pyhettejistä

Pyhätie is a statistical method for testing the goodness of fit of a theoretical distribution to an observed frequency distribution. The test was developed by Finnish statistician Einar Eriksson in the 1920s. It is an improvement over the older chi-square test method in that it takes into account the specific structure of the category system in place.

The method involves numerically calculating the amount of tied observations, as well as whether if the expected values in each cell are equal to an expected value. The test then calculates the degrees of freedom in a way that also takes into account the number of tied observations. A formula is then used to calculate the probability of observing the given frequency distribution under the assumption of the theoretical distribution.

Pyhätie is more accurate than the chi-square test for a smaller sample size due to its ability to account for tied observations. It has however been found that the method may be conservative, meaning it may provide a p-value that is higher than the true probability of the null hypothesis.