ABSTRACT
Nonexperiments (NEs) include any design that uses statistical controls. Any randomized field experiment (RFE) that uses statistical controls for pretest scores is partly a non experiment, though at its core it remains a randomized experiment. In the case of a simple single interrupted time series (SITS), there is a control only for the underlying trend. The regression portion itself is the most important part of the output. The first part of the table to examine is whether the parameter estimates associated with the treatment variables is are statistically significant. In many cases, if the n is large, the sampling distribution will be normal. Statistical significance is not the same as substantive significance. When coefficients are statistically significant, the next task is to interpret the corresponding estimated regression coefficients. The magnitude of the actual estimate of b helps the researcher to judge if the treatment is substantively significant.
