A researcher is investigating how observation of others might help five-year-old children solve a spatial reasoning puzzle. There are three samples of children in the experiment:

- An 'observer' group that watches an experimenter solve the puzzle, then tries the puzzle.

- An 'exposure' group that simply gets to look at the puzzle for a period of time before trying it.

- A 'no-exposure' group that is given the puzzle without any previous exposure at all.

The researcher performs three two-sample t-tests to compare the mean time to complete the puzzle between the observer and exposure groups; between the exposure and no-exposure groups; and between the observer and no-exposure groups. The difference between the 'observer' and 'no-exposure' groups is significant (p < .05). The researcher concludes that there are real differences among the groups. What is wrong with the researcher's conclusion?

Select one:

a.

The power of the experiment to detect differences between the three groups is unacceptably low.

b.

Running multiple t-tests results in an inflated Type I error rate; the researcher should have run a one-way ANOVA to see if the group means are significantly different.

c.

The researcher should have performed matched pairs t tests rather than two-sample t tests.