Cross-Commodity Perspective on Contracting: Evidenc e from Mississippi
to increase response rates (Pennings, Irwin and Good). A Dillman three-wave survey tech-
nique was used to mitigate non-response bias: the first wave included a questionnaire and
cover letter; the second wave consisted of a reminder card; the third wave another copy of
the questionnaire and a second cover letter.
4 Results
4.1 Descriptive Statistics
Out of the original 1,000 surveys mailed, 56 were returned due to an incorrect address,
leaving an effective sample of 944. An additional 67 were returned as out of business and
16 were returned as out of scope (prison farms, research farms, etc.), leaving a final viable
sample of 861. A total of 361 responses were returned, giving a response rate of 38%. After
accounting for incomplete responses, there were 265 usable responses or a usable response
rate of 31%.
The descriptive statistics of the relevant variables for the full sample are shown in
Table 3. The CVP variable is the coe¢cient of variation of price for the primary product
produced. The coe∏icienl of variation is a measure of price risk, which was derived from
the respondents expectations of price (Keefer and Bodily). Thus, the CVP is an expected
price risk variable for the primary crop produced. On average, respondents expected prices
to vary by 20.41%. HOURS represents the number of hours per week spent gathering and
analyzing market information, which is a transactions cost. Respondents reported spending
an average 1.17 hours per week on this task. The measure of diversification used in this
survey was off-farm income (OFFINC). On average, respondents reported that 53.89% of