The task of the researcher is to disprove the null hypothesis, or the claim that there is no difference between, for example, two levels of something. Since it's usually impractical to inspect all instances of something in a population, we select a sample which - if valid - is representative of the population at a level of confidence that we are willing to accept (typically 95 percent). We run the study, come up with findings, and report that we are 95 percent confident that the findings we came up with by looking at the sample are the same findings we would have come up with if we had looked at the entire population. Alternately, there is only a five percent chance that the findings we came up with would have appeared if no real difference exists between the two levels. Further, when we say a finding is statistically significant at the .05 level, we mean that if we were to run the study or investigation n number of times, we would come up with the same finding 95 percent of the time. Level of confidence is kind of like the amount of reasonable doubt we are willing to accept.
 |
| Robert and Shana ParkeHarrison |