"The conversionxl study is from mid 2014 so the results may be dated."
The first study I linked to is from less than a year ago and finds the exact same results.
No, it did not find the exact same results. CXL found that it was horrible in 2014. The first one from less than a year ago found that it just needs to be labeled to be more effective (with "effective" defined as interactions), and that it definitely has a place as a space saving menu icon on mobile apps. And 1-2 years from now you probably won't need a menu label. Paricularly on mobile where it is already becoming the de facto space saving standard.
The conversionxl study is from mid 2014 so the results may be dated.