Statistical Validity

Internal validity: how well the study was performed

External validity: how generalisable the results are to the rest of the population

Statistical validity:

• Significance → p value

• Suitability of the statistical test

• Relevance

• Accuracy

Assumptions for statistical testing

• The sample is a random sample

• Both variables are of interval/ratio level. If not, an alternative is needed. Either one or both variables are ordinal or both variables are nominal

  • Pearson correlation cannot be used if the variables are ordinal. Spearman correlation is needed for ordinal variables

  • Spearman correlation coefficient: can calculate correlations for a non-linear (monotone) relationship

Chi-squared test

Used to test the strength of the relationship between categorical variables

• Determines to what degree the distributions between the two variables are dependent on each other

• Compares the observed frequencies to the expected frequencies