Do people with some names tend to give apps higher ratings than people with other names? Through some fairly straightforward analysis on over two million iOS app reviews, we've concluded that the answer is yes.
Carmelas, for example, tend to give ratings nearly a full star higher (4.49 stars on average) than Randals (3.61 stars).
For the analysis, we first broke out reviewers by their user names and averaged each reviewer's app ratings. We then parsed out their first names from their user names and averaged the user averages together by first name. Not all user names for the 18.7 million reviews in our database have easily identifiable first names in them, but we were able to find 2.1 million that do -- a nice big data set.
We're all pretty happy, on average. That said, some are measurably more satisfied than others, at least when using apps. The rating differences, while not huge, are measurable and significant, so we made a visualization where you can see names ranked by relative review positivity. Here's the main takeaway: Though a Randal isn't likely, on average, to leave a negative rating, there is a higher chance that a Randal will bring down the average rating for an app than a Carmela, who will most likely bring it up.
Can we use this sentiment data as a proxy for people's general demeanors? Are people with certain names generally more content than others? This data by itself doesn't say. Perhaps there's a connection between names and age or socioeconomic factors, which are in turn tied to other things such as the kinds of apps people use or how easily they're satisfied by their apps.
The causes for the sentiment differences aren't clear, but it's an interesting phenomenon nonetheless.
Emoji art supplied by Emoji One.