Initially, associations between potential covariates were assessed by univariate log-binomial regression models; variables with p < 0.2 were included in the full model.
Data concerning adherence will be analysed by multivariate binomial regression with a latent variable with a level for each practice.
In addition, to more efficiently exploit the information on CAM usage provided by participants, we used a negative binomial regression model with the dependent variable being the number of different types of CAMs used by the individual (all independent variables similar to those of the logistic model described previously).
HRQOL is measured in days, which is a count variable; therefore, negative binomial regression models with 95%% confidence intervals (CIs) were used to analyze the association between caregiving status, gender, age, education level, financial status, social support, long-term illness and recent days with poor health or without work capacity.
We estimated the effect of open access on citation counts using a negative binomial regression model with the same set of explanatory variables previously described.
Table 4 Negative binomial regression with startups as the dependent variable Estimate Std.
In the negative binomial regression the following variables were retained in the model: rain with one and two months time lag, spraying, net used, minimum temperature the previous month, distance to the marsh, altitude and presence of animals in the house.
We used linear binomial regression with frequent versus rare use of organic food as outcome variable and characteristics of the respondent as independent variables.
These variables will be entered into a negative binomial regression model along with the product of these dummy vectors and treatment (i.e., interaction effects).
