One-way Anova SPSS

One-way ANOVA (analysis of variance) is a structural test procedure and is the same as the t-test in that it compares group differences (or condition differences). The advantage of ANOVA is that it can compare more than two groups.

Dieses Bild hat ein leeres Alt-Attribut. Der Dateiname ist cover_anova.jpg

For example, an experiment is to be conducted to determine which diet is most effective in reducing body weight. For this purpose, we divide the test subjects into three groups. The first group does without carbohydrates, the second group tries the pineapple diet of the trade magazine “Brigitte” , while group number three applies the Stone Age Paleo diet.

The dependent variable (AV) is body weight measured at two measurement time points (before and after diet). The diet form is the independent variable (UV) with three expressions. In this experiment, a fourth control group could be included. A possible result could be that the choice of diet has a significant influence on weight loss and that abstaining from carbohydrates (group 1) results in significantly more weight loss than the other diets.

The experimental design that distinguishes the means of dependent variables several groups is called a between-subjects design. The ANOVA itself gives us only a p-value, which may or may not be statistically significant. Significance means that the mean of at least one group is not just randomly different from the other groups. Which group or groups this is exactly can only be determined with further testing, the so-called post-hocPost Hoc Post-hoc-Analysen sind Analysen, die nach der Durchführung einer Studie durchgeführt werden, um die Ergebnisse der Studie zu interpretieren und zu vertiefen. analyses. Step-by-step instructions are given in the following tutorials.


  • Independent measurements: the analysis of variance compares different measurement objects, in our case test subjects, at different points in time. The groups or conditions (UV) must not influence each other, i.e. a test subject must not be in more than one group under any circumstances.
  • Scale levels: the independent variable (time of measurement) is nominally scaled, while the dependent variable (weight) is at least interval scaled.
  • Outliers: ANOVA is sensitive to outliers, so the data set should be checked for outliers.
  • Normal distribution: the dependent variable (weight) in each group should be normally distributed as much as possible – but not necessarily. The calculation in SPSS will be explained step by step later.
  • Homoscedasticity: the variances of all groups should be approximately equal to avoid rejecting a null hypothesis that actually was. The calculation of homoscedasticity will be explained step by step later.

Calculate Single Factor ANOVA in SPSS