At Notable Bookkeeping we believe a successful data analysis is attributed to how well it was planned. It will weed out the factors that do not have a relative effect on the response variable and identify the best value points for the factors that do have a relative effect on the response variable.
When conducting a data analysis you must first do the following:
- Establish the purpose by defining the problem.
- Identify the components of the data analysis.
- Design the data analysis plan.
- Conduct the data analysis.
- Analyze the data.
- Have someone audit your analysis and have it approved.
- Generate the report.
- Act on the results.
A well designed data analysis plan should provided unbiased results, utilize enough historical data and random sampling, have strong data quality assurance processes, and ensure past events are sufficiently similar to the current event or problem you are trying to predict or solve.
It is best to use all possible combinations of all selected value points of the factors to be investigated. This can determine the main effects of the factors manipulated on response variables and effects of factor interactions on response variables.