Puerto Rican Maternal and Infant Health Study (PRMIHS), 1994-1995 (ICPSR 36238)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Data Collection Notes View help for Data Collection Notes

Study Purpose View help for Study Purpose

To identify the determinants of infant health outcomes among Puerto Ricans.

Sample View help for Sample

Personal interview data were collected from 2,763 mothers of Puerto Rican infants sampled from the 1994 and 1995 birth and infant death records of six United States vital statistics reporting areas (Connecticut, Florida, Massachusetts, New Jersey, New York City, Pennsylvania) and the Commonwealth of Puerto Rico. The included U.S. states were those with the greatest number of births to Puerto Rican women each year. In 1994 and 1995, 72.3 percent of all births to mainland Puerto Rican women occurred in the included states. Mothers of the sampled infants were located from the address information provided on the vital records and asked to participate in a Computer Assisted Personal Interview (CAPI). All study interviewers were bilingual, and the questionnaire was available in both Spanish and English. Roughly two-thirds of the interviews (1,946) were with mothers of infants sampled from the computerized birth certificate files maintained by the states and Puerto Rico. The remaining 817 interviews were with mothers of infants drawn from death certificates for infant deaths.

Independent live birth and infant death samples were drawn from birth certificates and death certificates. The birth sample was stratified by vital statistics reporting area, month, and infant birth weight (less than 2500 grams; 2500+ grams). The death sample was based on the full population of infant deaths.

Time Method View help for Time Method

Universe View help for Universe

Puerto Rican women who had a live birth or infant death in the study area (Connecticut, Florida, Massachusetts, New Jersey, New York City, Pennsylvania, and Puerto Rico) between July 1, 1994 and December 31, 1995.

Unit(s) of Observation View help for Unit(s) of Observation

Data Source View help for Data Source

Data Type(s) View help for Data Type(s)

Mode of Data Collection View help for Mode of Data Collection

Response Rates View help for Response Rates

Response rates for the birth and death samples were 79 percent and 74 percent, respectively.

Original Release Date View help for Original Release Date

2015-09-30

Version History View help for Version History

Weight View help for Weight

The data are not weighted. However, the dataset contains two weight variables (BIRTHWGT and DEATHWGT) which should be used in any analyses.

Birth Sample Respondent-level Weights (BIRTHWGT): All 1,946 respondents were assigned a final analysis weight which was the product of the basic sampling weight and two adjustments -- a nonresponse weight and a post-stratification weight. Nonresponse weights adjust for differential completion rates across demographic subgroups. To determine this weight, a file containing data for all respondents was weighted by the basic sampling weight. To create weighting class cells, the weighted file was then cross-tabulated by geographic area (Puerto Rico/New York City/the 5 states combined), birth weight (low/normal), and mother's age, marital status, and education. The nonresponse weight is equal to the inverse of the weighted response rate in each cell of the cross-tabulation. Next, a post-stratification weight was calculated by weighting the respondent file by the product of the sampling weight and the nonresponse weight. The same cross-tabulation described for the nonresponse weight was produced for the sample file and for the population file (the 1994 and 1995 Natality Files for the U.S. and Puerto Rico). The goal of post-stratification is to align weighted sample totals to the known population totals for specified variables. The post-stratification weight adjustment is equal to P/p for all cases in a given cell, where P refers to the proportion of the population in a given cell and p to the proportion of the weighted sample for the same cell. The final weight, to be used when analyzing data from the respondents' interviews (or when analyzing data from both the interview and the birth certificate), was the product of all three weights -- sampling, nonresponse, and post-stratification.

Death Sample Weights (DEATHWGT): Since the death sample was selected with certainty, all cases in the death sample were assigned a basic sampling weight equal to one. The only weight calculated for the death sample, and thus the final weight assigned to the respondents in the death sample, was a nonresponse weight.

The birth and death samples can be combined to conduct analyses of infant mortality. Users should be aware that a small group of infants drawn into the live birth sample died before their first birthday (n = 79). These cases should be included in analyses based solely on the birth sample; however, they should be considered death cases in analyses of infant mortality. They have valid values for both the birth and death sample weights. For analyses based on the birth sample, the 79 overlapping cases should be weighted by the birth sample weight. For analyses of infant mortality based on the combined birth and death samples, the 79 overlapping cases should be assigned the death sample weight.

Note: Post-stratification was based only on geographic location, birth weight, and the mother's age, marital status, and education.