could note laws currently in the dataset that should be removed, a sheet asking researchers to note their general sources of information, and a locked data validation tab. The workbooks summarised relevant research guidance at the top of each column. The country folders contained five subfolders and two spreadsheet documents. Two folders corresponded to the laws and draft laws, each of which had subfolders for each category of law. A third folder was for case law, a fourth for translations, and the fifth was for secondary sources. One spreadsheet listed the laws currently in the ADRD dataset and the second was the new data collection workbook. For further refinement: While Google Drive and Docs satisfied our needs for an easily accessible and configurable tool – especially for being able to share documents among several users and track comments between them – there was at least one researcher who had trouble negotiating the folder structure and creating links to shared files. In addition, we used available data verification features to populate dropdown menus from one spreadsheet to another. This worked seamlessly when connecting original laws to their translations, for example, but not as well when connecting articles of law to primary or secondary legislation. For example, on the key provisions worksheet, researchers were asked to enter relevant articles. These entries populated a dropdown menu in the case law spreadsheet. But when a researcher wanted to indicate which article was relevant to the case law, they would sometimes see two articles with the same number but from different laws and not know which to choose, potentially leading to documentation errors. In future iterations of the workbook, we will explore tools that would make it more difficult for researchers to make these and other kinds of coding errors. Finally, organisations that prefer not to use Google products for security reasons may also want to adapt the workbook to other tools. Recruiting and orienting researchers Earlier data collection was conducted by volunteers and journalists, but not legal experts. Because the refined methodology relied much more on an understanding of law and legal systems, we prioritised working with lawyers preferably with expertise in the countries they were researching, or at least in the region. We launched a 10-day call for legal researchers47 and although our timeline was short we received 16 applications, among them researchers 47 SMEX Seeks Legal Researchers for Arab Digital Rights Database. https://docs.google.com/document/d/1SGW9STW-tx5Y34LmfHoS JhOmnlobfQcsuGNmLIGvam0/edit who had worked on the previous versions. Twelve candidates were contracted to do one round of research and one round of peer review. Some candidates took on more than one country. Researchers came from Egypt, Jordan, Lebanon, Morocco, Oman, Palestine, Sudan, Syria, Tunisia and Yemen, as well as the US and France. Researchers were asked to attend one of two one-hour virtual orientation sessions48 held by SMEX and led by legal adviser Jansen. Before the orientation session, researchers were able to review the data collection workbook and the research guidance and make suggestions for refinements. The sessions began with an overview of the scope of work and then relied on researchers to ask questions to clarify any unclear guidance. They also noted specificities within national legal systems that would pose challenges to capturing data in the format we had provided. For example, it was noted that in some jurisdictions, amendments are issued separately from the laws to which they apply, rather than integrated into a reissued law. This, plus questions about whether regulations should also be included, resulted in adding a column that qualified laws as either primary or secondary. Researchers raised concerns about different definitions of case law, which was clarified as referring to “judicial decisions and other jurisprudence that constitutes an authoritative interpretation of the law.”49 Also with regard to case law, some researchers relayed that in their jurisdictions the names of the parties are not used to name the cases. To create unique case names, researchers were asked to assign unofficial names to the cases. These notes and others were captured in an addendum to the research guidance document (available in the Resources section at the end of this article) called ADRD Workbook Updates Doc.50 After the sessions, a Google Group mailing list was set up where researchers could ask questions during the data collection process and further refine the research guidance as needed.51 More active researchers posed sporadic queries to the mailing list, but many remained quiet, making it necessary to follow up on an individual basis, which was burdensome given that one person was managing 12 researchers and 22 workbooks. 48 Budget constraints prevented us from being able to host an inperson training workshop. 49 https://docs.google.com/document/d/11NAys-JDDiU4Ht4VLyV4H PH3dzVqBxKwrsvUYTIoLQE/edit 50 Ibid. 51 ADRD Summary Report, August 2017, submitted by Nani Jansen. Methodology / 15