These two explanations of nonsignificant findings are not very satisfying because both problems can be corrected before collecting any data. If you believe that college students are unique and the behavior of interest may not be seen in this population, you should use a different sample.

Many colleges are situated in larger communities and students should be able to recruit research subjects from these communities. Even if you cannot leave campus to collect your data, there are college employees who might be willing to participate in your research. The target sample size of your research should be set after taking into account the typical effect size reported in the literature.

If effects are usually pretty small, you know that you will need a very large sample in order to detect the effect. If effects are moderate to large, you can conduct your study with a relatively small sample. It is becoming more common for research articles to report effect sizes along with the results of statistical tests, so this information is more available. Most statistical textbooks give formulas for estimating effect sizes from published data.