Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
47 A number of conclusions and findings were derived from the writing of this synthesis. First, transit researchers are con- ducting their research in ways that indicate foresight and thoughtfulness. Specifically, transit researchers are already implementing many of the successful practices described in this report, such as: ⢠Starting simply with web-based surveys to learn the differences between such surveys and other survey methods. ⢠Attempting to compile databases of e-mails from cus- tomers and potential customers to use as a sampling source for research. ⢠Applying web-based survey methods in a multi-method survey environment to improve response rates by pro- viding response alternatives and to enable transit researchers to gain the benefits of web-based survey data and techniques. ⢠Researching the issue of coverage error and trying to minimize sampling bias in their studies. ⢠Remaining cautious but optimistic about including web-based surveys in their research programs as web- based survey methods and the Internet mature. Although web-based surveys have already made some inroads into the culture of transit research, it appears that use of web-based surveys will become more significant in tran- sit research in the next few years. Based on the survey con- ducted for this synthesis, only 39% of researchers indicated they are conducting web-based research. However, as noted in the report, 70% of those who are not conducting research believed that they probably would be doing so within the next two years. Therefore, it is still early for web-based sur- veys in transit; however, in the near future, it appears web- based survey methods will both be used by a significantly larger proportion of transit researchers and likely be used more often and more prominently by researchers who are currently using web-based surveys. Owing to the likely increase in the use of web-based sur- veys, a major conclusion from this synthesis is that transit researchers should be measuring the Internet penetration among their target populations so that coverage error can be understood and mitigated. Another conclusion, along with coverage measurement, is that transit agencies and other transit researchers should be collecting databases of target population e-mail addresses, particularly those of current riders. Developing a representa- tive sample database of riders and potential riders in the tran- sit agencyâs geographical area combined with the strengths of web-based survey methods can give transit researchers powerful options to conduct effective research quickly and inexpensively. Web-based surveys have limitations as well as strengths, as do all survey methods. Therefore, the strengths and limitations of all applicable survey methods should be weighed together when undertaking transit research. For example, although web-based surveys have more coverage error compared with telephone surveys, they generally have better nonresponse error than telephone surveys. Both types of errors should be considered when conducting research. Furthermore, survey methods can be used in conjunction with each other to ensure the best research possible. For example, using targeted web- based surveys for certain hard-to-reach populations such as high-income, busy professionals, while using other survey methods to reach other populations in the sampling frame, is often beneficial to the study and still cost-effective. Ideas for further research that have been generated through this synthesis effort include understanding clearly how much more effective it is to add an additional survey method to a study and under what conditions. Whereas adding an addi- tional survey method can be more convenient for respondents of paper surveys, the additional cost of adding other survey methods to a study may or may not be worthwhile; therefore, research in this area could be quite interesting. Another issue that has been difficult to quantify is the effect of spam filters on survey invitations. It would be helpful to research how to avoid spam filters and to understand what types of populations are likely to block survey invitations and whether this generates a problematic nonresponse bias. The issue of establishing credibility using web surveys is another topic worthy of future research. Research into how to protect potential respondentsâ information so that they are assured that their data will be transmitted securely, kept con- fidential, and not sold to third parties is very important to enhance the credibility of web-based research. The ability to prove to respondents that the survey agency requesting their participation is a legitimate organization and not a front for some sort of scam is also critical. Some of these credibility CHAPTER SEVEN CONCLUSIONS
48 define, but only to a certain level. If there is interest in the transit community, developing a very detailed and compre- hensive tool kit for exactly how to develop a web-based research program could be undertaken as a future research endeavor. issues were raised in this synthesis; however, further research could be helpful to cover this topic more comprehensively. Finally, the topic of web-based surveys in transit contains many issues that this synthesis has been able to identify and
