Thread by @2weekSR

A small bonus final Tweetorial for #WorldEBHCDay, this time for our qualitatively-oriented friends :) Topic: using the Word Frequency Analyser (WFA) to analyse comments (e.g. on a survey, petition, etc.)... qualitative analysis made (a little bit) easier! @WorldEBHCDay (1/9)

Step 1: make a new endnote library with a 1 reference (your name as an author, e.g.). Copy/paste all of the text you wish to analyse (e.g. free-text survey responses, multiple comments) into the abstract field, & hit save. (2/9)

Step 2: turn this Endnote library into an xml library: click on the 1 reference which has all the free text/comments pasted in the abstract (to highlight it), then click file-export, save as: XML (important!), output style doesn’t matter. Xml library is ready! (3/9)

Step 3: create an account at Systematic Review Accelerator: sr-accelerator.com/#/ , log in & click ‘view your libraries.’ (4/9)

Step 4: upload the xml library you have just created to the Systematic Review Accelerator by clicking: tools, import refs, ‘select file’, hit upload. It’ll show you the reference you just uploaded. (5/9)

Step 5: click the tick box next to that reference & in the upper right corner of the screen, click tools & select the “word-frequency analysis” option. (6/9)

Step 6: when picking your options, it's easiest to leave the deburr, ignore common terms & ignore numbers options checked (green). The option “ignore duplicate words” has to remain white i.e. untoggled (this is crucial). Hit ‘perform frequency analysis’ button (7/9)

Step 7: WFA will tell you how often individual words/phrases appear in your comment set; in this sample set of comments, issues around ‘family’ ‘PPE’ ‘staff’ ‘uncertainty’ were mentioned by respondents most often (if you guessed these were comments on COVID, you are right!) (8/9)

Step 8: … and that’s how you get the Word Frequency Analyser to do (some of) the hard work for you, when doing a qualitative analysis on e.g. a set of survey comments. It won’t do *all* of the hard work for you, but then again, it won’t demand a co-authorship, either 😉 (9/9)