CDC Home

## Variance Estimation

### Purpose

This module introduces the basic concepts of variance (sampling error) estimation for NHANES data.  You will learn how the complex survey design of NHANES and clustering of the data affect variance estimation, which methods are appropriate to use when calculating variance for NHANES data, and how to calculate degrees of freedom and construct confidence limits for NHANES estimates.

### Comparison of Variance Estimation in NHANES I and Continuous NHANES

The contents of the Variance Estimation module in the Continuous NHANES tutorial also apply to the NHANES I data. Therefore, links to this module in the basic tutorial are provided here for your reference. The following text points out the key differences between NHANES I and continuous NHANES when it comes to variance estimation.

#### Methods of Variance Estimation

As with continuous NHANES, the Taylor Linearization Method is the recommended method for variance estimation that incorporates the complex survey sample design (see the Analytic Guidelines for full reference).

#### Survey Design Variables

PSEUDO primary sampling units (PSU) and stratification variables are provided in NHANES I, as opposed to the masked variance units (MVU) and strata provided Continuous NHANES. However, this difference does not affect how you use the statistical program in variance estimation. The PSU and Strata variables provided in the files are only to be used for analyses on stands 1-65 or stands 1-100.  You will need to recode the PSU / Strata variables for analyses on stands 1-35 or 66-100.

#### Estimate Stability and Sample Design

As mentioned in the basic tutorial, variance stability is greatly increased by using larger sample sizes.  For the same reasons it is suggested you use the Strata / PSU, and weighting variables provided when doing analyses on stands 1-65 or 1-100.  This increases the stability of the estimate and the stability of the estimated variance.

You can compare estimates from stands 1-35 to stands 66-100 if you wish to examine trends during the time of the survey.  You will have to use the weights provided and recode the Strata / PSU as instructed in the Documentation, under Variance Estimation.

Other than the differences pointed out above, the rest of the principles covered in the basic tutorial apply to NHANES I as well. Please consult the Continuous NHANES Web tutorial on this topic.

#### Variables

The Strata and PSU variable names necessary to specify the sample design are N1BM0196 for the PSU variable and N1BM0194 for the Strata variable. These are to be used when analyzing data from stands 1-65 or 1-100.  However, when analyzing data from stands 1-35 or 66-100 it will be necessary to recode the Strata / PSU codes according to the instructions outlined in the Documentation, specifically the section on Variance Estimation.

Go to Continuous NHANES Web Tutorial Variance Estimation module