Re: PPM vs. Diary
>
> How does Arbitron get the diaries to people?
Arbitron uses SSI listings for each MSA, using a random interval sort in proportion to the number of numbers needed. In addition, they do a percentage of RDD (random digit dial) calls to get unlisted phones int he sample. They will do the same eventually with cellular exchanges.
If an answer is obtained, the person is asked to participate. Certain demographic data is obtained, to make sure the quotas are not already filled on those cells. Diaries are mailed, with cash incentive, one per family member.
A follow up call verifies receipt and encourages participation. Another follow up is made after the weekend, Another on the last day, and another to insure return of the complete diary.
>
> How do the PPMs get to people?
They are placed in homes like Nielsen boxes, with an Arbitron rep installing them.
Recruit is different. The PPM is a panel, while the diary method is a proportionally balanced poll. Arbitron recruits in advance for PPM, and if anyone drops from the panel, they are replaced by an exact equivalent. The panel at all times is totally proportional on all stratification varibles, since the panel members can remain active up to 1 year, not just one week as with the diary. A market that has 30,000 completed diaries a year in 4 surveys will only have about 2500 metered persons... but they keep the meter up to a year, or till they are expelled or drop out.
>
> If, as I suspect, Arbitron is contacting people randomly by
> phone, then this does support my statement that the system
> is inherently flawed in todays technological world and will
> never accurately reflect the true listening preferences of
> the Metroplex, much less the rest of the country.
Parallel samples using street intercepts, etc. verify that there is not a significant difference. Even unlisted phone holder behaviour is similar to listed phone behaviour. Remember, this is a sample, and you do not have to get everybody in it.
>
> If the PPMs are doled out based upon random phone calls,
> then I will state that the same flaw that causes Arbitron to
> be inaccurate will be inherent in the distribution of PPMs.
The phone, including cellulars and unlisted, is the most accurate method of doing polling and research know, short of a cnesus (which is not a poll, anyhow)
>
> Also, how sensitive is the PPM? Will it record that I was
> listening to KKDA at 5:19pm yesterday, when in fact it was
> the car next to me blaring it's radio?
If you can hear it loud enough to understand, the PPM will register. However, a 5 minute span of detections is necessary to get credit in that quarter hour. So walking by someone's cubicle or the example you make will nto produce a credit for any station.
>
> I am serious in these questions, and really enjoy debating
> this to death with you, David.
>
Cool. I have followed the PPM since it was a big (big, as in the size of an old Motorola Star Tac phone) black box in a lab ont he third floor of Arbitron in Columbia... maybe 10 years ago. I also follow its implementation in Canada at the BBM, and have lots of comparitive experience in Mexico City where I consult and there is Arbitron with diaries, IBOPE with in-home comincidentals and INRA with a mixed intercept and coincidental method.
>
> How does Arbitron get the diaries to people?
Arbitron uses SSI listings for each MSA, using a random interval sort in proportion to the number of numbers needed. In addition, they do a percentage of RDD (random digit dial) calls to get unlisted phones int he sample. They will do the same eventually with cellular exchanges.
If an answer is obtained, the person is asked to participate. Certain demographic data is obtained, to make sure the quotas are not already filled on those cells. Diaries are mailed, with cash incentive, one per family member.
A follow up call verifies receipt and encourages participation. Another follow up is made after the weekend, Another on the last day, and another to insure return of the complete diary.
>
> How do the PPMs get to people?
They are placed in homes like Nielsen boxes, with an Arbitron rep installing them.
Recruit is different. The PPM is a panel, while the diary method is a proportionally balanced poll. Arbitron recruits in advance for PPM, and if anyone drops from the panel, they are replaced by an exact equivalent. The panel at all times is totally proportional on all stratification varibles, since the panel members can remain active up to 1 year, not just one week as with the diary. A market that has 30,000 completed diaries a year in 4 surveys will only have about 2500 metered persons... but they keep the meter up to a year, or till they are expelled or drop out.
>
> If, as I suspect, Arbitron is contacting people randomly by
> phone, then this does support my statement that the system
> is inherently flawed in todays technological world and will
> never accurately reflect the true listening preferences of
> the Metroplex, much less the rest of the country.
Parallel samples using street intercepts, etc. verify that there is not a significant difference. Even unlisted phone holder behaviour is similar to listed phone behaviour. Remember, this is a sample, and you do not have to get everybody in it.
>
> If the PPMs are doled out based upon random phone calls,
> then I will state that the same flaw that causes Arbitron to
> be inaccurate will be inherent in the distribution of PPMs.
The phone, including cellulars and unlisted, is the most accurate method of doing polling and research know, short of a cnesus (which is not a poll, anyhow)
>
> Also, how sensitive is the PPM? Will it record that I was
> listening to KKDA at 5:19pm yesterday, when in fact it was
> the car next to me blaring it's radio?
If you can hear it loud enough to understand, the PPM will register. However, a 5 minute span of detections is necessary to get credit in that quarter hour. So walking by someone's cubicle or the example you make will nto produce a credit for any station.
>
> I am serious in these questions, and really enjoy debating
> this to death with you, David.
>
Cool. I have followed the PPM since it was a big (big, as in the size of an old Motorola Star Tac phone) black box in a lab ont he third floor of Arbitron in Columbia... maybe 10 years ago. I also follow its implementation in Canada at the BBM, and have lots of comparitive experience in Mexico City where I consult and there is Arbitron with diaries, IBOPE with in-home comincidentals and INRA with a mixed intercept and coincidental method.