Calculate Variance, Analyze Subgroups, and Calculate Degrees of Freedom

Purpose

It is often necessary to determine the precision of a point estimate.  The variance of an estimate, or its standard error, is generally used to calculate this precision, and degrees of freedom are used to measure the reliability of a variance estimate.  Variance estimation requires special procedures, and even more so when analyzing subgroups.  This module introduces the basic concepts of variance estimation for NHANES data.  You will learn how the multi-stage probability sample design of NHANES, including the geographic clustering of the sampled individuals, affect variance estimation and which methods are appropriate to use when calculating variance for NHANES data.  You will also learn how to use SUDAAN and SAS syntax to properly analyze subgroups and determine degrees of freedom. 

 

Task 1: Calculate Variances

This task explains the concept of variance and its importance, and shows how to estimate variances appropriately in NHANES analyses.

• Key Concepts about Calculating Variances in NHANES
• How to Specify the Survey Design to Obtain Appropriate Variance Estimates Using SUDAAN
• How to Specify the Survey Design to Obtain Appropriate Variance Estimates Using SAS

 

Task 2: Analyze Subgroups

Because of the nature of the NHANES survey design, subgroup analysis is more involved than it would be with a simple, unweighted random sample.  This task covers examination of subgroups for basic NHANES dietary analyses in SAS and SUDAAN.

• Key Concepts about Analyzing Subgroups
• How to Specify the Survey Design to Analyze Subgroups in SUDAAN
• How to Specify the Survey Design to Analyze Subgroups in SAS

 

Task 3: Calculate Degrees of Freedom

Accurate determination of degrees of freedom is important for performing statistical tests and calculating confidence limits.

• Key Concepts about Calculating Degrees of Freedom