Question

. Two students (one from Atlanta, one from Great Britain) converge at a research station in Watamu, Kenya. They have each independently signed onto a research project to study the behavior of troops of Syke’s moneys at the Arabuko-Sokoke Forest. Beginning the following morning, they each spend their days observing a different troop. At the end of 6 months, they report their findings to their research director back in the Netherlands: the troops differ significantly in the mean amount of time they spend engaged in social behavior. The mean +/- SE for one troop is 5.43 +/- 0.61 hours per day, while the mean for the other is 7.73 +/- 0.44. Based on their experimental design and choice of statistics, is their conclusion valid? Why or why not?

Answer 1

The mean of one group is significantly different from the other, which suggests one of the researchers was constantly exposed to troops spending more time engaged in social behaviour. This suggests the troops selection/assignment during experiment design was possibly not randomized properly.

Even if the troop selection was randomized, the different mean values could be because of outliers in the observed data. In that case, median could be a better statistic than mean.