may lead to a reduced response rate (Cobanoglu et al., 2001; Dibb et al., 2001). On the other hand, the quality and appearance of an e-mail questionnaire may not be as good as those of a mailed questionnaire (Ilieva et al., 2002). However, the new technologies will overcome this barrier for e-mail surveys. Scrolling is another issue, especially for web-based survey pages with many items or many item alternatives. Scrolling, viewing additional pages, and using extra key clicks to move safely from one page to next can make survey tasks longer and more tedious (Miller, 2001). Depending on the design, surveys can be posted on the Web through e-mail or they can be posted on the Web and the URL provided to respondents who have already been approached (Ilieva et al., 2002). However, the respondents may not have: the hardware and software that will enable them to download, the ability to read and upload a foreign file, the knowledge of how to execute the various response steps, and who may fear computer viruses (Dommeyer and Moriarty, 2000).

CONCLUSIONS On-line research has already been discussed both enthusiastically and critically, in many different ways. As seen, on-line research is a fast moving area. On-line research will surely replace some of the qualitative and quantitative research work currently done in person or by telephone. However, this must not be interpreted as a total replacement of traditional research methodologies by on-line research. Practitioners and researchers will

17

continue to do in-person and telephone research; but the Internet enables researchers to do many things that they could not afford before: survey huge samples of people; survey even small sub-samples of the population; obtain much richer verbal replies; show images and moving images as well as audio; high transmission speed, response speed and response rate; affordable costs; researcher’s control of the sample; and data are directly loaded in the data analysis software, thus saving time and resources associated with the data entry process. However, there are also some drawbacks of on-line research which are, or will be, prevented by the enhanced technology: lack of measurement control and standardization; limited access to the Internet (sample coverage), especially for some parts of the world; respondent refusals, non-response, inaccurate response and incomplete survey response; and reliability and validity issues. The evaluation of on-line research shows that in the near future, when the various systems become more compatible and more and more people have access to the Internet, on-line research is very likely to emerge as a major means of conducting market and consumer research. Since the developing technologies may not be available to all members of a population, a single method would eliminate the chance of some members being selected. Therefore it is recommended that researchers use a mixed-mode method for surveys in order to overcome the problem of sample coverage. A combination of on-line and traditional techniques will positively affect the response rate, the representation of the respondents and data quality. In addition to decreasing costs and providing more timely data, a mixed mode strategy can reduce coverage error. E-mail surveys generate better response rates than Web-based surveys and they generate greater researcher control over the sample of respondents, avoiding multiple entries to the survey by the same person. But on the other hand, questionnaires are better

18

displayed, more attractive and easier to fill out when displayed in Web-based surveys. Therefore, establishing contact through e-mail and providing the questionnaire in HTML format (special text program) or sending the Web site address (URL) combines the advantages of e-mail and Web-based surveys and optimizes the use of data on-line collection. Researchers should determine whether on-line surveys can benefit from personalized techniques, post paid incentives, or the offering of survey results to the respondents. Furthermore, on-line researchers who work in cooperation with Web site designers, statisticians, database managers, and computer programmers can learn to employ appropriate and efficient methods for collecting and analyzing data. Academicians must begin to focus on why response rate, response speed and other methodological issues vary among e-mail surveys. Furthermore, the issue of spamming should be eliminated by asking the permission of potential participants. In addition, e-mail and web-based research can be successful in different branches of marketing. For example, on-line research can be a very attractive method for researchers in international marketing because of the low costs and fast response rates. Ethical issues also need caution when conducting on-line research. Market researchers have an obligation to conduct research in a responsible manner, obtain the permission of the respondents, provide the security of the data, and protect individual privacy. The future for the consumer and marker research will necessitate the integration of both on-line and traditional research. As a final note about the place of information technology in consumer research, it can be said that the methodologies of on-line research and traditional research are very similar; therefore on-line research can be accepted as a new mode of data collection. Hence, both on-line and traditional research should act hand in hand.

19

REFERENCES Bowers, Diane (1999). “FAQs on Online Research. ” Marketing Research. Winter 1998/Spring 1999, pp. 45-48. Brennan, M., Rae, N. and Parackal, M. (1999). “Survey-based Experimental Research via the Web: Some Observations. ” Marketing Bulletin. 10 (May), pp. 83-92. CASRO (2000). Frequently Asked Questions about Conducting Online Research: New Methodologies for Traditional Techniques. Available [online] at: http://www.casro.org/faq.cfm Churchill, G. (1999). Marketing Research: Methodological Foundations. 7th edition. Fort Worth: Dryden Press. Cobanoglu, C., B. Warde, and P. Moreo (2001). “A Comparison of Mail, Fax and Web-based Survey Methods. ” International Journal of Market Research. 43 (4), pp. 441452. Curasi, C. F. (2001). “A Critical Exploration of Face-to-Face Interviewing vs. Computer-Mediated Interviewing. ” International Journal of Market Research. 43 (4), pp. 361-375. Dibb, S., A. Rushmer and P. Stern (2001). “New Survey Medium: Collecting Marketing Data with E-mail and the World Wide Web. ” Journal of Targeting, Measurement and Analysis for Marketing. 10 (1), pp. 17-25. Dommeyer, C. and E. Moriarty (2000). “Comparing Two Forms of an E-Mail Survey: Embedded vs. Attached. ” International Journal of Market Research. 42 (1), pp. 39-50. Forrester Research (2003). Consumer Technographics 1998-2003. July 25. Available [online] at http://www.forrester.com/Research Glazer, R. (1991). “Marketing in an Information-Intensive Environment: Strategic Implications of Knowledge as an Asset. ” Journal of Marketing. 55 (Oct.), pp. 1-19.

20

Ilieva, J., S. Baron, and N. Healey (2002). “Online Surveys in Marketing Research: Pros and Cons. ” International Journal of Market Research. 44 (3), pp. 361-376. Johnson, E. J. (2001). “Digitizing Consumer Research.” Journal of Consumer Research. 28 (Sept.), pp. 331-336. Kozinets, R. (2002). “The Field behind the Screen: Using Netnography for Marketing Research in Online Communities.” Journal of Marketing Research. 39, pp. 61-72. Mehta, R. and E. Sivadas (1995). “Comparing Response Rates and Response Content in Mail vs. Electronic Mail Surveys.” Journal of Market Research Society. 37, pp. 429439. Miller, T. (2000). “Marketing Research and The Information Industry”. A.C. Nielsen Center for Marketing Research. Available [online] at: http://www.bus.wisc.edu/nielsencenter/research/rcfordistribution.pdf

Miller, T., D. Rake, T. Sumimoto, and P. Hollman (2001). “Reliability and Comparability of Choice-Based Measures: Online and Paper and Pencil Methods of Administration.” A.C. Nielsen Center for Marketing Research. Available [online] at: http://www.bus.wisc.edu/nielsencenter.com

Miller, T. (2001). “Can We Trust the Data of Online Research?” Marketing Research. Summer, pp. 26-32. Miller, T. and P. Dickson (2001). “On-line Market Research.” International Journal of Electronic Commerce. 5 (3), pp. 139-167. Montaya-Weiss, M., A. Massey and D. Clapper (1998). “Online Focus Groups: Conceptual Issues and A Research Tool”. European Journal of Marketing. 32 (7/8), pp. 713-723. Nancarrow, Clive, Pallister, John and Brace, Ian (2001). “A New Research Medium, New Research Populations and Seven Deadly Sins for Internet Researchers.” Qualitative Market Research: An International Journal. 4 (3), pp. 136-149. 21

Nielsen-NetRatings (2003). “Global Net Population Increases”. Available [online] at: http://www.netratings.com

Opperman, M. (1995). “E-mail Surveys-Potentials and Pitfalls”. Marketing Research. 7 (3), pp. 29-33. Schaefer, D. And D. Dillman (1998). “Development of a Standard E-Mail Methodology”. Public Opinion Quarterly. 62, pp. 378-397. Sheehan, K. and S. McMillan (1999). “Response Variation in E-mail Surveys: An Exploration”. Journal of Advertising Research. July-August, pp. 45-54. Taylor, H. (2000). “Does Internet Research Work?” International Journal of Market Research. 42 (1), pp. 51-63. Totty, M. (2003). “E-Commerce: Selling Strategies, Demographics: The Masses Have Arrived, and E-Commerce Will Never Be the Same”. Wall Street Journal. Jan. 27. R8.

22