Quasi-Experimental Research Question 3

Quasi-Experimental Question #3

There are two possible quasi-experimental approaches to this problem, and each has possible problems.

If you use a time-series design, you can use sales data from several years to ascertain the impact on sales during the particularly harsh winter, but it is likely that other variables will be uncontrolled by such a design. For example, economic fluctuations from year to year are common, and these fluctuations may be much more potent in affecting consumer spending than the weather.

If, on the other hand, you compare the stores in the Northeast and Midwest with stores in other regions, you are likely confounding your results with differences in economic growth in the various communities.

There is likely to be no way to completely eliminate confounding. Therefore, it would be best to gather both sales data from your stores and measures of other variables that may confound the results, such as increases or decreases in population, unemployment rates, measures of economic activity, and so on.