How to analyze it?
Tests of statistical inference are used to analyze experimental results. The purpose of using these tests is to determine how likely it is that the result obtained in the study is due to chance or reflects a real difference. When using inferential statistics to analyze the results, the following steps should be taken:
- Determine the type of experimental design used.
- Determine the number and levels of the independent and dependent variables.
- Determine the type of statistical test. Inferential statistic tests can be divided into two types: Parametric and Non-parametric. Parametric tests make certain assumptions about the parameters of the population from which the sample has been drawn. In contrast, Non-parametric tests do not specify any assumptions about the nature of the population. Parametric tests are more powerful (i.e., more likely that the test will detect a difference) and robust, thus it is always preferable to use them instead of non-parametric tests. However, given their assumptions, parametric tests can only be used when the following parameters are met:
The dependent variables have equal units of measurement (e.g., temperature, weight, time);
The data is drawn from a normally distributed population. This can be checked through statistical analysis of the data, using the Kolmogorov-Smirnov and the Shapiro-Wilk Tests;
There is similar variability between sets of scores that have been collected. This can be checked through statistical analysis of the data, using the Levene’s Test.
If these three conditions are met, parametric tests should be used as they provide a more accurate estimate of the probability that the result is due to chance instead of reflecting a real difference.
- Choose the specific statistical test to analyze data. This decision should be based on the experimental design used in the study, the number and type of independent and dependent variables, and the type of statistical test.
- Determine if the level of significance reached is “acceptable” to reject the hypothesis that the obtained difference is due to chance. By convention, if the level of significance reached is below 5 percent or less, the hypothesis that the differences obtained are due to chance is rejected.