Doing the quintiles, part 1: Posting frequency
As noted in Public Library Blogs: 252 Examples, the 252 blogs studied had a total of 5,976 posts during the study period (March 1, 2007 through May 31, 2007: 92 days). That’s an average (mean) of 23.7 posts per blog, or about one every 3.9 days–which I called “a bit less than two posts per week.” The median was considerably lower: 12 posts, or a little less than one per week.
Here are the quintile figures. I needed to account for the two extra blogs (5 into 252 leaves a remainder of 2), and as it happens there were 52 blogs with fewer than 16 and more than 9 posts, so I included the extras in the middle quintile.
- Q1: Most posts: From 33 posts (one every 2.8 days) to 400 (4.3 per day).
Average: 73.9 (four posts every five days)
Median: 58 (one post every 1.6 days).
This quintile accounts for 61.8% of all posts.
- Q2: More posts: From 16 posts (roughly one every six days) to 32 (one every 2.9 days).
Average: 22.5 (roughly one post every four days)
Median: 22 (same)
This quintile accounts for 18.84% of all posts.
- Q3: Average frequency: From ten posts (one every nine days) to 15 (one every six days)
Average: 12.2 (just under one per week)
Median: 12 (same)
This quintile accounts for 10.61% of all posts.
- Q4: Fewer posts: From five posts (one every 18 days) to nine (one every ten days).
Average: 7.06 (one every thirteen days).
Median: 7 (same).
This quintile accounts for 5.91% of all posts.
- Q5: Fewest posts: From two posts (one every 46 days) to five (one every 18 days)–noting that 15 blogs had five posts each, making it impossible to break quintiles at a number break.
Average: 3.38 posts (just over one per month)
Median: 3 (one per month).
This quintile accounts for 2.83% of all posts.
It’s worth noting that Q5 hits the lower limit of the study: Blogs with one post were excluded.
In an extreme case, if you took a sample of 50 public library blogs and happened to get all blogs from Q1, and then took another sample and happened to get all blogs from Q5, you’d have a truly ridiculous situation: One group would show more than twenty times as many posts as the other group!
With a distribution like this, it’s interesting to see how closely it approaches the Pareto assumption. The answer: Not very. To get 80% of the posts, you need to include 38% of the blogs.