NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Center for Health Statistics (US). Health, United States, 2011: With Special Feature on Socioeconomic Status and Health. Hyattsville (MD): National Center for Health Statistics (US); 2012 May.

Cover of Health, United States, 2011

Health, United States, 2011: With Special Feature on Socioeconomic Status and Health.

Show details

Technical Notes

Data Sources and Comparability

Data for the Health, United States, 2011 Chartbook come from many surveys and data systems and cover a broad range of years. Detailed descriptions of the data sources represented in the Chartbook are provided in Appendix I—Data Sources. Additional information clarifying and qualifying the data are included in the table notes and in Appendix II—Definitions and Methods.

Data Presentation

Many measures in the Chartbook are shown for people in specific age groups because of the strong effect of age on most health outcomes. Some estimates are age-adjusted using the age distribution of the 2000 standard population, and this is noted in the data tables that accompany the charts (see Appendix II, Age adjustment). Age-adjusted rates are computed to eliminate differences in observed rates that result from age differences in population composition. For some charts, data years are combined to increase sample size and the reliability of the estimates. Some charts present time trends, while others focus on differences in estimates among population subgroups for the most recent time point available. Trends are generally shown on a linear scale to emphasize absolute differences over time. The time trends for the overall mortality measures are shown on a logarithmic (log) scale to emphasize the rate of change and to enable measures with large differences in magnitude to be shown on the same chart. Point estimates and standard errors for Figures 1–21 are available in the Trend Table and Excel spreadsheet specified in each figure note. Data tables with point estimates and standard errors accompany the Special Feature Figures 22–41. Some data tables contain additional data that were not graphed because of space considerations.

Statistical Testing

Trends in rates can be described in many ways. For trend analyses presented in the Chartbook, the statistical significance of increases or decreases in the estimates during the entire time period shown was assessed at the 0.05 level using weighted least squares regression, performed using the National Cancer Institute’s Joinpoint software. Regression models describing relative changes over the period were fit to the log of the estimates, with the number of joinpoints limited to zero. For a test of the SES gradient, linear regression models were fit using the estimate, and the number of joinpoints was limited to zero. For more information on Joinpoint, see: For analyses that show two time periods, differences between the two periods were assessed for statistical significance at the 0.05 level using two-sided significance tests (z test).

Terms used in the text such as “similar,” “stable,” and “no difference” indicate that the statistics being compared were not significantly different. Lack of comment regarding the difference between statistics does not necessarily suggest that the difference was tested and found to be not significant. Because statistically significant differences or trends are partly a function of sample size (the larger the sample, the smaller the change that can be detected), even statistically significant differences or trends do not necessarily have public health significance (1).

Overall estimates generally have relatively small sampling errors, but estimates for certain population subgroups may be based on small numbers and have relatively large sampling errors. Numbers of deaths from the National Vital Statistics System represent complete counts and therefore are not subject to sampling error. However, they are subject to random variation, which means that the number of events that actually occur in a given year may be considered as one of a large series of possible results that could have arisen under the same circumstances. When the number of events is small and the probability of such an event is small, considerable caution must be observed in interpreting the conditions described by the figures. Estimates that are unreliable because of large sampling errors or small numbers of events have been noted with an asterisk. The criteria used to designate or suppress unreliable estimates are indicated in the notes to the applicable tables or charts.

For NCHS surveys, point estimates and their corresponding variances were calculated using the SUDAAN software package (2), which takes into consideration the complex survey design. Standard errors for other surveys or datasets were computed using the methodology recommended by the programs providing the data or were provided directly by those programs.


National Center for Chronic Disease Prevention and Health Promotion. Youth Risk Behavior Survey (YRBS): Interpretation of YRBS trend data. 2010. Available from: http://www​​/yrbs/pdf​/YRBS_trend_interpretation.pdf.
SUDAAN, release 10.0.1 [computer software]. Research Triangle Park NC: RTI International; 2009.


Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...