Methodology

This section describes the criteria for status and policy indicator selection, data sources and limitations, grading and ranking, and modifications from the 2004 Report Card, along with information about the demographic data sources. 

Status Indicator Methodology

Criteria for Indicator Selection
Health status indicators were selected based primarily on whether they had a significant impact on women’s quality of life, functioning, and well-being, and whether they affected a large number of women generally or a large number of women in a specific population and/or age group.  Additional criteria were whether the indicator could be affected through intervention, prevention or improvement; was potentially measurable; was commonly used or there existed consensus on use; or reflected an emerging important issue where the problem was increasing in prevalence, incidence, or severity.

Data Sources and Limitations
The Report Card uses the best data available for each indicator and, wherever possible, data sets that are comparable for the states and the nation.  Thus, for certain indicators, while more recent data might be available for the nation, the Report Card uses older national data that is comparable with state-level data, as noted in the data source notes.  With few exceptions, the information presented in the Report Card is based upon data collected at the state level and reported by sex.  Exceptions include a few indicators based on data not reported by sex, but where general population data were viewed as a reliable reflection of women’s health status (such as the number of people living in medically underserved areas). 
 
Some national data on key measures of women’s health are included, even though no reliable state data are available for these state indicators (i.e., osteoporosis, unintended pregnancies, and violence against women), given their importance to women’s health.  When available, data are also presented by race, by ethnicity and by age, but these data are collected inconsistently by states and national surveys (see also the Demographic Data Sources section that follows for more on race and ethnicity data).  Although these data are generally from the same data source as the overall data, there may be differences in methodologies and data years, which are explained in the data source notes. Data collection for the status indicators ended in August 2007. 

Grading and Ranking

Benchmarks
In devising the grading process, Report Card staff and advisors endeavored to find the best benchmarks available for each indicator.  For most indicators, the Report Card grades the nation and states against benchmarks drawn primarily from the health objectives set for the nation by the U.S. Department of Health and Human Services’ Healthy People 2010 agenda, which provides a roadmap for where the nation’s health should be by the year 2010.  In cases where there is no Healthy People benchmark, states are graded against another benchmark decided upon by the Report Card authors with the input of experts.  In some cases, no appropriate benchmark was found and the states are ranked and not graded.  Benchmarks and sources are found in the table below.

As part of the ongoing Healthy People 2010 process, the U.S. Department of Health and Human Services conducts a midcourse review to assess the status of national health objectives.  This review assesses data trends during the first half of the decade, considers new science and available data, and revises objectives accordingly. In December 2006, the midcourse review of the Healthy People 2010 objectives was released, with some revisions to baseline data that affect the benchmarks that the Report Card uses. (5)  However, in order to maintain a consistent and comparable set of data and benchmarks, the Report Card authors have maintained the benchmarks set in the original release of the Healthy People 2010 agenda.

Grading
The nation and states are graded as follows.  First, the raw data for each indicator is expressed as a percentage difference from the benchmark for that indicator.  Next, the percentage differences from the benchmarks are scaled to range between 0 and 100, in order to account for the differences in the magnitude and the range of each indicator.  For example, the Healthy People 2010 benchmark for increasing the consumption of fruits and vegetables is 50%, while the benchmark for increasing blood cholesterol testing is 80%.  The range for each of these indicators is markedly different; for example, the consumption of fruits and vegetables indicator has a range that is 33 %larger than the range for the blood cholesterol testing indicator.  Scaling the percentage differences from a given benchmark addresses this problem (further information on how raw data was converted to scaled scores is available at this note). (6)

Once the states are assigned scaled scores, they are graded based on those scores.  A state that meets the benchmark receives a score of 100 and a grade of "Satisfactory." A state that receives a score of between 75 and 99 receives a "Satisfactory Minus," a state that receives a score of between 55 and 74 receives an "Unsatisfactory," and a state that receives a score of below 55 receives a "Fail." The worst state receives a score of 0.  A score of 50 means that a state's performance is halfway between the worst state and the benchmark.  The 75 and 55 cutoff scores were determined by a panel of experts and were chosen to reflect how far states are from the benchmarks, recognizing that states still have several years to achieve the Healthy People 2010 benchmarks.  Nonetheless, a few states are already meeting these standards, as reflected in receiving the grade of "Satisfactory".  A state's overall score was computed by averaging the scores on the 27 individual indicators for which there are appropriate benchmarks and state data.  Each state's overall score was then used to determine both the overall grade and the rank for the state.  Each status indicator grade is given equal weight in calculating the total grade.  The nation is graded in the same manner.

 

Modifications from previous Report Cards
This Report Card presents the 2004 overall data for each indicator in addition to the 2007 overall data to show where states have made progress or not on these health status indicators. However, where there are not updated data available, or where the Report Card uses a source that is different from and/or not comparable to the source used in 2004, the 2004 data are not shown.

Policy Indicator Methodology

Criteria for Indicator Selection
The policy indicators examine state policies and programs important to women’s health—whether statutes, regulations, executive orders, or other manifestations of state policies and programs. The criteria used to select the indicators for state health policies are similar to those used to select the health status indicators.  State policy indicators were selected based on whether they addressed and could have a significant positive impact on the critical women’s health issues reflected in the status indicators and whether they were measurable and comparable across states.

While the status and policy indicators are closely connected, some state policy indicators are included even though there is no status indicator that correlates directly to those policies.  In cases where there were no reliable data for every state describing the extent of a major women’s health problem (e.g. domestic violence) the Report Card included state policies that address that problem.

Data Sources and Limitations
Generally, the Report Card includes state health policy information that was collected from published or online sources, as noted in the data source.

Adopting the policies covered by the indicators can improve women’s health, but the states' actual implementation is a crucial component in determining whether and how much the policies impact women's health. Generally, the Report Card does not explore the effectiveness of state implementation efforts or subsequent judicial actions because such data are not routinely or consistently available. Sources did not always note delayed effective dates of policies (e.g., a statute was passed in 2005, but not effective until 2006). Since it could not be reasonably determined that sources identified delayed effective dates uniformly (e.g., that some states with delayed effective dates were not identified) and since the adoption of the relevant policy still demonstrates some state commitment, the 2007 Report Card considers a state to be in the relevant category regardless of effective date. Data collection for the policy indicators ended in May 2007.

Evaluating the Policies
States are compared, but not graded, on the policy indicators.  In contrast to the status indicators—where basic data were available, although with serious gaps—the absence of consistently collected policy data precluded meaningful comparisons of the states in key policies areas, such as health program budget expenditures.  For all the policy indicators, the strength of each state's policy is indicated on the state report card pages by the designations "Meets Policy," "Limited Policy," "Weak Policy" and "No/Harmful Policy."  With certain indicators, the 2007 Report Card's lowest category is called "Harmful Policy" in order to recognize that states can adopt policies which are just as harmful (and in some cases more so) as having no policy at all.

The Report Card authors determined the categorizations for each of the policies after research and input from experts.  Some policies have all four categories, others have three or two categories depending most often on the scope of the policy.  Each state’s performance in the 2007 Report Card is compared to its performance in the 2004 Report Card on every indicator, except where noted.

Modifications from the 2004 Report Card
Whenever possible, this fourth Report Card uses updated information from the same source or sources that were used for the indicators in the 2004 Report Card. If those sources were not available, the 2007 Report Card uses other reliable sources. When no such updates were available, the 2007 Report Card either eliminated the indicator entirely or, when the data could still be meaningful, included the indicator data from the 2004 Report Card.  Comparisons between data in the 2004 and 2007 Report Cards reflect 2004 information with any corrections to the 2004 data as noted.
 
In an effort to present the most accurate and timely policy issues important to women, modifications were made to this Report Card. Four indicators have been eliminated due to lack of data that could still be meaningful. These are maternity stays, private insurance smoking cessation coverage, restrictions on second-hand smoke and monitoring environmentally-related diseases.  One indicator was eliminated for more substantive reasons. Due to recent changes in federal law, the 2007 Report Card does not include an evaluation of state bans on abortion procedures. In April 2007, the Supreme Court upheld a federal law that prohibits a medically-approved abortion method with no exception to protect a woman’s health. Since this is a federal ban, it is in place in every state in the nation. In response to the Supreme Court decision, state laws on this issue are in flux. It is expected that this indicator will be revisited in the next Report Card. As replacement for this indicator, the 2007 Report Card now evaluates insurance coverage of abortions.

Demographic Data Sources
The demographic profile for each state and the nation as a whole includes data that provide the context for the Report Card status and policy indicators.  Most of the demographic data presented here derive from the most recent two years (2006 and 2007) of the U.S. Census Bureau’s Current Population Survey (CPS) (except as noted), and the most recent year (2005) of the U.S. Census Bureau’s American Community Survey and Bridged-Race Postcensal Population Estimates as described below.  The most recent two years of CPS data are used to increase the sample size of women in the analysis and to improve accuracy, especially for smaller states.

Listed below are the sources for the specific demographic data on the national and state report card pages.

Population of Females by Race, by Age, and Total (% and #), 2005.
EXPLANATION: This measure includes females of all ages as a percentage of the total civilian, non-institutionalized (9) population of the state.  Data by race and ethnicity are in the following categories:  White (non-Hispanic), Black (10) (non-Hispanic), Native American/Alaskan Native (non-Hispanic), Asian/Pacific Islander (non-Hispanic), and Hispanic.  Data by age reflect the percentage of females in the following age categories:  18-44, 45-64, 65 and older.
SOURCE: U.S. Census Bureau and National Center for Health Statistics, Estimates of the July 1, 2000-July 1, 2005 United States resident population from the Vintage 2005 postcensal series by year, county, age, sex, race, and Hispanic origin (unpublished data available at http://www.cdc.gov/nchs/about/major/dvs/popbridge/popbridge.htm, analyzed by Quality Resource Systems, Inc.).

The concepts of race and ethnicity are in transition in American society and in the statistics that measure it.  Starting with Census 2000, the Census Bureau has followed guidelines that allow respondents to specify multiple races in collecting and issuing data. (11)  Since this report draws upon a combination of data from a variety of sources, there is incomparability across data systems.  Additionally, within a given data system, the change in race standards results in incomparability across time, making it difficult of perform trend analyses.  While states transition to the new OMB reporting guidelines, the U.S. Census Bureau in collaboration with the National Center for Health Statistics has developed a bridging methodology to ascribe the multiple-race group population counts to single-race categories as reported in this report and the earlier 2004 report.  For a complete description of the methodology used, reference the site noted above.

The statistical and health reporting agencies that provide data for this report vary widely in their collection and reporting of race and Hispanic identity. Hispanic identification is considered to be an ethnicity rather than a race, and the Census Bureau asks separate race and Hispanic questions. In processing the bridge population data for this report, Hispanic identity is given priority in determining respondents' race and ethnicity. That is, anyone who reported herself or himself as Hispanic was counted as Hispanic regardless of what race she or he reported.  Thus a Black Hispanic would be reported only as Hispanic.  The Report Card presents parallel sets of columns that apply to race and Hispanic origin.  Columns marked as "non-Hispanic" omit all Hispanics regardless of race.  Columns that lack that label include Hispanics with whatever race that they report.  The "White" column, for example, includes everyone who reported White as a race.  The "White non-Hispanic" column omits Hispanics.

Households Headed by Single Women (% and #), 2006 and 2007.
EXPLANATION: This measure includes households headed by a woman with no spouse present.
SOURCE: SOURCE: U.S. Bureau of Labor Statistics and U.S. Census Bureau, Current Population Survey (CPS), "Annual Social and Economic Supplement" (ASEC) 2006, 2007 (databases) (unpublished data available at http://www.census.gov/hhes/www/cpstc/cps_table_creator.html, analyzed by the National Women’s Law Center).

Lesbian-Headed Households (% and #), 2005.
EXPLANATION: Lesbian-Headed Households are households where the householder is female and there is another female whose relationship to the householder is reported as "unmarried partner."  Since Census 2000 asks no direct question about sexual orientation, this household relationship item has been used to estimate lesbian-headed households.  This estimate represents a count of households headed by women who have same-sex partners, cohabit, report one of the couple as the household head, and report the other as an unmarried partner.  Women in this group are distinguished by their willingness to report these combined characteristics in the Census and may not represent all women living in this status. (12)  It is estimated that 44.1 percent of lesbians live as cohabiting partners like this respondent group. (13)  Applying that rate to the U.S. total of lesbian-headed households implies that there were at least 1.3 million lesbians in the U.S. as of 2000.  If the assumption is made that four percent of U.S. women age 18 and older are lesbians, (14) then there were up to 4.3 million lesbians in 2000.
SOURCE: U.S. Census Bureau.  2005 American Community Survey Detailed Tables. American FactFinder. Available at http://factfinder.census.gov/, analyzed by Quality Resource Systems, Inc.  Data are from 2005 and found in Table B11009.

Median Earnings for Women ($), 2005.
EXPLANATION: This measure includes wages, salaries, self-employment income, and farm income for civilian, non-institutionalized women age 16 and older who reported full-time, full-year employment. The median income divides the income distribution into two equal parts; half fall above the median and half fall below.
SOURCE: U.S. Census Bureau. 2005 American Community Survey Detailed Tables. American FactFinder. Available at http://factfinder.census.gov/, analyzed by Quality Resource Systems, Inc. Data are from 2005 and found in Table B24042.

Women with Disabilities Affecting Workforce Participation (% and #), 2005 and 2006.
EXPLANATION: This measure includes civilian, non-institutionalized women ages 18 to 64 who meet one of the following conditions: (a) not in the labor force because of disability; (b) labor force participation in the past year has been limited by disability or illness; (c) covered by Medicare in the past year; (d) receive Supplemental Security Income.
SOURCE: U.S. Bureau of Labor Statistics and U.S. Census Bureau, Current Population Survey (CPS), "Annual Social and Economic Supplement" (ASEC) 2005, 2006 (databases) (unpublished data available at http://www.census.gov/hhes/www/cpstc/cps_table_creator.html, analyzed by Quality Resource Systems, Inc.).

Linguistically Isolated Households (% and #), 2005.
EXPLANATION: A household living in "linguistic isolation," as defined by the Census, is a household in which no person age 14 and older speaks only English, and no person age 14 and older who speaks a language other than English speaks English "very well." In other words, all members age 14 and older have at least some difficulty with English. This measure includes all linguistically isolated households as a percentage of the total number of households in a state.
SOURCE: U.S. Census Bureau.  2005 American Community Survey Detailed Tables.  American FactFinder.  Available at http://factfinder.census.gov/, analyzed by Quality Resource Systems, Inc.  Data are from 2005 and found in Table C16002.

Births Attended by Midwife (%), 2001.
EXPLANATION: This measure includes the percentage of live births attended by a midwife using data reported on birth certificates.  Although the percentage of birth records that contains missing information for the attendant is very small (less than one percent), there is some evidence that midwife-attended births are under-reported on the birth certificates. (15)
SOURCE: Centers for Disease Control and Prevention, National Center for Health Statistics, "Birth Attendant by State of Residence of Mother, United States, Live Births, 2001" (unpublished data analysis by Decision Demographics).

Women Residing in Urban and Rural Areas (% and #), 2000.
EXPLANATION: Urban women include females of all ages who live in urban areas.  Urban areas are densely settled, contiguous areas delineated by the Census Bureau that exceed specified size and density criteria.  "Urban clusters" have 2,500 to 49,999 people and a core that exceeds 1,000 people per square mile.  "Urbanized areas" have at least 50,000 people in a densely settled area.  Women who live in urban clusters and urbanized areas are classified as urban, while all other women are classified as rural.
SOURCE: Census 2000 (Washington: U.S. Census Bureau, 2000).

Women with Some College or Associate Degree (% and #), 2006 and 2007.
EXPLANATION: This measure includes the percentage of civilian, non-institutionalized women age 25 and older who have one or more years of college but no degree, and civilian, non-institutionalized women age 25 and older who have attained an Associate degree.
SOURCE: U.S. Bureau of Labor Statistics and U.S. Census Bureau, Current Population Survey (CPS), "Annual Social and Economic Supplement" (ASEC) 2006, 2007 (databases) (unpublished data available at http://www.census.gov/hhes/www/cpstc/cps_table_creator.html, analyzed by the National Women’s Law Center).

Women with a Bachelor’s Degree or Higher (% and #), 2005 and 2006.
EXPLANATION: This measure includes the percentage of civilian, non-institutionalized women age 25 and older who have attained a bachelor’s, master’s, doctorate, or professional degree.
SOURCE: U.S. Bureau of Labor Statistics and U.S. Census Bureau, Current Population Survey (CPS), "Annual Social and Economic Supplement" (ASEC) 2006, 2007 (databases) (unpublished data available at http://www.census.gov/hhes/www/cpstc/cps_table_creator.html, analyzed by the National Women’s Law Center).

Footnotes

1. For this indicator, the complementary data are presented.

2. Healthy People 2010 contains separate objectives and benchmarks for fruits and for vegetables – HP2010 Objective 19-5 is to increase the proportion of persons age two and older who consume at least two daily servings of fruit to 75% and HP2010 objective 19-6 is to increase the proportion of persons age two and older who consume at least three daily servings of vegetables to 50%.  The data are published as one grouping for fruits and vegetables.  Therefore, the benchmark is adapted from these two and is 50%.

3. The benchmarks for these three indicators were adapted to be more applicable to women, to address concerns that the Healthy People 2010 benchmarks for these indicators are based on data for men and women combined.  The timing of trends in these diseases has historically been different for men and women, so the use of benchmarks based on men and women combined could be misleading about the current status of women specifically.  For example, the lung cancer epidemic appeared in men well before the 1950s but began to reverse by the late 1980s.  In women, however, it appeared later and continued to climb in most states into the 1990s; only now is the lung cancer epidemic slowing in women, at least in some states. American Cancer Society, Cancer Facts and Figures 2003 (Atlanta: American Cancer Society, 2003), 13, available at http://www.cancer.org/downloads/STT/CAFF2003PWSecured.pdf.

For these reasons, the Report Card employs benchmarks more applicable to women for these three disease indicators, using the same principles as the Healthy People 2010 target-setting standard of "better than the best." Using this Healthy People 2010 standard, the benchmark for each disease is set as the rate in the state that currently has the lowest death rate.  The benchmarks for these indicators are based on the data presented in the 2007 Report Card (whereas the 2004 Report Card used data presented in that edition).

The Healthy People 2010 benchmarks for men and women combined for these indicators are as follows: coronary heart disease—reduce deaths to 166 per 100,000; stroke—reduce deaths to 48 per 100,000; lung cancer—reduce deaths to 44.9 per 100,000.

4. The Healthy People 2010 goal is to increase life expectancy, but no specific target is provided.  The Report Card adopted Japan’s life expectancy for women as a benchmark, since it is a highly industrialized nation with the highest life expectancy for women.

5. These changes were primarily due to baseline data revisions after the November 2000 publication of Healthy People 2010. The Healthy People 2010 benchmarks used by the Report Card that were affected by the change are: Healthy People 2010 Objective 3-2, Reduce the lung cancer death rate (Target: 43.3 deaths per 100,000 population, revised from 44.9); Healthy People 2010 Objective 3-3, Reduce the breast cancer death rate (Target: 21.3 deaths per 100,000 females, revised from 22.3); Healthy People 2010 Objective 12-7, Reduce stroke deaths (Target: 50 deaths per 100,000 population, revised from 48); Healthy People 2010 Objective 12-9, Reduce the proportion of adults with high blood pressure (Target: 14%, revised from 16%); Healthy People 2010 Objective 16-4, Reduce maternal deaths (Target: 4.3 maternal deaths per 100,000 live births, revised from 3.3); Healthy People 2010 Objective 26-11c, Reduce the proportion of adults aged 18 and older engaging in binge drinking during the past month (Target: 13.4%, revised from 6.0 due to change in data source for the baseline data). More information is available at http://www.healthypeople.gov/data/midcourse/default.htm.

6. See, e.g., "Grading," in National Center for Public Policy and Higher Education, Measuring Up 2000: The State-by-State Report Card for Higher Education 2000, available at http://measuringup.highereducation.org/2000/articles/grading.cfm. The goal was to scale the scores so that they could range between 0-100 for each indicator.  The following description uses the consumption of fruits and vegetables indicator as an example to illustrate how the scaled scores were calculated for each indicator.  The benchmark for the consumption of fruits and vegetables indicator is 50 percent.  The state performances range from a high of 38.6 percent in the District of Columbia to a low of 16.8 percent in Oklahoma.  The first step in calculating the scaled score is to express these raw data as a percentage difference from the benchmark.  The District of Columbia is 22.8 percent short of the benchmark and Oklahoma is 66.4 percent short of the benchmark.  The next step is to determine the number that, when multiplied by the worst state's percentage difference from the benchmark, equals 100.  In the case of the consumption of fruits and vegetables, this figure is 1.506 (66.4 x 1.506 = 100). Each state's percentage difference from the benchmark is then multiplied by this figure.  At this stage, the District of Columbia would receive 34.3 (22.8 x  1.506 = 34.3), and Oklahoma would receive 100 (66.4 x 1.506 = 100).  Finally, the resulting number is subtracted from 100 to achieve the scaled score.  The District of Columbia’s scaled score for the consumption of fruits and vegetables is 65.7 (100 – 34.3  = 65.7) and Oklahoma’s scaled score is 0 (100 – 100 = 0).  (States that meet or exceed the benchmark receive a scaled score of 100.)

7. In recognition of the fact that the nation should be making progress toward the benchmark goals of Healthy People 2010, this edition of the Report Card uses a new, more stringent grading scale.  In the 2004 edition, the grading cut-off points were 70% for the grade of "S-" and 50% for the grade of  "U". Discrepancies apparent between the minimum performance value and states' grades are due to rounding.

8. For this indicator, the complementary data were used for grading in order to be consistent with the relevant benchmarks listed in the benchmark chart above.

9. The term "institutionalized population" as used in the Report Card includes persons "under formally authorized, supervised care or custody, such as in federal or state prisons; local jails; federal detention centers; juvenile institutions; nursing, convalescent, and rest homes for the aged and dependent; and homes, schools, hospitals or wards for the physically handicapped, mentally retarded, or mentally ill."  U.S. Census Bureau, Census of Population and Housing, 1990: Summary Tape File 3, Technical Documentation(Washington: U.S. Census Bureau, 1992) [CD-ROM].

10. In its use of "Black" and "African American," this Report Card has attempted to follow the source material's usage, wherever possible.  There is some confusion over the use of the two terms, with sources using them inconsistently.  Black women are primarily "African American," the term commonly used to describe the descendants of Africans brought to the United States as slaves.  There is, however, increasing diversity among Blacks, with foreign-born Blacks accounting for six percent of all Blacks in the United States.  Most other Blacks in America are of Caribbean descent, coming from island nations including the Dominican Republic, Haiti, Jamaica, and Trinidad and Tobago.  Recent immigrants from African countries account for less than four percent of all U.S. immigrants between 1981 and 1998, but there is some indication that these numbers are increasing.  National Women’s Law Center and others, Making the Grade on Women’s Health: A National and State-by-State Report Card, 2001 (Washington: National Women’s Law Center, 2001), 201.

11. In 1997 and 2000 the Office of Management and Budget issued its "Revisions to the Standards for the Classification of Federal Data on Race and Ethnicity" (Federal Register Notice October 30, 1997) and "Provisional Guidance on the Implementation of the 1997 Standards for Federal Data on Race and Ethnicity" (December 15, 2000) which direct that in the federal collection and reporting of data by race, respondents be permitted to specify multiple races.

12. David Smith and Gary Gates, in their Human Rights Campaign report, "Gay and Lesbian Families in the United States: Same-Sex Unmarried Partner Households," August 22, 2001, available at http://www.urban.org/UploadedPDF/1000491_gl_partner_households.pdf, analyzed these data for both lesbians and gays and found that the geographic and other response patterns correspond with what is known of the lesbian community. 

13. Dan Black and others, "Demographics of the Gay and Lesbian Population in the United States: Evidence from Available Systematic Data Sources," Demography 37 (May 2000), 139-154.

14. A reasonable maximum from other evidence accumulated by Gates, et al.

15. According to the results of the 1994 membership survey of the American College of Nurse Midwives, about six percent of midwives reported that they were not identified as the attendant at delivery for some births that they attended.  L.V. Walsh and others, "Findings of the American College of Nurse-Midwives, Annual Membership Survey, 1993 and 1994," Journal of Nurse Midwifery 41 (1996), 230-235.