Describe the issue or problem
One of our editors reported an issue with some of the figures in the Editorial Activity dashboard, specifically the percentages in the Total > Submissions Declined and the Total > Rejection Rate columns. See screenshot below for details.
Total > Submissions Declined:
The percentage in the Total > Submissions Declined field does not seem to reflect the editorial activity correctly. According to the dashboard,330 items were submitted and 100 were accepted, so how can there be 317 articles declined? The 214 “other” submissions were all published and none declined, so where did the 317 articles declined come from?
Total > Rejection Rate:
It’s been my understanding that the Acceptance Rate and the Rejection Rate do not include the journal’s 214 imported submissions. If that is the case, the Total > Acceptance Rate of ~30% seems to be correct (330 total non-imported submissions received, 100 of which were accepted), but the Total > Rejection Rate of 96% seems to be incorrect. I would have expected a Total Rejection rate of something nearer 70%, not 96%. Is this an error, and, if it is not, where is that figure coming from?
Here is the screenshot of the Editorial Workflow dashboard, captured today:
Curious if you have articles imported via QuickSubmit or NativeXML and if this might be affecting this?
This looks to me like it is possibly this bug: Improve Editorial Activity calculations · Issue #9813 · pkp/pkp-lib · GitHub
However, this should’ve been fixed as of 3.4.0-6, and if you’re running a later version, it should be accounted for. @bozana - any ideas?
@rcgillis It was 7 years ago that the journal came over to us from Human Kinetics, so I don’t remember whether the 214 articles were added through Quick Submit or through an XML import. I checked the articles and the Activity Logs are blank, so I suspect that it was done by XML. We were using OJS 3.1.1.2 at the time, if that makes any difference.
This did seem like a variant on the issues that I reported in Editorial activity statistics look strange - OJS 3.4.0.4 - #34 by bozana and it had been my understanding from that post that the fix had been applied to 3.4.0.6 and beyond, but we’re still seeing odd calculations in 3.4.0.8, so I wanted to report it.
@rcgillis Upon further investigation, it doesn’t look like Quick Submit vs. NativeXML import makes a difference in this case. Another much smaller journal that was added at around the same time is also showing the calculation problem, and those 40 back articles were Quick Submitted. See screenshot below for details.
I checked more of our other journals, including ones that started on OJS and didn’t have any back content, and there is something wrong with the calculations on each.
Hmmm… When I look into the code everything seems fine to me… So maybe we can try to understand the first case you have mentioned:
Maybe you can try to see if that all is true using the SQL:
Run this to see the result for initially declined submissions (exchanging ??? with that journal ID):
select DISTINCT s.submission_id, pi.publication_id, ed.decision
from `submissions` as `s`
left join `publications` as `pi` on `pi`.`publication_id` = (
select `pi2`.`publication_id`
from `publications` as `pi2`
where `pi2`.`submission_id` = `s`.`submission_id` and `pi2`.`status` = 3
order by `pi2`.`date_published` asc limit 1
)
left join `edit_decisions` as `ed` on `s`.`submission_id` = `ed`.`submission_id`
where `s`.`context_id` in (???) and `s`.`submission_progress` = '' and (`pi`.`date_published` is null or CAST(s.date_submitted AS DATE) <= pi.date_published) and `ed`.`decision` in (8) and `s`.`status` = 4;
And run this to see the result for after review declined submissions (exchanging ??? with that journal ID):
select DISTINCT s.submission_id, pi.publication_id, ed.decision
from `submissions` as `s`
left join `publications` as `pi` on `pi`.`publication_id` = (
select `pi2`.`publication_id`
from `publications` as `pi2`
where `pi2`.`submission_id` = `s`.`submission_id` and `pi2`.`status` = 3
order by `pi2`.`date_published` asc limit 1
)
left join `edit_decisions` as `ed` on `s`.`submission_id` = `ed`.`submission_id`
where `s`.`context_id` in (???) and `s`.`submission_progress` = '' and (`pi`.`date_published` is null or CAST(s.date_submitted AS DATE) <= pi.date_published) and `ed`.`decision` in (6) and `s`.`status` = 4;
This way you can see what submissions are considered declined on the editorial activity dashboard and double check if it is correct.
I would be also interested to know what you see in the column publication_id in the results. – It could be that already published submission is later declined…
Those all looks like declined and not published. Could you take a look at few of those submissions to see what actually happened there, what are those submissions and how are they declined? – to find an answer on your question how come there are so many declined…
@bozana, I double-checked a random sampling of 11 articles from the two queries done for the larger journal and noted below whether the rejection happened before or after peer review. None appear to have been published, then declined.
Please note that the relative percentages are probably not representative.
(Note: It would be really nice if Archived items could be filtered by whether they were declined!)
