Unpacking Family Child Care Performance in QRIS: A Latent Growth Mixture Modeling Approach

• In Event: Early Care and Education Quality Indicators, Implementation, and Policy

Abstract

Introduction

Quality rating and improvement systems (QRIS) have become an important part of early care and education systems. Many states struggle to engage licensed family child care (FCC) professionals in QRIS, particularly at higher rating levels (NCECQA, 2019). Hypothesized reasons include misaligned rating criteria and financial incentives that do not cover the full costs of quality improvement (Author, 2022; Bromer et al., 2021; Hallam et al., 2017; Tonyan et al., 2017). Using longitudinal administrative data from one midwestern state, this paper explores QRIS growth trajectories and predictors of change over time.

Research Questions and Hypotheses

1. Are there different trajectories of provider performance on QRIS ratings over time?
2. Are some quality indicators more likely to change than others?
3. What characteristics predict different trajectories of QRIS performance?

I hypothesized that several growth trajectories would emerge, with at least one of growth and one of stagnation over time; that changes would be driven by structural features of quality; and that changes would be concentrated among providers caring for subsidy-eligible children.

Methods

The data source is an administrative dataset of 2667 FCC programs’ subsidy and QRIS data at three timepoints (November 2015, 2017, 2019). During this period, the state had a hybrid point-block system with five levels; one star indicates being out of regulatory compliance and three stars is the threshold for full subsidy reimbursement within a tiered system. A latent growth mixture model (LGMM) was fitted to programs’ QRIS points data using a process of class enumeration and listwise deletion (sample N=892). Descriptive statistics and one-way ANOVAs were run using most likely class assignments. Analyses currently underway will address missing data and further explore predictors of class membership using the three-step method.

Results

The LGMM revealed three latent classes of QRIS growth trajectories (Figure 1). Class 1 (N=365) sustained less than two points on average, mostly remaining at star level 2. Class 2 (N=275) moved from seven to 11 points on average, with many providers moving from star 2 to 3. Class 3 (N=252) made moderate gains from 19 to 20 points on average, with some moving up to star 4 or 5. Classes were well-separated on three of the four subscales (Education and Training, Learning Environment and Curriculum, Business and Professional Practices) with no significant differences on Health and Wellness. Class 2 providers’ improvements were driven by gains on Education and Training and Business and Professional Practices. Among other significant differences (Table 1), providers in Classes 2 and 3 cared for more subsidy-eligible children.

Findings indicate that while some FCC providers successfully move up in QRIS ratings, the majority tend to stay at the lower levels of the system and/or do not make significant gains over time. Convergence with explanatory qualitative findings about QRIS engagement (collected from FCC providers in this state in 2020) will be discussed. This work has implications for redesigning QRIS to be more responsive to home-based settings and for building a more methodologically robust body of literature about quality improvement in FCC.

Author

• Samantha Melvin, Erikson Institute
  Presenting Author