Interpretation of research data: hypothesis testing

The application of statistical tests to the evaluation of hypotheses is discussed. The statistical test is used to determine whether or not a hypothesis is correct by telling the researcher how likely it is that the results of an experiment are due to chance alone. Generally, a null hypothesis is set up stating that there is no difference between the control and experimental samples. The data are than collected and analyzed, and the null hypothesis is either accepted or rejected. The researcher, before beginning actual experimentation, should establish a level of significance to indicate how certain he wishes to be that the results are not due to chance alone. Traditionally, this level is set at 95% or 99% (expressed as a level of significance of 0.05 or 0.01, respectively). In using these tests, two types of error are possible. The null hypothesis can be rejected when it is in fact true (Type I error), or it can be accepted when it is false (Type II error). Since all statistical tests are based on certain assumptions concerning the data, the test applied to experimental results must be chosen to fit the data (e.g., collection methods, distribution type). for these reasons, statistical tests should be chosen before the experiment, and the experimental procedures should be tailored to fit the statistical test so that the validity of the analysis will be maximized.