Rating Methodology

The other day I was pondering over the fact that my ratings range from 2 to 5 stars on a 5 star scale – I’ve never rated a book lower than 2 stars, even when I can’t bring myself to finish them. Then I thought about how I’ve tended to rate books lower than the average rating I see on Goodreads, and I began to wonder if I should rethink how I approach rating the books.

As many of you know, and as is posted on my parent-level Book Reviews page, I rate such that a book starts at 4 stars, and I dock points and add them as things irritate me or impress me, respectively. This means that if I liked a book well enough, it gets a rating of 4/5 — if it has truly exceptional moments, I’ll push it up to 4.5/5 or 5/5, but if there are parts of it that bother me, I’ll dock points. Generally, it’s easier to have points docked than have them added.

So since I’ve been thinking about this a fair amount lately, and since I’m getting a master’s degree in Applied Statistics, I decided to do some analysis on my ratings versus another reliable source of rating information: Goodreads.

I collected information on the ratings of all 37 books I’ve rated so far, I calculated a weighted average based on the number of people who rated that book, and I compared the overall sample average to my rating average. It’s true that my average rating is slightly lower than that of Goodreads (which I’m considering to be representative of all readers), but not enough different to be considered significant.

And since there’s not a statistically significant difference between my ratings and the ratings on Goodreads, I’m going to continue sticking to my status quo, since it seems to work for me. I might, however, try to make it a little easier to earn points so that it’s more fair with the ease of losing points. Maybe it will balance out better.

Posted on September 9, 2013, in Miscellaneous. Bookmark the permalink. 2 Comments.

  1. Nice use of stats in an unusual context. I think people underrate the use of intelligently analysed figures when they’re available. Then again, I used to use them in my last job, so I’m a bit biased. I recommend this for novelty stat use, on similar grounds:

    • That was a thoroughly enjoyable read — thank you very much for the link. I absolutely love seeing this kind of analysis done, especially when people aren’t afraid to go into the details and show the graphs and state their correct conclusions and limitations. Especially when there seems to be this idea that this kind of analysis isn’t useful (or, if you ask my co-workers) normal for this kind of context… Or interesting, for that matter.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: