We have pre-loaded assignments with expected quantity as 1. We are noticing that we are receiving duplicate interviews (different interview keys) for the same assignments. This hasn’t happened during a questionnaire version update (last update to the questionnaire was in the first week of march).
@klaus has already reported this case and this should have been partially fixed in a recent hotfix 21.5.3. As noted, we have so far suspected the multiple interview creation from the map dashboard as the vulnerability and addressed that issue. If there is any remaining scenario when the multiple interviews are still created, please do let us know the steps.
Your screenshot indicates the interviews were initiated on May 24, which is about the same days as in Klaus’ report. Have you had any recent reports of the same problem in the version 21.5.3+ ?
We do not use map dashboard. All the assigned with duplicate interviews were reassigned to the interviewers on or before 19 May.
We installed 21.05.3 on 30 May 2021. Please find an example where the interviews were created after this date. Probably running 21.05.5. Reassigned to the interviewer on 17 May 2021
We have the 32 assignments with 64 duplicate interviews. assign_datetime is the date and time of assigning the assignment to the interviewer and create_datetime is the date and time of creating the interview.
interview__key
assignment__id
assign_datetime
create_datetime
99-49-86-39
41412
18 Nov 2020 4:58:53 AM
18 Nov 2020 7:01:17 AM
27-95-59-32
41412
18 Nov 2020 4:58:53 AM
18 Nov 2020 7:00:51 AM
79-41-33-22
25126
22 Jan 2021 11:06:04 AM
24 Jan 2021 6:52:03 AM
41-41-45-53
25126
22 Jan 2021 11:06:04 AM
24 Jan 2021 7:03:03 AM
78-66-28-85
31476
20 Feb 2021 7:25:17 AM
20 Feb 2021 10:37:15 AM
73-33-35-19
31476
20 Feb 2021 7:25:17 AM
20 Feb 2021 12:03:51 PM
47-42-35-87
28321
03 Mar 2021 9:03:52 AM
08 Mar 2021 4:25:49 AM
49-08-63-10
28321
03 Mar 2021 9:03:52 AM
08 Mar 2021 4:13:41 AM
48-35-53-14
42169
25 Mar 2021 11:18:42 AM
26 Mar 2021 11:56:27 AM
62-72-60-40
42169
25 Mar 2021 11:18:42 AM
26 Mar 2021 12:02:11 PM
59-52-99-51
47144
08 May 2021 3:53:39 AM
24 May 2021 5:37:33 AM
64-10-07-16
47144
08 May 2021 3:53:39 AM
24 May 2021 5:35:20 AM
11-76-78-96
47197
08 May 2021 3:53:39 AM
22 May 2021 11:48:57 AM
31-79-81-42
47197
08 May 2021 3:53:39 AM
22 May 2021 6:52:42 AM
59-12-63-29
46891
08 May 2021 3:53:44 AM
24 May 2021 6:07:57 AM
12-97-46-85
46891
08 May 2021 3:53:44 AM
24 May 2021 6:08:47 AM
48-45-15-06
47032
08 May 2021 3:53:44 AM
22 May 2021 6:19:12 AM
47-81-93-84
47032
08 May 2021 3:53:44 AM
22 May 2021 6:20:20 AM
32-28-84-11
45840
08 May 2021 3:53:55 AM
24 May 2021 6:27:07 AM
15-50-65-07
45840
08 May 2021 3:53:55 AM
24 May 2021 12:00:44 PM
34-22-19-32
45887
08 May 2021 3:53:55 AM
24 May 2021 5:54:34 AM
17-68-09-53
45887
08 May 2021 3:53:55 AM
24 May 2021 5:53:15 AM
39-89-80-43
45889
08 May 2021 3:53:55 AM
24 May 2021 12:07:57 PM
16-01-83-66
45889
08 May 2021 3:53:55 AM
24 May 2021 12:09:54 PM
75-53-03-88
45895
08 May 2021 3:53:55 AM
22 May 2021 6:07:49 AM
94-48-11-99
45895
08 May 2021 3:53:55 AM
22 May 2021 6:10:23 AM
58-59-64-07
45699
08 May 2021 3:54:00 AM
24 May 2021 12:29:32 PM
20-58-61-08
45699
08 May 2021 3:54:00 AM
24 May 2021 12:28:37 PM
64-85-42-64
46710
13 May 2021 9:48:26 AM
19 May 2021 8:39:30 AM
36-10-33-71
46710
13 May 2021 9:48:26 AM
19 May 2021 8:38:24 AM
87-19-59-20
46431
17 May 2021 4:50:32 AM
29 May 2021 1:20:21 PM
21-33-57-97
46431
17 May 2021 4:50:32 AM
29 May 2021 12:54:38 PM
46-39-72-00
46383
17 May 2021 4:50:41 AM
31 May 2021 1:00:05 PM
79-33-81-11
46383
17 May 2021 4:50:41 AM
31 May 2021 12:58:13 PM
80-25-95-37
46278
17 May 2021 4:50:57 AM
20 May 2021 10:34:08 AM
26-84-73-58
46278
17 May 2021 4:50:57 AM
20 May 2021 10:32:46 AM
45-13-94-84
46280
17 May 2021 4:50:57 AM
19 May 2021 12:18:37 PM
95-31-81-64
46280
17 May 2021 4:50:57 AM
19 May 2021 12:21:59 PM
64-93-15-53
46250
17 May 2021 4:51:02 AM
19 May 2021 11:34:22 AM
88-45-06-99
46250
17 May 2021 4:51:02 AM
19 May 2021 11:46:00 AM
11-26-41-41
46221
17 May 2021 4:51:06 AM
30 May 2021 10:40:31 AM
00-29-32-39
46221
17 May 2021 4:51:06 AM
30 May 2021 10:28:51 AM
07-33-82-00
46189
17 May 2021 4:51:07 AM
19 May 2021 11:27:41 AM
71-95-41-39
46189
17 May 2021 4:51:07 AM
19 May 2021 11:24:39 AM
65-78-46-81
46204
17 May 2021 4:51:07 AM
26 May 2021 10:25:01 AM
49-76-58-85
46204
17 May 2021 4:51:07 AM
26 May 2021 10:56:31 AM
85-86-02-63
46166
17 May 2021 4:51:11 AM
22 May 2021 11:24:19 AM
42-02-89-10
46166
17 May 2021 4:51:11 AM
22 May 2021 11:15:05 AM
82-27-81-92
46173
17 May 2021 4:51:11 AM
19 May 2021 12:04:31 PM
95-46-77-36
46173
17 May 2021 4:51:11 AM
19 May 2021 12:10:10 PM
90-85-52-79
46175
17 May 2021 4:51:11 AM
22 May 2021 5:22:45 AM
10-89-63-43
46175
17 May 2021 4:51:11 AM
22 May 2021 5:21:03 AM
27-77-02-55
46131
17 May 2021 4:51:16 AM
07 Jun 2021 7:30:26 AM
68-97-75-47
46131
17 May 2021 4:51:16 AM
07 Jun 2021 7:21:08 AM
64-37-49-06
46125
17 May 2021 4:51:21 AM
26 May 2021 10:11:50 AM
92-12-20-58
46125
17 May 2021 4:51:21 AM
26 May 2021 10:13:29 AM
70-91-72-23
46078
19 May 2021 4:17:25 AM
20 May 2021 6:44:39 AM
69-14-02-52
46078
19 May 2021 4:17:25 AM
20 May 2021 6:43:17 AM
24-31-55-46
46068
19 May 2021 4:17:30 AM
20 May 2021 5:37:48 AM
06-92-11-77
46068
19 May 2021 4:17:30 AM
20 May 2021 5:27:54 AM
15-12-58-90
46060
19 May 2021 4:17:31 AM
19 May 2021 1:29:42 PM
72-73-46-15
46060
19 May 2021 4:17:31 AM
19 May 2021 1:27:07 PM
30-78-54-97
46010
19 May 2021 4:17:36 AM
19 May 2021 1:20:23 PM
99-38-15-61
46010
19 May 2021 4:17:36 AM
19 May 2021 1:22:37 PM
If it is any help, here is the schedule of updating our SS server.
Is there any update on this issue? We still see duplicate interviews from the same assignment id (assignment are for 1 interview only). We are still on 21.5.7. Has 21.06 touched on this issue? I couldn’t spot a mention in the release notes.
I spotted this issue again in our current survey where two interviews were generated from an assignment that expects only 1 interview. We are on ver 22.02.6.
There are several valid scenarios that could produce this result and of cause there could be a bug.
To help us to understand better the source of the problem could please clarify:
Did you change the size? (More than one initially and reduced to one)
Were those interviews created on tablet or directly on HQ? (Was synchronization involved?)
Did you reassign this assignment? (The first interviewer synced with tablet and created interview, assignment was reassigned, the second interviewer creates on tablet the second one)
Do you use Supervisor app? (More complex scenarios)
Did you migrate assignment between different versions? (Tablet sync during migration)
How much time has passed between those interviews creation moments? (Double click at some point)
We feel that the reassignment on 16 Jun to Rd07 shouldn’t have been allowed as the assignment was already received by tablet by the previous interviewer.
Another scenario were we observed duplicates was when interviews were received by the tablet but the interviewer accessed the web interface and opened an interview to have a look at them. This created two interviews. Possible solution may be to not permit creation of interviews on web-interface once assignments have been synced to the tablets.
@ashwinikalantri , case 1 described in your message is exactly the scenario described in the Excessive interviews article:
This has always been the case for many years, and there are some recommendations there. But I agree the assignments page may have benefitted from the same safety mechanism as implemented in Interviews reassignment:
Your case 2 is probably missing some details. There is a safety lock which prevents the interviewers to edit the same interview on the tablet and on the web:
Yes, this is exactly our scenario. The checkbox suggested by you will definitely help mitigate this issue, but Humans are stupid !! The Excessive Interviews article talks about Recommendations. These are not fool proof and we may and do skip steps. I would love to see the system enforce the expected behaviour - Collected interviews are never more than Expected.
This could be done is a few ways:
Only re-assign assignments that haven’t been received by a tablet.
If there is a need to re-assign a received assignment, that it should only be possible to re-assign them to the supervisor.
The supervisor/HQ can assign to another interviewer only when the first interviewer has synchronised their tablet and the assignment removed from their tablet OR if an interview was created, it was sent back to the server.
This forces the first interviewer to release the assignment before the second interviewer can get it.
Regarding the scenario 2, following are the timestamps:
Assignment No 18182 (along with 499 other assignments) created on 9 Jun 2022 at 16:31 and assigned to the supervisor.
Assignment re-assigned to the interviewer on 13 Jun 2022 at 15:28
Interviewer tablet synchronised at 13 Jun 2022 at 15:29
The interviewer logs in to the web interface to glance at the interviews he has for the next day and starts a few interviews (including Assignment 18182) to look at the pre-loaded data. (This should not be permitted as the assignments were already in the tablet.)
The interview (75-40-63-66) for Assignment No 18182 was created in the web-interface on 13 Jun 2022 at 17:04 and has the status Interviewer Assigned.
The interviewer goes to the field the next day and starts the interview (88-59-46-15) again on his tablet on 14 Jun 2022 at 09:19.
Again, most users are not technology savvy or understand the functioning of the SuSo platform. They only perform selected tasks. Training them is not always feasible. They would screw up even after training. Putting these checks in the system saves monitoring and training resources. It also gives me more confidence that nothing unexpected is happening that I should be worried about.