5.26 of Survey Solutions released on Feb. 08. 2018 comes with several
new features and numerous important improvements.
1. Online Tester tool allows instant preview of the designed questionnaires in the browser.
2. Interviewers can be transferred between different teams.
3. Descriptive meta information can be added to the questionnaires being designed.
4. New tab “Shared with me” in Designer for shared questionnaires.
5. Enhancements in data export.
6. Objective measurement of time of recalculation of expressions.
7. Other improvements and bugfixes.
The 5th Survey Solutions User Group online meeting took place on 14th February 2018.
The Survey Solutions Designer now provides an instantly accessible tool to test questionnaires online: the Online Tester. This allows more streamlined testing of the questionnaires without the need to transfer them to the tablet. Among the other advantages is automatic access to the most recent version of the Tester, without the need to update it on the tablet and possibility to utilize the tester functionality for the users that have no access to the Google Play.
To start the Online Tester tool, click the Test button located next to the Compile button.
Online Tester provides an accurate impression of the questionnaire functioning during the preparation for the web interviews, which don’t utilize the tablets. Users that utilize Online Tester for testing their questionnaires should keep in mind that the performance of the devices (computer and tablet) is different, and in order to have an accurate impression of the performance of the tablet interviews they need to be tested on the actual device (and best on the same model that will be used for the actual data collection). See also the performance measurements for tablet’s response time below.
Currently the Online Tester does not show the computed value of variables during testing. Questionnaires that have syntax errors in conditions or other compilation errors can’t be tested.
Reassign interviewers between the teams
Reassigning an interviewer from one team to another is a useful team management action to bring a team’s size in accordance with it’s workload, eliminate bottlenecks and solve other practical problems. We have added the reassigning as a possible action of the headquarters users in v.5.26 and it can be completed in one of the following two modes:
a) interviewer is being moved to another team, taking all the assignments and interviews with him/her;
b) interviewer is being moved to another team, passing his/her existing interviews and assignments to the original supervisor.
Frequently, the reallocation of the interviewers between the teams is an action taken after e.g. a training or pilot, in which case the interviewers have no assignments and no interviews for which they are responsible. The reassignment at this stage is easy and safe, and there is no difference between the two modes.
If the assignments have been made already, one has to decide whether the interviewer is moved with his workload or not. Usually, if the interviewer is reassigned to the team working in a different geographical area, there is no possibility to continue the work started in the original area, and such work is becoming the responsibility of the supervisor. If the new team is working in the same area, the interviewer still might complete the assignments and so may be transferred with his existing workload.
If the assignments are conducted during the data collection activities, the reassignment should be done with caution to avoid data loss! If the interviewer is reassigned with a rejected interview, this interview will not be accepted from him when he tries to complete it (since it becomes the responsibility of the supervisor immediately after the reassignment).
Questionnaire meta information
A new tab in the questionnaire designer now provides access to more than a dozen meta information fields, which can be filled out by the designer to describe the questionnaire (type of survey, covered country(ies), period of survey, responsible organization, etc).
We recommend our users provide as much information as possible in these fields to properly classify their composed instruments.
The meta information entered here also appears in the PDF preview of the questionnaire for documentation purposes.
User Folders: Own and Shared Questionnaires
We have added a new tab in the Designer “Questionnaires shared with me” in addition to “My questionnaires” and “Public questionnaires”. All of the questionnaires that are shared with you by your co-authors will be shown there. “My questionnaires” shows only those questionnaires for which you are the owner.
Changes to Export of Data
The following changes have been introduced to the exported data:
- During the export to statistical packages Survey Solutions inserts an extended missing value where the answer could have been provided by the interviewer, but wasn’t. Specifically, for SPSS system file format (*.sav), this value is now marked as a missing value. For Stata binary file format (*.dta) this value is replaced with Stata’s extended missing value .a. Both SPSS and Stata will automatically account for missinges in computation of averages, tabulations, and other analyses. There are no extended missing values for tab-delimited export and the user will need to properly account for the special value -999,999,999 in their analyses based on the tab-delimited files.
- Types of the variables are now more rigidly determined by the questionnaire. You should expect now that even if all content of a string variable is numeric, it should export as string (earlier it could be saved as a numeric column).
- The API for data export now supports specifying parameters from and to, to restrict the exported data to only the interviews on which any change occurred in the specified window.
Time of the Expressions Calculation
To be able to get objective measurements of the responsiveness of the questionnaire on a tablet one can activate the execution time measurement in the Tester App’s settings: Show elapsed time per answer.
In this mode the Tester app will report the time it took to recalculate all of the expressions as a result of the user providing or changing any answer. Time is reported at the bottom of the screen in seconds or milliseconds. Note that this is only the calculation time, which does not include the time necessary to redraw the screen, so total response time will be larger.
Smaller values correspond to faster reactions and better interviewer experience. If you are getting large values, either simplify the conditions (e.g. validations) in the questionnaire, or seek for better hardware.
- Improvement in speed of calculations: v5.26 includes a new algorithm to reduce the re-calculation of all the user-written expressions (enabling conditions, validation conditions, etc) to gain performance and to make sure the tablet remains responsive even when working with very complex questionnaires.
- We have restricted the roster variable name (identifier assigned to roster) to 28 characters long. This is done to make sure that the automatically generated ID (rostername__ID) is a valid variable name (fits into Stata’s 32-characters limit). If your older questionnaire made use of longer roster variable names, you may need to rename them.
- We improved the presentation of the header of an interview when it is being reviewed by the supervisors/headquarter users. It is now presented in more compact, yet informative way:
- We have added support of Chinese language in the Designer:
- The new minimum version of .Net framework for local servers is now .Net 4.7.1. Refer to Microsoft’s web site to acquire the .Net 4.7.1 if you need to update your installation: https://www.microsoft.com/net/download/dotnet-framework-runtime
- We are about to retire the old presentation of interview details (link: old interview details) that we have retained to smoothen the transition. For all users we recommend to utilize the new interview details view immediately and catch up with supervisors’ training as necessary.