To characterise the form of sexual selection imposed on cuticular hydrocarbons we used separate multiple regressions to estimate the vector of linear selection gradients, and the matrix of quadratic and correlational selection gradients for each measure of male attractiveness [ 31].
The association between MHL and each of the health outcomes including self-rated mental health, physical health, social health, depression and anxiety were analysed using separate multiple linear regression models adjusted for all demographics.
The relationship between change in weight and change in each outcome variable was assessed using separate multiple linear regression analyses, each with corresponding clinical change score as a dependent variable, the corresponding baseline score and number of weeks in the study as covariates, and the following independent variables: weight change, treatment group assignment, and gender.
We used separate multiple linear regression models to estimate β coefficients and 95% confidence intervals (CIs) for individual PBDE congeners with detection frequencies > 50% and ΣPBDEs in relation to each thyroid hormone in maternal and cord sera.
We used a separate multiple regression for each of the four traits: health, dominance, sociability, and knowledge.
Each cortisol assessment (in separate multiple regression equations) was regressed on these covariates and the residuals saved for use in subsequent analyses.
Subsequently, two separate multiple regression models will be used to examine the association of the predictor variables with BMI.
These components were used in subsequent analysis as predictors in three separate multiple regression models with PDT, SDT and TDT as outcome variables (table 3).
To control for potential differences in age and gender distributions in the two samples, the US general population benchmark sample was matched to the age and gender of the acute phase sample using separate least squares multiple regression models for each SF-12v2 domain and both PCS and MCS scores.
Separate multiple regressions were conducted for men and women.
Secondly, the nature of interactions was explored with separate multiple regressions for each level of chronicity.
