Katarina Richardson and Gerard J. van den Berg, "Duration Dependence Versus Unobserved Heterogeneity in Treatment Effects: Swedish Labor Market Training and the Transition Rate to Employment", Journal of Applied Econometrics, Vol. 28, No. 2, 2013, pp. 325-351. Note: Because of the accented characters, which are not handled correctly in this file, there is also a PDF version of this readme file. The data used in this article are freely available for scientific research, and we would be delighted to help any researcher who is interested in the data. The data consist of a sample of individuals who are longitudinally followed in two Swedish administrative population register files: 1. HANDEL (from the official employment offices). 2. AKSTAT (from the unemployment insurance fund). Hence, the sample contains records from these registers that are merged at the individual level. The sample construction is explained in the paper. The sample contains 16467 individuals who in total experienced 28451 unemployment spells. Swedish administrative population registers are subject to the Swedish law on data protection. It is not allowed to distribute individual data from these registers without permission from the relevant authorities. We now explain how such data can be accessed for scientific research. The IFAU (Institute for Labour Market Policy Evaluation) is a research institute of the Swedish Ministry of Employment, situated in Uppsala. Affiliates of IFAU have access to register databases containing individual records from certain merged Swedish administrative population registers. Since we as authors are affiliated to IFAU, this enabled us to use the data for our article, having signed data protection and confidentiality statements. Non-affiliates of IFAU can also obtain access to the databases we used. This requires them to submit a request to the Director (see below) of IFAU. Subsequently, they must visit IFAU, sign the relevant statements, and work with the requested data on a local computer. This is a non-bureaucratic procedure. IFAU is delighted to help non-affiliated researchers with data access and with travel accommodation, to provide hospitality for the reseachers, and to enable computer usage. Subsequently, in addition, we as authors are delighted to offer direct access to the constructed data files used in the econometric analysis for the article and to provide the programs used to create the relevant variables in the econometric analyses from the original register database. Alternatively, non-affiliates could try to obtain access to the original data registers through a different Swedish institution with access to the original registers. Below we report the relevant names and addresses and data files. IFAU Kyrkogardsgatan 6 Box 513 751 20 Uppsala Tel: 018 - 471 70 70 Fax: 018 - 471 70 71 E-post: ifau@ifau.uu.se Director IFAU: Professor Olof Aslund Tel: 018 - 471 70 89 E-post: olof.aslund@ifau.uu.se Coordinator data registers: Helge Bennmarker Tel: 018 - 471 70 88 E-post: helge.bennmarker@ifau.uu.se Supervision of data storage and coordinator IT equipment: Jorgen Moen Tel: 018 - 471 70 77 E-post: jorgen.moen@ifau.uu.se half.dta is the name of the STATA data file constructed by us from the original data registers, for the econometric analysis for the article. Its size is 2,793,822 bytes. The number of variables is 34. Each line is an unemployment spell; hence it contains 28451 observations. The STATA data file stor.dta is essentially the same, except that it contains the full population of individuals who satisfy the criteria to be included in the sample. As explained in the paper, it contains 500,960 individuals. The file size is 140,809,868 bytes. The number of variables is 54. Each line is an unemployment spell; hence it contains 869,142 observations. Both files are available at IFAU. We and the Supervisor of data storage have access to them. Katarina Richardson Gerard J. van den Berg September, 2011