Most research is undertaken in the hope of demonstrating a relationship
between variables. To accomplish this goal, a research study often maximizes
the differences for one of the variables to increase the likelihood of revealing
a relationship with a second variable. In particular, researchers often exaggerate
the differences between treatment conditions to increase the chance that
the scores obtained in one treatment are noticeably different from the scores obtained in another treatment. To evaluate the effects of temperature on learning,
for example, a researcher probably would not compare a 70-degree room
and a 72-degree room. The study has a greater chance of success if it involves
comparison of 70 degrees and 90 degrees. Although the larger temperature
difference is likely to reveal a relationship between temperature and learning,
the researcher should be cautious about generalizing the result to a normal
classroom situation in which 20-degree temperature changes are unlikely.