Estimation and use of nested merge ratios for audience measurement
Methods, apparatus, systems and articles of manufacture (e.g., physical storage media) to estimate and use nested merge ratios (e.g., nested Fréchet ratios) for audience measurement are disclosed herein. Example apparatus disclosed herein are to identify a first merged media event for which a first census audience value is to be determined, the first merged media event based on a union including a second merged media event. Disclosed example apparatus are also to determine whether the first merged media event satisfies a mutual exclusivity condition. Disclosed example apparatus are further to, in response to the first merged media event satisfying the mutual exclusivity condition, estimate the first census audience value for the first merged media event based on at least a merge ratio and a second census audience value previously estimated for the second merged media event.
1. An audience measurement system comprising a processor and a memory, the audience measurement system configured to perform a set of operations comprising:
receiving, from a plurality of meters via a network, metering data for a plurality of media events;
storing the metering data in a datastore managed by an audience measurement entity;
obtaining, from one or more service providers, return path data for the plurality of media events;
identifying a first merged media event for which a first deduplicated reach value is to be determined, the first merged media event based on a union including a second merged media event;
determining whether the first merged media event satisfies a mutual exclusivity condition;
in response to the first merged media event satisfying the mutual exclusivity condition, estimating the first deduplicated reach value for the first merged media event based on at least a merge ratio and a second deduplicated reach value previously estimated for the second merged media event, wherein the estimating is carried out in a manner that satisfies logical constraints on the first deduplicated reach value, and wherein the logical constraints are derived from a union inequality; and
outputting the first deduplicated reach value.
2. The audience measurement system of claim 1 , wherein the merge ratio is based on a first panel audience value obtained for the first merged media event and a plurality of panel audience values obtained for media events included in the union, the plurality of panel audience values including a second panel audience value obtained for the second merged media event.
3. The audience measurement system of claim 2 , wherein the merge ratio is further based on a universe estimate obtained for a panel associated with the first panel audience value and the plurality of panel audience values.
4. The audience measurement system of claim 3 , wherein the set of operations further comprises:
determining a lower bound and an upper bound for the first panel audience value based on the plurality of panel audience values and the universe estimate; and
determining the merge ratio based on the first panel audience value, the lower bound and the upper bound.
5. The audience measurement system of claim 4 , wherein the set of operations further comprises dividing a first difference between the first panel audience value and the lower bound by a second difference between the upper bound and the lower bound to determine the merge ratio.
6. The audience measurement system of claim 1 , wherein the union further includes a third individual media event, and the set of operations further comprises estimating the first deduplicated reach value for the first merged media event based on at least the merge ratio, the second deduplicated reach value previously estimated for the second merged media event, a third deduplicated reach value obtained for the third individual media event, and a universe estimate obtained for a population associated with the third deduplicated reach value.
7. The audience measurement system of claim 6 , wherein the set of operations further comprises:
determining a lower bound and an upper bound for the first deduplicated reach value based on the second deduplicated reach value previously estimated for the second merged media event, the third deduplicated reach value obtained for the third individual media event, and the universe estimate; and
adding the lower bound to a product of the merge ratio and a difference between the upper bound and the lower bound to estimate the first deduplicated reach value.
8. The audience measurement system of claim 1 , wherein the set of operations further comprises outputting at least one of an instruction or a signal in response to the first merged media event not satisfying the mutual exclusivity condition, the at least one of the instruction or the signal to cause the audience measurement system to at least one of halt or skip audience measurement processing associated with the first merged media event.
9. A non-transitory computer readable medium comprising computer readable instructions that, when executed, cause one or more processors to at least:
receive, from a plurality of meters via a network, metering data for a plurality of media events;
store the metering data in a datastore managed by an audience measurement entity;
obtain, from one or more service providers, return path data for the plurality of media events;
identify a first merged media event for which a first deduplicated reach is to be determined, the first merged media event based on a union including a second merged media event;
determine whether the first merged media event satisfies a mutual exclusivity condition;
in response to the first merged media event satisfying the mutual exclusivity condition, estimate the first deduplicated reach value for the first merged media event based on at least a merge ratio and a second deduplicated reach value previously estimated for the second merged media event, wherein the estimating is carried out in a manner that satisfies logical constraints on the first deduplicated reach value, and wherein the logical constraints are derived from a union inequality; and
output the first deduplicated reach value.
10. The non-transitory computer readable medium of claim 9 , wherein the merge ratio is based on a first panel audience value obtained for the first merged media event and a plurality of panel audience values obtained for media events included in the union, the plurality of panel audience values including a second panel audience value obtained for the second merged media event.
11. The non-transitory computer readable medium of claim 10 , wherein the merge ratio is further based on a universe estimate obtained for a panel associated with the first panel audience value and the plurality of panel audience values.
12. The non-transitory computer readable medium of claim 11 , wherein the instructions are to cause the one or more processors to:
determine a lower bound and an upper bound for the first panel audience value based on the plurality of panel audience values and the universe estimate; and
determine the merge ratio based on the first panel audience value, the lower bound and the upper bound.
13. The non-transitory computer readable medium of claim 12 , wherein the instructions are to cause the one or more processors to divide a first difference between the first panel audience value and the lower bound by a second difference between the upper bound and the lower bound to determine the merge ratio.
14. The non-transitory computer readable medium of claim 9 , wherein the union further includes a third individual media event, and the instructions are to cause the one or more processors to estimate the first deduplicated reach value for the first merged media event based on at least the merge ratio, the second deduplicated reach value previously estimated for the second merged media event, a third deduplicated reach value obtained for the third individual media event, and a universe estimate obtained for a population associated with the third deduplicated reach value.
15. The non-transitory computer readable medium of claim 14 , wherein the instructions are to cause the one or more processors to:
determine a lower bound and an upper bound for the first deduplicated reach value based on the second deduplicated reach value previously estimated for the second merged media event, the third deduplicated reach value obtained for the third individual media event, and the universe estimate; and
add the lower bound to a product of the merge ratio and a difference between the upper bound and the lower bound to estimate the first deduplicated reach value.
16. The non-transitory computer readable medium of claim 9 , wherein the instructions include first instructions that are to cause the one or more processors to output at least one of a second instruction or a signal to an audience measurement system in response to the first merged media event not satisfying the mutual exclusivity condition, the at least one of the second instruction or the signal to cause the audience measurement system to at least one of halt or skip audience measurement processing associated with the first merged media event.
17. A method comprising:
receiving, from a plurality of meters via a network, metering data for a plurality of media events;
storing the metering data in a datastore managed by an audience measurement entity;
obtaining, from one or more service providers, return path data for the plurality of media events;
identifying a first merged media event for which a first deduplicated reach value is to be determined, the first merged media event based on a union including a second merged media event;
determining, by executing an instruction with at least one processor, whether the first merged media event satisfies a mutual exclusivity condition; and
in response to the first merged media event satisfying the mutual exclusivity condition, estimating, by executing an instruction with the at least one processor, the first deduplicated reach value for the first merged media event based on at least a merge ratio and a second deduplicated reach value previously estimated for the second merged media event, wherein the estimating is carried out in a manner that satisfies logical constraints on the first deduplicated reach value, and wherein the logical constraints are derived from a union inequality; and
outputting the first deduplicated reach value.
18. The method of claim 17 , wherein the merge ratio is based on a first panel audience value obtained for the first merged media event and a plurality of panel audience values obtained for media events included in the union, the plurality of panel audience values including a second panel audience value obtained for the second merged media event.
19. The method of claim 18 , wherein the merge ratio is further based on a universe estimate obtained for a panel associated with the first panel audience value and the plurality of panel audience values.
20. The method of claim 19 , further including:
determining a lower bound and an upper bound for the first panel audience value based on the plurality of panel audience values and the universe estimate; and
determining the merge ratio based on the first panel audience value, the lower bound and the upper bound.