prev next front |1 |2 |3 |4 |5 |6 |7 |8 |9 |10 |11 |12 |13 |14 |15 |16 |17 |18 |19 |20 |21 |review

Search for most updated materials ↑

 
A statistic that is particularly useful for assessing whether the studies are all homogenous is the Q-statistic. The hypotheses we test here are:

The null hypothesis states that the studies are homogeneous

The purpose of this test is to see if the null hypothesis holds. For each model, Q-statistic is calculated. The Q-statistic has a chi-square distribution with degree of freedom N where N = (Number of Studies - 1).

For homogeneous studies, there is essentially no difference between a fixed effects model and the random-effects model.