Childcare Landscape Study - Flipbook - Page 23
Davidson County Child Care Landscape Study Results
10
analyses; we refer to these survey respondents as the home-based analytic sample. The number
of home-based providers in greater Davidson County overall and in our sample is too small to
assess the representativeness of the sample.
Analytic Approach
The following section describes our approach to data analyses conducted for each
component of Phase 2 of the Davidson County Child Care Landscape Study. For this phase,
we use Child Care Provider Survey data to describe the greater Davidson County child care
landscape, including information on child care supply and availability and characteristics of
the child care workforce.
Broadly speaking, we use the survey data reported on the Child Care Provider Survey and,
where possible, make population-level estimates for center-based programs using strategies to
ensure the estimates are representative of the full population of child care, not just the survey
respondent programs. Where population-level estimates are not possible, we present data from
survey respondents. All home-based results are presented for the survey sample only, based on
the small sample size.
Because the briefs in the Davidson County Child Care Landscape Study draw on different
research objectives, analytic approaches are organized by analytic construct rather than by
individual brief.
Calculating Weights for Center-Based Estimates
For all center-based child care measures for which we create population-level estimates, we
use a weighting strategy to create estimates that are representative of the full population of
child care programs.
To account for oversampling of larger programs in the analytic sample, centers were
grouped into licensed capacity size categories (listed in Table 2). Post-stratification weights
were calculated for each size category as the ratio of the number of eligible centers in the
population to the number represented in the analytic sample. These weights were applied to
programs within each category so that the weighted distribution of centers by size aligns
with the population of licensed providers.
Weights were applied differently depending on the type of measure analyzed. For measures
reported as counts (e.g., child care slots or educator counts), weights were applied directly to
provider-level values and summed to generate regional totals. For workforce characteristics
originally collected at the educator level (e.g., wages, education level, and years of
experience), measures were first aggregated to the center level before weights were applied,
so that centers with larger staff did not disproportionately influence the population estimates.
For all supply and workforce constructs, distributions were examined for extreme values.
Sensitivity analyses were conducted to assess the influence of outliers on population-level
estimates. When a single provider disproportionately influenced results, estimates were
recalculated excluding the outlier(s). Final reported estimates reflect these adjustments
where appropriate.
Prenatal-to-3 Policy Impact Center 2025
www.pn3policy.org
