Experimental psychology: methods and applications

¶ … convincing test would be for a different researcher to replicate the first researcher's findings with the two rats who performed better. In the first experiment, the two rats likely did perform better and this would be confirmed with statistical testing. However, there were likely other rats that performed worse, and this could be confirmed with statistical tests as well. Simply put, it is to be EXPECTED that some test subjects will perform better (and worse) than the average. In a common analogy, have you ever thrown three or four sixes in a row in a dice game? This is certainly more than statistically expected per throw, however, over an increasing number of throws, it is statistically more probable. As such, replicating the results with the same two rats would indicate that the initial results were more than simple chance.

Question 2. One problem with the study is that the experiment is not blind. That is, the researcher knows where the food is. As such, it is possible that some rats are picking up on unintentional cues from the researcher to find the food.

Question 3. An alternative interpretation for the study's findings is that the two rats that performed better are somehow improving their performance based on some extraneous variable that was not accounted for. This variable could be, for example, the researcher (who gives the rats some cue as to the location of the food), or the weight of the food causing a minute slant in the maze (so the rats know the food is downhill).

Question 4. If I was the second researcher, I would first make the study blind. I would have more than one researcher run the rats through the maze (and compare findings by researcher). In addition, I would use several different mazes (in case an extraneous variable is maze-dependent).

Question 76.

Design the study, including determining numbers needed for statistical significance, and choose participants.

Offer a low carbohydrate diet to one study group, and a normal (average carbohydrate diet) to another group.

Test mental performance using several different tests (be sure to do the studies blind).

Perform statistical tests to determine if the results are statistically significant.