Guide overview
The essential idea
The August 2021-August 2023 public demographics file contains 11,933 interviewed participants. Of these, 8,860 were both interviewed and examined in a Mobile Examination Center, while 3,073 were interviewed only. These are unweighted file counts. They describe participant status in DEMO_L; they are not population percentages or survey response rates.
Practical framework
What to apply
11,933 interviewed
What does the full demographics-file count represent?
- DEMO_L includes 11,933 participants with a completed household interview.
- The demographic file provides participant status, demographics, interview and exam weights, and masked variance variables.
- Interview-derived variables may be available for people who did not complete the examination.
- The correct analytical denominator still depends on the variables and eligibility rules used.
8,860 interviewed and examined
Who can contribute to many examination and laboratory analyses?
- RIDSTATR equals 2 for participants who completed both the interview and MEC examination.
- This group represents 74.2 percent of the 11,933 interviewed records as an unweighted file ratio.
- Many physical examination and laboratory variables require MEC participation, but component-specific eligibility can narrow the sample further.
- MEC or subsample weights may be required when the analysis includes these measurements.
3,073 interviewed only
Why do some participant records lack examination measures?
- RIDSTATR equals 1 for 3,073 participants who were interviewed but not MEC examined.
- This is 25.8 percent of interviewed DEMO_L records as an unweighted file ratio.
- Their interview information can remain analytically relevant when the study uses interview-only variables.
- They generally cannot contribute values that required the MEC examination or a later component subsample.
Analysis consequence
How should participant status affect the study workflow?
- Define the denominator after mapping every variable to its collection component and eligible population.
- Report the unweighted participant flow from interviewed to eligible, measured, complete-case, and analyzed records.
- Use the weight corresponding to the smallest applicable sample represented by the analysis variables.
- Describe missingness by reason where the public documentation and data permit.
Before you finish
A short quality review
- RIDSTATR has been examined before defining the analytical sample.
- Interview-only and MEC-collected measures are not treated as if they share the same denominator.
- Unweighted participant counts are reported separately from weighted population estimates.
- The selected weight matches the most restrictive variable or subsample.
- Structural absence, component ineligibility, and item missingness are distinguished.
- The manuscript does not describe these file ratios as official response rates.
Counts are not national estimates. The 74.2 percent and 25.8 percent figures are simple ratios calculated from published DEMO_L frequencies. Population inference requires the appropriate NHANES weights and design variables.