Issues in Web-based Research

        The use of the internet to complete surveys and experiments involves a number of issues.  There are many issues to keep in mind when conduction any research project.  Some issues are common to any survey method (e.g., response rate, participant selection), though there are some special considerations when conducting Web-based projects.
        The information below would be beneficial as a synopsis of these issues.  This is not an attempt at a complete survey of the literature pertaining to Web-based research methods.  Rather, it should serve as an introduction to these issues, with a useful bibliography included.



Contents:
1) Issues in Web-based research
2) A general bibliography for an Introduction to Web Experimenting - Good overviews of methods and issues in Web-based research.


Page 1

1) Advantages to Web-based research.
2) Disadvantages to Web-based research.
3) What kinds of studies can you do on-line?
4) Distributing your instrument.
5) Housing your lab study on-line.
6) Response rates and attrition in Web-based research.

Page 2

7) Issues of privacy, anonymity, and confidentiality.
8) Using searchable databases on the Internet.
9) Similarity of Internet- and lab-based studies.
10) Using your Web servers to your advantage.
11) Maximizing responses to Web-based surveys.

* Issues in Web-based research - Page 1 *

1) Advantages to Web-based research
        Conducting a Web-based research project has a number of advantages to laboratory work (as opposed to simply using the internet/server to house your research materials; see here).  The most salient advantages include access to a wider range of participants than a lab study usually allows (see here for issues regarding similarity of lab and Web-based samples), savings on time (e.g., no data entry) and financial resources (e.g., postage, copying, stationary, etc.), and the possibility of interactive survey materials (Schmidt, 1997).  Depending on the type of project, the possible advantages will vary.  However, these advantages are tempered w/ a healthy list of disadvantages as well.
        Using Web-based storehouses for research projects (such as the Web Experiment Psychology Lab or PsychExps) also allows easy dissemination of materials for other researchers.

     * Further reading:
        Reips, U.-D. (2000). The web experiment method:  Advantages, disadvantages, and solutions. In M. H. Birnbaum (Ed.), Psychological Experiments on the Internet (pp. 89-114). San Diego, CA:  Academic Press.
        Schmidt, W. C. (1997). World-Wide Web survey research:  Benefits, potential problems, and solutions. Behavior Research Methods, Instruments, and Computers, 29, 274-279.
        Stanton, J. M.  (1998). An empirical assessment of data collection using the Internet. Personnel Psychology, 51, 709-725.

2) Disadvantages to Web-based research
        Disadvantages to Web-based research frequently involve a lose of control by the experimenter.  For example, when conducting Web-based research, the experimenter is unavailable to answer questions, deal with concerns, or troubleshoot problems.  While the dynamic nature of the internet allows communication between researcher and respondent, help may not be available in a timely manner.  Also, integrity of the data becomes a bigger issue, with increased possibility of multiple submissions, incomplete responses, and data security.

     * Further reading:
        Reips, U.-D. (2000). The web experiment method:  Advantages, disadvantages, and solutions. In M. H. Birnbaum (Ed.), Psychological Experiments on the Internet (pp. 89-114). San Diego, CA:  Academic Press.
          Schmidt, W. C. (1997). World-Wide Web survey research:  Benefits, potential problems, and solutions. Behavior Research Methods, Instruments, and Computers, 29, 274-279.
        Stanton, J. M.  (1998). An empirical assessment of data collection using the Internet. Personnel Psychology, 51, 709-725.

3)  What kinds of studies can you do on-line?
            The variety of tools now available allow researchers freedom in devising and constructing web experiments.  Surveys can be easily constructed and implemented with a variety of formats (e.g., SurveyWiz, WWW Survey Assistant).  WEXTOR allows for complex experimental designs involving between-, within-, and quasi-experimental variables.  With WEXTOR, you use web pages to present stimuli and collect dependent measures.  Using PsychExps and the Authorware software, the sky is practically the limit.  Authorware can create complex surveys (with skip patterns and real-time data presentation to the participant), record latencies in milliseconds, present a variety of graphic and audio stimuli, and do real-time data checking.

4) Distributing your instrument
        Common methods for disseminating surveys are embedding the survey in an email message or including the necessary URL in an email message (see below for response rate information).  If your study is posted in a research web site (e.g., PsychExps or the Web Experiment Psychology Lab), you will get some traffic just from random visitors.  These sites frequently report up-to-date statistics on visitors to the site.  As of November, 2000, the WEPL was linked from approximately 1,000 other web pages and averaged over 4,500 visitors per month (Reips, 2001).  It is not uncommon to email notices to relevant listservs, though it is advisable to check on the listserv's policy or check with the moderator first.
        There are definite cost advantages to to Web-based distribution rather than other methods (e.g., saving on postage and copying, saving interviewer costs).  Businesses are increasingly using Local Area Networks (LANs) to collect evaluations and surveys (Thompson, Surface, Martin, & Sanders, 2003).  This advantages must be tempered with the disadvantages of the noncoverage problem and reduced response rates to Web-based instruments.  While there's no question more and more people are jumping online, this sample is not representative of the general public.

        * Further reading:
        Reips, U.-D. (2002). Internet-based psychological experimenting:  Five dos and five don'ts. Social Science Computer Review, 20, 241-249.

5) Housing your lab study online
        Though you may not want to distribute your survey/experiment online for the whole world to complete, you can still house your instrument online with several advantages.  First, it centralizes your data collection instrument and storage, eliminating the loading and troubleshooting of problems on multiple machines.  Also, it simplifies data collection if the same study is being conducted at multiple locations (e.g., other lab rooms or other institutions).  For example, when collaborating with a colleague at another institution, data collected with the same instrument is stored in the same place.  Similarly, it eliminates the need for purchasing multiple licenses of the same software package for loading on multiple computers.  Disadvantages include the possibility of server malfunctions (e.g., if the server goes down, it's a global data collection problem rather than a localized problem) and a lack of control should server problems occur.

     * Further reading:
            McGraw, K. O., Tew, M. D., & Williams, J. E. (2000). An online psychology laboratory. In M. H. Birnbaum (Ed.), Psychological Experiments on the Internet (pp. 219-233). San Diego: Academic Press.
            Reips, U.-R. (2001). The web experimental psychology lab:  Five years of data collection on the internet. Behavior Research Methods, Instruments, & Computers, 33, 201-211.

6) Response rates and attrition to Web-based research
        Studies conducted to this point have shown emailed (and email announced) surveys, show lower response rates than "snail" mail or telephone surveys (for a review, see Cook, Heath, & Thompson, 2000).  In fact, one early study compare e-mail to "snail" mail response rates showed a drop in response rates between 1995 and 1998.  For both years, mail responses exceeded email responses, but there was a more dramatic difference in 1998 than 1995 (Bachamann, Elfrink, & Vazzana, 1999/2000).  The authors speculate that the increased use of email for for communication, which includes junk mail, helped lower the response rates.  This is a problem that is only increasing.
        Related is the issue of drop-out rate.  With no experimenter/interviewer watching, and often little incentive, incompletion rates can be high.  While response and attribution rates are a concern, who drops out is also a problem.  Inadequate research has assessed what specific aspects of web studies could reduce attrition and increase valid responses.  However, a meta-analysis of response rates to Web-based surveys analyzed 68 surveys (from 49 studies) and found 1) number of contacts, 2) use of personalized contact letters, and 3) use of precontact increased response rates.  Notably, offering incentives tended to decrease response rates (Cook, Heath, & Thompson, 2000).

     * Further reading:
        Bachmann, D., Elfrink, J., & Vazzana, G. (1999/2000). E-mail and snail mail face off in a rematch. Marketing Research, 11, 10-15.
        Cook, C., Heath, F., & Thompson, R. L. (2000). A meta-analysis of response rates in web- and internet-based surveys. Educational and Psychological Measurement, 60, 821-836.
       Knapp, F., & Heidingsfelder, M. (2001). Drop-out analysis:  Effects of the survey design. In U.-D. Reips, & M. Bosnjak (Eds.), Dimensions of Internet Science (pp. 221-230). Lengerich: Pabst Science Publishers.
        Truell, A. D., Bartlett, J. E., & Alexander, M. W. (2002). Response rate, speed, and completeness: A comparison of Internet-based and mail surveys. Behavior Research Methods, Instruments, & Computers, 34, 46-49.
        Welker, M. (2001). E-mail surveys:  Non-response figures reflected. In U.-D. Reips, & M. Bosnjak (Eds.), Dimensions of Internet Science (pp. 231-237). Lengerich: Pabst Science Publishers.


General Bibliography




If you have questions or problems, please let me know.
Last updated: 8/2/03