Methods and apparatus to determine synthetic respondent level data
Methods, apparatus, systems, and articles of manufacture are disclosed to generate synthetic respondent level data. An example apparatus includes: a comparator to compare a target rating to a computed rating to determine a comparison result, the computed rating determined based on a seed panel, the seed panel including monitored panelists associated with return path data; a seed panelist data adjuster to adjust the seed panel based on the comparison result to reduce an error between the target rating and the computed rating; and a household data adjuster to add tuning without viewing data to households of the adjusted seed panel, the tuning without viewing data for a first one of the households to represent monitored data corresponding to a media presentation device of the first one of the households being on while a media output device in communication with the media presentation device is off.
1. An apparatus comprising:
a comparator to compare a target rating to a computed rating to determine a comparison result, the computed rating determined based on a seed panel, the seed panel including monitored panelists associated with return path data;
a seed panelist data adjuster to adjust the seed panel based on the comparison result to reduce an error between the target rating and the computed rating; and
a household data adjuster to add tuning without viewing data to households of the adjusted seed panel, the tuning without viewing data for a first one of the households to represent monitored data corresponding to a media presentation device of the first one of the households being on while a media output device in communication with the media presentation device is off.
2. The apparatus of claim 1 , wherein the household data adjuster is to generate the tuning without viewing data by, when the tuning without viewing data is below a threshold:
selecting a first household of a first panelist in the adjusted seed panel that (i) affects a relative error between when viewing of the first household is modified and when the viewing of the first household is not modified and (ii) affects the relative error less than a second household, and the first household included in synthetic respondent level-data; and
removing the first household from the synthetic respondent level-data to increase the tuning without viewing data.
3. The apparatus of claim 1 , wherein the seed panel data adjuster is to adjust the seed panel by:
selecting a panelist from seed panelists in the seed panel that affects the error between the target rating and the computed rating; and
adjusting a weight for the selected panelist.
4. The apparatus of claim 1 , wherein:
the target rating is determined based on the return path data and corresponds to a first percentage of persons of a first demographic exposed to first media at a first time; and
the computed rating corresponds to a second percentage of seed panelists exposed to the first media at the first time.
5. The apparatus of claim 1 , further including a data assigner to select the monitored panelists to be in the seed panel based on demographics associated with a population associated with the return path data.
6. The apparatus of claim 5 , wherein the data assigner is to select a first group of the monitored panelists to be seed panelists based on a comparison of the monitored panelists and the population associated with the return path data, and further including a weighter to weight the seed panelists based on a universe estimate of media users, the universe estimate including the monitored panelists and the population.
7. The apparatus of claim 1 , further including a seed panelist replicator to at least one of replicate or down-sample seed panelists.
8. The apparatus of claim 1 , wherein the seed panel data adjuster is to adjust the seed panel based on a viewership error corresponding to the seed panel, the viewership error based on a constraint corresponding to the target rating.
9. The apparatus of claim 8 , wherein the seed panelist data adjuster is to adjust the seed panel based on the viewership error corresponding to the seed panel by:
when the viewership error does not satisfy a first threshold:
selecting a first panelist from seed panelists in the seed panel that (i) affects relative error between when viewing of the first panelist is modified and when the viewing of the first panelist is not modified and (ii) affects the relative error less than a second panelist; and
removing the first panelist from an audience corresponding to the target rating; and
when the viewership error does not satisfy a second threshold different than the first threshold:
selecting a third panelist from the seed panelists in the seed panel that (i) affects the relative error between when viewing of the second panelist is modified and when viewing of the second panelist is not modified and (ii) affects the relative error less than a fourth panelist; and
adding the first panelist from the audience corresponding to the target rating.
10. An apparatus comprising:
memory including computer readable instructions; and
a processor to execute the instructions to at least:
compare a target rating to a computed rating to determine a comparison result, the computed rating determined based on a seed panel, the seed panel including monitored panelists associated with return path data;
adjust the seed panel based on the comparison result to reduce an error between the target rating and the computed rating; and
add tuning without viewing data to households of the adjusted seed panel, the tuning without viewing data for a first one of the households to represent monitored data corresponding to a media presentation device of the first one of the households being on while a media output device in communication with the media presentation device is off.
11. The apparatus of claim 10 , wherein the processor to generate the tuning without viewing data by, when the tuning without viewing data is below a threshold:
selecting a first household of a first panelist in the adjusted seed panel that (i) affects a relative error between when viewing of the first household is modified and when the viewing of the first household is not modified and (ii) affects the relative error less than a second household, and the first household included in synthetic respondent level-data; and
removing the first household from the synthetic respondent level-data to increase the tuning without viewing data.
12. The apparatus of claim 10 , wherein the processor is to adjust the seed panel by:
selecting a panelist from seed panelists in the seed panel that affects the error between the target rating and the computed rating; and
adjusting a weight for the selected panelist.
13. The apparatus of claim 10 , wherein:
the target rating is determined based on the return path data and corresponds to a first percentage of persons of a first demographic exposed to first media at a first time; and
the computed rating corresponds to a second percentage of seed panelists exposed to the first media at the first time.
14. An apparatus comprising:
means for comparing a target rating to a computed rating to determine a comparison result, the computed rating determined based on a seed panel, the seed panel including monitored panelists associated with return path data;
means for adjusting the seed panel based on the comparison result to reduce an error between the target rating and the computed rating; and
means for adding tuning without viewing data to households of the adjusted seed panel, the tuning without viewing data for a first one of the households to represent monitored data corresponding to a media presentation device of the first one of the households being on while a media output device in communication with the media presentation device is off.
15. The apparatus of claim 14 , wherein the adding means is to generate the tuning without viewing data by, when the tuning without viewing data is below a threshold:
selecting a first household of a first panelist in the adjusted seed panel that (i) affects a relative error between when viewing of the first household is modified and when the viewing of the first household is not modified and (ii) affects the relative error less than a second household, and the first household included in synthetic respondent level-data; and
removing the first household from the synthetic respondent level-data to increase the tuning without viewing data.
16. The apparatus of claim 14 , wherein the means for adjusting is to adjust the seed panel by:
selecting a panelist from seed panelists in the seed panel that affects the error between the target rating and the computed rating; and
adjusting a weight for the selected panelist.
17. The apparatus of claim 14 , wherein:
the target rating is determined based on the return path data and corresponds to a first percentage of persons of a first demographic exposed to first media at a first time; and
the computed rating corresponds to a second percentage of seed panelists exposed to the first media at the first time.
18. A tangible computer readable storage medium comprising instructions which, when executed cause one or more processors to at least:
compare a target rating to a computed rating to determine a comparison result, the computed rating determined based on a seed panel, the seed panel including monitored panelists associated with return path data;
adjust the seed panel based on the comparison result to reduce an error between the target rating and the computed rating; and
add tuning without viewing data to households of the adjusted seed panel, the tuning without viewing data for a first one of the households to represent monitored data corresponding to a media presentation device of the first one of the households being on while a media output device in communication with the media presentation device is off.
19. The computer readable storage medium of claim 18 , wherein the instructions cause the one or more processors to generate the tuning without viewing data by, when the tuning without viewing data is below a threshold:
selecting a first household of a first panelist in the adjusted seed panel that (i) affects a relative error between when viewing of the first household is modified and when the viewing of the first household is not modified and (ii) affects the relative error less than a second household, and the first household included in synthetic respondent level-data; and
removing the first household from the synthetic respondent level-data to increase the tuning without viewing data.
20. The computer readable storage medium of claim 18 , wherein the instructions cause the one or more processors to adjust the seed panel by:
selecting a panelist from seed panelists in the seed panel that affects the error between the target rating and the computed rating; and
adjusting a weight for the selected panelist.
21. The computer readable storage medium of claim 18 , wherein:
the target rating is determined based on the return path data and corresponds to a first percentage of persons of a first demographic exposed to first media at a first time; and
the computed rating corresponds to a second percentage of seed panelists exposed to the first media at the first time.
22. The computer readable storage medium of claim 18 , wherein the instructions cause the one or more processors to select the monitored panelists to be in the seed panel based on demographics associated with a population associated with the return path data.
23. The computer readable storage medium of claim 22 , wherein the instructions cause the one or more processors to:
select a first group of the monitored panelists to be seed panelists based on a comparison of the monitored panelists and the population associated with the return path data; and
weight the seed panelists based on a universe estimate of media users, the universe estimate including the monitored panelists and the population.
24. The computer readable storage medium of claim 18 , wherein the instructions cause the one or more processors to at least one of replicate or down-sample seed panelists.
25. The computer readable storage medium of claim 18 , wherein the instructions cause the one or more processors to adjust the seed panel based on a viewership error corresponding to the seed panel, the viewership error based on a constraint corresponding to the target rating.
26. The computer readable storage medium of claim 25 , wherein the instructions cause the one or more processors to adjust the seed panel based on the viewership error corresponding to the seed panel by:
when the viewership error does not satisfy a first threshold:
selecting a first panelist from seed panelists in the seed panel that (i) affects relative error between when viewing of the first panelist is modified and when the viewing of the first panelist is not modified and (ii) affects the relative error less than a second panelist; and
removing the first panelist from an audience corresponding to the target rating; and
when the viewership error does not satisfy a second threshold different than the first threshold:
selecting a third panelist from the seed panelists in the seed panel that (i) affects the relative error between when viewing of the second panelist is modified and when viewing of the second panelist is not modified and (ii) affects the relative error less than a fourth panelist; and
adding the first panelist from the audience corresponding to the target rating.