Pkp-lib: [OJS] Upgrading to OJS3: reviewers can see the names of the authors

Created on 13 Jan 2017  路  9Comments  路  Source: pkp/pkp-lib

After upgrading to OJS3 the review_method variable in review_assignments is handled differently, but the old values in that field are not updated during the migration. When adding a new assignment in OJS2 the review_method is set to '0'. OJS3 only uses values '1', '2' and '3'.

To make things worse, if you edit an review assigment in OJS3 it shows as if the review would be double blind although it is only showing that value, because it is the first item on the list.

The result seems to be that at least in some cases after upgrading to OJS3 the reviewers can see the names of the authors. We already had such situation.

If no better migration is possible, then the upgrade script should set that value to 2 for all reviews just in case.

I fixed this by updating the review_assignments assignments table with UPDATE review_assignments SET review_method = '2'

This is related to an older issue, but so important that I decided to make an independent issue instead
https://github.com/pkp/pkp-lib/issues/2173

Most helpful comment

Thanks a lot @asmecher. I added the constant name and as soon as the tests are finished I will merge and cherry-pick...

All 9 comments

I'll take a look...

Thanks!
Probably the value for double-blind reviews in OJS2 was '0'? In OJS3 that value does not exist anymore.

Edit: can't figure out where the value is set in OJS2. I started to think that our OJS2 installation was faulty in some way, but then I saw this: http://forum.pkp.sfu.ca/t/ojs-2-4-8-1-review-problem-missing-upload-file-field/19996/9

The user has the same value 0 in review_method in OJS2.

Edit2: definitely not an issue with just our installation. The same thing happens with a fresh 2.4.8. installation when upgrading to 3.0.1.

Yes, I think it would be best to set the value of the review_method to 2 (double-blind) when migrating, because, there was no such option in OJS 2.4.x, where a journal manager/editor could define the review method i.e. review method wasn't used. @asmecher, do you think this would be OK?

Alec said yes, so I will soon make a PR with a fix...

@asmecher, could you take a look at the PR above?

Just one minor comment, otherwise looks good, @bozana!

Thanks a lot @asmecher. I added the constant name and as soon as the tests are finished I will merge and cherry-pick...

Everything merged, thus the issue can be closed, I think.

Was this page helpful?
0 / 5 - 0 ratings