2004 was a better year for Globe accuracy
EVERY JANUARY I write an errors-in-review column tallying the Globe's corrected mistakes for the year just ended. Last year's column concluded with my view that, if nothing else, the record-high number of mistakes in '03 gave the Globe ample opportunity to improve in '04.
Mercifully, it has.
The 2004 tally is just in, and it shows that the number of corrections required over the past year dropped by almost 200 -- to 1,031 -- reversing the sharply upward trend of the previous three years. But even with the drop, there were still more corrections in the year just ended than in any other recent year except 2003.
Reporters made more errors, editors made fewer, and about 10 percent of all corrections were for page one stories. There was a dip in the number of stories that required multiple corrections, but there were still 77 of them. Overall, City/Region logged the most corrections -- approaching a quarter of the total -- which is not unusual, given Metro reporters' higher production rates and tighter deadlines. Once again "misidentification" was the most common type of error.
So what does all this tell us? That the Globe staff was more accuracy-conscious in '03 than the year before? Probably. That it was still more error-prone than it was in, say, the 1990s, when corrections hovered around 675 -- or just that, with time, it's more willing to run corrections? Maybe, on both counts.
It's impossible to know what drives the annual fluctuations, although a lower tally is certainly a good thing, given that the paper corrects all known errors.
The tally has its limitations. Other qualities -- such as fairness and public service -- are probably better measures of a paper's overall character. But because accuracy is the bedrock of all else, the annual corrections tally tells us something important. Even allowing for humans' fallible nature, most of the 1,031 errors corrected last year were preventable. The Globe's initiative to reduce errors, now in its second year, still has a way to go.
As part of that initiative, Globe editor Martin Baron has instituted an auditing system whereby sources on stories are periodically called to see if they found the story accurate. The Globe has also set up special phone and e-mail lines for readers to report errors (see Page A2), and Baron frequently reminds the staff of the need for accuracy. All reports of errors now require attention by a department head, part of a concerted effort to correct even small mistakes that might have gone unacknowledged in the past.
As the staff has become more committed to fact-checking, he says, the paper has also grown more generous in its corrections policy, and he points to a recent example: A Nov. 11 clarification noting that a story should not have used Bechtel/Parsons as shorthand for Bechtel/Parsons Brinckerhoff. "I'm not sure we were obligated to do that," he says, since the shorter version is common parlance in Boston. "But we did it anyway."
Indeed, most of the corrections are, if not that minor, relatively routine, even if they sometimes pose inconvenience for readers -- as in the play that was listed to run through Sunday but had actually closed last week. Mixed in are some far more serious -- of a wrong "fact" that was essential to the story's premise; a main character misidentified; a criminal charge presumed but not filed.
A bit of context: While there were 1,031 known errors, there were at least 58,922 stories written in 2004.
The analysis of types of errors shows "misidentifications" led with 293, followed by "misstatements" at 166 and "misrepresentations" at 111. The most troubling category -- "misquotes" -- remained almost unchanged from last year at 11. There were 98 corrections of misspellings, although the paper does not attempt to correct all misspellings or grammatical errors.
The vast majority of mistakes were handled with a conventional correction on A2. A few dozen were addressed with notes of omission, clarification, or an editor's explanation.
The detailed annual analysis, prepared by Globe library chief Elisabeth Tuite and her staff, shows that nearly two thirds -- or 618 -- of the year's corrections were the result of reporting errors, up substantially from last year, while editing errors, 241, were down. Designers, researchers, photographers, sources, and production problems were responsible for the rest.
Most of the reporters who had any errors at all had only one; a dozen or so had between five and nine. One had more.
Some mistakes are truly unavoidable -- such as the time a writer took the spelling of an author's name from the cover of her publisher's book proof. Trouble was, the publisher had misspelled its own author's name. Thus, another correction was added to the annual tally, an exercise not without its ironies.
"The more you run, the worse it looks for you," says Baron, "although you are being more rigorous" -- and doing the right thing.
The ombudsman represents the readers. Her opinions and conclusions are her own. Phone 617-929-3020 or, to leave a message, 929-3022. Our e-mail address is firstname.lastname@example.org.