NAA Revamps Newsprint Statistics

By: Jim Rosenberg Newspaper Association of America reporting of U.S. dailies' newsprint consumption, receipt, and inventory now relies on a new model in arriving at figures for the year 2002 and for at least the opening months of this year. NAA advises that results will not be "strictly comparable to prior years."

NAA Newsprint Analyst Nick Nocella said such updates to the procedure will "happen more often," at least every two years. Until the change, the NAA used a method adopted in 1991, when, with a smaller staff, it shrunk its survey to 160 members from 575. Nocella said the larger survey and an end-of-year census of all members "was becoming very expensive."

Software used by the association now crunches numbers from a more recent 1998-circulation database to derive figures for sampled papers and extrapolate for all U.S. dailies. Data is collected from a "stratified random sample of newspapers, representing more than 50% of [U.S.] daily circulation," according to the NAA's "Statistical Program Profile." The program calculates consumption-to-circulation ratios for the reporting papers, then uses them to estimate the statistics for all other papers, based on their circulation sizes.

Dailies are divided into Northeast, South, North Central, and West regions and into 11 circulation-based strata that begin in 5,000-copy increments, range to 150,000-copy increments, and end at over-500,000 copies. The program profile states that data analysis has shown the number of Sundays in a month and changes in basis weight or web width "are not statistically related to consumption."

Data-gathering and statistical techniques developed by Francis Ponti, a statistician at George Washington University who is under contract with the NAA, were approved by the North American Newsprint Producers Association (NANPA) and the Pulp and Paper Products Council. The latter, formerly a part of the Canadian Pulp and Paper Assocation, also provides statistics to NANPA and shares with NAA the cost of the analyses.

In another change, Puerto Rico (in the database since 1991) was dropped from the survey. The Pulp and Paper Products Council was unaware of Puerto Rico's inclusion until now, and had not been reporting it in its U.S. figures, according to Nocella. "Our new database does not include Puerto Rico," he said. Exclusion of the island commonwealth will affect figures reported for the South region.

Whereas regional data precision is within +/-10%, national totals' precision is within +/-2%, according to the NAA. The program profile states that the sample was designed to yield estimates with a 95% confidence level.

Under the updated system, U.S. dailies consumed 695,000 metric tons of newsprint in December and 8,128 million tons in all of 2002. At 676,000 tons, preliminary January figures are 1.7% higher than restated figures for January 2002. Preliminary numbers also show publishers' 926,000-ton January inventories 5.8% higher than restated year-earlier figures.