Design a student project that interests you. This candidate is an identical twin, so she chose a data set that looks at the determines of IQ in identical twins. The data set has only 10 pairs of twins, so the results are not statistically significant; the candidate notes this and shows the t-values and p-values. Actual twin studies are more sophisticated than this project. For instance, twin studies look at differences within in each pair, so a study of identical twins can’t show an effect for sex, since each pair is two males or two females. This candidate looked at the population of 20 people; that is fine for a student project.

The moral: some candidates are concerned that their data are too sp**** for statistically significant results. If you have a larger sample, use it. If not, do the project on a small sample and show the tests of significance.