In the wake of David Sifry’s latest State of the Blogosphere post, a number of thoughtful commentators are challenging Sifry’s estimate of 50 million blogs. A principal line of argument revolves around whether the Technorati numbers create a distorted, inflated picture by including inactive blogs. Some argue that the actual size of the active blogosphere is much smaller, perhaps closer to 1.6 million blogs.

It is important to gain a better understanding of the actual participation in the blogosphere by developing a more sophisticated, refined data on the distribution and actual behaviour of bloggers.

Having said that, some of the discussion reminds me of the debate that political pollsters often engage in. They focus on differences in the results between their competing polls and argue at length about whose methodology is superior.

On the other hand, political pros – the people who actually use the polls – assume from the start that methodology will vary between pollsters and that these different methodologies will yield different results at any given time. What they focus on is the trends over time between polls with known and consistent methodology. This analysis enables them to understand what is really going on and to look for the drivers of behaviour, not simply the manifestation of that behaviour.

When I look at David Sifry’s (or anybody’s) stats, I look for consistency of methodology and trends over time. The ongoing addition of new bloggers. The accelaration of the discussion through both new posts and comments. And the understanding that many bloggers post rarely or abandon their blogs altogether. That’s the real value for me.

And of course, once I’ve drawn everything I can from his results, I will look for other data sets that will show me other things (e.g. active blogs.)

So, I hope the conversation continues. I will be an active follower of it (and occasional participant.) But let’s remember, there is real value in following Sifry’s results over time, regardless of whether we agree with the details of the snapshot at any particular time.