Arbitron: Sample size in PPM much bigger than diary

Since sample size issues are becoming such a big deal right now in the Houston and Philadelphia PPM markets, we followed up on a story idea from Kathy Crawford, MindShare President/Local Broadcast: "Tell us about sample size in PPM vs. diary."

We went to Arbitron spokesperson Thom Mocarsky, who noted how much bigger the sample size in PPM really is vs. diary, and supplied charts for Houston and Philly to show the data.

In the diary world, there's a quarterly report that's the average of 12 1-week measurements. So somebody in the first week of January says what they're listening to; another person in the second week of January tells Arbitron what they've listened to, all the way to the end of March. Mocarsky notes that in diaries, "January listening gets diluted by February listening, which gets dilutes by March listening. So in terms of any given month, it's tough to know what's going on."

Even Arbitrends is a rolling average of the previous three months. So while dairy  the sample size is "larger" than PPM for the time period being measured, for each month, it is actually equivalent or smaller. Note line 5 on the Excel chart for October 2007-P12+ is much higher for the PPM sample than diary for one month (quarterly number divided by three).

The PPM Average Daily column is the average number of people that contributed to the report on any given day-note in Houston diary was 1/12th the sample size of PPM: 1220 vs. 342.

"That's what makes PPM so much more valuable," says Mocarsky. "First off, for October there is a lot of confidence there because it's got more sample. You get more reliability, you get more trendability-stuff doesn't bounce around as much. It gets delivered much more quickly and the time period is much more discrete. In other words, I know what happened in the month of October, not the average of the previous three months. A lot can happen from January to March."

Add to: del.icio.us | Digg