The slippery slope of modern medical reporting : part 3

Abstract

Part 3 of the article. With few exceptions, top scientists publish fewer, but much more important papers – Nobel Prize winners in physics, such as Albert Einstein or Richard Feynman are examples. Why keep repeating studies unless one is unhappy with their findings? If one continually repeats studies, one is trying to get random chance to back them up, rather than letting science confirm their effectiveness. There may be 1000 studies showing a positive result and 950 showing a negative result, yet the “positives” are considered to prevail. Physicians often think this slight preponderance “proves it works”. The more studies performed, the greater the random chance of success when there should be failure. In medical statistics, studies are given a “statistical significance” rating, which is the level of confidence in the results. It answers the question: how much of the results are based on chance? A 95% confidence level is often used to show that a certain effect works, but also means there is a 5% probability that the result is due to chance alone – i.e., the “positive finding” would actually be false. A higher level of statistical significance raising confidence to 99% means much more money must be spent in the study, requiring more subjects, and also possibly entailing much more failure.