Common use of Oversampling Clause in Contracts

Oversampling. Oversampling was employed for some subgroups of interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "oversampled" are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "not oversampled". As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "cost" of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the NHIS, Hispanic households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. The oversampling approach for both Panels 10 and 11 was the same. From among the NHIS households eligible for MEPS, reflecting the oversampling of Hispanics and blacks described above, three strata were constructed for sampling purposes for MEPS. One stratum contained households with Asians and those "predicted to be poor", a second contained households with black members that were not among the households in the first stratum, while the third stratum contained all remaining households. All households in the "Asian/Predicted Poor" stratum were selected with certainty. The sampling rate for the black stratum was three-fourths. The sampling rate for the “other” stratum was about 53 percent for Panel 10 and 56 percent for Panel 11. Within strata, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled for the NHIS, as described above, households with Hispanics were also included at disproportionately high rates (oversampled) among the households selected at the roughly 50% rate. Thus, for MEPS, households that were oversampled were those containing Hispanics, Blacks, Asians, and those predicted to be poor.

Oversampling. Oversampling ▇▇▇▇▇▇▇▇▇▇▇▇ was employed for some selected subgroups of policy-level interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be is discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "“oversampled" ” are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "“not oversampled"”. As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "“cost" ” of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve achieved if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the earlier sample design of the NHIS, Hispanic households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. For the new NHIS sample design Asians are also oversampled. The estimated overall oversampling rates associated with each of the three minorities have not yet been reported. The oversampling approach approaches and the sampling domains used for both MEPS Panels 10 13 and 11 was 14 for subsampling among the sameNHIS respondents eligible for MEPS were slightly different. From among the NHIS households eligible for MEPSthe Panel 13, reflecting the oversampling of Hispanics and blacks described above, three four domains (strata) were established in a hierarchical sequence but in essence there were only two sampling strata were constructed for sampling purposes for MEPSemployed. One The first stratum contained households with Asians and those "“predicted to be poor"”, a second stratum contained households with Hispanics not assigned to the first stratum, a third stratum contained households with black members that were in households not among the households in assigned to the first stratumtwo strata, while the third fourth stratum contained all remaining households. All households in the "strata of “Asian/Predicted Poor" stratum ”, Hispanics, and Blacks were selected sampled with certainty. The sampling rate for the black stratum was three-fourths. The sampling rate for the “other” stratum was about 53 57 percent. Similar to Panel 13, there were also four strata established for Panel 14, with the only difference that the first stratum only contained households with Asians. Those households “predicted to be poor” did not constitute a separate sampling domain for Panel 14. All households in the Asian stratum were sampled with certainty. The sampling rate was about 87 percent for Panel 10 and 56 the Hispanics stratum, about 90 percent for Panel 11the Blacks stratum, and about 72 percent for the “other” stratum. Within stratastrata for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled for With the NHIS, as described above, households with Hispanics were also included at disproportionately high rates (oversampled) among the households selected at the roughly 50% rate. Thus, for MEPSsubsampling, households that were oversampled for calendar year 2009 were those the households containing Hispanics, Blacks, and Asians, and those predicted to be poorbased on their NHIS membership.

Oversampling. Oversampling was employed for some subgroups of interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "oversampled" are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "not oversampled". As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "cost" of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the NHIS, NHIS Hispanic households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. The oversampling approach for both differed somewhat between Panels 10 6 and 11 was the same7 as, beginning with Panel 7, MEPS has begun to oversample Asians and those predicted to be poor. From among the NHIS households eligible for MEPS, reflecting the oversampling of Hispanics and blacks described above, three two strata were constructed for sampling purposes for MEPSpurposes. One stratum contained households with Asians and those "predicted to be poor", a second contained households with black members that were not among the households in the first stratum, " while the third other stratum contained all remaining households. All households in the "Asian/Predicted Poor" stratum were selected with certainty. The sampling rate for certainty while roughly 50% of the black other stratum was three-fourths. The sampling rate for the “other” stratum was about 53 percent for Panel 10 and 56 percent for Panel 11. Within strata, responding NHIS households were selected for MEPS MEPS, using a systematic sample selection procedure from among those eligible. Because Hispanics and blacks had been oversampled for the NHIS, NHIS as described above, households with Hispanics and Blacks were also included at disproportionately high rates (oversampled) among the households selected at the roughly 50% rate. Thus, for MEPS, households that were oversampled were those containing Hispanics, Blacks, Asians, and those predicted to be poor.

Oversampling. Oversampling ▇▇▇▇▇▇▇▇▇▇▇▇ was employed for some selected subgroups of policy-level interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be discussedis discussed below. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "oversampled" are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "not oversampled". As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "cost" of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the earlier sample design of the NHIS, Hispanic households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. For the new NHIS sample design Asians are also oversampled. The oversampling rates of the three minorities have not yet been reported. The oversampling approach used for both MEPS among the NHIS respondents eligible for MEPS subsampling differed substantially between Panels 10 11 and 11 was 12 although the samesame groups were targeted. From among the NHIS households eligible for the Panel 11 component of MEPS, reflecting the oversampling of Hispanics and blacks described above, three strata were constructed for sampling purposes for MEPS. One stratum contained households with Asians and those "predicted to be poor", a second contained households with black members that were not among the households in the first stratum, while the third stratum contained all remaining households. All households in the "Asian/Predicted Poor" stratum were selected with certainty. The sampling rate for the black stratum was three-fourths. The sampling rate for the “other” stratum was about 53 percent 56 percent. From among the NHIS households eligible for the Panel 10 12, four strata were established in a hierarchical sequence but in essence there were only two sampling strata employed. The first stratum contained households with Asians and 56 percent those "predicted to be poor", a second stratum contained households with Hispanics not assigned to the first stratum, a third stratum contained households with black members in households not assigned to the first two strata, while the fourth stratum contained all remaining households. All households in the "Asian/Predicted Poor" stratum were selected with certainty while the sampling rates for the remaining three strata were essentially the same, roughly 90 percent. Panel 1112 was the first year in which Hispanics formed a separate sample domain for MEPS. Within stratastrata for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled for With the NHIS, as described above, households with Hispanics were also included at disproportionately high rates (oversampled) among the households selected at the roughly 50% rate. Thus, for MEPSsubsampling, households that were oversampled for MEPS were those the households containing Hispanics, Blacks, Asians, and those predicted to be poorpoor based on their NHIS membership.

Oversampling. Oversampling ▇▇▇▇▇▇▇▇▇▇▇▇ was employed for some selected subgroups of policy-level interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be is discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "“oversampled" ” are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "“not oversampled"”. As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "“cost" ” of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve achieved if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the earlier sample design of the NHIS, Hispanic households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. For the new NHIS sample design Asians are also oversampled. The estimated overall oversampling rates associated with each of the three minorities have not yet been reported. The oversampling approach approaches and the sampling domains used for both MEPS Panels 10 14 and 11 was the same. From 15 for subsampling among the NHIS households respondents eligible for MEPS, reflecting the oversampling of Hispanics and blacks described above, three strata MEPS were constructed for similar. The same sample domains were employed but with slightly different sampling purposes for MEPSrates. One Four domains (strata) were established in a hierarchical sequence. The first stratum contained households with Asians and those "predicted to be poor"Asians, a second stratum contained households with black members that were Hispanics not among the households in assigned to the first stratum, a third stratum contained households with Black members not assigned to the first two strata, while the third fourth stratum contained all remaining households. All For Panel 14, all households in the "Asian/Predicted Poor" Asian stratum were selected sampled with certainty. The sampling rate was about 87 percent for the black stratum was three-fourths. The sampling rate Hispanic stratum, about 90 percent for the Black stratum, and about 72 percent for the “other” stratum. For Panel 15 again all households in the Asian stratum were sampled with certainty. The sampling rate was about 53 76 percent for Panel 10 and 56 the Hispanic stratum, about 86 percent for Panel 11the Black stratum, and about 61 percent for the “other” stratum. Within stratastrata for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled for With the NHIS, as described above, households with Hispanics were also included at disproportionately high rates (oversampled) among the households selected at the roughly 50% rate. Thus, for MEPSsubsampling, households that were oversampled for calendar year 2010 were those the households containing Hispanics, Blacks, and Asians, and those predicted to be poorbased on their NHIS membership.

Oversampling. Oversampling ▇▇▇▇▇▇▇▇▇▇▇▇ was employed for some selected subgroups of policy-level interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be is discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "“oversampled" ” are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "“not oversampled"”. As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "“cost" ” of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve achieved if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the earlier sample design of the NHIS, Hispanic households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. For the new NHIS sample design Asians are also oversampled. The estimated overall oversampling rates associated with each of the three minorities have not yet been reported. The oversampling approach approaches and the sampling domains used for both MEPS Panels 10 15 and 11 was the same. From 16 for subsampling among the NHIS households respondents eligible for MEPS, reflecting the oversampling of Hispanics and blacks described above, three strata MEPS were constructed for sampling purposes for MEPSsomewhat different. One For Panel 15 four domains (strata) were established in a hierarchical sequence. The first stratum contained households with Asians and those "predicted to be poor"Asians, a second stratum contained households with black members that were Hispanics not among the households in assigned to the first stratum, while a third stratum contained households with Black members not assigned to the third first two strata, and the fourth stratum contained all remaining households. All For Panel 16 the corresponding hierarchical ordering of the domains for sampling among NHIS responding households was: households where an individual was identified in the NHIS as having cancer via the NHIS sampled adult questionnaire; Asian; Hispanic; Black; Other households characterized as “complete” for the NHIS; and Other households characterized as “partial complete” for the NHIS. For Panel 15, all households in the "Asian/Predicted Poor" stratum Asian domain were selected sampled with certainty. The sampling rate was about 76 percent for the black stratum was three-fourths. The sampling rate Hispanic domain, about 86 percent for the Black domain, and about 61 percent for the “otherOther” stratum domain. For Panel 16 the corresponding sampling rates for all the domains except those characterized as “Other” were all 1 (i.e., all households assigned to those domains were included in MEPS). The “Other, complete” domain was sampled at a rate of about 53 79 percent for Panel 10 and 56 percent for Panel 11while the “Other, partial complete” domain was sampled at a rate of about 46 percent. Within stratastrata (domains) for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled For all but those strata involving “Others”, the selection was with equal probability. Within strata involving “Others” (for Panel 15, one stratum; for Panel 16, two strata) the NHIS, as described above, households selection was with Hispanics were also included at disproportionately high rates probability proportionate to size (oversampledpps) among where the households selected at size measure was the roughly 50% rateinverse of the NHIS initial probability of selection. Thus, for MEPSThe pps sampling was undertaken to help reduce the variability in the MEPS weights incurred due to the variability of the NHIS sampling rates. With the subsampling, households that were oversampled for MEPS in calendar year 2011 were those containing Hispanicsresponding households in the NHIS identified as having members with cancer or whose race/ethnicity was Hispanic, BlacksBlack, Asiansor Asian. Again, note that not all NHIS households where a member had cancer were identified as such - the member with cancer had to have been randomly selected to complete the NHIS “sampled adult” questionnaire and those predicted to be poorhave self-identified as having cancer in response to questions from that questionnaire.

Oversampling. Oversampling was employed for some subgroups of interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "oversampled" are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "not oversampled". As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "cost" of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the NHIS, NHIS Hispanic households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. The oversampling approach differed slightly between Panels 8 and 9. For both panels, NHIS responding households containing Asians and those predicted to be poor were oversampled for both Panels 10 and 11 was the sameMEPS. This practice began with MEPS Panel 8. However, in addition, for Panel 9, NHIS responding households with black members were sampled at higher rates than all other except those with Asian members or those predicted to be poor. From among the NHIS households eligible for MEPS, reflecting the oversampling of Hispanics and blacks described above, three two strata were constructed for sampling purposes for MEPSpurposes. One stratum contained households with Asians and those "predicted to be poor", a second contained households with black members that were not among the households in the first stratum, " while the third other stratum contained all remaining households. All households in the "Asian/Predicted Poor" stratum were selected with certainty. The sampling rate for certainty while roughly two thirds of the black “black” stratum was three-fourths. The sampling rate for and 50% of the “other” stratum was about 53 percent for Panel 10 and 56 percent for Panel 11. Within strata, responding NHIS households were selected for MEPS MEPS, using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled for the NHIS, NHIS as described above, households with Hispanics were also included at disproportionately high rates (oversampled) among the households selected at the roughly 50% rate. Thus, for MEPS, households that were oversampled were those containing Hispanics, Blacks, Asians, and those predicted to be poor.

Oversampling. Oversampling ▇▇▇▇▇▇▇▇▇▇▇▇ was employed for some selected subgroups of policy-level interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be is discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "“oversampled" ” are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "“not oversampled"”. As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "“cost" ” of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve achieved if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the earlier sample design of the NHIS, Hispanic households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. For the new NHIS sample design Asians are also oversampled. The estimated overall oversampling rates associated with each of the three minorities have not yet been reported. The oversampling approach approaches used for MEPS Panels 12 and 13 for subsampling among the NHIS respondents eligible for MEPS was somewhat similar. The sampling domains are the same for both Panels 10 panels and 11 was the samesubsampling rates are all high except for the “Other” domain for Panel 13. From among the NHIS households eligible for MEPSthe Panel 12, reflecting the oversampling of Hispanics and blacks described above, three four domains (strata) were established in a hierarchical sequence but in essence there were only two sampling strata were constructed for sampling purposes for MEPSemployed. One The first stratum contained households with Asians and those "“predicted to be poor"”, a second stratum contained households with Hispanics not assigned to the first stratum, a third stratum contained households with black members that were in households not among the households in assigned to the first stratumtwo strata, while the third fourth stratum contained all remaining households. All households in the "“Asian/Predicted Poor" ” stratum were selected with certainty. The certainty while the sampling rate rates for the black stratum remaining three strata were essentially the same, roughly 90 percent. Panel 12 was three-fourthsthe first year in which Hispanics formed a separate sample domain for MEPS. For Panel 13 there were also in essence only two strata as the domains associated with Asians, the “predicted to be poor”, Hispanics, and Blacks all were sampled with certainty. The sampling rate for the “other” stratum was about 53 percent for Panel 10 and 56 percent for Panel 1157 percent. Within stratastrata for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled for With the NHIS, as described above, households with Hispanics were also included at disproportionately high rates (oversampled) among the households selected at the roughly 50% rate. Thus, for MEPSsubsampling, households that were oversampled for calendar year 2008 were those the households containing Hispanics, Blacks, Asians, and those predicted to be poorpoor based on their NHIS membership.

Oversampling. Oversampling was employed for some subgroups is a feature of interest the MEPS sample design, helping to help increase the precision of estimates associated with members for some subgroups of those subgroupsinterest. Before going into detailsdetails related to MEPS, the concept of oversampling will be discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for specific subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "oversampled" are represented at disproportionately high rates an oversampled subgroup comprises a higher proportion of the sample than it represents in the samplegeneral population. Sample weights help ensure that population estimates account for this are not distorted by a disproportionate contribution from oversampled subgroups, as the base . Base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "not oversampled". As mentioned above, oversampling a subgroup is done implemented to increase the sample sizes and thus improve the precision of survey estimates for that particular subgroupsubgroups of the population. The "“cost" ” of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve achieved if the same overall sample size were selected without any oversampling. For The oversampling of Hispanic, Black, and Asian households for the NHIS carries over to MEPS through the set of NHIS responding households eligible for sample selection for MEPS, some of . In the oversampling was achieved NHIS under the old sample design utilized through its linkage to the NHIS. For the NHIS2005, Hispanic households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black Black households under the old design was roughly 1.5 to 1. The oversampling approach for both Panels 10 and 11 was the same. From among Under the NHIS households eligible sample design employed through 2015 (which is the sample design applicable for MEPSMEPS Panels 20 and 21), reflecting the oversampling of Asians, as well as Hispanics and blacks described aboveBlacks, are oversampled. The average oversampling rates for the three strata were constructed for sampling purposes for MEPSminority groups have not yet been reported. One stratum contained households with Asians For both Panel 20 and those "predicted to be poor"Panel 21, a second contained households with black members that were not among the all households in the first stratumAsian, Hispanic, and Black domains were sampled with certainty (i.e., all households assigned to those domains were included in the MEPS). For Panel 20, the “Other, complete” domain was sampled at a rate of about 84 percent while the third stratum contained all remaining households“Other, partial complete” domain was sampled at a rate of about 53 percent. All households in For Panel 21, the "Asian/Predicted Poor" stratum were selected with certainty. The corresponding sampling rate for the black stratum was three-fourths. The sampling rate rates for the “otherOther, complete” stratum was domain and the “Other, partial complete” domain were about 53 81 percent for Panel 10 and 56 percent for Panel 1149 percent, respectively. Within stratastrata (domains) for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled For the “non-Other” strata households were all selected with certainty. Within strata involving “Others” (two strata for both panels) the NHIS, as described above, households selection was with Hispanics were also included at disproportionately high rates probability proportionate to size (oversampledpps) among where the households selected at size measure was the roughly 50% rateinverse of the NHIS initial probability of selection. Thus, for MEPSThe pps sampling was undertaken to help reduce the variability in the MEPS weights incurred due to the variability of the NHIS sampling rates. With the subsampling, households that were oversampled for MEPS in calendar year 2016 were those containing Hispanicsresponding households in the NHIS identified as having members whose race/ethnicity was Hispanic, BlacksBlack, Asiansor Asian for both panels. Typically, sample allocations across sample domains change from one MEPS panel to another. The sample domains used may also vary by panel although this was not the case for Panel 20 and Panel 21. When one compares unweighted measures (e.g., response rates) between panels and years, one should take into account such differences. If, for example, members of one domain have a lower propensity to respond than those predicted to be poorof another domain, then if that domain has been allocated a higher proportion of the sample, the corresponding panel may have a lower unweighted response rate simply because of the differences in sample allocation. Within each domain (sample stratum) systematic samples of the MEPS-eligible households were selected from among the NHIS household respondents made available for MEPS sample selection purposes.

Oversampling. Oversampling ▇▇▇▇▇▇▇▇▇▇▇▇ was employed for some selected subgroups of policy-level interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be is discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "“oversampled" ” are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "“not oversampled"”. As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "“cost" ” of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve achieved if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the NHIS, Hispanic households were oversampled at a rate of roughly 2 to 1. That iswhere Hispanics, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not Blacks, and Asians are oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. The oversampling approach for both MEPS Panels 10 18 and 11 was 19 used the same. From same sampling domains (strata) for subsampling among the NHIS households respondents eligible for MEPS, reflecting . Five domains were established in a hierarchical sequence: the oversampling of Hispanics and blacks described above, three strata were constructed for sampling purposes for MEPS. One stratum first domain contained households with Asians and those "predicted to be poor", a Asians; the second domain contained households with black Hispanics not assigned to the first domain; the third domain contained households with Black members that were not among assigned to domains one and two; the fourth domain contained those remaining households (“Other” households) characterized as “complete” for the NHIS; and the fifth domain contained all remaining “Other” households, those characterized as “partial complete” for the NHIS. In terms of sampling rates, for Panel 18 all households in the first stratumAsian, while the third stratum contained all remaining households. All households in the "Asian/Predicted Poor" stratum Hispanic, and Black domains were selected sampled with certainty. The sampling rate for the black stratum was three-fourths. The sampling rate about 63 percent for the “otherOther complete” stratum was domain, and about 53 43 percent for the “Other partial complete” domain. For Panel 10 and 56 19 the corresponding sampling rates for all the domains except those characterized as “Other” were 1 (i.e., all households assigned to those domains were included in MEPS). The “Other, complete” domain was sampled at a rate of about 66 percent for Panel 11while the “Other, partial complete” domain was sampled at a rate of about 42 percent. Within stratastrata (domains) for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled For all but those strata involving “Others”, the selection was with equal probability. Within the two strata involving “Others” (for both Panel 18 and Panel 19) the NHIS, as described above, households sample was selected with Hispanics were also included at disproportionately high rates probability proportionate to size (oversampledpps) among where the households selected at size measure was the roughly 50% rateinverse of the NHIS initial probability of selection. Thus, for MEPSThe pps sampling was undertaken to help reduce the variability in the MEPS weights incurred due to the variability of the NHIS sampling rates. With the subsampling, households that were oversampled for MEPS in calendar year 2014 were those containing Hispanicsresponding households in the NHIS identified as having members whose race/ethnicity was Hispanic, BlacksBlack, Asians, and those predicted to be pooror Asian.

Oversampling. Oversampling ▇▇▇▇▇▇▇▇▇▇▇▇ was employed for some selected subgroups of policy-level interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be is discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "“oversampled" ” are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "“not oversampled"”. As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "“cost" ” of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve achieved if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS, where Hispanics, Blacks, and Asians are oversampled. The oversampling approaches and the sampling domains used for MEPS Panels 16 and 17 for subsampling among the NHIS respondents eligible for MEPS differed only in that a “cancer” domain was used for Panel 16. For the NHIS, Hispanic Panel 16 six domains (strata) were established in a hierarchical sequence. The first stratum contained households were oversampled at a rate of roughly 2 to 1. That is, the probability of selecting a Hispanic household for participation where an individual was identified in the NHIS was roughly twice that for households in the general population that were not oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. The oversampling approach for both Panels 10 and 11 was the same. From among as having cancer via the NHIS households eligible for MEPSsampled adult questionnaire, reflecting the oversampling of Hispanics and blacks described above, three strata were constructed for sampling purposes for MEPS. One a second stratum contained households with Asians and those "predicted to be poor"Asians, a second third stratum contained households with black Hispanics not assigned to the second stratum, a fourth stratum contained households with Black members that were not among assigned to strata two and three, the fifth stratum contained Other households characterized as “complete” for the NHIS; and the sixth stratum contained Other households characterized as “partial complete” for the NHIS. For Panel 17, after eliminating the cancer domain, the same hierarchical ordering was used. In terms of sampling rates, for Panel 16 all households in the first stratumcancer, while the third stratum contained all remaining households. All households in the "Asian/Predicted Poor" stratum , Hispanic, and Black domains were selected sampled with certainty. The sampling rate for the black stratum was three-fourths. The sampling rate about 79 percent for the “otherOther complete” stratum was domain, and about 53 46 percent for the “Other partial complete” domain. For Panel 10 and 56 17 the corresponding sampling rates for all the domains except those characterized as “Other” were 1 (i.e., all households assigned to those domains were included in MEPS). The “Other, complete” domain was sampled at a rate of about 51 percent for Panel 11while the “Other, partial complete” domain was sampled at a rate of about 40 percent. Within stratastrata (domains) for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled For all but those strata involving “Others”, the selection was with equal probability. Within the two strata involving “Others” (for both Panel 16 and Panel 17) the NHIS, as described above, households sample was selected with Hispanics were also included at disproportionately high rates probability proportionate to size (oversampledpps) among where the households selected at size measure was the roughly 50% rateinverse of the NHIS initial probability of selection. Thus, for MEPSThe pps sampling was undertaken to help reduce the variability in the MEPS weights incurred due to the variability of the NHIS sampling rates. With the subsampling, households that were oversampled for MEPS in calendar year 2012 were those containing Hispanicsresponding households in the NHIS identified as having members with cancer or whose race/ethnicity was Hispanic, BlacksBlack, Asiansor Asian. Again, note that not all NHIS households where a member had cancer were identified as such - the member with cancer had to have been randomly selected to complete the NHIS “sampled adult” questionnaire and those predicted to be poorhave self-identified as having cancer in response to questions from that questionnaire.

Oversampling. Oversampling ▇▇▇▇▇▇▇▇▇▇▇▇ was employed for some selected subgroups of policy-level interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be is discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "“oversampled" ” are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "“not oversampled"”. As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "“cost" ” of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve achieved if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the NHIS, Hispanic households were oversampled at a rate of roughly 2 to 1. That iswhere Hispanics, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not Blacks, and Asians are oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. The oversampling approach for both MEPS Panels 10 19 and 11 was 20 used the same. From same sampling domains (strata) for subsampling among the NHIS households respondents eligible for MEPS, reflecting . Five domains were established in a hierarchical sequence: the oversampling of Hispanics and blacks described above, three strata were constructed for sampling purposes for MEPS. One stratum first domain contained households with Asians and those "predicted to be poor", a Asians; the second domain contained households with black Hispanics not assigned to the first domain; the third domain contained households with Black members that were not among assigned to domains one and two; the fourth domain contained those remaining households (“Other” households) characterized as “complete” for the NHIS; and the fifth domain contained all remaining “Other” households, those characterized as “partial complete” for the NHIS. In terms of sampling rates, for Panel 19 all households in the first stratumAsian, while the third stratum contained all remaining households. All households in the "Asian/Predicted Poor" stratum Hispanic, and Black domains were selected sampled with certainty. The sampling rate for the black stratum was three-fourths. The sampling rate about 66 percent for the “otherOther complete” stratum domain, and about 42 percent for the “Other partial complete” domain. For Panel 20 the corresponding sampling rates for all the domains except those characterized as “Other” were 1 (i.e., all households assigned to those domains were included in MEPS). The “Other, complete” domain was sampled at a rate of about 84 percent while the “Other, partial complete” domain was sampled at a rate of about 53 percent for Panel 10 and 56 percent for Panel 11percent. Within stratastrata (domains) for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled For all but those strata involving “Others”, the selection was with equal probability. Within the two strata involving “Others” (for both Panel 19 and Panel 20) the NHIS, as described above, households sample was selected with Hispanics were also included at disproportionately high rates probability proportionate to size (oversampledpps) among where the households selected at size measure was the roughly 50% rateinverse of the NHIS initial probability of selection. Thus, for MEPSThe pps sampling was undertaken to help reduce the variability in the MEPS weights incurred due to the variability of the NHIS sampling rates. With the subsampling, households that were oversampled for MEPS in calendar year 2015 were those containing Hispanicsresponding households in the NHIS identified as having members whose race/ethnicity was Hispanic, BlacksBlack, Asians, and those predicted to be pooror Asian.

Oversampling. Oversampling ▇▇▇▇▇▇▇▇▇▇▇▇ was employed for some selected subgroups of policy-level interest to help increase the precision of estimates associated with members of those subgroups. Before going into details, the concept of oversampling will be is discussed. In a sample where all persons in a population are selected with the same probability and survey coverage of the population is high, the sample distribution is expected to be proportionate to the population distribution. For example, if Hispanics represent 15 percent of the general population, one would expect roughly 15 percent of the persons sampled to be Hispanic. However, in order to improve the precision of estimates for subgroups of a population, one might decide to select samples from those subgroups at higher rates than the remainder of the population. Thus, one might select Hispanics at twice the rate (i.e., at double the probability) of persons not oversampled. As a result, subgroups that are "“oversampled" ” are represented at disproportionately high rates in the sample. Sample weights help ensure that population estimates account for this disproportionate contribution from oversampled subgroups, as the base sample weights for oversampled groups will be smaller than for the portion of the population not oversampled. For example, if a subgroup is sampled at roughly twice the rate of sample selection for the remainder of the population not oversampled, members of the oversampled subgroup will receive base or initial sample weights (prior to nonresponse or poststratification adjustments) that are roughly half the size of the group "“not oversampled"”. As mentioned above, oversampling a subgroup is done to improve the precision of survey estimates for that particular subgroup. The "“cost" ” of oversampling is that the precision of estimates for the general population and subgroups not oversampled will be reduced to some extent compared to the precision one could have achieve achieved if the same overall sample size were selected without any oversampling. For MEPS, some of the oversampling was achieved through its linkage to the NHIS. For the NHIS, Hispanic households were oversampled at a rate of roughly 2 to 1. That iswhere Hispanics, the probability of selecting a Hispanic household for participation in the NHIS was roughly twice that for households in the general population that were not Blacks, and Asians are oversampled. The NHIS oversampling rate for black households was roughly 1.5 to 1. The oversampling approach for both MEPS Panels 10 17 and 11 was 18 used the same. From same sampling domains (strata) for subsampling among the NHIS households respondents eligible for MEPS, reflecting . Five domains were established in a hierarchical sequence: the oversampling of Hispanics and blacks described above, three strata were constructed for sampling purposes for MEPS. One stratum first domain contained households with Asians and those "predicted to be poor", a Asians; the second domain contained households with black Hispanics not assigned to the first domain; the third domain contained households with Black members that were not among assigned to domains one and two; the fourth domain contained those remaining households (“Other” households) characterized as “complete” for the NHIS; and the fifth domain contained all remaining “Other” households, those characterized as “partial complete” for the NHIS. In terms of sampling rates, for Panel 17 all households in the first stratumAsian, while the third stratum contained all remaining households. All households in the "Asian/Predicted Poor" stratum Hispanic, and Black domains were selected sampled with certainty. The sampling rate for the black stratum was three-fourths. The sampling rate about 51 percent for the “otherOther complete” stratum was domain, and about 53 40 percent for the “Other partial complete” domain. For Panel 10 and 56 18 the corresponding sampling rates for all the domains except those characterized as “Other” were 1 (i.e., all households assigned to those domains were included in MEPS). The “Other, complete” domain was sampled at a rate of about 63 percent for Panel 11while the “Other, partial complete” domain was sampled at a rate of about 43 percent. Within stratastrata (domains) for both panels, responding NHIS households were selected for MEPS using a systematic sample selection procedure from among those eligible. Because Hispanics had been oversampled For all but those strata involving “Others”, the selection was with equal probability. Within the two strata involving “Others” (for both Panel 17 and Panel 18) the NHIS, as described above, households sample was selected with Hispanics were also included at disproportionately high rates probability proportionate to size (oversampledpps) among where the households selected at size measure was the roughly 50% rateinverse of the NHIS initial probability of selection. Thus, for MEPSThe pps sampling was undertaken to help reduce the variability in the MEPS weights incurred due to the variability of the NHIS sampling rates. With the subsampling, households that were oversampled for MEPS in calendar year 2013 were those containing Hispanicsresponding households in the NHIS identified as having members whose race/ethnicity was Hispanic, BlacksBlack, Asians, and those predicted to be pooror Asian.