SECTION 4 - TECHNICAL APPENDIX SOURCES OF DATA Marriage and divorce statistics in this volume are based on information from two sources: Complete counts of events obtained from all States and data on characteristics from marriage and divorce certificates from States meeting certain reporting criteria. State and local officials annually provide complete counts of marriages and divorces by county of occurrence, the number of annulments granted in each State, and marriages by month of occurrence. From these counts marriage and divorce totals are derived. Rates are computed for each State, geographic division, region, registration area, and for the United States. Data on marriages and divorces and on the persons involved for the States that constitute the marriage-registration area (MRA) and the divorce- registration area (DRA) are based on information from two sources. For some States, samples of records are drawn by the National Center for Health Statistics (NCHS) from microfilm copies of the original certificates received from the State registration offices. Other States submit computer tapes to NCHS through the Vital Statistics Cooperative Program (VSCP). For these States the complete file is used. Statistical data for 1988 for marriages and divorces were provided through the VSCP by nine States--Illinois, Missouri, Nebraska, New Hampshire, New York (except New York City marriages), Rhode Island, South Carolina, Vermont, and Virginia. States that participated in the VSCP for marriage data only were Florida, Maine, Montana, and Wisconsin. Marriage and divorce statistics for the United States, for the registration areas, and for individual States are limited to events occurring during the year and registered within the specified area. Interlocutory divorce decrees or decrees "nisi" (which grant that the divorce takes permanent effect at a specific time in the future if the couple does not reconcile in the meantime) are included among the records of the States that grant them, usually for the year in which they were granted. Tabulations are by place of occurrence and include events occurring to nonresidents. Marriages or divorces of members of the Armed Forces or other U.S. nationals that occur outside the United States are excluded from the tabulations. United States refers to the 50 States and the District of Columbia. Alaska has been included in the U.S. tabulations since 1959 and Hawaii, since 1960. Data for Puerto Rico and the Virgin Islands were compiled for 1988 and preceding years except 1966-69 and 1972. Data for these areas were processed separately from the rest of the MRA and DRA and appear in section 3 of this volume. Sources of marriage and divorce data since 1867 are shown in table 4-1; a description of these sources can be found in section 7, Volume III, Vital Statistics of the United States, 1960. COUNTS OF EVENTS Nationwide counts In 1988 the total counts of marriages by State and county were obtained from central files of marriage records for 47 States and the District of Columbia. The total count of marriages for New York City was obtained from a 100-percent sample of microfilmed marriage certificates. In the States tech88.txt - Page 1 without central files--Arizona, New Mexico, and Oklahoma--counts were obtained from counties by State officials and reported to NCHS. New Mexico reported the number of premarital health examination forms issued rather than the number of marriages performed. Counts of marriages performed in 1988 were not received from one county in Kentucky. Beginning with final marriage statistics for 1978, nonlicensed (confidential) marriages registered in California were included in national and geographic totals and rates. Section 4213 of the California Civil Code allows unmarried couples who have been living together to be married confidentially without obtaining a marriage license or health certificate. In 1972 this section was amended to require county clerks to keep sealed records of these marriages and periodically to report the total number to the California State Department of Health Services. Since reporting began, nonlicensed marriages have increased rapidly, from 2,857 in 1973 to a peak of 90,494 in 1987 and declined slightly in 1988 (88,203). Total counts of divorces by State and county include decrees of absolute divorce and of annulment as well as decrees of marriage dissolution introduced in many States during the 1970's. These counts were obtained from central files of 46 States and the District of Columbia. The total count of divorces for the District of Columbia was obtained from a 100-percent sample of microfilmed divorce records. Either local or State officials provided the county totals for the four States without central files of divorce records. Local officials in Indiana and New Mexico supplied the county totals. State officials in Arizona and Oklahoma obtained the county totals and forwarded them to NCHS. Divorces and annulments in 1988 were not reported for one county or more in the following States: California (2 counties), Indiana (9 counties), Louisiana (34 counties), Michigan (1 county), Mississippi (9 counties), and New Mexico (2 counties). When marriages or divorces are not reported for counties, the following criteria are applied for the data shown in tables 1-5, 1-6, 1-53, 2-3, 2-69, and 3-2: 1. The State totals are published as reported even when there are nonreporting counties. 2. If the population of nonreporting areas in the latest available census is less than 1 percent of the State population, the State total does not have a footnote to indicate the data are incomplete. 3. If the population of nonreporting areas is 1.0-9.9 percent of the State population, data for the State and division have the footnote "Data are incomplete." Rates are computed and no estimates are made for the missing data. 4. If the population of nonreporting areas is 10 percent or more of the State population, the State and division totals have the footnote "Data are incomplete," and the marriage or divorce rate is not computed. Estimates are made for inclusion in the totals for the region and the United States, and rates are computed. Figures for the region and United States have the footnote "Figure includes estimates." These criteria and footnotes apply also to data for standard metropolitan statistical areas (SMSA's) shown in tables 1-52 and 2-68. tech88.txt - Page 2 Marriages and marriage rates with seasonal adjustments The seasonally adjusted marriage rates shown in table 1-4 are computed from the X-11 variant of Census Method II (U.S. Bureau of the Census, 1967). This is a ratio-to-moving-average method of adjustment that adjusts for the total number of days in the month and for varying number of specific days in a month, particularly Saturdays. Standard metropolitan statistical areas The SMSA's used in this report (tables 1-52 and 2-68) are those established by the U.S. Office of Management and Budget, using final 1980 census population counts, and used by the U.S. Bureau of the Census except in the New England States. Except in the New England States, an SMSA is a county or a group of contiguous counties containing a city of 50,000 inhabitants with a total metropolitan population of at least 100,000. In addition to the county or counties containing such a city or cities, contiguous counties are included in an SMSA if, according to specified criteria, they are essentially metropolitan in character and are socially and economically integrated with the central city or cities. In the New England States, the U.S. Office of Management and Budget uses towns and cities rather than counties as geographic components of SMSA's. However, NCHS cannot use the SMSA classification for these States because its data are not coded to identify all towns. Instead, the New England County Metropolitan Areas are used. These areas are established by the U.S. Office of Management and Budget and are made up of county units. DATA FROM MARRIAGE AND DIVORCE CERTIFICATES Registration areas Registration areas for collecting marriage statistics were established in 1957 and for divorce statistics, in 1958. These areas include States with adequate programs for collecting marriage and divorce statistics. Criteria for participation in the registration areas are 1. A central file of marriage or divorce records 2. A statistical report conforming closely in content to the U.S. Standard License and Certificate of Marriage (figure 4-1) or U.S. Standard Certificate of Divorce, Dissolution of Marriage or Annulment (figure 4-2) 3. Regular reporting to the State office by all local areas in which marriages or divorces are recorded 4. Tests for completeness of marriage or divorce registration and completeness of reporting of items on the records carried out in cooperation with NCHS. In 1988, the MRA (figure 4-3) comprised 42 States, the District of Columbia, the independent registration area of New York City, Puerto Rico, and the Virgin Islands. The DRA (figure 4-4) included 31 States, the District of Columbia (added in 1986), and the Virgin Islands. Table 4-2 shows when States and independent registration areas established central files and when they were admitted to the MRA and DRA. tech88.txt - Page 3 In the statistical tabulations in this volume, marriage-registration area refers only to the 42 States and the District of Columbia. Divorce- registration area refers to the 31 States and the District of Columbia. Marriages in the MRA accounted for 81 percent of all marriages in the United States (77 percent excluding California nonlicensed marriages) in 1988, and divorces in the DRA accounted for 49 percent of all divorces. Standard certificates The U.S. Standard License and Certificate of Marriage (figure 4-1) and the U.S. Standard Certificate of Divorce, Dissolution of Marriage or Annulment (figure 4-2), issued by the Public Health Service, serve as the principal means of obtaining comparable data from documents used to collect information on marriages and divorces in the United States. Model reporting forms, prepared in close collaboration with the Public Health Conference on Records and Statistics, were first recommended to the States in 1954 for adoption January 1, 1955, and stood without revision through 1967. Revised standard certificates were recommended to the States for use beginning January 1, 1968. A description of the original standard certificate can be found in the Technical Appendix, Vital Statistics of the United States, 1960, Volume III; a description of the first revision can be found in the Technical Appendix of the same volume for 1968. A second revision of each of the standard certificates was recommended to the States for use beginning January 1, 1978. The marriage certificate was revised to include, in addition to date of birth, an item requesting age and an item for both the title of the person performing the ceremony and the type of ceremony (civil or religious) in place of the item for officiant (civil or religious). On the divorce certificate, items requesting person to whom decree was granted and legal grounds for divorce were deleted; the item requesting specification of plaintiff was changed to petitioner; and the items referring to children involved in divorce were changed to request number of children ever born alive of this marriage and number of children under 18 in this family. Each State and independent registration area determines the form and content of its vital records. Consequently, the records vary in certain details. Although modified in each State as required by particular needs or by special provisions of the State vital statistics law, the marriage and divorce certificates of most States in the MRA and DRA conform closely to the standard certificates. Table 4-3 lists items appearing on the U.S. Standard License and Certificate of Marriage and shows which States had the items on their forms that were used by NCHS in 1988. Table 4-4 shows the same information for the U.S. Standard Certificate of Divorce, Dissolution of Marriage or Annulment. Characteristics of marriages and divorces The general rules used to classify and code items for marriages and divorces are set forth in NCHS Instruction Manual, Parts 6 and 7, respectively, for 1988. The classification of certain important items is discussed below. The dates of birth, marriage, separation, and divorce are used to calculate the age at marriage, interval to remarriage, age at divorce, tech88.txt - Page 4 duration of marriage, duration of marriage to separation, and duration of separation to decree. The State of marriage, State of residence, and State of birth are used to derive residency and nativity status. Information about previous marriages of the spouses is obtained from two items--number of this marriage and previous marital status. First marriages and remarriages (referred to as marriage order) can be identified if either item is stated. Thirty-four States requested both items or similar information on their marriage certificates. The District of Columbia requested information on former marriages, from which the number of the marriage and previous marital status were derived. Colorado, Idaho, Maryland, and Oregon requested only previous marital status. For these States the number of not-stated cases for number of this marriage was reduced by the number of cases assumed to be first marriages because previous marital status was stated as single. Michigan and Ohio requested only the number of times previously married. Not-stated cases for previous marital status were reduced by the number of persons assumed to be single because they had been previously married "0" times. South Carolina requested number of this marriage. Not-stated cases were reduced by those assumed single because the number was "1." During computer processing of the marriage data, values were allocated for missing ages of bride and groom. Reporting of these items has been almost 100 percent complete since 1960. No allocation was done during computer processing of divorce data. However, before age-specific divorce rates were calculated, figures for age not stated were distributed. History of reporting Many tables in this volume contain data for several years. The number of States reporting certain items can change from year to year because States join or drop out of the registration areas or States in the registration areas revise their certificates and add or delete items. Tables 4-5 and 4-6 show the history of reporting items of statistical information coded from marriage and divorce records for each State that has ever been in the registration areas. As mentioned in a previous section, a major revision of the standard certificates occurred in 1968. Therefore, the history of reporting begins with 1968 and ends with data year 1988. Items reported by all States in the registration areas for all years were omitted from Tables 4-5 and 4-6. However, because the States in the registration areas changed over the period, the States reporting these items were not consistent. Following is a brief summary of reporting for each marriage item. ž Date of marriage; date of birth and age at marriage of bride and groom; State of marriage; and State of residence of bride and groom: 1968-The 39 States and the District of Columbia that were in the MRA reported these items. 1971-Minnesota and South Carolina joined the MRA and reported. 1979-Colorado joined the MRA and reported. 1988-All States that were in the MRA and the District of Columbia reported these items. tech88.txt - Page 5 ž State of birth of bride and groom: 1969-NCHS began coding State of birth of bride and groom. The 39 States that were in the MRA reported this item. The District of Columbia did not report. 1971-Minnesota and South Carolina joined the MRA and reported. 1979-Colorado joined the MRA and reported. 1988-All 42 States that were in the MRA reported State of birth of bride and groom. The District of Columbia did not report. ž Type of ceremony: 1972-NCHS began coding type of ceremony. The following 37 States and the District of Columbia reported this item: Alabama, Alaska, California, Connecticut, Delaware, Florida, Georgia, Hawaii, Idaho, Illinois, Indiana, Kansas, Louisiana, Maine, Maryland, Massachusetts, Michigan, Minnesota, Mississippi, Missouri, Montana, Nebraska, New Hampshire, New Jersey, New York, North Carolina, Oregon, Pennsylvania, Rhode Island, South Carolina, South Dakota, Tennessee, Utah, Vermont, Virginia, Wisconsin, and Wyoming. 1979-Colorado joined the MRA and reported. West Virginia began to report. 1985-Kentucky began to report. 1987-Nebraska ceased to report. 1988-Thirty-nine States and the District of Columbia reported type of ceremony. ž Race: 1968-The following 35 States and the District of Columbia reported this item: Alabama, Alaska, Connecticut, Delaware, Florida, Georgia, Hawaii, Idaho, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maine, Maryland, Massachusetts, Mississippi, Missouri, Montana, Nebraska, New Hampshire, New Jersey, North Carolina, Oregon, Pennsylvania, Rhode Island, South Dakota, Tennessee, Utah, Vermont, Virginia, West Virginia, Wisconsin, and Wyoming. 1971-Minnesota and South Carolina joined the MRA and reported. Maryland ceased to report. 1972-Massachusetts ceased to report. 1975-The District of Columbia ceased to report. 1977-Iowa ceased to report. 1988-Thirty-four States reported race. ž Education: 1970-NCHS began coding education. The following 14 States reported this item: California, Hawaii, Illinois, Kansas, Louisiana, Nebraska, New Hampshire, North Carolina, Rhode Island, Tennessee, Utah, Vermont, Virginia, and Wyoming. 1971-Minnesota and South Carolina joined the MRA and reported. 1972-Iowa began to report. 1974-South Carolina ceased to report. 1975-Missouri began to report. tech88.txt - Page 6 1976-Montana began to report. Minnesota ceased to report. 1978-Iowa ceased to report. Maine and Wisconsin began to report. 1979-Mississippi began to report. 1981-Connecticut began to report. 1984-Kentucky began to report. 1988-Alabama and Indiana began to report; thus for 1988, 23 States reported education. ž Previous marital status: 1968-The 39 States and the District of Columbia that were in the MRA reported this item. For Michigan and Ohio, previous marital status was derived from the "number of times previously married" item. 1971-Minnesota and South Carolina joined the MRA. Minnesota reported previous marital status while the previous marital status for South Carolina was derived from the "number of this marriage" item. 1977-Iowa ceased to report previous marital status and number of this marriage, so no information on marital history was available for Iowa. 1979-Colorado joined the MRA and reported. 1988-All MRA States except Iowa reported previous marital status or other marital history items from which previous marital status information was derived. However, most tables that include previous marital status in this volume used only the 38 States and the District of Columbia that specifically reported previous marital status. ž Number of this marriage: 1968-The 39 States and the District of Columbia that were in the MRA reported this item. For Idaho, Maryland, and Oregon, number of this marriage was derived from the "previous marital status " item. 1971-Minnesota and South Carolina joined the MRA and reported. 1977-Iowa ceased to report previous marital status and number of this marriage, so no information on marital history was available for Iowa. 1979-Colorado joined the MRA and reported previous marital status from which number of this marriage was derived. 1988-All MRA States except Iowa reported number of this marriage or previous marital status from which number of this marriage was derived. However, most tables that include number of this marriage in this volume used only the 37 States and the District of Columbia that specifically reported number of this marriage. ž Marriage order: Marriage order is a variable derived from either the number of this marriage or the previous marital status item. Therefore, States that report either item are included. The history of reporting of marriage order is similar to that of previous marital status or number of this marriage. For tables involving marriage order in this volume, the District of Columbia and all MRA States except Iowa are included. tech88.txt - Page 7 ž Date last marriage ended: 1970-NCHS began coding date last marriage ended. The 39 States and the District of Columbia that were in the MRA reported this item. 1971-Minnesota and South Carolina joined the MRA and reported. 1978-Iowa ceased to report. 1979-Colorado joined the MRA and reported. 1988-The District of Columbia and all MRA States except Iowa reported date last marriage ended. Following is a brief summary of reporting for each divorce item. ž Date of divorce; number of children under 18; date of birth/age at decree of husband and wife; and State of divorce: 1968-The 26 States that were in the DRA reported these items. 1969-Kentucky and New York joined the DRA and reported. 1971-South Carolina joined the DRA and reported. 1978-California dropped out of the DRA. 1979-Massachusetts and New Hampshire joined the DRA and reported. 1981-Delaware joined the DRA and reported. 1986-The District of Columbia joined the DRA and reported. 1988-All 31 DRA States and the District of Columbia reported these items. ž State of residence: 1976-NCHS began coding State of residence. The 29 States that were in the DRA reported this item. 1978-California dropped out of the DRA. 1979-Massachusetts and New Hampshire joined the DRA and reported. 1981-Delaware joined the DRA and reported. 1986-The District of Columbia joined the DRA and reported. 1988-All 31 DRA States and the District of Columbia reported State of residence. ž State of marriage: 1972-NCHS began coding State of marriage. The following 29 States reported this item: Alabama, Alaska, California, Connecticut, Georgia, Hawaii, Idaho, Illinois, Iowa, Kansas, Kentucky, Maryland, Michigan, Missouri, Montana, Nebraska, New York, Ohio, Oregon, Pennsylvania, Rhode Island, South Carolina, South Dakota, Tennessee, Utah, Vermont, Virginia, Wisconsin, and Wyoming. 1978-California dropped out of the DRA. 1979-New Hampshire joined the DRA and reported. 1981-Delaware joined the DRA and reported. 1988-Thirty States reported State of marriage. ž Number of children ever born alive of this marriage: 1970-NCHS began coding number of children ever born alive of this marriage. The following 19 States reported this item: Alaska, tech88.txt - Page 8 California, Connecticut, Hawaii, Idaho, Illinois, Iowa, Kansas, Maryland, Michigan, Nebraska, New York, Pennsylvania, Rhode Island, Tennessee, Utah, Vermont, Wisconsin, and Wyoming. 1974-Missouri began to report. 1976-Oregon began to report. 1977-Georgia began to report. 1978-California dropped out of the DRA. Alaska, Connecticut, Georgia, Maryland, and Oregon ceased to report. Alabama, Montana, and Virginia began to report. 1979-Alaska, Connecticut, and Maryland began to report. Massachusetts and New Hampshire joined the DRA and reported. 1980-Maryland ceased to report. 1981-Maryland began to report. 1982-Georgia began to report. 1988-Twenty-five States reported number of children ever born alive of this marriage. ž Date couple separated: 1970-NCHS began coding date couple separated. The following 19 States reported this item: Alaska, California, Connecticut, Hawaii, Idaho, Illinois, Kansas, Michigan, Missouri, Nebraska, New York, Rhode Island, South Dakota, Tennessee, Utah, Vermont, Virginia, Wisconsin, and Wyoming. 1972-Iowa began to report. 1977-Montana began to report. 1978-California dropped out of the DRA. Alabama and Montana began to report. 1979-New Hampshire joined the DRA and reported. South Dakota ceased to report. 1988-Twenty-one States reported date couple separated. ž Number of this marriage: 1968-The following 25 States reported this item: Alabama, Alaska, California, Connecticut, Georgia, Hawaii, Idaho, Illinois, Iowa, Kansas, Maryland, Michigan, Missouri, Montana, Ohio, Oregon, Pennsylvania, Rhode Island, South Dakota, Tennessee, Utah, Vermont, Virginia, Wisconsin, and Wyoming. 1969-Kentucky and New York joined the DRA and reported. 1971-South Carolina joined the DRA and reported. 1978-California dropped out of the DRA. Nebraska began to report. 1979-Massachusetts and New Hampshire joined the DRA and reported. South Dakota ceased to report. 1981-Delaware joined the DRA and reported. 1986-The District of Columbia joined the DRA and reported. 1988-Thirty States and the District of Columbia reported number of this marriage. ž Education: 1970-NCHS began coding education. The following 15 States reported this item: Alaska, California, Connecticut, Hawaii, Illinois, Kansas, tech88.txt - Page 9 Michigan, Nebraska, New York, Rhode Island, Tennessee, Utah, Vermont, Virginia, and Wyoming. 1972-Iowa began to report. 1974-Missouri began to report. 1978-California dropped out of the DRA. Alabama, Montana, and Wisconsin began to report. 1979-New Hampshire joined the DRA and reported. 1982-Georgia began to report. 1988-Twenty-one States reported education. ž Race: 1968-The following 25 States reported this item: Alabama, Alaska, California, Connecticut, Georgia, Hawaii, Idaho, Illinois, Iowa, Kansas, Maryland, Michigan, Missouri, Montana, Nebraska, Oregon, Pennsylvania, Rhode Island, South Dakota, Tennessee, Utah, Vermont, Virginia, Wisconsin, and Wyoming. 1969-Kentucky and New York joined the DRA and reported. 1970-Michigan ceased to report. 1971-South Carolina joined the DRA and reported. 1978-California dropped out of the DRA. 1981-Delaware joined the DRA and reported. 1986-The District of Columbia joined the DRA and reported. 1988-Twenty-seven States and the District of Columbia reported race. ž If previously married, how many ended by death; if previously married, how many ended by divorce: 1970-NCHS began coding how last marriage ended. The following 16 States reported this item: Alaska, California, Connecticut, Hawaii, Illinois, Iowa, Kansas, Michigan, New York, Rhode Island, South Dakota, Tennessee, Utah, Vermont, Wisconsin, and Wyoming. 1971-Georgia and Wisconsin ceased to report. South Dakota ceased to report if previous marriage ended by death. 1972-South Dakota ceased to report if previous marriage ended by divorce. Wyoming ceased to report if previous marriage ended by death. 1973-Wyoming began to report if previous marriage ended by death. 1974-Missouri began to report. 1975-Alabama, Georgia, Idaho, Kentucky, Maryland, Montana, Ohio, Oregon, Pennsylvania, South Carolina, South Dakota, Virginia, and Wisconsin began to report. 1978-California dropped out of the DRA. Nebraska began to report. 1979-Massachusetts and New Hampshire joined the DRA and reported. South Dakota ceased to report. 1981-Delaware joined the DRA and reported. 1986-The District of Columbia joined the DRA and reported. 1988-Thirty States and the District of Columbia reported how last marriage ended. tech88.txt - Page 10 Children involved in divorce To obtain national estimates for 1970-88, the average number of children under 18 years of age per decree for the DRA was multiplied by the national divorce total. For 1965-69 the average number of children per decree was computed for the 16 States that had reported children with a satisfactory degree of completeness in 1965 (Alaska, Hawaii, Idaho, Iowa, Kansas, Michigan, Missouri, Montana, Nebraska, Ohio, Oregon, South Dakota, Tennessee, Virginia, Wisconsin, and Wyoming); the national divorce total was multiplied by this average. This method also was used to revise figures for 1961-64, originally estimated by using a different method. Consequently, data for 1961-64 shown in table 2-11 differ from figures appearing in Volume III, Vital Statistics of the United States before 1965. In 1960 the estimate was based on sample data for the United States. For earlier years total counts from States reporting the number of children were used to estimate the children involved in divorce for the United States. Table 2-24 shows the number of children under 18 years of age involved in divorce by race. The race of a child for this table is inferred from the race of the husband and wife. When both spouses are the same race, the child is assigned to that racial category. When the husband is white and the wife is not, the child is assigned to the wife's race. When the husband is not white, the child is assigned to the husband's race. When the race of only one spouse is stated, the child is assigned to the stated race. When the race of both spouses is not stated, the child is assigned to race not stated. Sampling procedures The registration area samples were designed to yield estimates of State totals as well as frequency distributions by characteristics of the spouses. A sampling rate was designated for each of the States so that the selected sample would consist of at least 2,500 records. Five different sampling rates were used--all records, 1/2, 1/5, 1/10, and 1/20. All records were automatically included in the samples for States submitting data through the VSCProgram. Overall, about 42 percent of all marriages in the MRA and about 49 percent of all divorces in the DRA were included in the samples for 1988. 1988. The sampling procedures assure for each State a relative variance of 4 percent or less on a 1-percent estimate, that is, on an estimate of a characteristic that comprises 1 percent of the total events in the State. The sampling rates, number of events, and sample cases by State for 1988 are shown in tables 4-7 (marriages) and 4-8 (divorces). Area totals The estimated totals based on sample data differ slightly from the figures obtained from complete counts of events furnished by the State and local offices. When possible major differences are reconciled before the sample data are weighted and tabulated. However, even in States where all records are included in the sample, the sample estimate may differ from the State count because of differences between NCHS and State cutoff dates in accepting records. The sample estimate of marriages in the MRA differed from the complete count in 1988 by 4.5 percent. However, when California nonlicensed marriages tech88.txt - Page 11 were excluded from the count, the sample estimate differed from the complete count by less than 1 percent. Nonlicensed marriages were not available for inclusion in the MRA marriage sample (see earlier discussion in "Nationwide counts"). The 5-percent sample of California marriages for 1988 was drawn from 140,945 licensed marriages. The corresponding weight for California sample records was based on this total. Because no information is available on the characteristics of persons obtaining nonlicensed marriages, no estimate is available for the bias introduced in sample estimates by the omission of the marriages. The sample estimate of marriages differed from the complete count by less than 3 percent for all States except for California (39 percent) when nonlicensed marriages were included in the count. The sample estimate of divorces in the DRA differed from the complete count by less than 0.2 percent. It differed by 2 percent or less in all States. Sampling errors Statistics for States where all records were used have no sampling error. For other States each statistic has a sampling error (or standard error) denoted by å(x). The sampling error is primarily a measure of variation that occurs by chance because information from a sample of records is used rather than from the entire population of records. The sampling error can be computed because within each State or registration area the probability that a record will be chosen for the sample is known (either all records, 1/2, 1/5, 1/10, or 1/20, as shown in tables 4-7 and 4-8). For about 68 percent of all possible samples of equal size, the estimate from the sample would differ from the value obtained from a complete count by less than the standard error. For about 95 percent of all samples, the difference would be less than twice the standard error. Sampling errors are shown in table 4-9 for marriages and table 4-10 for divorces. To facilitate comparisons among registration States with widely varying totals of marriages and divorces, the sampling errors for the registration areas and participating States are shown for frequencies expressed as percents of State or registration area totals. For example, table 1-38 shows 19,029 marriages to brides 65 years of age and over in the MRA for 1988. This estimated frequency is about 1 percent of all brides in the MRA (1,852,275). In table 4-9 under the column "1 or 99" in the row for the total MRA, the sampling error is 393. Thus chances are about 68 of 100 that the actual number of marriages to brides 65 years of age and over is within one standard error of the estimate from this particular sample (19,029 ń 393), or between 18,636 and 19,422. The interval from 18,636 to 19,422 is called a 68-percent confidence interval for the actual number of marriages of brides 65 years of age and over. The sampling error of a given frequency of marriages or divorces in a particular State can be calculated from the formula ~ ŚÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ å (xij) = ū N1(N1-n1)pijqij ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ n1-1 where tech88.txt - Page 12 i = a State or registration area j = a subclass, such as an age group of brides Ni = the sample estimate of events for State i (table 4-7 or 4-8) ni = the number of sample records in State i (table 4-7 or 4-8) xij = estimated frequency of the characteristic under study Pij = xij/Ni, estimated proportion of all cases with the characteristic j to be studied for State i (shown in table 4-9 or 4-10 as a percent under "1 or 99," "2 or 98," "3 or 97," and so forth) qij = 1-pij. In this formula, when the number of sample records, ni, is the same as the total marriages, Ni, the numerator becomes zero, and there is no sampling error for States that have all their records processed. Because the values of pij and qij are complementary, their product (pijqij) remains the same if their values are reversed. For instance, the estimated sampling error for an estimate of 5 percent is the same as the sampling error for 95 percent. Therefore, the sampling error for both values is listed in one column in tables 4-9 and 4-10 under the heading "5 or 95." For example, to calculate the sampling error for previously widowed Alabama brides Ni = 44,610, the estimated number of Alabama marriages (table 4-7) ni = 4,461, the number of sample records for Alabama (table 4-7) xij = 1,890, the estimated frequency of previously widowed Alabama brides (table 1-29) pij = the number of previously widowed Alabama brides expressed as a percent of all marriages: 1,890/44,610 = 0.04 or 4 percent qij = 1-pij = 0.96 or 96 percent and, ~ ŚÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ å (1890) = ū 44,610(40,149) (.04) (.96) ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ = 124 4,460 So chances are 68 of 100 that the actual number of previously widowed Alabama brides is between 1,766 and 2,014 (1,890 ń 124). The estimated sampling error for the MRA can be computed from the formula ŚÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ ~ ³43 å (xi) = ū Nk(Nk-nk) ä ÄÄÄÄÄÄÄÄÄÄPkjqkj nk-1 k=1 tech88.txt - Page 13 where 43 = the 42 States and the District of Columbia in the MRA Nk = the sample estimate of events for the kth MRA State nk = the number of sample records in the same State Pkj = the estimated percent of all cases in State k with the characteristic j qkj = 1-pkj. Using the same formula, sampling errors for the DRA shown in table 4-10 were computed by summing over the 31 States and the District of Columbia in the DRA. Many of the tables in this volume show percent distributions rather than frequencies. The sampling error for such a percent can be computed by dividing the sampling error found in table 4-9 or table 4-10 by the estimated total number of events for the area. For example, table 1-11 shows that 4.3 percent of Alabama brides were previously widowed. In table 4-9 under the column "4 or 96" the Alabama sampling error is 124. Divide 124 by 44,610, the estimated number of marriages in Alabama as shown in the stub of the same table. The result is 0.3 percent, the sampling error for an estimated 4 percent of 1988 Alabama marriages. Thus chances are about 68 of 100 that the actual percent of previously widowed Alabama brides is (4.3 ń 0.3), or between 4.0 and 4.6 percent. The following formula is used to compute a sampling error for the proportion of events in a subclass of a subgroup, where the number of events in the subgroup is also subject to sampling error: ~ ŚÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ ~ ³~ ~ å (x/y) = (1/y) ū åż(x)-(x/y)ż åż(y) ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ where y = the estimated frequency of the subgroup x = the estimated frequency of the subclass of the subgroup ~ å(y) = the sampling error of y from table 4-9 or 4-10 ~ å(x) = the sampling error of x from table 4-9 or 4-10. To illustrate, for the sampling error for the proportion of all previously married Michigan brides who were 45-54 years of age, let x = 3,080 previously married Michigan brides who were 45-54 years of age (table 1-35) y = 25,580 previously married Michigan brides tech88.txt - Page 14 x/y = 0.120 ~ ~ å(x) = å(3,080) = 234 from "4 or 96" in table 4-9, as x is 0.04 of all Michigan brides (x/N = 3,080/75,300 = 0.04) ~ ~ å(y) = å(25,580) = 547 by interpolation between "25 or 75" and "50" in table 4-9, as y is 0.34 of all Michigan brides (y/N= 25,580/75,300 = 0.34) and, ~ ( 1 ) ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ å (x/y) = (ÄÄÄÄÄÄÄ) ū 234ż-(.120)ż(547)ż (25,580 ) = 0.009 or 0.9 percent. So chances are 68 of 100 that the actual proportion of previously married Michigan brides who were 45-54 years of age is (0.120 ń 0.009), or between 11.1 percent and 12.9 percent. The statistical significance of the difference between two proportions may be tested by the following formula: ________________________ ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ ū ~ ~ åż (x1/y1) + åż (x2/y2) where the x's and y's represent the same subgroups as above, but of two populations, and where the denominator, under the square root sign, represents the sampling error of the difference between the proportions x1/y1 and x2/y2 shown in the numerator of the formula. If this ratio is 2 or larger, the difference between proportions is considered significant at the 0.05 level. For example, 12.0 percent of previously married brides in Michigan were 45-54 years of age, with a sampling error of 0.9 percentage point. Using the same procedure for previously married brides in New Jersey, 0.136 or 13.6 percent were 45-54 years of age, with a sampling error of 0.011 or 1.1 percentage points. Substituting these values in the formula above gives the following results: .136-.120 ________________________ .016 ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ = ÄÄÄÄ = 1.1 ū (.011)ż+(.009)ż .014 Because the difference between the proportions divided by the sampling error is less than 2, this difference of 1.6 percentage points between remarried brides 45-54 years of age in Michigan and those brides in New Jersey is expected to occur by chance more than once in 20 times; the difference is not statistically significant at the 0.05 level. Therefore, it cannot be concluded that more of the previously married brides were 45-54 years of age in New Jersey than in Michigan. tech88.txt - Page 15 Information not reported Nonresponse errors in marriage and divorce data may arise from any of the following three sources. 1. The item is not on the certificate of a particular reporting area. For example, the item "color" was removed from the 1971 revision of the Massachusetts marriage record. 2. Sometimes when a State revises its form to include additional items, copies of the old form are used by registration offices until their supply is exhausted. 3. Although the item may be on the form, there may be no response, or the response may be illegible or unusable. Items not answered for any of these three reasons are tabulated as "not stated." Tables 4-11 and 4-12 show the proportions of not-stated cases for specified characteristics of marriages and divorces for 1988. A major proportion of the not-stated cases for the MRA and DRA was due the absence of items on some of the reporting forms (not reportable). On the divorce records, information about the husband and wife (age, race, and so forth) typically has been less completely reported than information pertaining to the decree (the month of divorce and the petitioner). Other sources of error Inconsistent responses--Occasionally, for marriages and divorces, the responses to related items are inconsistent. For example, the previous marital status of the bride or groom, as shown on the marriage certificate, may not be consistent with the stated number of times married or the number of this marriage. Such differences are reconciled before tabulation. Ambiguous wording of items--Some of the items are worded so that they do not yield information easily classified in standardized categories. This often leads to errors in classification, most frequently of geographic place names (residence, place of marriage, and place of birth) and of previous marriages. Microfilm errors--Errors in the arrangement of the records received on microfilm in NCHS sometimes occur when the numbers on the records are out of sequence, missing, or duplicated, or when records that are not part of the universe being sampled are erroneously included in the file. Processing errors--Errors may arise in coding, keying, or tabulating the data. However, 100 percent of sample coding was verified; data already on computer tape were edited for impossible codes and for inconsistencies. Generally, for most of the data from 1960 onward, the completeness and year-to-year consistency are evidence of the basic reliability and coverage of the marriage- and divorce-registration systems. COMPUTATION OF RATES Population bases The rates shown in this report were computed on the basis of population statistics prepared by the U.S. Bureau of the Census. Rates for the census tech88.txt - Page 16 years 1940-80 are based on the population enumerated as of April 1 in the censuses of those years. Rates for all other years are based on the estimated midyear (July 1) population for the respective years. Rates are based on population estimates including Armed Forces stationed in the United States but excluding Armed Forces abroad, except divorce rates for 1941-46 which are based on populations including Armed Forces abroad. Since 1980, rates by previous marital status for the MRA have been based on the unmarried population 15 years of age and over rather than 14 years and over, which had been the standard until 1980. Marriages of brides and grooms under 15 years of age were included in data for the MRA totals and the youngest age group. The resident population for the United States for 1920-88, the MRA for 1957-88, and the DRA for 1958-88 is shown in tables 4-13 and 4-14. The sources for these populations are shown in table A. The resident population estimate for 1988 for each State is shown in table 4-15. The unmarried population by age, marital status, and sex for the MRA is shown in table 4-16, and the married population by age and sex for the DRA is shown in table 4-17. Rates for 1971-79, except those by marital status, have been revised based on revised population estimates that are consistent with the 1980 census levels and therefore may differ from rates published in the volumes for those years. The 1980 census counted about 5.5 million more persons than had been estimated for April 1, 1980. The revised estimates for the United States are published by the U.S. Bureau of the Census in the Current Population Reports, Series P-25, Number 1022. Population estimates by month are based on data published in the Current Population Reports, Series P-25, Number 1023. Unpublished revised estimates for States were obtained from the U.S. Bureau of the Census. tech88.txt - Page 17 Table A. Sources for Resident Population: United States, 1920-88, and States in Marriage- and Divorce-Registration Areas, 1957-88 ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ Area and Year | Source ÄÄÄÄÄÄÄÄÄÄÄÄÄÄ|ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ United States | | 1988----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 1045, Jan. 1990. 1986-87-------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 1022, Mar. 1988. 1985----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 1000, July 1986. 1983-84-------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 965, Mar. 1985. 1982----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 949, May 1984. 1981----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 929, May 1983. 1980----------| U.S. Bureau of the Census, U.S. Census of Population: | 1980, General Population Characteristics, PC(80)-1, | A1-A52, 1983. 1971-79-------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 917, July 1982. 1970----------| U.S. Bureau of the Census, U.S. Census of Population: | 1970, Number of Inhabitants, Final Report, PC(1)-A1, | United States Summary, 1971. 1969----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 441, Mar. 19, 1970. 1968----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 416, Feb. 17, 1969. 1967----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 385, Feb. 14, 1968. 1966----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 352, Nov. 18, 1966. 1960-65-------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 321, Nov. 30, 1965. 1951-59-------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 310, June 30, 1965. 1940-50-------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 499, May 1973. 1920-39-------| National Office of Vital Statistics, Vital Statistics | Rates in the United States, 1900-1940, 1947. | tech88.txt - Page 18 Table A. Sources for Resident Population: United States, 1920-88, and States in Marriage- and Divorce-Registration Areas, 1957-88 ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ Area and Year | Source ÄÄÄÄÄÄÄÄÄÄÄÄÄÄ|ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ States in | marriage- and | divorce- | registration | areas | | 1988----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 1044, Aug. 1989. 1986-87-------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 1024, May 1988. 1985----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 1010, Sept. 1987. 1983-84-------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 970, June 1985. 1982----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 951, May 1984. 1981----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 930, April 1983. 1980----------| U.S. Bureau of the Census, U.S. Census of Population: 1980, | General Population Characteristics, PC(80)-1, A1-A52, 1983. 1971-79-------| U.S. Bureau of the Census, unpublished data. 1970----------| U.S. Bureau of the Census, U.S. Census of Population: | 1970, Number of Inhabitants, Final Report, PC(1)-A1, | United States Summary, 1971. 1969----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 436, Jan. 7, 1970. 1968----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 414, Jan. 28, 1969. 1967----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 380, Nov. 24, 1967. 1965-66-------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 354, Dec. 8, 1966. 1964----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 324, Jan. 20, 1966. 1963----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 289, Aug, 31, 1964. 1962----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 273, Oct. 4, 1963. 1961----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 239, Dec. 13, 1961. 1960----------| National Office of Vital Statistics, Vital Statistics | Rates in the United States, 1900-1940, 1947. 1959----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 212, Jan. 26, 1960. 1958----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 208, Dec. 7, 1959. 1957----------| U.S. Bureau of the Census, Current Population Reports, | Series P-25, No. 186, Oct. 27, 1958. ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ tech88.txt - Page 19