Citation Guidelines

It is the responsibility of all registered users of the Human Mortality Database (HMD) to give proper credit to the institutions and/or individuals who have created the data being provided or transmitted here. Proper citation is necessary whether the data in question are quoted in formal publications or in other contexts. The following comments are intended to assist users by providing information about HMD data sources and by offering suggestions for appropriate citation procedures. Please note, however, that ultimate responsibility in these matters resides with the user alone, under the terms of the User Agreement.


Raw Data versus HMD Estimates

It is important to distinguish between two types of data included in the HMD. First, there are several kinds of raw data for each population, which serve as inputs for all subsequent calculations. Second, there are data of various sorts that are original to the HMD, having been generated according to techniques that are fully described in our Methods Protocol. In the former case, it is important to acknowledge the institutions or individuals responsible for creating and/or publishing the raw data: in these circumstances, the HMD is merely transmitting the data from its source to the user. Thus, it is only in the second case that it is appropriate to give credit to the HMD as the source (i.e., author) of the information provided here. Please note, however, that a portion of the data processed by the HMD can be identical to the original (raw) data series (e.g. birth counts). In such situations, it is appropriate to cite first the original source of the data, but also to acknowledge the HMD as the intermediary.


Identifying the Correct Data Source

The user can successfully determine whether a particular number is being transmitted from an external source or whether it is original to the HMD using the following guidelines:

Input Data

Generally, the Input Data have been extracted from other sources, which are clearly documented using a system of reference codes. The HMD should not be cited as the source of such information, but only as an intermediary that has transmitted the data from its source to the user.

Complete Data Series

  1. Births: All birth counts come from an external source, and the HMD is merely transmitting the information to the user. The user should identify and cite the original source, merely acknowledging the HMD as the intermediary. There are two ways of identifying the original source of such data. First, the user may consult a complete list of sources for a given population by means of the link to "Data sources" found near the top of each country page. This list is organized by data type (births, deaths, etc.). In most cases, the source of birth counts reported in HMD should be readily apparent. However, if questions remain after consulting this list, the user should consult the Input Data, in which the source of every raw data point is clearly noted.

  2. Deaths: The death counts presented in the "Complete Data Series" section of country page may or may not have been manipulated in various ways. For citation purposes, we must distinguish between two cases:

    1. Some death counts are presented without alteration. Thus, they are identical to the Input Data, and the proper citation procedure is the one given above for birth data. For death data by Lexis triangle, this case is relatively rare: it occurs only when the original source contains death counts classified by single year of age, year of occurrence, and year of birth, and where there are no reports of deaths with age unknown (for a given calendar year).

    2. Most death counts undergo various alterations, which may be slight or substantial in nature. Briefly, there are two kinds of manipulations applied to raw death counts in the HMD: i) assigning deaths of unknown age to specific ages, and ii) splitting aggregate death counts (in various formats) into data organized by Lexis triangle. In these situations, the numbers are original creations of the HMD, which should thus be acknowledged as the source of the data.

    The data for a given population may comprise one or both types of death counts (i.e., unaltered raw data and/or original HMD estimates). The simplest method of distinguishing between these two types is to observe whether a number contains some fractional deaths. That is, raw data are expressed as a whole number of deaths, whereas estimated death counts typically contain some fractional element. This method is not infallible, as the fractional part of an estimate could be zero just by chance: since death data are formatted with 2 decimal points, this should occur at random with a frequency of 0.01. However, by inspection of surrounding data points, there should in general be no ambiguity about this issue. In cases of lingering doubt, the user should consult the HMD Input Database and examine the raw data directly.

  3. Population size: For ages 80 and above, most of the population estimates reported in the "Complete Data Series" section are original creations of the HMD and should be cited accordingly. One exception pertains to cohorts who would be aged 80-89 at the end of observation (see Figure 6 of the Methods Protocol). A second exception includes a few countries (i.e., Sweden, Denmark, Norway, Finland, and Iceland) for which we use official population estimates for ages 90 and older at the end of the data series (see footnote 16 of Methods Protocol).

    Below age 80, population estimates may come directly from an external source, or they may be unique to the HMD. As with death counts, it is possible to determine whether a number comes from an external source or was created expressly for the HMD by observing its decimal part: whole numbers usually indicate that a number is taken from some other source, whereas a non-zero fraction (i.e., anything other than "00" after the decimal point) is a sign that the number is a product of the HMD itself.

  4. Exposure-to-risk, death rates, life tables, or life expectancy at birth: All such data are original creations of the HMD, which should thus be cited as the unique source of the information.


Choosing a Bibliographic Citation

HMD Estimates

As described above, proper citation practice requires that a user determine whether numbers obtained through the HMD come from an external source or whether such data are original creations unique to this database. If the user concludes that the data in question should be attributed to the HMD itself, we suggest the following format for a bibliographic citation:

HMD. Human Mortality Database. Max Planck Institute for Demographic Research (Germany), University of California, Berkeley (USA), and French Institute for Demographic Studies (France). Available at www.mortality.org (data downloaded on [date]).

If you choose another citation format, please be sure that it includes the full name of the database (Human Mortality Database), the full names of all institutional sponsors (Max Planck Institute for Demographic Research, University of California, Berkeley, and French Institute for Demographic Studies), our Internet address www.mortality.org, and the date accessed.

HCD Estimates

For the cause-of-death component of the HMD specifically, we suggest the following format for the bibliographic citation:

Human Cause-of-death Data series. Human Mortality Database. French Institute for Demographic Studies (France), Max Planck Institute for Demographic Research (Germany) and the University of California, Berkeley (USA). Available at www.mortality.org (data downloaded on [date]).

STMF estimates

Short-Term Mortality Fluctuations data series (STMF). Human Mortality Database. Max Planck Institute for Demographic Research (Germany), University of California, Berkeley (USA), and French Institute for Demographic Studies (France). Available at www.mortality.org

Data from external sources

For data from an external source, which are merely transmitted via the HMD, the user should choose an appropriate citation based on standard practice. The citation given in the list of "Data sources" for each population may be used as a guide, but it should not be considered authoritative. It is appropriate in these situations to acknowledge the HMD as the intermediary through which the data were obtained. For example, a bibliographic citation for death counts in the United States might be as follows:

National Center for Health Statistics. Vital Statistics of the United States, Volume II: Mortality, Part A. Washington, D.C.: Government Printing Office, various years. (Data obtained through the Human Mortality Database, www.mortality.org, on [date].)


User's responsibility

Finally, please remember that proper citation of data obtained through the HMD is the responsibility of the user alone, as stated clearly in the User Agreement. However, if you still have questions or comments about citation procedures after reading these guidelines, please write to us at hmd@mortality.org.