Friends at Intelliseek wondered whether I’d get any different results on my LpkC metric if I performed the search using their Blogpulse tool. The results I got were sufficiently different to be intriguing, but not so different that I’m questioning the validity of the metric.
The top ten of US and UK newspapers ranked by the Blogpulse LpkC is as follows:
x The Scotsman – 154.39
Christian Science Monitor – 147.28
x The Guardian – 122.38
New York Times – 96.46
The Washington Post – 94.87
The Washington Times – 91.31
San Francisco Chronicle – 66.58
x The Independent – 46.04
Boston Globe – 42.22
The Seattle Times/Post-Intelligencer – 32.51
(x – UK newspaper)
Those are precisely the same ten papers that rank highest in the Technorati study, though the ranking is somewhat different. The Scotsman, Edinburgh’s paper, leaps from 6th to 1st between the Technorati and Blogpulse studies. The Washington Times moves up from 9th to 6th, and the Boston Globe slides from 7th to 9th. Otherwise, the ordering of papers is unchanged.
Blogpulse clearly turned up more results than Technorati, and not just on Scottish newspapers. The mean increased from 4.45 to 11.87, the median from 1.41 to 6.36. (As I mentioned in an earlier post, I was having difficulties getting results for all papers with Technorati. I suspect once I figure out how to compensate for that problem, Technorati will turn up more results.)
The Wall Street Journal fares significantly worse in Blogpulse numbers than on Technorati – using the opinionjournal.com, the Journal ranks 69th, with an LpkC of 7.49, slightly above the median.
I’m looking forward to meeting Matt Hurst and Natalie Glance of Intelliseek at the Chiba conference in a couple of days. I’ll try to figure out if they’re secretly Scottish and skewing the statistics to benefit Edinburg’s papers…
I’ll post the results shortly. The hotel I’m in is blocking all secure ports, and I’m only set up to access my servers via SSL. (Grr.)
Having done some exploratory work on news feed aggregation I’ve noticed that there is great variance in coverage. The Scotsmans
OPML file contains more than 1000 feeds, some very specific and some very general. I’m wondering if LpkC needs some modification to account for citation opportunities presented by papers.
The next step for LpkC is to do some analysis of how many RSS feeds each newspaper offers and when they began offering them. My suspicion is that papers that embraced RSS early and well are going to have high LpkC, and that this factor may be even more important than the pay firewall issue…
Then again, perhaps you’ve unintentionally written a Scottish bias into your search tools, Matt… I know that if I were to write a blogging search engine, it would index Berkshire County newspapers extremely well… :-)
Guilty as charged – sort of. I did put in that bias
originally. But then when I went to expand data for
other specific areas (e.g. New Zealand, Singapore, Egypt)
I found very differening results in terms of rss
availability. The Scotsman case is strange as it is really
a local paper (and sort of a national paper) but has
a large number of feeds (== large number of citation opportunities). It is also interesting to think about metrics for stories – e.g. there are many publication points for a specific Reuters article in different papers.