NPHI NHIS Technical Information

Overview

This page provides additional technical information that can help data users understand and analyze Native Hawaiian and Pacific Islander National Health Interview Survey (NHPI) NHIS data.

NCHS staff used the same survey questionnaires for the NHPI NHIS and the 2014 NHIS. NHPI NHIS data collection procedures were similar to the NHIS procedures. These similarities allowed for data from the NHPI NHIS to be compared with data from the annual NHIS. However, the 2014 NHPI NHIS was a separate survey using a separate sample design.

• Trained U.S. Census Bureau interviewers collected data for the NHPI NHIS.
• Interviews were conducted face-to-face or over the phone.

How the NHPI NHIS survey instrument differs from the NHIS

The instrument used for the NHPI NHIS was identical to the 2014 NHIS survey instrument, except—

1. The sample selection logic was modified so that only households with at least one NHPI member were asked to participate in the full interview (see Sample Design and Allocation section).
2. Data used to link to NHIS data or other HHS datasets, such as Social Security numbers, were not collected.
   1. Linking is not permitted under the legislative authority governing NHPI NHIS.
   2. NHIS regularly collects data that can be matched to other HHS datasets so the two datasets can be linked in a new, richer dataset.
      

More detailed information can be found in the NHPI NHIS Survey Description.

Additional files are available on the NHPI NHIS Questionnaires, Datasets, and Documentation page.

Racial identity

Not all people surveyed in the 2014 NHPI NHIS data file have an NHPI racial identity. Of the 11,085 surveyed, 8,697 identified as NHPI alone or in combination with one or more other race.

• Use the NHPI_ANY variable
  • To determine which people have an NHPI racial identity.
  • In combination with the RACERPI2 variable to distinguish people who are NHPI only and those who are NHPI and one or more other race.
• Use the MRACRPI2 variable to identify the main or primary race named for people by multiple races.

It is not possible to determine detailed NHPI racial identity using the public use data files. To reduce the risk of disclosing confidential respondent information, such variables are unavailable on the NHPI NHIS public use data file.

More than 40 detailed NHPI race variables have been created and are available on the restricted NHPI NHIS data files.

Geography

The risk of disclosing confidential information is higher when publicly released datasets include detailed geography variables and survey observations. Neither NHIS nor NHPI NHIS publicly releases state identifiers or other geographic variables to minimize this risk.

Variable detail

Some NHPI NHIS survey variables were suppressed to prevent disclosing the identity of the survey subjects. To conduct analyses using the suppressed variables, you must apply to use the restricted NHPI NHIS data file.

Linking data sets

Data users cannot link NHPI NHIS data to other Department of Health and Human Services datasets. The legislative authority governing the NHPI NHIS expressly prohibits linking NHPI NHIS data with other data. For this reason, NHPI NHIS did not collect the information—Medicare numbers and the last digits of the Social Security number—needed to link these data to other datasets.

It is not appropriate to combine the NHPI NHIS sample with the regular NHIS sample. The two come from different sample designs.

For more information about the NHIS area frame-based sample design, see the 2014 NHIS Survey Description. For information about the American Community Survey frame used for the NHPI NHIS sample design, see the NHPI NHIS Survey Description.

Comparisons

To calculate percentages and standard errors—

• Use the appropriate weights and variance estimation variables.
• Use a statistical software package designed to analyze complex samples.
• Calculate the percentages and standard errors separately for each population using the appropriate analytical dataset.
  

For more information about how to do this, see the Variance Estimation Method for Public Use Data appendix in both—

• NHPI NHIS Survey Description document
• 2014 NHIS Survey Description document