Oops! About that Journal Report….

PEJ’s statistical slip

Last week, a report from the venerable Project for Excellence in Journalism (PEJ) created something of a stir with a quantitative analysis of how The Wall Street Journal’s front-page coverage has changed by topic since Rupert Murdoch took over Dow Jones in mid-December. Jim Romenesko, the Internet’s premier media news aggregator, quoted the report’s key phrase: “In the first four months of Murdoch’s stewardship, the Journal’s front page has clearly shifted focus…”

That may be true, but PEJ, an arm of the Pew Research Center, whose bread and butter comes from such statistics, is now re-running its numbers. It turns out the center was not comparing apples to apples; or rather, it was comparing four apples to three. The study, quite mistakenly, did not compare the first four months of Murdoch’s tenure to the preceding four months-it compared the four months before with only three after (August-December 12 before New Corp., and then December-March 13 on the flip side). When CJR pointed out the discrepancy to the report’s lead author, Mark Jurkowitz, he was surprised and a bit chagrined. As a result the center is making new calculations, which should be ready today, based on the three-month periods immediately before and after Murdoch’s ascension.

According to Jurkowitz, the new calculations may not change the numbers too much, and he’s probably right, but we suspect they may make a difference in a few news categories.

The report’s biggest finding was that the Journal’s trademark business coverage had fallen off by half, from roughly thirty to roughly fourteen percent of total front-page stories, while presidential campaign coverage jumped from roughly five to eighteen percent. Once PEJ makes its corrections-by subtracting a month’s worth of output from the pre-Murdoch period-those two numbers probably won’t change much, because the shift in news focus is already relatively large. (One might argue, however, that the spike in electoral coverage is not due soley to any “changing news agenda,” but also to the fact that the primaries began soon after Murdoch took charge. It might have been more meaningful to compare the Journal’s election coverage under Murdoch to the same period in during the 2004 presidential race.)

The revise of the original PEJ analysis is more likely to affect the smaller news categories, which count for a few percentage points of the total. Take, for example, the report’s figures on science/technology and environmental coverage, respectively. Jurkowitz’s analysis found that during the first three months of Murdoch’s tenure, science and technology coverage dropped from 1.9 percent of the coverage to 1.1, and that environmental coverage dropped from 3.1 to one percent, compared to the four months immediately preceding his takeover. These finding prompted American University professor Matthew Nisbet, who runs the popular blog, Framing Science, to criticize the dearth of science and environment coverage in the Journal and in The New York Times, for which PEJ also tallied front-page coverage by topic:

… the NY Times and the Wall Street Journal are the papers of record for the country… Given the absence of front page coverage, is it any wonder that climate change legislation is stalled in Congress or that climate change and other science issues are barely mentioned on the presidential campaign trail? Or that in polls, climate change ranks dead last among 22 issues as a political priority?

While I wholeheartedly agree with Nisbet that the media often give short shrift to these subjects, PEJ’s conclusion that their coverage has declined in the Journal may be overstated. This is especially true for the sci/tech category, which only dropped from the baseline 1.9 percent to 1.1 percent. If you want to know why, read the rest of this graph (if you are number-phobic, or merely trusting, skip to the next one). Still here? So, if we take one quarter of the 1.9 percent baseline (i.e. one month’s worth of output) we get roughly 0.5 percent of the total coverage. If we then add that on to the truncated three-month, post-Murdoch period at the Journal we might expect the 1.1 percent of sci/tech coverage to grow to perhaps 1.6 percent, which would render the discrepancy with the 1.9 percent baseline statistically insignificant (i.e. the kind of variation one might expect to occur naturally from month-to-month even without a change in ownership). Now, the disjointed chronology might not mask the bigger drop in environmental coverage, from 3.1 percent to 1 percent of the total front-page articles-given another month, the second figure might climb to only 1.75 percent-but that’s still a change for the worse.

In the end, PEJ made an honest mistake, which almost certainly didn’t lead to gross mischaracterization. Still, those who were interested in the original analysis should check PEJ’s Web site today for more accurate numbers.

Even then, however, don’t read too much into the findings. There are a couple other limitations to the PEJ analysis. First and foremost is the fact that the project only looked at the front page “every other week day,” when it could have (and should have) made daily tallies. Second, three months is not a long time. Almost every graph will show a lot of static on such a timescale. In other words, analysis of the full years before and after Murdoch’s acquisition will say a lot more about how the paper has changed.

Has America ever needed a media watchdog more than now? Help us by joining CJR today.

Curtis Brainard writes on science and environment reporting. Follow him on Twitter @cbrainard.