Sample Size for Frequentist and Bayesian Statistical Intervals

Sample Size for Frequentist and Bayesian Statistical Intervals

Design considerations for Confidence, Prediction and Credible Intervals

Statistical intervals are a very useful tool for quantifying the uncertainty of an estimate used in a clinical trial design. However, if using a statistical interval in a clinical trial it is valuable to understand the expected precision of the interval based on considerations such as the sample size. 

The most common statistical interval associated with sample size determination is the frequentist confidence interval but sample size estimates can be found for other intervals that are also useful in certain scenarios such as prediction intervals, when considering future observations, and Bayesian credible intervals, when you want to formally incorporate uncertainty into the design.

In this free webinar, we will cover:

• An Overview of Common Statistical Intervals

• Sample Size for Confidence Intervals

• Sample Size for Prediction and Tolerance Intervals

• Sample Size for Bayesian Credible Intervals

Sample Size for Frequentist and Bayesian Statistical Intervals

There is always a degree of uncertainty in our parameter estimates used in statistical tests and models. Statistical intervals are tools that enable us to explore and quantify this uncertainty. However, their precision will typically depend on a number of factors including the sample size.

Therefore on a practical level, it is important for a trialist to know what sample size is required to achieve the required target precision of a particular interval.

Confidence intervals are the most widely used type of statistical interval which gives the expected uncertainty for a parameter under repeated sampling. Sample size methods for confidence intervals are developed for a wide variety of endpoints, regression models and study designs. These sample size methods typically target a specific level of precision in the estimate.

However, sample size determination methods also exist for other less common interval types. For example, sample size methods are available for Prediction Intervals, which provide an interval for future samples, and Tolerance Intervals, an interval for expected population percentiles. In addition, Bayesian Credible Intervals are the most common type of Bayesian interval which some argue have a more useful interpretation than standard confidence intervals. 

Watch this tutorial as we outline a number of popular statistical intervals and how they compare. We will also consider the specific requirements for sample size determination for each type of interval.About nQuery
nQuery helps make your clinical trials faster, less costly and more successful. It is an end-to-end platform covering Frequentist, Bayesian, and Adaptive designs with 1000+ sample size procedures. Start your unlimited 14-day trial.