Statistical Sampling and Validity

Statistical Methodology Design and Critique

It has been seen that often different statisticians differ in their methodologies used for SPSS Data Analysis, in spite of having nearly same desired results. This creates a sort of confusion amongst the clients. Silver Lake Consulting, therefore, provides detailed third party reports on optimal methodology design with analysis of the validity and reliability of the used model. To do this effectively, we possess the required intuitive understanding of the statistical methodologies. This way, we try our best to bridge the gap between statistics and real-life. After finalizing the basic structure and general aims, we test the reliability and statistical power features of the standard model. An ideal statistical method has broad reliability and validity with high statistical power. It is robust under varying underlying assumptions. Then, we provide the client with a comprehensive report about the model used and associated shortcomings if any, and give suggestions for potential improvements.

Subsample Selection Criteria and Validity

We have helped several firms in settling statistical disputes over subsample validity, and also with tactical subsample selection to reduce the possibility of a dispute. One of the most common issues we have been seeing in practical is that of minimum required sample size. Therefore, we conduct a statistical power analysis to determine the minimum sample size, and then randomly select the sample from the population to meet the power analysis criteria. The other things to consider while doing a power analysis for sample size determination are:

• The general approach to find the apt sample size assumes the sampling design to be a simple random sample.
• More complex designs must consider the subpopulations variances before estimating the variability in the population.
• The sample size should be apt for the planned analysis.
• There must be scope for some adjustment in the sample size to accommodate a comparative analysis of subgroups.
• Skewed distributions may cause serious deviations from normality thereby requiring a larger sample.
• In case of a dispute about the validity of the subsample, a data analysis is done in reverse. The prior analysis was called an ‘A Priori’ Power Analysis, while the latter is called a ‘Post-Hoc’ Power Analysis.
• The demographics of the sample are checked with the population to find discrepancies, and the effective ‘power’ is calculated.

Confidentiality

We understand the significance and necessity of confidentiality of our clients. We sign a Non-Disclosure Agreement with our clients, immediately upon contact. This ensures the security of your information from any third parties.