1. Comment on influence of study design, sampling strategy and scale of measurement of outcome variable, on your choice of analytical approach
Study design: the study design is cross section study design. This design with a binary exposure will present these results:
For numerical outcome: differences in means and t-test.
For binary outcome: risk ratio/odd ratio (OR).
For rate outcome: rate ratio.
In this study results are going to be expressed as adjusted odds ratios and 95% confidence intervals. With statistical significance at the conventional p < 0.05 level (two-tailed).
Sampling strategy: the sampling strategy in this study is random sampling. The influence of random sampling in this study is eliminating selection bias in subject assignments. Therefore, the results of the study would meet the generalizability and can be applied to the general population if the process of randomisation done correctly.
Outcome variable: the outcome variable in this study is vitamin D status which was measured by (25 hydroxyvitamin D [25 (OH) D] concentrations). This variable is dichotomous variable [vitamin D insufficiency: serum 25 (OH) D <50nmol/L and vitamin D Normal: serum 25 (OH) D >50nmol/L]. Therefore, the Binary logistic regressions will used this is because it allows one to fit a number of factors (dichotomous such as the status of vitamin D: insufficient and normal) and covariates simultaneously. Since this model is a logistic regression, model the log (OR) will be used.
2. Rationale for choice of any exploratory analytical approaches
3. Rationale for choice of any modelling approaches (both statistical method, and model building approach)
4. List of specific models to be fitted and what question you expect to answer with each
5. Assumptions to be tested for each analytical approach [and suggested alternative strategies if assumptions are not met]
6. Criteria that will be used for reporting significant associations
7. Dummy tables and/or graphs depicting the style of presentation of results from your proposed analyses. Remember that the format (headings and presentation) must be planned to a high level of reader-friendliness for the report.