I'm trying to get the new data, but I'm having trouble opening the .sas7bdat (with SAS saying it's either not a SAS data set or "unable to load table") and the .csv file seems to have only text as values for many variables. Any help? Thanks!
Hello - I would like to reiterate Pixie's request.
Is there any way to get a data file for stata or R file (or maybe a stata dictionary file) that has the coded numerical responses rather than text responses? As currently provided, the csv file with all text labels will require massive amounts of labeling and recording to get the data in any kind of working state for analyses.
Hello all. We looked extensively into this issue. Switching to the .csv file was only a partial solution. It is easy to load .csv into any statistical program. However, some of the fields are text formatted, rather than numeric. This seems to be the way that these values were entered originally. We do not, at the moment, have funding to hire some research assistants to convert the text to numeric values, and there is no easy automated way to do this. However, we are working on a posting a dataset.
We never had the funding to export the dataset and extensively test the dataset in formats other than R and CSV. Both have been extensively tested. The only issue is that quirks that may be associated with the validity of mortality records at different ages may lead to incorrect results in analyses that do not control for age. One example of such an analysis would be the estimation of absolute mortality risk (rather than relative mortality risk). For example, the data should not be used to calculate life expectancy using a life table. If this is desired, you should compute relative risks prior to estimation, and then apply these risks to validated age-specific mortality rates.
If anyone in the community has assembled a dataset in a format other than R, please forward to us for further testing and validation.