HelpForum

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

HomeHomeUsing EPPI-Revi...Using EPPI-Revi...Questions about...Questions about...Bug Report ForumBug Report Forum
Previous
 
Next
New Post
06/07/2018 21:11
 

Is there a place in the forums to report bugs?

I ran the "Assign documents to specified codes randomly function" a little while ago and I just recently realized that it didn't allocation citations into groups randomly. Rather, it created 10 groups and alphabetically allocated the citations to the various groups by author name. Which is really unfortunate because my team didn't realize (We would sort citations by Title when screening) until just now.

 

I tried to replicate this bug but was unable to do so. I re-randomly assigned documents into groups and it worked properly today.  

 
New Post
12/07/2018 12:00
 

Hello William,

Thanks for letting us know.

I have had a closer look at that function and I'm not sure how it could end up alphabeticaly arrange based on what is happening in the code behind.

In the creating the groups you decide what items to include in the group using the 'Select from' option. Is it possible that you may have selected just the items on a particular page. For example, if all of the items on the selected page (say, page 1) started with the letters 'A' and 'B' then all of the items in that group could appear to be alphbetical (with respect to all of the items in your review).

If you can recreate what you saw we can have a closer look at it.

Best regards,

Jeff

 
New Post
16/07/2018 20:59
 

Thanks for the reply Jeff.

I did not only select the items on a particular page. I had created a unique code for 5409 citations. I then selected this code for the randomization process. I utlized the search function to varify that the original 5409 citations were found in the 10 groups.

(Group 1 OR Group 2 OR Group 3... OR Group 10) = 5409 original citations.

There is some randomization in the groups but there is obviously some order to the groups that are not accountable to chance.

Here are a list of some examples:

1. Group 5: Author names A-H

2. Group 6: Author names A- L

3. Group 7: Author names A-R

4. Group 8: Author names E-R

5. Group 9: Author names L-Z

6. Group 10: Author names N-Z

The likelihold that this many groups do not have the full spectrum of author names is extrememly unlikely given a sample of 5409 citations.

Not sure what else I can add here. It appears as the randomization failed in groups >5.

Cheers

 
New Post
17/07/2018 16:53
 

Hello William,

I had a look at your review (is it ID:14245?) and tried to match up what you are describing with what I am seeing.
Are the 10 groups the ones I see in the codeset called '__________'? (the second codeset in your review). Each group has 504 items (except the last one with 540). This adds up to 5409.
If it is those 10 groups then I'm seeing something a bit different. Each of those groups appear to have a fairly even mix of items from A to Z.

Group 1: A - Z
Group 2: A - Z
Group 3: A - Z
Group 4: A - Z
Group 5: A - Z
Group 6: A - Z
Group 7: A - Z
Group 8: A - Z
Group 9: A - Z
Group 10: A - Z

If I am looking at the wrong groups please let me know.

Best regards,
Jeff

 
New Post
17/07/2018 21:50
 

Hi Jeff,

That is the second randomized group I created when I tried to replicate the bug. As you noted, it is well randomized.

The initial not-so random groups are under the codeset "Citations for Screening (Start Here)" and under the parent code "April 30 Search Results."

Cheers,

Wil

 
New Post
18/07/2018 14:48
 

Hello Wil,

I had a look at those groups and they are as you have described. I then looked in the database to see if there was anything unusual about the codes and how they were generated but was unable to see anything that might explain how the items in the groups ended up in that order. The attribute_ids for the items in the later groups are consecutive suggesting they are in the order they were generated.
We will need to investigate this a bit further. We appreciate your bringing it to our attention.

Best regards,
Jeff

 
New Post
28/07/2018 00:34
 

Good afternoon Jeff,

Just wondering if you've made any progress on this. Also, what do you think about my suggestion for a bug report forum group?

I came across another bug in which the "Refresh Search List" function doesn't work. I.e. If I update codes on documents and rerun the searches the hit counts don't update. Rather, I have to manually redo the searches (and use the combine tool again)after the documents have been coded to get them to show up in the search results.

Cheers,

Wil

 
Previous
 
Next
HomeHomeUsing EPPI-Revi...Using EPPI-Revi...Questions about...Questions about...Bug Report ForumBug Report Forum


Copyright 2021 by EPPI-Centre :: Privacy Statement :: Terms Of Use :: Site Map :: Login
Home::Help::EPPI-Mapper::RIS Export::About::Account Manager