Survey Solutions. Headquarters Manual. Computational Tools team, Development Research Group

Transcription

1 Survey Solutions Headquarters Manual Computational Tools team, Development Research Group 1

2 Table of Contents Introduction... 3 Changing the Interface Language... 3 Mozilla Firefox... 4 Microsoft Internet Explorer... 6 The Components of the Headquarters Software... 7 Reports... 7 By quantity... 7 By template... 8 By supervisor... 9 By map By quantity By speed Interviews Search for survey cases Review survey cases Delete survey cases Teams and Roles Supervisor accounts Interviewer accounts Headquarter accounts (applicable to administrators only) Observer accounts (applicable to administrators only) Batch creation of user accounts Survey Setup Import a template Create survey cases Data Export Interview Comments Interview Actions Data Export Encryption (applicable to administrators) Appendix 1: Overview of the survey workflow

3 Introduction The headquarters software is a suite of connected tools for the administrator and headquarters users: to track the overall progress of the survey (Reports), to review completed interviews (Interviews), to manage the human resources (Teams and Roles), to specify survey instruments, create survey assignments with those instruments(survey Setup), and to export the data collected from these assignments (Data Export). The difference between the headquarters and administrator users is that the administrator can create headquarters/observer user accounts, archive user accounts, and delete questionnaire templates from the headquarters server. The following section will provide an overview of each tool Reports, Interviews, Teams and Roles, Survey Setup, and Data Export and of how each tool works. The subsequent section will explore common actions the headquarters must take with these tools and, through that optic, delve into each component in more detail. Changing the Interface Language Survey Solutions interface for Headquarter and Supervisor users can be presented in different languages Russian, Chinese, or French. This article describes how to switch the interface to be presented in an alternative language. Recall that both HQ and Supervisor users work via a web browser. Switching the interface is done by telling the browser to request the content in a particular language. Note that this affects both Survey Solutions and other sites that are visited with the same browser. The remainder describes step-by-step instructions for the particular browsers you may be using. Other browsers likely to have similar functionality. Google Chrome The following instructions are based on Google Chrome version Start Google Chrome browser; 2. Click on the drop down menu and select Settings item; 3. Click the Show advanced settings... link at the bottom; 4. Click Language and input settings button; 3

4 5. Click the Add button in the Languages list to add a language you need. 6. Drag and drop that language to the top of the list. 7. Click Done. 8. Close and restart the browser. Mozilla Firefox The following instructions are based on version of Mozilla Firefox. 4

5 1. Start the Mozilla Firefox browser. 2. Click on the drop down menu and select Options item; 3. Select Content tab. 4. In the Languages section, click the Choose... button. 5. Select a new language you want to add from the drop-down menu and click the Add button. 6. Adjust the order of the languages to have the desired language at the top. 7. Click OK. 8. Close and restart the browser. 5

6 Microsoft Internet Explorer The following instructions are based on the version of Microsoft Internet Explorer. 1. Start Internet Explorer. 2. Click the gear icon to access the drop-down menu. 3. Select Internet Options. 4. In the General tab click Languages button: 5. Click the Add... button to add the desired language. 6

7 6. Use the Move up and Move down buttons to change the order of the languages to have the desired language at the top of the list. 7. Click OK. 8. Close and restart the browser. If you are using a different browser the steps to take are likely very similar. See also the W3C Internationalization page for the additional information. The Components of the Headquarters Software Reports Reports provide headquarters with a summary of the status of fieldwork of progress by questionnaire, by team or by geography. By quantity 7

8 Figure 1: List of report types in menu bar By template The Surveys and Statuses report summarizes the progress of each survey that headquarters is overseeing (i.e., data collection with a given template). The anatomy of the report is as follows. The column labeled Survey Name contains the name of each survey. The subsequent columns correspond to each possible status of a survey assignment in the Survey Solutions survey workflow. (For a detailed explanation of the workflow and the statuses, see Appendix 1.) Each cell in the report captures the number of assignments that are in a given state for a given survey (e.g., number of Complete assignments for Household Survey 2016 ). Figure 2: Survey and Statuses report sample By default, this report summarizes the status of survey cases over all supervisors it reports the total sum of survey cases in each status, summing over all supervisors. This report can be filtered to provide a summary or focus on one supervisor at a time. To do this, click on the drop-down menu below Supervisors filter and select a name. 8

9 Figure 3: Surveys and Statuses tab page Doing so will yield a summary of progress for only the selected supervisor (Figure 4). In other words, filtering allows headquarters to focus on the progress of an individual supervisor. Figure 4: Surveys and Statuses tab page filtered by supervisor By supervisor The Team and Statuses report provides a status report for each supervisor. The first column of the report lists all supervisors under headquarters supervision. The subsequent columns correspond to each possible status of a survey assignment in the Survey Solutions workflow. Each cell captures the number of assignments in a particular status for a specific supervisor (Figure 5). By default, the Teams and Statuses report provides a global summary of progress at the supervisor level. Headquarters can filter the report in order to obtain a more granular view of progress by questionnaire template. To do this, click on the drop-down menu located below Template. Then, select a particular template to see each team s progress for that template alone. 9

10 Figure 5: Teams and Statuses tab page. Doing so will generate a summary of progress for the chosen template. Figure 6: Teams and Statuses tab page filtered by template By map The Map Report shows the geographic location of all completed questionnaires on Google Maps. 10

11 Figure 7: Map report sample To use this report, first select Maps Report from the Reports menu. This will open the interface for creating this type of report. Then, select the questionnaire you want to map from the Questionnaire drop-down menu. Next, select the version of the questionnaire. Some questionnaires may have only one version, while others may have multiple versions. For questionnaires with only one version, the user must select version 1, the only version option available in the drop-down menu. Finally, select the geo-location variable from the questionnaire that you wish to plot on the map, and click on the red Add markers to map button. This will plot completed questionnaires on a map 1, showing their geographical dispersion and providing some sense of where survey teams have and have not reached. 1 Includes survey cases with one of the following statuses: Completed, ApprovedBySupervisor, RejectedBySupervisor, ApprovedByHeadquarters, or RejectedBySupervisor. Only those survey cases that supervisors have synchronized to headquarters will be included in the map report. 11

12 Figure 8: Steps to create a maps report By quantity The Quantity report tab provides the following reports for each questionnaire template: 12

13 Number of completed interviews: Provides the number of completed interviews for each data collection team. Interviews that have been rejected and then completed are counted as an additional completed interview. Figure 9: Number of completed interviews report sample Number of interview transactions by HQ: Provides the number of headquarters approvals and rejections for each data collection team. Figure 10: Number of interview transactions by HQ report sample Number of interviews approved by HQ: Provides the number of headquarters approved interviews for each data collection team. 13

14 Figure 11: Number of interviews approved by HQ report sample Number of interview transactions by supervisor: Provides the number of supervisor approvals and rejections on interviews collected by each data collection team. Figure 12: Number of interview transactions by supervisor report sample By default, the Quantity report delivers reports at the supervisor level. However, you can obtain a more granular view across one data collection team by selecting a supervisor from the Team column. 14

15 To use these report, first select Quantity from the Reports menu. Then, select the Type of report and the questionnaire template. Select a start date for the report (From). Next, define the time interval-- day, week, or month (Period), and the number of time intervals that you would like to display in the report (Period count). The maximum period count that can be defined is 12. By speed The Speed report tab provides the following reports for each questionnaire template: Average interview duration: Provides the average time difference between the first recorded answer on a tablet and when the complete button is selected. Figure 13: Average interview duration report sample 15

16 Average supervisor processing time: Provides the average time difference between the moment the complete button is pressed and the supervisor s approval or rejection of that questionnaire. Figure 14: Average supervisor processing time report sample Average HQ processing time: Provides the average time difference between a supervisor approval of a questionnaire and a HQ approval or rejection of that questionnaire. Figure 15: Average HQ processing time report sample 16

17 Average case assignment duration: Provides the average time difference between survey case assignment to an interviewer and the moment an interviewer sets the survey case as complete on a tablet. Figure 16: Average case assignment duration report sample Average overall case processing time: Provides the average time difference between survey case assignment and survey case approval by HQ. Figure 17: Average overall case processing time report sample 17

18 Similar to the Quantity report, the Speed report by default provides the report at the supervisor level. However, you can obtain a more granular view across one data collection team by selecting a supervisor from the Team column. To use these reports, first select Speed from the Reports menu. Then, select the Type of report and the questionnaire template. Select a start date for the report (From). Next, define the time interval-- day, week, or month (Period), and the number of time intervals that you would like to display in the report (Period count). The maximum period count that can be defined is 12. Interviews While the Reports tab is designed for tracking progress, the Interview tab is designed for taking action. Under the Interview tab, headquarters may search for survey cases, review them for quality assurance, and delete them to fix problems. The interview tab consists of the following tools: 1. Database of survey cases. The rows are survey case entries. The columns capture details about survey cases, such as the respondent s address, interview status, and an indicator that the interviewer has or has not received the survey case. 2. Filters for querying the database. The database can be filtered to show only the survey cases of interest for example, those with a particular status or those done by a particular supervisor. 3. Toolbar with a Delete, Approve, and Reject button, and a search bar to navigate through the survey cases quickly. Survey cases can be deleted, approved, or rejected individually or as a group by selecting the check box next to each case. If you would like to select all the survey cases displayed on that page, click on the first check box at the top of the list. 18

19 4. Review button to perform quality assurance. Completed survey cases can be reviewed in detail by clicking the green button associated with the case. This will open another interface that will be discussed below. Figure 18: The four tools of the Interviews tab Search for survey cases The Interview tab contains a detailed list of survey assignments those that headquarters has assigned to a supervisor, those that a supervisor has assigned to enumerators, and those that supervisors have approved and sent for headquarters review. The details can be used to search for particular survey cases. The interview panel in the lower right-hand quadrant of the screen presents a list of interviews and information about them where the respondent is located (Pre-filed Questions), which supervisor is responsible (Responsible), when the assignment was last updated (Last Updated), whether the interview contains errors (Has Errors), what status the interview has (Status), whether the interviewer has received the assignment (Received by Interviewer), and in the case of census mode surveys whether the case has been created on the device by the interviewer (Created on Client). 19

20 Figure 19: Interviewer Panel There are three mechanisms for finding a survey. The first is by clicking on the up (^) and down (v) arrows associated with any column of the information in the lower right-hand quadrant. This will sort the list of surveys in ascending (^) or descending (v) order by that chosen column. Figure 20: Arrows to navigate the list of surveys The second mechanism is through the filters located on the left-hand side of the screen. To narrow down the list of survey cases, headquarters may filter by the questionnaire (Template), team member responsible (Responsible), and/or the status of the interview (Status). Figure 21: Other filter options on the left-hand side of the screen The third mechanism is through the search field located in the toolbar. Use keywords to narrow down the list of survey cases. 20

21 Review survey cases The questionnaire review interface consists of five components: 1. Navigation pane. Displays a table of contents for the questionnaire, where you can use each entry in this pane to see a particular part of the questionnaire. Clicking on the top entry, the name of the questionnaire, brings the full questionnaire back into focus. 2. Filters. These buttons can be used to see questions that have certain attributes for example, comments or flags. 3. Questionnaire review pane. The main part of the screen displays the questions asked in the questionnaire, in the first column, and the answers given, in the second column. 4. Comment pane. This pane appears when a question is clicked. This shows the conversation associated with the selected item. 5. Approve/reject buttons. These buttons determine whether a questionnaire returns to the supervisor or is integrated into the final data set. Figure 22: Five components of the questionnaire review interface Headquarters plays a critical quality control role in the Survey Solutions workflow. Supervisors supervise enumerators, reviewing questionnaires the latter complete. Headquarters supervises supervisors, reviewing the questionnaires the latter approve. To do this, first headquarters clicks on the green button associated with any questionnaire whose status is ApprovedBySupervisor. 21

22 Then, headquarters reviews the questionnaire for completeness, consistency, and plausibility of answers using the questionnaire review interface. The questionnaire pane, located in the middle of the screen, contains the questions and their answers. Groups, on the left-hand part of the screen, allow headquarters to see different parts of the questionnaire when any element of the pane is clicked. The top heading corresponds to the complete questionnaire. Bolded headings lead to modules (chapters) within the questionnaire. Indented headings are linked to either groups of questions or rows of rosters. Filters, arrayed atop the questionnaire pane, facilitate the display of questions that have particular attributes. The All filter presents all questions. The commented and flag filters display questions with comments or flags, respectively. The answered questions, as the name suggests, are all those that have an answer. The invalid filter displays questions that failed a validation check (e.g., range, consistency with other responses, etc.). The supervisor s filter yields special questions that only the supervisor sees and can answer. The enabled filter returns questions that are available to be answered based on the skip logic of the questionnaire and answers provided to key questions. Next, headquarters either approves or rejects the questionnaire by clicking the appropriate button. The headquarters choice determines whether the assignment is included in the survey database (Approve) or returned to the supervisor for correction (Reject). To unapprove any interview in the Approved by Headquarters status select this interview(s) by checking the checkbox next to it and click the Unapprove button. The interview will be reverted to the Approved by Supervisor status, and the headquarters user will have a possibility to reject it back to the supervisor if necessary. You can apply this operation to multiple interviews if necessary. Delete survey cases Deletion is a dangerous tool reserved for rare occasions for example, when errors arise in survey sampling or in creating survey cases. 22

23 Because deletion is dangerous, only the administrator has the authority to delete selected interviews. It is only allowed for interviews with the status SupervisorAssigned or InterviewerAssigned. Those with the former status have been assigned to a supervisor, but not yet assigned by the supervisor to an interviewer. Those with the latter status have been assigned to an interviewer, but have not been completed. In both cases, headquarters should exercise extreme caution. Once deleted, an interview cannot be restored. Hence, administrator should only delete interviews that have been carefully confirmed as errors to be deleted. To delete an interview, first find an eligible case. The most efficient strategy for doing so may be to filter the questionnaires based on status so that only those with status SupervisorAssigned or InterviewerAssigned appear. (See the section Search for survey cases above for more details.) Next, select the case(s) to delete by clicking in the check box to the left of the case. Make sure the correct cases have been selected before proceeding. Then, click on the Delete interview button located immediately above the interview panel. This will permanently delete the selected survey cases. Teams and Roles The Teams and Roles tab provides tools for managing the human resources responsible for data collection creating, editing, and deleting user accounts for supervisors, interviewers, headquarters (administrator), and observers (administrator). An administrator can manage all user accounts under this tab. On the other hand, a headquarters user is directed to the supervisor account page after clicking on the Teams and Roles tab. Supervisor accounts Every survey needs supervisors. To add a supervisor, click on the white Add supervisor button at the top of the Supervisors tab, and enter all the requisite information. To send an to a user, click on the user s address. The default client setup on your computer will start to facilitate sending s to that user. s are not sent automatically and it is up to the headquarter user to compose the message. 23

24 Figure 23: Supervisors tab page Figure 24: Form to add a supervisor To modify the details of an existing supervisor account, click on Total Interviewers link associated with a supervisor in order to perform any of the following actions: change the password for the supervisor account (e.g., if the password is forgotten) or lock all access to the account (e.g., if the supervisor is fired due to poor performance). Administrators can archive supervisor accounts. By archiving a supervisor, you are also archiving all the interviewers on that supervisors team. Supervisors and their data collection team will no longer be able to sync with headquarters. To archive a supervisor, click on the red button associated with that supervisor. 24

25 Archived interviewer accounts can be restored. Click on the Archived supervisors button. Then, click on the green addition button associated with that supervisor. Interviewer accounts Each supervisor needs interviewers. To manage the interviewer s in a supervisor s team, click on the number of interviewers associated with that supervisor. Under this option you can add interviewers for a given supervisor s team, modify their accounts, and monitor synchronization. To add a new interviewer to a given supervisor s team, click on the white Create button. Figure 25: Interviewers screen displayed after clicking on total interviewers link for a supervisor To send an to a user, click on the user s address. The default client setup on your computer will start to facilitate sending s to that user. s are not sent automatically and it is up to the supervisor or the headquarter user to compose the message. To modify an interviewer s account change the password or lock the account click on the interviewer s name. 25

26 Figure 26: Form for editing an interviewer s account To monitor an interviewer, check the activation status in the Connected to device column and the synchronization log. The Connected to device column displays the status of the tablet s activation or in other words it will show if an interviewer has successfully activated a tablet using their login credentials. Administrators can also archive individual interviewer accounts. An interviewer will no longer be able to sync with headquarters. To archive an interviewer, select the interviewer s supervisor. Then, click on the red button associated with that interviewer. 26

27 Archived interviewer accounts can be restored. Click on the Archived Interviewers button on the supervisor page. Then, click on the green addition button associated with that interviewer. Headquarter accounts (applicable to administrators only) Every survey needs one or more headquarters user. To manage headquarters, click on Headquarters in the drop down menu under the Teams and Roles menu. To add a headquarters user, click on the white Add headquarters button at the top of the page, and enter all the requisite information. 27

28 To modify a headquarters account change the password or lock the account click on the blue Edit button associated with that headquarters account. Observer accounts (applicable to administrators only) An observer account provides individuals, such as managerial personnel, the means to monitor the activity (shadow) of another user in the system, without being able to damage the system by unintentionally deleting or modifying an interview assignment. To manage observers, click on Observers in the drop down menu under Teams and Roles. 28

29 To add an observer user, click on the white Add observers button at the top of the page, and enter all the requisite information. To modify an observers account change the password or lock the account click on the observer s name. 29

30 Batch creation of user accounts Up to 10,000 user accounts can be created from a single tab-delimited file, which must include the login, password, and role of the user, and may include additional attributes, like or phone. Batch creation of user accounts comes handy when: the same accounts need to be replicated on a different server; migrating from a different CAPI system with established accounts; the survey is so huge that manual creation of accounts via a web interface would take a significant amount of time. Creating user accounts in batch mode is available to both the administrator and the headquarter users. For new supervisor accounts the role must be specified as Supervisor (case-sensitive). For interviewer accounts the role must be specified as Interviewer (case-sensitive), and in addition the login of the supervisor must be specified, which determines the team where the interviewer account will be added. Interviewers may be added to existing supervisors, or to new supervisors mentioned anywhere in the users list. When creating new users with a batch upload feature, select the Batch user upload menu item of the Teams and Roles menu. Survey Solutions will present the layout of the file it expects: 30

31 The fields login, password and role are mandatory, and must be filled out for every user in the list. Other fields are optional. Click the Choose File button to locate the file on your drive which contains the list of the users and click the Upload button to upload it to the Survey Solutions server. 31

32 Survey Solutions will validate the uploaded list of the users report if any problems are found: empty required fields (login, password, role); passwords not conforming to the security requirements (too short, etc); duplicate user accounts; unknown supervisors (all interviewers created must become part of the team of a supervisor known to the system); and other checks. If the program didn t find any problems with the list of users, click the Create users button to start creating accounts in the batch mode. Depending on the number of users and server performance this process may take a few minutes. Survey Setup This tab enables headquarters to manage creation, assignment, and export of survey cases and has dedicated tools for each task. Let us consider each tool in turn. The questionnaire tab consists of several tools: 1. List of questionnaires. This section is a list of questionnaires that have been imported for headquarters use accompanied by information on the mode of survey deployment, survey version, creation date and the most recent date that the questionnaire was used for data collection. 2. Import template button. This button starts the process for downloading questionnaires from the Designer site to the headquarters software. 3. Create survey cases buttons. These buttons provide two ways of adding survey cases for a given questionnaire: one at a time, or in a larger batch. 4. Copy questionnaire template buttons. These buttons allow you to copy the questionnaire templates previously imported to the server. 32

33 5. Delete questionnaire button. This feature is only available to the administrator. The red button is located under Actions for each questionnaire and it should be used in special circumstances with extreme caution, as it deletes the imported questionnaire and all survey cases from headquarters. Figure 27: Four components of the questionnaires tab Import a template When clicked, the Import template button prompts the user to connect to the Designer site and to select a questionnaire template for import. To connect to Designer, headquarters must enter a valid login and password (as pictured below). Note that these login credentials are for the Designer site, not the headquarters software. Figure 28: Accessing the designer tool After connecting to the Designer site, the user must select a questionnaire template to import. To do so, the user must first find a questionnaire, either by scrolling through the list of questionnaires available or using the search box. To choose a questionnaire for import, click on the blue Import button on the line associated with that questionnaire. 33

34 Figure 29: List of templates in the designer tool Once import of the selected questionnaire is complete, the name of that questionnaire appears in the list of questionnaires on the Questionnaires tab. If this is the first time the questionnaire has been imported, the headquarters software will add (version 1) to the end of the questionnaire s name. Each time a new version of the same questionnaire is imported, the software will add a new entry for this questionnaire with a new version number (e.g. version 2, version 3, etc.). Figure 30: Visualizing different versions of the same template listed in the Questionnaires tab page Create survey cases To create survey cases, headquarters has two different tools: the +New Interview button, which creates survey cases one at a time; and the Batch upload button, which creates several survey cases at once. One case at a time After clicking the +New Interview button, the user is prompted to provide identifying information for the survey case typically name and geographic location of the respondent and to select a supervisor to be responsible for the interviewing case. The set of identifying information, which corresponds to the questions marked as Identifying on the Designer site, must be provided by headquarters when creating a case and will appear on the interviewer s dashboard when assigned. 34

35 The user is also prompted to designate a supervisor as responsible for the case (Responsible). Once assigned to a supervisor, the survey case will appear on the supervisor s dashboard and will be available to assign to one of the supervisor s data collection team. Figure 31: Creation of one interview Several cases at a time While the +New Interview button requires the user to enter case address information manually, the Batch upload method relies on an external.tab file that contains that same information for each case to be created and potentially other information to include into those cases as well. The headquarters software offers two ways of using external.tab files to pre-fill survey cases in a batch upload: (1) a sample mode, typically used for cross-sectional surveys, which loads only respondent address information (e.g., region, enumeration area, village, name of household head); and (2) an advanced mode, typically used for panel surveys, which loads not only information on the respondent s address, but also information for any other field of the questionnaire (e.g., name and age all of household members from last wave of data collection). 35

36 Figure 32: Main window of the batch upload procedure. Sample mode Used for most surveys, the sample mode involves uploading a tab delimited (.tab) file that contains data for all pre-filled questions. To do this, first create a.tab file with variable names in the first row and variable values in subsequent rows (see Figure 24 for an example). The variable names in the top row of the.tab file must match those in the Designer template. The format of those variables must match those specified in Designer. Moreover, no variable values may be left blank. Instead, blank values should be recorded as codes that fall within the variable s valid range (e.g., using -9 for blank entries in a numeric field only if -9 is in the admissible range). Figure 33: Example of a.tab file to be used during the sample procedure. 36

37 Then, upload the data in the.tab file. To do so, first choose the.tab file to be uploaded by clicking on the white Choose File and then click on the blue Upload button (Figure 23) to open a new window (Figure 25). Figure 34: Checking for data inconsistencies Next, click on the blue Verify button. The headquarters software will confirm that the uploaded.tab file contains valid contents and follows the format that the headquarters software expects. If there are any error messages, check the.tab file and make the necessary changes. Otherwise, proceed to the next step. Then, select which supervisor will be responsible for this batch of survey cases. To do so, choose the name of a supervisor from the Responsible drop-down menu. Note that in this method headquarters must upload one.tab file per supervisor. However, you can upload one.tab for the survey sample by adding a column in the.tab file with the heading _responsible and populating all the fields with the appropriate interviewer or supervisor usernames. 37

38 Figure 35: Example of a.tab file with interviewer and supervisor assignments. Advanced mode Advanced mode, used primarily for panel surveys, is generally similar to the sample mode but involves a few additional steps. For advanced mode, you need to first create a.tab file or set of.tab files. To do this, simply click on the link marked download template archive. A.zip file containing one.tab file for each hierarchical level in the questionnaire will download (e.g., one file containing all householdlevel variables, another file containing all household member-level variables, and still another file containing all consumption item-level variables). In other words, this provides headquarters with a template to fill with any pre-existing information about the survey cases that will be created. Figure 36: Creation of.tab file for each hierarchy in the questionnaire template. 38

39 While the number of such template files depends on the questionnaire, the structures of these template files follow the same format. Each file is composed of: Id, ParentId, and all other variables. The Id is meant to be a unique identifier for the template file 2 (e.g. a simple serial number). The ParentId is an identifier that links a lower-level item to a higher-level item (e.g., serial number of household to which a plot is linked agricultural plot) 3. See the image below for a graphical explanation. The other variables in the file correspond to the variables present at that hierarchical level of observation. Figure 37: Examples of.tab file format for advanced mode. Once the.tab file has been created, fill it with three sets of information. First, headquarters must provide the same set of respondent address information as outlined in the previous section that is, information for the fields marked in the Designer tool as Pre-filled. Second, headquarters may fill in any pre-existing information about respondents, using any field in the questionnaire. This would be done by filling in the columns for any variables, in any one of the.tab files that make up the questionnaire template. Third, headquarters must include identifiers in the Id and ParentID fields so that 2 The Id must be unique within the.tab file being uploaded. 3 The ParentId does not need to be unique within the.tab file, but must instead match an observation in its parent file (e.g., the ParentId of a household member must map to the Id for the household). 39

40 each record is unique and that the headquarters software can link each child.tab file to its parent file. The Id, typically a simple serial number, serves as a unique identifier for each record that is, a computer identifier that is distinct from the case identifiers used for locating respondents. The ParentId links records at one level of hierarchy with records from another level of hierarchy for example, an ID that links household members with the household to which they belong. After all.tab files have been filled in this way, zip all the.tab files into one.zip file. Then, upload the.zip file using the Choose file and Upload buttons in the Advanced data upload mode rubric. Figure 38: Uploading a.zip file with all filled in.tab files. As in the sample mode, the headquarter software will then evaluate the uploaded files for validity and consistency with the questionnaire template. If any errors are identified, correct the files. Otherwise proceed to the next step and assign the batch of cases to a supervisor. Figure 39: Uploaded.zip file. 40

41 Note that headquarters must either upload one set of.tab files per supervisor (i.e., one.zip file containing one or more.tab files) or one.tab file for all supervisors where each survey case has been assigned a supervisor within the file. Copy a questionnaire template To copy a questionnaire in headquarters, click on the copy button in the survey setup screen: If necessary, a new name can be given to the survey initiated from a cloned questionnaire, such as LFS June 2016, LFS July 2016, and so on. Data Export The Data Export tab enables headquarters to export: Data for all survey cases or by survey case status (interviewer assigned, completed, approved by supervisor, or approved by headquarters Binary data Metadata Paradata 41

42 To export data, first select the questionnaire template from the Template drop down menu on the left. Then, select the status for the survey cases from the drop down menu. Click on the refresh button for the type of data file that you would like to download. You can download all the data or by survey case status in a.sav (for SPSS),.dat (for STATA), or.tab file. Click on the download button after the update is completed. If you are thinking about committing your collected data into a public repository you are likely to be requested for a DDI description of your data. Data Documentation Initiative is an international project to create a standard for meta data in social sciences. You can export an xml file with meta information about your data for processing in DDI compatible software, for example Nesstar Publisher. The following information is exported: the list of data files, variables, their types, labels, question texts, interviewer instructions, etc. If you choose the.tab file format you will download a.zip file containing one or more.tab files accompanied by the same number of Stata.do files (e.g. household data.tab and household data.do). The number of such files is a function of the structure of the questionnaire, meaning that exporting produces one file per distinct level of hierarchy in the questionnaire for example, one file for household-level variables, one file for household member-level variables, and one file for asset-level or consumption item-level variables. Whatever the number of files, each.tab is accompanied by a Stata.do file that reads the.tab file into Stata format and that applies variable and value labels (that match the contents of variable labels and answer option fields in the Designer). Additionally, the.zip file will contain interview comments and interview actions files. Interview Comments Comments added to interviews or individual questions by users during the interview process are exported in a file named interview_comments. The file contains the information in the originator of 42

43 the comment and his role, the date and time when the comment was left, and the particular interview, or a question within an interview that it relates to, along with the text of the comment. Interview Actions Simple paradata for each survey case is exported in a file named interview_actions. The file is structured as follows: 1) InterviewId. Unique identification number of the survey case, which is assigned by the system automatically. 2) Action. The action taken. 3) Originator. The username of the individual that has taken the action. 4) Role. The assigned role of the individual, which can be Interviewer, Supervisor, Administrator or Headquarters. 5) Date/Time. Timestamp of the action, Records the date and time of the action if action occurs on the tablet (e.g. Set interview as Complete ) then the timestamp recorded will be based on the tablet date/time settings. On other hand, if action occurs on the server (e.g. Set Interview as Approved by Supervisor ), then the timestamp recorded will be based on the server date/time settings. If the World Bank hosts the server, the date/time will be in UTC (Coordinated Universal Time). Data Export Encryption (applicable to administrators) This setting allows administrators to password protect data files. 43

44 In order to enable this feature, open Settings from the user drop down menu at the top right of the page. Then, select the enable encryption check box. The password for the data export files will automatically be generated, and can be reset by clicking on the refresh button. 44

45 Appendix 1: Overview of the survey workflow To understand headquarters role, it is necessary to understand the survey workflow (Figure 1). Figure 1: Overview of the survey workflow Headquarters, designated by the building at the top of the figure, determines the households that need to be interviewed and their assignments across team supervisors. Team supervisors, denoted by the laptops in the middle of Figure 1, first receive these survey assignments (Figure 2) and then allocate them to the members of their team (Figure 3). Figure 2: Supervisors receive assignments Figure 3: Supervisors allocate assignments Interviewers, represented by hands holding a tablet, receive the assignments from the team supervisors, collect data for those assignments, and send completed assignments back to supervisors for review (Figure 4). 45

46 Figure 4: Interviewers send completed questionnaires to supervisors Team supervisors, having received completed questionnaires, review these questionnaires to confirm that all questions are answered and that answers are accurate, coherent, and plausible. After reviewing these completed assignments, team supervisors either approve or reject them (Figure 5). If a team supervisor approves a completed assignment received from an interviewer, the assignment is sent to headquarters, as represented by the paper with a checkmark in Figure 5. If a team supervisor rejects a completed assignment received from an interviewer, the assignment is returned to the interviewer initially responsible for completing it, as represented by the paper with an X in Figure 5. Receiving the rejected assignment, the interviewer must either correct it or provide explanatory notes on strange or implausible answers. When the assignments are corrected, the interviewer sends them back to the team supervisor for approval or rejection, a process that continues until the assignments are completed with the highest level of quality according to the team supervisor. Figure 5: Supervisors reject or approve questionnaires As an assignment moves through the survey workflow, it takes on a different status at each stage, as illustrated in Figure 6. 46

47 When headquarters sends an assignment to the supervisor, that assignment is considered as assigned to the supervisor (in the terminology of the software, SupervisorAssigned). When the supervisor allocates an assignment to a member of their team, that assignment is considered as assigned to an interviewer (InterviewerAssigned). When an interviewer completes an assignment and sends it to the supervisor, that assignment is considered completed by the interviewer (Completed). When a supervisor approves an assignment, that assignment is considered as approved by the supervisor (in the terminology of the software, ApprovedBySupervisor). When a supervisor rejects an assignment, that assignment is considered as rejected by the supervisor (in the terminology of the software, RejectedBySupervisor) When headquarters rejects an assignment, that assignment is considered as rejected by headquarters (in the terminology of the software, RejectedByHeadquarters) When headquarters approves an assignment, that assignment is considered as approved by headquarters (in the terminology of the software, ApproveByHeadquarters) Figure 6: Status of an assignment in each stage of the survey workflow 47