DESIGNING A STUDY

Even when the population of interest seems to be well defined, you may not actually be able to study it. If you are evaluating a new method for weight loss, you would ideally like to draw conclusions about how well it works for all overweight people. You cannot really study all overweight people, though, or even a group that is typical of all overweight people. People who do not want to lose weight or who have been disheartened by past efforts to reduce may not agree to try yet another method. You will probably be able to try out your new method only on people who want to lose weight and who have not given up trying. These people, not all overweight people, form your population.

Remember that a population defined realistically in this way may be different from the ideal population. For example, the population in your weight loss study may be lighter, younger , or healthier than the ideal population of all overweight people. Therefore, your conclusions from studying people who want to lose weight do not necessarily apply to people who are not motivated. For example, the treatment may have some unpleasant consequences, such as making people want to chew on the nearest thing available, such as gum, a pencil, or the corner of a desk. People who really want to lose weight may be willing to put up with such minor inconveniences in order to reach their goal. People who do not care much about their weight probably will not be. Thus, the new treatment may work quite differently for those who are motivated versus those who are not.