Top: John S. Allen's
Up: Table of Contents
Previous: Literature review
Next: Aggregate Analysis
COLLECTION OF DATA
DESIGN OF QUESTIONNAIRE
It was determined that the scope of data collection efforts should be nationwide in order to discover if there were substantial differences in riding characteristics and distances traveled by bicyclists living in urban areas versus rural areas, hilly or mountainous areas versus flat terrain, cold climates versus warm climates, and other similar comparisons. However, only essential personal data and travel information would be asked of each respondent. The idea was not to overburden the effort with unnecessary data requests. Lengthy questionnaires tend to compromise the quality of the data that are being requested.
The national headquarters office of the League of American Wheelmen, located in Palatine, Illinois (a Chicago suburb) was preparing to mail a copy of the League's Annual Report to each of its approximately 9,000 members in Spring 1975. Included with the mailing would be a ballot to elect next year's Regional Directors, and a postage paid return envelope. After the idea of the questionnaire was discussed with the L.A.W. Board of Directors, an agreement was reached that allowed the form to be included in the mailing. The questionnaires would be collected and forwarded to the researcher for analysis. The Board felt that the data collected would also be beneficial to the League because they would have the opportunity to learn more about their rapidly growing membership.
The questionnaire was designed in cooperation with L.A.W., and with the idea that a simple checklist type of response would probably yield the greatest number of returns with the most accuracy. However, this checklist format could not be followed throughout because discrete numerical data on trip purpose and overall bicycle travel were desired. It was hoped that the enclosure of the questionnaire with the ballot would increase the number of responses received. Similarly, L.A.W. thought that people responding to the questionnaire would be more likely to cast their ballots, and mail both items together.
The instructions on the questionnaire stated that only persons 16 years of age or older were to respond. This was in order to obtain data only from those members who were most likely to have an automobile available for their use in addition to a bicycle. The instructions also requested that the most active rider fill out the form. Although family memberships are available in L.A.W., only a very small percentage of the total membership is in this category. Therefore, in some cases, only one member of a bicycling household filled out the questionnaire, probably a male. This would likely lead to a male bias in the final sample. Nonetheless, it was felt that there would be a sufficient number of female respondents to identify any significant differences between the sexes.
The information asked of each respondent was divided into two broad categories. The first pertained to personal information including geographical location of the respondent, and the second category included questions relating to the respondent's bicycling activities, as well as any accident experiences encountered in the last year. Table 1 summarizes the information that was requested of each respondent.
The questionnaire was pre-tested on 18 individuals in the Washington, D.C., area who were considered very knowledgeable on the subject of bicycling, including planning, design, and safety-related issues. They were, in addition; regular bicycle users, by the definition in the questionnaire. The persons tested included U.S. Department of Transportation staff working in the field of bicycle transport and representatives of the local bicycle clubs. Comments and :suggestions from these individuals were considered and appropriate deletions, additions, and rewording were accomplished. In particular, many subjects felt they could not adequately respond to the question involving "accidents". The researcher decided to use the term, "serious fall or collision", and allow the respondent to indicate severity. The final version of the questionnaire is presented in Appendix A.
INFORMATION REQUESTED ON QUESTIONNAIRE
Category I - Personal Information
Category II - Bicycling Information
PERFORMANCE OF SURVEY
On March 15, 1975, the questionnaire and the ballot were sent to 8,405 members of L.A.W. who were on the mailing list as of late February. It was stated in the mailing that all returns had to be received at L.A.W. headquarters by April 15 in order to be counted. The final number of ballots received was 4,342, or 51.7 percent of all those mailed. With the ballots, 3,270 questionnaires (38.9 percent of the total sent out) were returned that were used for this research. An additional 38 forms were unable to be used due to incomplete information provided for some basic questions. Also, 618 of the questionnaires were received after the data were analyzed. Therefore, 3,926 questionnaires were actually returned (46.7 percent).
Geographically, the returns represented all 50 states, although some states in the vicinity of the L.A.W. headquarters near Chicago have a very large membership and were somewhat over-represented in terms of each state's population. Map 1 displays the level of response from each state. The League of American Wheelmen is broken down into 13 different Regions for administrative purposes. The breakdown is based in part on geography and part on the number of members. The number of mailings that went to each Region was noted by L.A.W., so that tabulations of returns could be recorded. A complete summary of the number of returns and percentages by Region is included in Appendix B.
Map 1: Level of response from each state
A requirement of the questionnaire was that it should be as simple to code as possible in order to facilitate the processing of data. Any thought to make the form self-codable for the respondents was quickly rejected because of the necessary complexity expected in any instructions that might have to be provided. Also, the time required to carefully spot check every fifth or tenth return might better be used in actually coding the data. An 80-column format was developed where all the information provided could be coded on one form and then keypunched. An example of the coding form used is presented in Appendix C. Each response was carefully edited and coded before being keypunched and verified. After the data deck was produced for each day's returns, special edit programs were run to spot any abnormal or non-sensical data.
Difficulties arose with certain questions on the form. Questions 15 through 17 requested information on the respondent's most recent collision or fall. The individual was required to answer this set of questions only if he responded positively to question 14, "Have you had a collision or serious fall in the last year on your bicycle?" Evidently, many respondents did not notice the comment to continue answering at question 18 if the response was "No" to the question just discussed. Others wrote on their form that they understood the instructions but wished to provide information on past collisions or falls. This problem was solved by only considering positive responses to question 14 for establishing accident rates. The additional information provided by some of the respondents who did answer questions 15 through 17 was used when investigating accident patterns.
There also was a minor problem with the question that inquired as to the minimum temperature at which the respondent did not usually ride his bicycle. It appears that many cyclists in the colder climates of the country ride often when the temperature is 0 degrees or even below. As this was not expected, no provision was made for coding sub-zero temperatures. Consequently, all responses of zero degrees or below were coded as 01 in the appropriate columns. Therefore, the mean temperatures for some northern states discussed in this report are somewhat higher than what was actually reported.
Top of page