Tommy L. Brown

Factors Affecting Response Rates to Natural Resource - Focused Mail Surveys: Empirical Evidence of Declining Rates Over Time

We analyzed factors that could affect response rates for 105 mail surveys conducted by the Human Dimensions Research Unit at Cornell University between 1971 and 2000. The surveys addressed diverse natural resource-based topics and involved a variety of study populations. We tested the null hypothesis that response rates to natural resource-focused mail surveys are not changing over time. We found the best multiple regression model (adjusted R 2 =.558) to explain differences in response rates incorporated five independent variables: (1) saliency of study topic to surveyed population, (2) the month of the first mailing, (3) the size (font) of print type, (4) the number of complex questions included in the survey, and (5) year of survey implementation. All other variables held constant, the model suggests that response rates dropped on average 0.77% per year over the 30-year period. The scatterplot of response rate versus time suggests a gradual decline in response rates over time.

Response to mail surveys on resource-based recreation topics: a behavioral model and an empirical analysis.

Abstract The ability to predict response rates to mail surveys is important in estimating sample size and in assessing the probable need for a nonrespondent follow‐up study to improve the accuracy of population estimates. Predictive models of response rates to mail surveys are scarce, and none has examined leisure or recreation resource‐related studies. This report starts with a behavioral model that takes the potential respondents perspective. On the basis of the model, variables affecting response rates are hypothesized, operationalized, and tested in 38 studies, each of which used standard mail survey techniques including a cover letter, printed questionnaire booklet, and three reminder letters. Saliency of the topic in combination with the type of survey audience, the amount of hypothetical questions used, the number of pages, the height of the type used, and the month of survey implementation were all statistically significant predictors of response rate.

Assessing the relative importance of recall bias and nonresponse bias and adjusting for those biases in statewide angler surveys

Abstract Accurate information about angler effort and expenditures, obtained at a reasonable cost, is very important to fisheries managers and planners. Several sources of systematic bias in measurement can affect estimates of participation. We used regression analysis to examine nonresponse and recall bias. We compared respondents’ estimates of days fished based on a 12‐month‐recall mail survey, its nonrespondent phone follow‐up, and a quarterly phone survey and found that both nonresponse and recall bias significantly and almost equally influenced estimates of days fished. We calculated a combined recall and nonresponse bias correction factor of 29% for deriving estimates of days fished based on the 12‐month‐recall mail questionnaire. Further work is needed to ascertain the generalizability of this correction factor.

Gender-Biased Data in Survey Research Regarding Wildlife

Underrepresentation of females among respondents to mail surveys focused on wildlife is a growing concern for survey researchers. In an effort to better understand why females are underrepresented, we analyzed the sample frame of a wildlife-focused survey for which a notable disparity existed between the number of male and female respondents in relation to actual population figures. An analysis of the sample frame used for this study indicated that male bias existed that likely contributed to female underrepresentation among survey respondents. Suggestions are made for improving sampling frames to minimize gender bias.

Using diaries to estimate fishing effort and fish consumption: A contemporary assessment

Abstract This paper examines both the strengths and weaknesses of the diary method, focusing on a recent use of the method for measuring both fishing participation and fish consumption. A diary was sent to the 1,202 anglers willing to keep a record of fish consumption and fishing trips to Lake Ontario and other New York waters in 1992. Participants were contacted by telephone every three months to retrieve information recorded in the diary. A subsample of participants was sent a postcard every three weeks between telephone calls to serve as a reminder of the project. Diaries were returned at the end of the year and compared with information received during the telephone follow‐ups. We found no evidence for a conditioning effect bias, but did find an item‐mismatch rate of 5–7% between information gathered during the telephone follow‐ups as compared with that written in the diary. The postcards were successful at alerting participants to potential recording errors on a gross level. Possibly, more use could ...

Item response bias in angler expenditures.

Item response bias associated with the reporting of recreation-related expenditures was examined in a survey of anglers fishing in New York State. Nonresponse to the at-location expenditure item wa...