Oscar Jr. Was Here

Comments on Comments on the Preliminary, Shoddy Analysis...

[moved to a new post for space reasons]
[edited extensively]

Charles G. Hill wrote:

I'm not sure exactly when you're supposed to add those ten to achieve the desired effect, and I have some general qualms about futzing with blogrolls — the ninety or so blogs I list are there because they are regularly read, not because I think I stand to gain anything from their presence...

I certainly didn't intend to imply that adding any random, additional blogroll links at any time will increase one's incoming links. My interpretation of these data is that websites with a high-quality blogroll (where quantity of good links can add to quality) are recognized as more valuable to readers and are, therefore, more likely to be linked. If I am correct, adding random sites to the 90 or so that Mr. Hill finds most read-worthy would not increase the value of his weblog to his readers and, at the least, not increase the number of linking sites. On the other hand, if he identified 10 additional weblogs that both he and his readers found worth visiting, Mr. Hill might expect a few more incoming links.

To give an example of this, even when I knew I'd read everything on a given day on InstaPundit, I would often return to the site because I knew that many of the weblogs that I liked to visit (and whose url's I hadn't yet memorized) were linked on his site.

Bigwig wrote (in the comments below):

[Y]ou need to factor in the age of the blog, as well. Those lists tend to grow over time.

Perhaps, but I'm not so sure. I wrote, "Obviously, I'm not in MBE territory yet", recognizing the fact that this site has not been around long enough (or, more likely, is not good enough) to fulfill the prediction implied by my analysis. Similarly, Cyberangel, who (in Mr. Hill's comments) wrote, "Ha! I defy you both & have no one linked to me! Yet. :)" appears to have been around for only three-ish months.

However, once a weblog is established (however that may be defined), as those in the MBE would seem to be, I see no reason to expect that the ratio of ingoing-to-outgoing links will necessarily increase over time. At an extreme, while InstaPundit may be reluctant to add more links, new websites like mine will continue to link to him. But are most established sites close to that extreme?

Next, in my comments, Mr. Hill wrote, "As long as you don't factor in the age of the blogger. I'm already depressed." Please don't be depressed, at least in this respect: I hope to test next the number of incoming links based on webloggers' age (controlling, if possible, for the age of the blogs themselves). I expect that my hypothesis, that blogger age correlates positively with incoming links (and hits), will be borne out. I may even seek testimony from Mr. Quick, Mr. Den Beste and Mr. Reynolds. (Thanks to Frank J. of IMAO for the idea!)

Finally, for tonight, Kevin McGehee (who, note, is on my blogroll) wrote (in Mr. Hill's comments):

To my utter mystification, and in utter confoundment of Oscar's number-crunching, I am linked by a great many more blogs than are in my blogroll.

I live to contradict the statisticians.

Well, outliers exits, and Mr. McGehee may well be one. (However, he may not realize how many incoming links he's forgone by failing to link to this site.) But, of course, I would never predict that all the sites on the MBE would have more outgoing than incoming links. Even at the heights of the blogosphere, not everyone can be above average within the sample, so I don't consider my analysis confounded. For what it's worth, InstaPundit (second to Dave Winer) receives 515 more incoming links than can be explained by his outgoing generosity. As noted below, I'm down about 135 at this point.

Posted by oscarjr at January 30, 2003 11:35 PM | TrackBack