Upgrade StorageWhy were these articles deleted from our dataset? Confused.
Hi all,
I am a first year graduate student in a social sciences PhD program. My advisor is working on a meta-analysis, and my job has been to clean, sort, and manage a dataset on Zotero with 330,000 articles extracted from the databases based on her research questions. We have been manually coding the dataset until recently when my PI told me that she wanted to do "batch exclusions" and search for keywords in the dataset and then delete large swaths of articles. We deleted 72,000+ articles. Then deleted them from the trash. Unfortunately, after we went back to check, 12 seminal articles related to my PI's research questions were deleted from the large dataset.
My PI wants me to propose a solution for the dataset going forward. She is very frustrated that the keywords eliminated some of the articles central to her research questions. Does anyone have any ideas for us going forward? I'm at a loss.
Thanks in advance!
I am a first year graduate student in a social sciences PhD program. My advisor is working on a meta-analysis, and my job has been to clean, sort, and manage a dataset on Zotero with 330,000 articles extracted from the databases based on her research questions. We have been manually coding the dataset until recently when my PI told me that she wanted to do "batch exclusions" and search for keywords in the dataset and then delete large swaths of articles. We deleted 72,000+ articles. Then deleted them from the trash. Unfortunately, after we went back to check, 12 seminal articles related to my PI's research questions were deleted from the large dataset.
My PI wants me to propose a solution for the dataset going forward. She is very frustrated that the keywords eliminated some of the articles central to her research questions. Does anyone have any ideas for us going forward? I'm at a loss.
Thanks in advance!
I don’t recommend ever deleting any item when doing a meta-analysis. Instead, add a tag for “rejected” and set it as a colored tag (I use Red for this) to clearly indicate these have been rejected. When you reject an item, also add a tag labeling why it was rejected (I start each of these label tags with * to keep them together).
In general, I would strongly recommend against using keyword searches to exclude items from your meta-analysis. This is highly like to result in false positives, as you found here.