Help & FAQ

If you used any of the Harmonized Data or Codebooks in a written analysis, then please include the following acknowledgement in your written work (We also ask that you send an email to papers@g2aging.org to inform our team of any written analysis) :

"This analysis uses data or information from the Harmonized [Study] dataset and Codebook, Version [Letter] as of [Month & Year] developed by the Gateway to Global Aging Data. The development of the Harmonized [Study] was funded by the National Institute on Aging (R01 AG030153, RC2 AG036619, 1R03AG043052). For more information, please refer to www.g2aging.org."

If you used a working paper on cross-country comparability, then please cite the paper; for example,

"Jain U, Min J, Lee J. Harmonization of cross-national studies of aging to the Health and Retirement Study - user guide: Family transfer - informal care. University of Southern California, CESR-Schaeffer Working Paper Series No. 2016-008. Published January 2016."

If you used a graph, table, or map, then please cite the following:

"This graph uses data from the Gateway to Global Aging Data (g2aging.org). The Gateway to Global Aging Data is funded by the National Institute on Aging (R01 AG030153)."

If you used cross-study or longitudinal concordance information, recent presentations, or other tools on the Gateway to Global Aging Data, then please cite the following:

"Gateway to Global Aging Data, Produced by the Program on Global Aging, Health & Policy, University of Southern California with funding from the National Institute on Aging (R01 AG030153)."

We are constantly updating our existing Harmonized datasets and adding new Harmonized datasets. If you don't see a Harmonized version of a study which is important to your research, please write to us at help@g2aging.org and we will provide you our best estimate of when that Harmonized dataset will be available.

The easiest way to register for most of the HRS family studies is actually on this website. First, you should register for the Gateway to Global Aging Data by selecting Register on the top right of any page. If you have already registered, then make sure you're logged in to the Gateway website. Once logged in, go to the Download Data and Links page, and select the link at the top of the page that says "To register and access data for any of the HRS-family studies, click here".

You can now select the studies you're interested in, fill out the information, and submit the form. Once submitted, you will receive an email. This may contain documents (for TILDA, SHARE & JSTAR) that you need to print, sign, and email back to the email(s) provided in the email text. If you do not want to use this time-saving tool, you can also follow the links on the top row (Links to Download Survey Data) of the Download Data and Links page to the website of the study you're interested in and register there.

The Gateway does not generally provide data to download. Most studies included in the Gateway distribute their own data and the Harmonized datasets we create. Instead, the Gateway seeks to provide as much information as possible about each survey's questionnaires, methodology, design, and about their similarities and differences. We also provide the codebooks and Stata creation code for each Harmonized dataset on our website. For links to download data or the Harmonized datasets please see our Download Data and Links page. The Gateway does distribute study data for the Longitudinal Aging Study in India (LASI) and the Costa Rican Longevity and Healthy Aging Study (CRELES). To download LASI data, see our LASI Downloads page. To download the Harmonized CRELES, visit our CRELES Downloads page.

We currently produce internationally harmonized datasets for many studies. Many of these Harmonized datasets are distributed by the original study along with their study data. For studies that do not distribute Harmonized datasets, we distribute a Stata program for users to download which transforms the original survey data into the Harmonized dataset. For more information on downloading a Harmonized dataset or the Stata code to create a Harmonized dataset, please see our Download Data and Links page.

Currently, the Harmonized HRS, Harmonized MHAS, Harmonized ELSA, Harmonized CRELES, Harmonized TILDA, and Harmonized LASI are distributed in Stata, SAS, and SPSS datasets. Any Harmonized dataset which is generated using Stata Creation code or is downloaded in Stata format can be converted for use in other statistical packages using a program such as Stat/Transfer. If you do not have access to Stat/Transfer, you may be able to read the .dta dataset into your stat package using an “import” or “get” function. When reading a .dta dataset into another package, it is best to first save the dataset in a Stata version 12 format using the command saveold.

The Gateway's Stata Creation code pulls original survey variables directly into Stata's working memory to create Harmonized variables. To pull in the correct original survey variable from the survey data requires the specification of the exact file name of each original survey dataset. Many studies update the dataset names variable locations between release versions. If you are given a 'file not found' error message when running Stata, you may not be using the most recent release of the survey data. Please make sure you have the latest version of the survey data. If you still encounter an error, please write to us at help@g2aging.org and we will help you as quickly as possible.

Some versions of Stata only allow users to read fewer variables into working memory than are in some of our Harmonized datasets (e.g. Stata/IE). All versions of Stata will allow users to pull select variables into Stata from a dataset with more variables than it could read at once. You can identify the variables you would like to use by searching or browsing for your items of interest on the Surveys at a Glance page, or you can download and search through the codebook that accompanies the Harmonized dataset on the Download Data and Links page. You can create a smaller dataset for your personal use by updating the variable names, filepath, and dataset name in the following Stata code: use variable1 variable2 variable3 using "filepath\H_dataset.dta".

It is very simple to merge the Harmonized datasets with the original study data using the unique identifiers employed by the study. You can identify the variables from the original study data you would like to use by searching or browsing for your items of interest on the Surveys at a Glance page, or you can look through the original survey questionnaire or datasets. In Stata, you can merge in these orginal survey variables with the Harmonized data using the following Stata code:

merge 1:1 studyID using "filepath\dataset_name.dta", keepusing(variable1 variable2 variable3).

It is important to remember that all Harmonized datasets are individual-level, where each record is one person, but original survey data files can also be couple, household, community, or child-level datasets. All possible identifiers from each study are kept as part of the Harmonized dataset to allow for the merger to original survey datafiles which are not necessarily also individual-level. If the original survey dataset is not individual-level, then you will need to change the merge from 1:1 to m:1, 1:m, or m:m and use the appropriate identifier rather than the unique individual identifier. This method would also work when merging variables between the Harmonized HRS, RAND HRS, and RAND HRS Family datasets.

For our purposes, the RAND HRS can be thought of as the original harmonized dataset for the HRS, which was created and is maintained at the RAND Corporation. We, at the Gateway to Global Aging Data, have built our other Harmonized datasets (Harmonized MHAS, ELSA, SHARE, CRELES, KLoSA, JSTAR, TILDA, CHARLS, LASI) using the RAND HRS as a model and basis for comparison (as you can see in the Harmonized codebooks, we have a section for each group of variables titled "Differences with the RAND HRS"). As such, the variables in our Harmonized datasets have been created to be comparable with the similarly named variables in the RAND HRS wherever possible. If there are differences in how the Harmonized variable was created or if the variables are similar but not strictly comparable, we explain that in each Harmonized codebook.

The Harmonized HRS is different from the RAND HRS because it contains variables that the RAND HRS doesn’t include. As we have worked on the different Harmonized datasets and heard from researchers and users, we became aware that there were many potential HRS variables that would be helpful, but that weren't created for the RAND HRS. To address this opportunity, we created these new variables for the HRS and released them in the Harmonized HRS. We plan to add as many variables into each of our Harmonized datasets to be comparable with the Harmonized HRS over time.