One Sample Proportion Test
One sample proportion tests are useful for scenarios where sample proportions are being compared to a baseline or expected value. For instance, this test could be used to assess if the proportion of target hits from operational testing surpasses a requirement threshold (e.g. 70% target hit requirement). The one sample proportion test power application is a tool to guide sample size selection. Click the following link to view the application: One Sample Proportion Test
Two Sample Proportion Test
Two sample proportion tests are useful for scenarios where evaluators are interested in comparing proportions from two samples. For instance, this test could be used to compare the proportion of target hits from a new system to the results from a test of a legacy system. The two sample proportion test power application is a tool design to guide sample size selection. Click the following link to view the application: Two Sample Proportion Test
Categorical DOE Application
The goal of Test & Evaluation is to assess the suitability and effectiveness of systems. Accomplishing this goal requires that we design valid tests that are likely to detect system degradation or improvements under test conditions. To this end, evaluators are often in interested in answer the questions “how large of a sample do I need for my study”, and “does my design have adequate power to detect effects of interest.” The Categorical DOE application allows the user to calculate power for a variety of test configurations and can support decisions related to sample size selection.
One Sample t-Test
One sample t-tests are useful for scenarios where sample data that are continuous (distance, speed, etc.) are being compared to a baseline or expected value and the population variance is unknown. For instance, this test could be used to assess if the average miss distance found during testing surpasses a requirement (e.g. requirement states average miss distance must be less than 10 yards). The one sample t test power application is a tool design to guide sample size selection. Click the following link to view the application: One Sample t-Test
Two Sample t-Test
Two sample t-tests are useful for scenarios when evaluators want to compare two sample means for continuous data (distance, speed, etc.). For instance, this test could be used to assess if the average miss distance for a new system is smaller than the average miss distance for a legacy system. The two sample t test power application is a tool design to guide sample size selection. Click the following link to view the application: Two Sample t-Test
Categorical Factors in a GLM
This application approximates power for effects in a generalized linear model for experiments with categorical factors: GLM Power
For Tom to deploy applications: R Studio Server