Next section: The null hypothesis

"Of course, things that happen extremely infrequently occasionally do happen. This might be one of those times. However, being skeptical does not mean being totally unwilling to accept new findings. The data are quite strong so I am forced to conclude that the drug is more effective than the placebo."Naturally, there is a large degree of subjectivity in deciding how unlikely results must be (given that only chance is operating) before it should be concluded that chance is not responsible for the effect. Traditionally, researchers have used either the 5% or the 1% significance levels. When using the 5% significance level, one concludes that the experimental treatment has a real effect if chance alone would produce a difference as large or larger than the one obtained only 5% of the time or less. A good portion of this text is devoted to methods used for figuring out the probability of outcomes when chance alone is operating.

Next section: The null hypothesis