Rate this Post

Thanks to Holidailies, I've been blogging a lot lately. I think I've had some pretty good posts. Unfortunately, I think those posts are getting harder for people to find, given the overall amount of posting activity.

I thought it might be nice to highlight some of the more interesting posts. One approach would be for me to flag the posts that I thought deserved attention. Another is to let my readers decide which posts are more interesting.

Since I recently moved to the Drupal content management system, it's possible for me to add new functions to the site such as content rating. So, the other day I added the fivestar module so visitors could rate my blog posts. Then, I setup a custom view of the most popular blog posts.

The first interesting problem is how to define "most popular". I can easily get an average of the stars rating and use that. But I don't think that's the right metric.

Consider two cases:

case   average stars total stars
A rated five stars by two people 5.0 10
B rated three stars by ten people 3.0 30

I'd argue that even though article A may be the better article (average score 5.0 versus average score 3.0), it is not as popular as article B. I think article B should be considered more popular because it stimulated more rating activity.

The total statistic seems to be a good measure to determine popularity, since it captures both the article quality and rating activity. That's what I'm using here.

So, if you are a regular visitor to my site, please consider clicking on the stars widget at the end of the article when you encounter a post of interest. I'll be curious to see if after a couple weeks, this turns out to be a good mechanism to highlight interesting content.

Comments

Comments have been closed for this entry.

deviation

When I'm looking at ratings for books and music, I'm very interested in the distribution. If twenty people gave something three stars, then it's solidly mediocre and it's likely not what I want. But if it has ten one-star ratings and ten five-star ratings, then that's very interesting to me. Often the low ratings are people who expected something other than what they bought, and I'll really like the product and perhaps appreciate that it's a bit avant-garde.

So I think you're right to distinguish popular from highly rated. Lots of votes indicate that people were interested to read the article and care enough about the topic to vote. Low votes will likely be more of an expression of disagreement than an indicator of your writing quality.