Date of Award

Spring 2018

Project Type


Program or Major

Natural Resources and Environmental Studies

Degree Name

Master of Science

First Advisor

John M Halstead

Second Advisor

Alberto Manalo

Third Advisor

Robert Robertson


This thesis uses primary survey data to identify and characterize the various types of New Hampshire food service establishments currently sourcing local food products ("local" being grown or raised in New England) and to assess the potential for increasing intermediate purchase of locally grown food products. Results from a recent direct-to-consumer survey revealed that consumers in northern New England have a negative propensity to consume produce that is purchased directly from growers or farmers markets, but consumers had a positive propensity to consume local and organically grown items overall. In other words, consumers in New England would like options for consuming local besides purchasing directly from farmers. Increasing local sourcing to intermediate channels, such as food service establishments, may help lower the opportunity cost of buying local food products for consumers.

Currently, there is an information gap between New Hampshire restaurants and local food producers. This study examines which variables contribute to the likelihood that a New Hampshire food service establishment will make a local food purchase. The implementation of a state-wide survey is used to help examine various restaurant characteristics, perceptions, and practices that affect purchasing decisions for local food products. Based on previous research, it is hypothesized that small-midsized restaurants (serving less than 1,750 meals per week) will be more likely to source local food products. The data set is drawn from a survey of New Hampshire establishments with help from hospitality associations, state agencies, and individual restaurateur contacts.

In the binary logistic regression model, a threshold parameter for the dependent variable is used. The dependent variable represents one of two values: y=1 when the respondent’s percentage of monthly local food purchases is 41%, and y=0 when the respondent’s percentage of monthly local food purchases is. The threshold parameter is based on the average percentage of monthly purchases of locally grown food products among foodservice establishments found by the Food Processing Center study. This screening process prevents establishments purchasing small percentages of local food from being classified as local buyers; therefore, the model identifies characteristics of only the major purchasers in the market.

The model estimates revealed a negative propensity to consume local for restaurants serving less than 750 meals per week. The results also implied that restaurants that have been making food purchasing decisions longer than two years have a negative propensity to buy local. The significant negative coefficients on “length of autonomy” may be capturing the aversion to changing time withstanding business practices. The composite variables, Impacts, and Production were found to have a significant and positive effect on the propensity to buy local. Advocating the importance of personally knowing who and where their food came from may help increase intermediate purchase of locally grown food products.