Osmond, C., Kwong, W.Y. and Fleming, T.P.
Statistical analysis of data in support of the Barker hypothesis, advantages of using random effects regression model in hierarchical data
Reproductive Biomedicine Online, 10, (2), .
Full text not available from this repository.
In a recent note by Walters and Edwards (2004), the authors argued that summary statistics should be used in analysing hierarchical data from our earlier analysis of a rat model of developmental programming and the Barker hypothesis (Kwong et al., 2000, 2004). We reiterate here why such a view is inappropriate. Hierarchical data merits multilevel analysis using a 'random effects' model to enable estimation of variances at different levels and easy assessment of other parameters in a complex data structure.
Actions (login required)