Double Screening - Questions about using EPPI-Reviewer

Forum (Archive)

This forum is kept largely for historic reasons and for our latest changes announcements. (It was focused around the older EPPI Reviewer version 4.)

There are many informative posts and answers to common questions, but you may find our videos and other resources more informative if you are an EPPI Reviewer WEB user.

Click here to search the forum. If you do have questions or require support, please email eppisupport@ucl.ac.uk.

<< Back to main Help page

24/10/2011 19:17

Faye Wolse

Joined: 24/10/2011

Posts: 8

Double Screening

Hi,

Our team would like to check our coding so we want to double screen a random group of references. However, when I mark the codeset for multiple data entry and assign the random allocations to two coders, we can still see each other's coding. Is there away to blind each user to the other's coding? Could it be an issue if one coder started coding before I switched the codeset over to multiple data entry?

Also, in the manual it says that items coded with the multiple data entry code set won't appear complete, however, all the references in my random allocation have the little green disc that indicates complete. I know why this is the case with single data entry (the random allocations appears within the codeset so are therefore already coded under at least one code), but why is this happening with the codeset marked for multiple data entry?

Thanks for your help,

Faye

25/10/2011 11:09

Jeff Brunton

Joined: 17/10/2011

Posts: 594

Re: Double Screening

Hello Faye,

If the items that you are speaking of were already coded when the codeset was set for single data entry they would have been marked as 'complete' at that time. Even if you change the codeset to multiple data entry the items will still be marked as 'complete'.

If you look in the Review statistics tab (bottom right corner) you can see which items are 'complete' and 'not complete'. If all of your coding is listed in the 'Completed coding' table then the original coding most likely took place while the codeset was in single data entry.

If you wanted to double screen items that have already been coded and 'completed' you would need to first mark those items as 'not complete'. As you are looking to do this on a random sample (rather than all of the items) you will need to manually mark each item in your sample as 'not complete'. You might first want to create an allocation code (in an allocation codeset) to identify the items in your sample. This allocation code can also be used to create a work assignment for the two coders. Once the items are marked as 'not complete' you should set the screening codeset for multiple data entry and create the work assignments. The second coder can then code the items and they will not see the other persons coding.

The double coding functions work under the assumption that two (or more) people will be coding the same item to begin with rather than one person coding and completing an item under single data entry and then a second person coming back to code it a second time.

Best regards,

Jeff

26/10/2011 07:23

Faye Wolse

Joined: 24/10/2011

Posts: 8

Re: Double Screening

Hi Jeff,

Maybe I read your reply wrong or maybe I did something wrong, but now all the references in the random allocation that my colleague had coded are gone. Here's what I did:

We created a random allocation of references (Random set 2) which was split into two groups (1 and 2). Then my colleague coded most of group 1. Subsequent to that we decided to use this random set to do a double screening; since my colleague had started, it would save time. Using the manual I couldn't get her coding to be hidden to me so I posted the above question. From you reply I went through and checked off all her previously coded references as not complete (I quickly realized I had to do group 2 even though she hadn't coded them, because we had unforunately listed them as a branch within our main codeset and so they were all showing as complete for that codeset automatically). Once that was done I went and set the whole code set to multiple data entry and then allocated groups 1 and 2 of Random set 2 to both her and I to code. But when I went back to show the list (by right clicking on the group and clicking "list all assigned this code") nothing came up. Three team members have signed in, including the woman who did the original coding and none of us can see the references that were initially randomly allocated to the groups. What happened? Are they still there? Can I recover them or do we have to start again?

I realize we went a rather unconventional route, but if there's a way we can fix it that would be very helpful!

Faye.

26/10/2011 10:59

Jeff Brunton

Joined: 17/10/2011

Posts: 594

Re: Double Screening Modified By Jeff Brunton on 26/10/2011 10:32:14

Hello Faye,

I'm having a bit of difficulty following what you have done. I've had a look at your review and looking in the Review statistics tab I can see lots of coding that has taken place for the different members of your team. Unless some coding has explicitly been removed (by unclicking checkboxes or selecting 'Remove selected items from this code'), it will still all be there.

It looks like you have set up allocation codes within your screening tool. This can be done but is possibly leading to some of the confusion. If you are switching between single and multiple data entry your allocation codes might be complete or incomplete depending on the state of the codeset when you assigned items to them.

To help you sort this out I will need to ask you questions specifically about your review so I will email you directly.

Best regards

Jeff

Page 1 of 1