Page 41

V. DETERMINING CHILD CARE MARKET RATES Child Care Market Rate Survey 2023 Page | 31

For example:  If there were 50 or more children in an age category, information from that age category was used to impute preliminary rates for the star levels with less than 50 children.  If there were fewer than 50 children in the age category within a county, but 50 or more children in a type (center or home) in the county, all of the children in that county in that type were used to impute the preliminary rates for that star level.  Finally, if there were not 50 or more children in a type within a county, all the children in the county were included to impute preliminary rates for the star levels with less than 50 children. For imputation levels of age category and type, the difference between the county’s raw 75th percentile (from the survey) for each combination of type, age category, and star level, and the corresponding state’s 75th percentile is determined for all categories at the imputation level. Then the mean difference is calculated, weighting each difference by the number of children in the category. The mean difference is then added to the state rate for the categories that need an imputed value. For the imputation level of county (combining centers and homes), the mean percent difference is used instead of the mean difference. Tables 13 through 15 show examples of three imputation methods. 2. Age Imputed Rates An age-imputed rate was determined when there were not fifty children at a particular star rating in an age category, but there were at least fifty children across all star levels for that age category. In the first example, the weighted mean difference was $276 (Table 19). The greater of the county’s 75th percentile rate and the age-imputed rate becomes the modeled rate. This methodology assures that counties were held harmless by never receiving a model rate less than the county’s 75th percentile rate. The final step is to determine an adjusted rate (final adjusted rate) by performing a weighted linear regression for each age category in each county to obtain a strictly increasing rate as the star rating increases.