BOX 3-1

Measuring Impacts

To evaluate the economic impacts of GE crops on adopters and non-adopters, the committee relied on the results of empirical analyses of farmer surveys and market data. Studies were peer reviewed, but the research approach and methods varied considerably with each study’s purposes and data. Each study has its own strengths and limitations. For example, some studies may use a different guideline in judging the significance (i.e., confidence level) of factors affecting the adoption of GE crops compared to other studies. The committee could not make the various studies comparable and accepted each set of findings as valid evidence. Some of the general approaches used to estimate economic impacts are explained here.

Empirical data. A comparison of means or averages is sometimes used to analyze results from experiments in which factors other than the item of interest are “controlled” by making them as similar as possible. For example, means of yield or pesticide use can be compared for two groups of soybean plots that are similar in soil type, rainfall, sunlight, and all other respects. One of the two groups is considered to have a treatment (e.g., soybean with a genetically engineered trait), and the other does not (e.g., conventional soy-bean). As an alternative to controlled experiments, the subjects that receive treatment and those that do not can be selected randomly with data collected through mail, phone, Internet, or personal surveys.

Survey data. Caution must be exercised in interpreting the results obtained by analyzing the differences in means from data from “uncontrolled experiments,” such as farm surveys. Conditions other than the “treatment” are not equal across the farms surveyed. For example, differences between mean estimates for yield and pesticide use from survey results cannot necessarily be attributed to the use of GE seeds because the different results are influenced by many other factors which are not controlled, including irrigation, weather, soil, nutrient and pest-management practices, other cropping practices, operator characteristics, and pest pressures.

Moreover, farmers are not assigned randomly to the two groups (adopters and nonadopters) but make the adoption choices themselves. Therefore, adopters and nonadopters may be systematically different as groups, and these differences may manifest themselves in farm performance. They could be confounded with differences due to the adoption of GE crops (i.e., the treatment). This situation, called self-selection, would bias the statistical results unless it is recognized and corrected.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement