Hardcover
-
PICK UP IN STORECheck Availability at Nearby Stores
Available within 2 business hours
Related collections and offers
Overview
The volume opens with an introduction to the survey and its methods. Contraceptive practice in 1970 is then compared with data for 1965, and an analysis is supplied of trends since 1955 in the attitudes of Roman Catholics.
Originally published in 1977.
The Princeton Legacy Library uses the latest print-on-demand technology to again make available previously out-of-print books from the distinguished backlist of Princeton University Press. These editions preserve the original texts of these important books while presenting them in durable paperback and hardcover editions. The goal of the Princeton Legacy Library is to vastly increase access to the rich scholarly heritage found in the thousands of books published by Princeton University Press since its founding in 1905.
Product Details
| ISBN-13: | 9780691643915 |
|---|---|
| Publisher: | Princeton University Press |
| Publication date: | 04/19/2016 |
| Series: | Office of Population Research , #1675 |
| Pages: | 398 |
| Product dimensions: | 6.30(w) x 9.30(h) x 1.40(d) |
Read an Excerpt
The Contraceptive Revolution
By Charles F. Westoff, Norman B. Ryder
PRINCETON UNIVERSITY PRESS
Copyright © 1977 Princeton University PressAll rights reserved.
ISBN: 978-0-691-09371-0
CHAPTER 1
Introduction
The publication of this report of the 1970 National Fertility Study (NFS) marks the last of a series of four national studies of U.S. fertility conducted by university organizations and personnel. The first two such studies, in 1955 and 1960, respectively, were organized jointly at the Survey Research Center of the University of Michigan and the Scripps Foundation for Research in Population Problems at Miami University, Oxford, Ohio. The 1965 National Fertility Study, successor to these, was based at the Office of Population Research, Princeton University and the Center for Demography and Ecology, University of Wisconsin (Madison). The present study has been conducted primarily at Princeton, although a substantial part of the early work occurred at Wisconsin. The fieldwork in 1970 was the responsibility of the Institute for Survey Research at Temple University.
The kind of information obtained in these studies has been deemed of sufficient importance that comparable biennial surveys are now being conducted on a regular basis by the National Center for Health Statistics of the U.S. Public Health Service.
These studies have provided all of the basic information available to date on trends in contraceptive practice in the United States and, with the exception of some recent data from the Current Population Survey of the U.S. Bureau of the Census, the time series of data on fertility expectations or intentions. In addition to their contribution to information about historical trends, the studies have each attempted to add to knowledge of the ways in which differences in human fertility are determined. This volume reflects our latest effort to extend knowledge of factors affecting various facets of fertility: fecundability, reproductive intentions, and contraceptive behavior.
The coverage of the population has been extended with each successive study. In 1955, the sample included only currently married white women 18-39 years of age. In 1960, the range was extended to include women aged 40-44 and a small sample of black women was added. In 1965, the sample size was increased considerably, the black population was double-sampled, and for that survey only, a sample of women aged 45-54 was added. In the NFS of 1970, the sample size was again enlarged and the universe was defined in terms of ever-married women. The main reason for including the formerly married is that a sample of currently married women misses a large segment of fertility experience by excluding the records of women who, at the time of the survey, were separated, divorced, or widowed. Their inclusion makes possible generalizations about all fertility except for the small fraction of experience of never-married women. The ideal, of course, would be to include all women of reproductive age, but the problems inherent in asking questions about fertility and contraception of never-married young girls have deterred us. The first such national study of teen-aged girls has since been conducted by John Kantner and Melvin Zelnik of Johns Hopkins University; the profession may be close to reconsidering this whole question. Indeed, the 1973 National Survey of Family Growth, conducted by the National Center for Health Statistics, has taken an important step in this direction by including "never-married women with children in the household" in their sample.
THE SAMPLE
As noted above, the Institute for Survey Research at Temple University assumed the responsibility in 1970 of drawing the sample, conducting the interviews, editing and coding, and preparing a basic data file. For the 1970 National Fertility Study, a national probability sample of ever-married women born since July 1, 1925 was drawn. Utilizing areal sampling methods, the sample was designed to be representative of all housing units in the conterminous United States, exclusive of military reservations. The sample was drawn in successive stages, beginning with the selection of primary sampling units from areas of approximately equal numbers of households. A first-stage sample of 126 areas was selected. Each primary sampling unit was then subdivided into 200 listing areas, each with approximately 50 housing units. Within each listing area, all housing units were listed by interviewers before any interviews were conducted. Households were then randomly selected in the office at a rate calculated to yield the desired sample size (about 6,500). Since we wanted a more adequate sample of blacks than would result from a self-weighting national sample, a sampling ratio for blacks double that for non-blacks was established. (Additional details are in Appendix A, which provides sampling errors.)
COMPLETION RATES
There are four components of noncoverage:
1. Failure to cover all housing units at the time of listing. The estimate was that 6 percent of housing units were not covered, most likely in multi-unit buildings located primarily in central cities and households in suburban areas that experienced rapid growth since 1960.
2. Failure to enumerate the members of sample households. This mainly occurred either because no one was at home or because the household listing request was refused. This type of noncoverage resulted in another loss of 6 percent.
3. Failure to find eligible respondents within households where listings were obtained. This problem may occur because of such factors as the unreliability of enumerating marital status, particularly for women not currently living with their husbands. The reporting of age is another source of error. Including this source of error with the other two types of noncoverage enumerated above brings the completion rate to 0.76.
4. Refusal or inability of eligible respondents to be interviewed. Of the 7,970 eligible respondents located, a total of 6,752 women were interviewed, for a rate of 0.85. The major loss was due to refusal, accounting for 60 percent of the 1,238 eligible respondents not interviewed. Other reasons included language barriers and respondents chronically not at home.
The net effect of these sources of noncoverage is that the estimated coverage of the entire study population is 64 percent.
The usual measure of nonresponse in national household surveys, however, is the study completion rate, or the proportion of those households listed in which an eligible respondent could be expected to be found that actually yield an interview. In this study, 38,839 households were listed. As determined by the criteria of sex, age, marital status, and race, eligible respondents were found in 22 percent of those households where household listings were obtained. Under optimal conditions, then, we would expect to obtain interviews with 8,467 eligible respondents. Owing to factors of nonresponse such as refused household screenings, refused or terminated interviews, and inability to find eligible respondents at home, completed interviews were obtained from 6,752 respondents. The study completion rate is thus computed as (6,752)/ (8,467) or 80 percent. This is lower than the 88 percent completion rate for the 1965 survey and is probably due to those factors which have led to the decline in response rates reported for survey research in general.
INTERVIEWERS
Because of the subject of the interview and the fact that the respondents were women, all interviewers recruited for this study were female. A total of 283 interviewers were trained in ten regional two-day conferences. Before receiving any assignment, each interviewer was required to conduct a practice interview, which was then reviewed.
The first day of the conference was devoted to ways of contacting the respondent, techniques for establishing rapport and for probing, and the importance of following instructions. It was essential that the interviewer read the questions verbatim and that she learn the rather complicated pattern of "Skip" instructions. All the various forms — the call report form, the screening form, and the progress report form — were described. The second day was spent in an item-by-item review of the questionnaire. The objectives of each question were explained and a demonstration of how each question was to be read was given.
Various validation procedures were employed to maintain the quality of the project. Letters were sent to every individual with whom a completed interview had ostensibly been conducted, which resulted in the rejection of some 60 interviews for one reason or another; these households were then assigned to another interviewer. Written requests for interviews were also mailed to households where an eligible respondent had refused to be interviewed or could not be reached for an interview. This procedure elicited an additional 148 interviews. A third form of validation was also conducted by telephone under special circumstances.
REPRESENTATIVENESS OF THE SAMPLE
How well does the 1970 NFS sample represent the universe it was intended to reflect? To explore this question, we have compared the distributions of various characteristics in the 1970 NFS sample with comparable distributions reported in the 1970 census. Throughout, we have assumed that the census is the more accurate of the two.
A couple of minor differences between the census and the fertility study should be noted first. The census is dated April 1, whereas the National Fertility Study interviews were centered on December 31. Secondly, the census is an enumeration, the fertility study a sample survey. Even though these two methodologies share certain common problems, they also face distinctly different difficulties.
Tables I-I and I-2 show comparisons between the census and the NFS for region and size of place of residence, respectively. These are the two stratifying variables used to obtain the primary sampling units. For all age, marital status, and racial comparisons, the region-of-residence distribution is very similar for the 1970 NFS and the census. Table I-2, however, shows consistently too few central city residents in the sample and, correspondingly, too many residents in the other three categories. The deficit of central city residents is somewhat greater among the older age groups than among the younger ones.
However, it should be noted that the census categories are not strictly comparable to those of the NFS. The sampling frame and the size of place of residence designations of the primary sampling units for the 1970 NFS were based on published results from the 1960 census — for the simple reason that the 1970 census results were not available in time. But the number of Standard Metropolitan Statistical Areas increased from 1960 to 1970, and there were annexations of suburban territory by central cities during the decade. Thus, some respondents classified by the 1970 NFS as not residing in the central cities of SMSA's actually would be so classified if the 1970 census classification scheme had been used.
Unfortunately, we do not know the extent to which the deficit of central city residents shown in Table I-2 is classificatory in nature. Our guess is that if the census and NFS classifications were completely comparable, there would still be a deficit of central city residents — albeit a somewhat smaller one. The reason is that housing units are more difficult to list in central city areas than in the suburbs, smaller cities, or rural areas. Furthermore, nonresponse rates are traditionally higher in central cities than in other areas, which suggests that there might be a shortage of working women as well. Table I-3 shows, for various age, racial, and marital status categories, the percentage of women in the labor force for the 1970 NFS and the census. With the exception of teen-aged, married whites, the proportions in the labor force are similar for the census and NFS. This indicates that the deficit of central city residents is not confined to working women.
In Table I-4, we examine whether the lack of central city residents extends to all racial, marital status, and age groups. The census classification used in Table I-4 is based on the concept of the urbanized area rather than the SMSA concept. (The required SMSA census tabulations are not available.) For both SMSA's and urbanized areas, the central cities are defined in essentially the same way. The two definitions differ with respect to nearby sparsely populated territory. For this reason, the NFS-census comparisons shown in Table I-4 are limited to the central city category. It can be seen that the proportion of central city residents in the National Fertility Study is lower than the proportion in the census for every comparison by race, marital status, and age.
Tables I-5 and I-6 present comparisons for two major independent variables, age and education. The age distributions of the four samples shown in Table I-5 are extremely close to those reported in the census. However, we find in Table I-6 that the NFS consistently indicates higher educational attainment than the census. Almost without exception, the proportion not finishing high school is lower in the NFS than in census reports. Similarly, the proportion reporting at least some college education is uniformly too high. It is possible that the small but persistent excess of women who attended college is related to the deficit of central city residents. However, it is also worth noting that a similar problem was reported for the 1955 GAF, 1960 GAF, and 1965 NFS studies.
Another classification used repeatedly in our study is marital status. This variable is also of interest because the 1970 NFS is the first U.S. fertility study to use an ever-married sample rather than a currently married one. Table I-7 shows the proportion of ever-married women who are currently married for both the NFS and the census. For white women, the proportion currently married is similar for both sets of data; where there is a difference, the NFS sample tends to have a higher proportion currently married. For black women, the proportion currently married in the NFS sample is consistently lower than in the census. Although the number of black women is smaller than the number of white women, the differences in the proportion currently married appear to be greater than could be attributed to sampling variability.
The final comparison involves a variable that is central to any fertility study; the number of children born. Table I-8 shows the percent distribution and mean number of children ever born to ever-married women by race and age. Almost without exception, the average number of children born to women in our sample is greater than that reported for comparable women by the census. For whites the difference tends to be small, averaging one-tenth of a birth. For blacks the differences tend to be larger — with a maximum of one-half a child for women aged 40-44. Examination of the detailed distributions suggests that the difference between the means is primarily the result of a deficit of zero-parity women, and for black women aged 40-44 there is an excess of women with seven or more children. A similar shortage of zero-parity women was reported in previous U.S. fertility studies and undoubtedly relates to the difficulty of reaching women without children. But it may also be related to the deficit of central city residents in the 1970 NFS.
THE QUESTIONNAIRE
Two forms of interview questionnaire were employed, one for the currently married and one for formerly married women. The main difference in content is that the questionnaire for the currently married women contained items on current contraceptive behavior and on fertility intentions excluded from the form for the formerly married women. Otherwise, the two questionnaires were identical.
The questionnaire for the 1970 study is similar to that of its predecessor in 1965. There were some important changes, however, particularly in the phrasing of certain key questions — for example, those designed to assess the incidence of unwanted fertility. The details of these and other modifications are given in the course of our report, where attempts are made to compare 1965 and 1970 data. As in 1965, when the 1960 and 1955 surveys were behind us, we were again struggling in 1970 to reconcile the need for improvement with the wish to provide comparability over time. When changes in wording are introduced to improve the clarity of a question, changes that have occurred in the course of time are obscured and their interpretation made difficult. It is a conflict between improving validity at the cost of an unknown reduction in comparability. There is probably some tendency on our part to exaggerate the importance of such changes; we hasten to add that there are many areas in which very little, if any, change was made — attitudes toward abortion and sterilization, for example, coital frequency, and many other aspects of the measurement of fertility.
(Continues...)
Excerpted from The Contraceptive Revolution by Charles F. Westoff, Norman B. Ryder. Copyright © 1977 Princeton University Press. Excerpted by permission of PRINCETON UNIVERSITY PRESS.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Table of Contents
- Frontmatter, pg. i
- CONTENTS, pg. v
- Preface, pg. vii
- I Introduction, pg. 1
- II Contraceptive Practice, pg. 15
- III The Pill and the IUD, pg. 31
- IV Contraception: The Timing of First Use and Coital Frequency, pg. 51
- V Demographic and Social Aspects of Contraceptive Sterilization, pg. 71
- VI Attitudes Toward Abortion, pg. 163
- VII Reproductive Intention, Contraceptive Use, and the Fertility of Nonusers, pg. 179
- VIII Contraceptive Efficacy, pg. 227
- IX Wanted and Unwanted Fertility in the United States: 1965 and 1970, pg. 249
- X Differential Fertility, pg. 277
- XI Fertility Differences by Employment Patterns and Role Attitudes, pg. 311
- XII Summary and Conclusions, pg. 333
- Appendix A. Sampling Errors for the 1970 National Fertility Study, pg. 357
- Appendix B. Measurement of Interval of First Use of Contraception, pg. 371
- Appendix C-1. Differences in Question Phrasing in 1965 and 1970, pg. 377
- Appendix C-2. The Age-at-Marriage Bias, pg. 379
- Appendix C-3. Errors in Reporting Unwanted Status of First and Second Births, pg. 381
- Appendix D. Related Publications, pg. 383
- Index, pg. 385