If the primary measure of a journal's value is its impact -- pretty layouts and a good Employment section and so on being presumably secondary -- and if the Impact Factor is a measure of impact, and if publishers are making a good faith effort to offer value for money -- then why is there no apparent relationship between IF and journal prices? - Bill Hooker

those graphs illustrate your point well - Jean-Claude Bradley

You should break the journals down more finely by subject area: IF varies between areas (e.g. bioinformatics is high, plant pathology low). Also, prices will depend on the number of subscriptions, so you might want to try factoring that in as well. - Bob O'Hara

@Bob: good point, though I wouldn't know how to subdivide the geography titles. Also, with only a couple hundred journals even in the larger Elsevier set I'm not sure further breakdown would be useful (you'd end up with very small sets). Not sure I understand your last sentence -- afaik, publishers don't take into account how many subs they have except in order to determine what price increase is needed to cover attrition. I've never seen 'em lower prices as subscriptions increase. - Bill Hooker

Awesome post, Bill! Very valuable! Thanks! - Björn Brembs

In Figure 3, prices for the non-IF journals look bimodal. I wonder if there is a correlation with field or proximity to another, (better-)ranked title? - Chris Cotsapas

On subject area, the ISI has categories, for example. You might only need 10 journals per subject area to get some reasonable results: have a look at random regression methods. On sales - less popular publications can sometimes be more expensive, because the print runs are expensive, and the economies of scale work against them. - Bob O'Hara

@Chris: it might be an artefact of the way I've graphed it. I'm trying to figure out a better way to plot the distributions for this and another, larger dataset -- I'll keep your question in mind. - Bill Hooker

@Bob: ISI categories are a good idea, but "random regression methods" -- you want me to (shudder) learn math??!! :-) I see now what you mean about subscriptions, but I don't think I can get hold of that information. As a crude proxy I could break out online-only, print+online and print-only (if there are any such), because marginal costs of reproduction are so much lower online. - Bill Hooker

No, not math - merely stats. :-) Any decent modern stats package should be able to do it. - Bob O'Hara

@Bob: is there any such package I can get hold of for free? I'm still unemployed so no access to usual resources. - Bill Hooker

There's R. I've never used it myself, but it get's good reviews from people I know that have: http://www.r-project.org/ - John Dupuis

I would recommend R too - it's what I use as my statistical work-horse. - Bob O'Hara

Ah, I had forgotten that R was Free. OK, I'll try it. It's well past time I learned some stats anyway. Don't expect to hear from me for at least a week... :-) - Bill Hooker

Feel free to pester me with questions! - Bob O'Hara

wrt R, use the gui package maybe until you get situated (Rcmdr) - Christina Pikas

Bob, you may come to regret that kind offer! :-) - Bill Hooker

Christ on a crutch! You people were kidding about R, right? Right? I got it running OK (the gui version, thanks Christina), and that's where the Easy stopped and the Horrible started. @Neil, calling that curve "steep" is like calling the known universe "big". - Bill Hooker