ou must: state your ideas clearly if you plan to evaluate them.<br > advice applies to any kind of work and especially to research de<br >statistical analysis. Before you begin working on design and anal<br >need to have a clearly defined topic to investigate.<br >3ESTION<br >may have a general suspicion that smoking less makes people<br >er. Or you may have an idea for a study method that will make pc<br >n more. Before you begin a study about such intuitions， you sin<br >ace vague concepts such as "feeling better" or "smoking less<br >ruing more" with definitions that describe measurements you<br >e and compare. You might replace "feeling better" with an obje(<br >lition such as "the subject experiences no pain for a week." Or<br >nt record the actual dosage of medication required to control pai<br >re interested in smoking， you need a lot of information to descrik<br >tt does each of the subjects smoke--a pipe， cigars， or cigarettes?<br >h tobacco do the subjects consume in a day? How long have they<br >king? Has the amount they smoke changed?<br >On the other hand， you must balance your scientific curiosity<br >practical problems of obtaining information. If you must tel<br >figs memory， you can t ask questions like "What did you hav~<br >ter ten years ago?" You must ask questions that people will be ab<br >ver accurately. If you re trying to show a relationship between<br >disease， for example， you can t rely on memory of what people al<br >vidual meals. Instead， you have to be satisfied with overall part<br > people can recall. Some information is simply not available to<br >ever much you d like to have it. It s better to recognize this<br >，re you begin a study than when you get your (tuestionnaires back<br >that people were not able to answer your favorite question. If<br >k about your topic in advance， you can substitute a better quesl<br >that will give you information you can use， even if it s not<br >rmation you wish you could have.<br >at Information Do You Need?<br >