Data collection and evaluation are central activities in a quality rating and improvement system (QRIS). Pressure to complete a system redesign or launch a new QRIS can make a focus on data and evaluation challenging. However, it is important to elevate data collection and evaluation in system planning rather than waiting until a challenge or question arises that is best addressed through evaluation. Prospective planning ensures that data collection activities are maximized and take into account future evaluation questions. Yet, it is never too late to engage in data collection and evaluation activities that can inform system improvement. This section poses questions and offers tools that can be used early—and later—in QRIS implementation to collect data and answer critical evaluation questions. Discussions on the use of data in planning and implementation are included in the Initial Design Process and Approaches to Implementation sections of this guide.
Quality Rating and Improvement System State Evaluations and Research (2018) from Child Care & Early Education Research Connections provides a comprehensive list of state QRIS evaluations and research in the Research Connections collection.
Print the Data Collection and Evaluation section of the QRIS Resource Guide.
All states have data systems that contain information on early and school-age care and education programs. In deciding what data to collect, states should first identify the questions it wants to answer and how the data will be used. Some data sources that may be helpful for QRIS include the following: licensing; registries of license-exempt providers; subsidy administration; practitioner and training/trainer registries; child care resource and referral (CCR&R) databases; technical assistance tracking systems; program profiles; classroom assessments; economic impact research studies; and Head Start, prekindergarten, and other education systems. An initial step in planning for a QRIS or designing an evaluation is to compile a list and description of existing state/territory data systems, including where they are located, how to access them, who has access to them, what information is collected in them, and how they interface with other data systems.
Data Resources Analysis for Decisionmaking
Completing an inventory of the available data at the beginning of the planning and design stages is a helpful first step. The information gathered during this process can then be used to guide decisions during the implementation phase. For example, data from the licensing system or Head Start Program Information Reports may help the QRIS design team determine, at least initially, which types of programs (center, home, prekindergarten, Head Start) to include in the QRIS and which and how many programs may be able to achieve the standards. Data from workforce studies or professional development registries can provide a needs assessment of scholarships and educational offerings. This information will help estimate participation rates and predict the resources necessary to support projected participation. Looking at these data elements may reveal existing information that can help document compliance with proposed standards. Reviewing an inventory of existing data can also help determine whether it is best to begin with a pilot and, if so, which programs to include.
Child care subsidy data can also be helpful. Examining these data may lead to the conclusion that tiered subsidy reimbursement will not be sufficient as a support of higher program quality for a number of reasons. For example, if only 20 percent of the enrollment of a typical program are children who receive child care subsidies, that may not be sufficient to support the cost of higher quality for the program as a whole. The balance of the cost must be passed on as tuition fees to other families. Or the enrollment may fluctuate enough that programs cannot rely on tiered subsidy reimbursement to maintain quality. Therefore, subsidy data may be a good indicator of the potential impact of tiered subsidy reimbursement, pointing out the need to explore additional provider incentives.
Data for QRIS Management
Using existing data systems can help make QRIS implementation more cost efficient and ensure consistency in data across systems. Adding reporting capacity or data elements or aligning data elements to an existing data system, such as licensing or a professional development registry, can be much less expensive than creating a new data collection and processing system specifically for QRIS. This may or may not be possible, depending on who administers the QRIS and what data systems can be tapped for the information. For example, if the existing data system is in a state agency and the QRIS will be operated outside of the state government structure, it may not be possible to use the state data system. Even when data exist in several separate systems, it may be cost-effective and ensure consistency if data can be transferred from one system to another, rather than entering all data anew for each child care program that wants to participate. For example, one QRIS requirement for participation might be a license in good standing or a license with no serious violations. It would be critical to have continuing, current information on the status of a license to produce reliable ratings. Similarly, if programs that participate in the QRIS are also rated or assessed by other entities, such as national accrediting organizations or the Head Start monitoring system, using data from those systems can make participation easier, more cost-effective, and more reliable. Linking to data in professional development registries or credentialing and certification systems is another cost-effective way to verify staff qualifications, ensure consistency, and eliminate duplicative work in the rating process.
In summary, an accurate inventory of existing data systems, their accessibility, accuracy, and reliability is helpful in determining QRIS system design. A good introduction to data elements, collection, management, and governance is found in the slides and videos of the Early Childhood Data: Building a Strong Foundation webinar series presented by Quality Initiatives Research and Evaluation Consortium (INQUIRE) in spring 2013. An overview of the use of data to monitor and evaluate QRIS in five states may be helpful in thinking about the broad perspective of using data (Caronongan, Kirby, Malone, & Boller, 2011).
States are increasingly relying on comprehensive data systems that they either purchase or develop to help with the administration of their QRIS. This section of the guide focuses on identifying the data needed and whether they can be collected from existing systems or if new data collection mechanisms need to be developed.
As a resource to state agencies, specific products, vendors and systems are referenced throughout this document. However, the Office of Child Care and the National Center on Early Childhood Quality Assurance do not endorse any non-Federal organization, publication, or resource.
Looking closely at each QRIS standard and determining how compliance will be verified, what data for documentation will be needed, who will review the data, and where data will be stored are essential steps in QRIS planning. New data may be needed to assign a rating or to guide follow-up activities, such as development of an improvement plan. For example, QRIS standards may require that all teaching staff receive training in a state’s early learning guidelines for a certain rating level. If completion of the training is collected in the professional development registry, it may be possible to import information from that system for the rating process. If the information is not currently collected, it may be necessary to develop a process for collecting that data, such as requiring program staff to document their trainings by submitting successful-completion certificates, requiring rating assessors to enter information into a new QRIS database, or asking early learning guideline trainers to input their class lists into the professional development registry. A thorough review of the rating assessment and monitoring process is needed to identify data to document compliance with QRIS standards. Once a QRIS is implemented, this data will also be invaluable in informing and guiding needed modifications.
Data systems are a valuable resource for staff who manage the QRIS provider support system. Two types of data may be useful to them: (1) data on supports for individuals working in the early and school-age care and education programs, and (2) data on supports for the programs that seek QRIS ratings.
Data on supports for individuals working in the programs are helpful in projecting and managing the cost for scholarships for staff education and any type of retention incentives, such as wage supplements. These data can also help determine the effectiveness of various supports. Is the education level of the staff across the state going up? Are there any geographic areas not using scholarships? If not, why? Answering these questions requires data that are specific to QRIS participation. If, for example, a state currently has a scholarship program that is available to all early and school-age care and education providers, knowing which of these staff work in programs that participate in the QRIS is crucial. These data, coupled with broader data on staff qualifications, can help identify trends and inform decisions regarding the capacity of practitioners to meet QRIS standards and how to best support continuous improvement.
Collecting data on technical assistance and other supports for programs is usually a more complex process than collecting data on individuals working in the programs. Often programs that participate in a QRIS have access to technical assistance, including consultation and coaching supports. These supports might be available to a broad group of programs, including those that do not participate in the QRIS. Thus, it is important to create data systems that identify which supports and how much of each is received by each program participating in the QRIS. It is important to think carefully about what data about program supports needs to be collected, including data on new supports that may be created and accessible only to programs participating in the QRIS.
The QRIS planning team should think carefully about how program support information will be used. Will the data identify participating programs that access supports and how often? Will it be used to determine the correlation between supports accessed and improvements in program ratings? Will it be used to manage the cost of such supports or to monitor the effectiveness of support service providers? Being clear about the projected use of data will help to define what is collected and how.
Collecting data on financial supports for programs that participate in QRIS, such as grants, bonus payments, tiered reimbursement, loans, or tax benefits, can help project and manage budgets. Again, it may be very useful to correlate data with the maintenance or improvement of ratings. This will help identify which supports are most critical.
In many states, the QRIS becomes an organizing framework for a wide range of program and practitioner supports designed to promote quality improvement. States have moved from providing technical assistance and financial supports that are believed to improve child care quality to using the QRIS to track whether these supports are actually associated with quality improvements.
The exploration of what data might be needed is best done early in the process and with a broad view to future needs. In the planning and design phase, considering how to verify the standards has become increasingly important to states. Assessing the impact of key interventions to assist programs in improving quality is critical to project management. Within the rating process, it is becoming crucial to coordinate assessment of the QRIS ratings across sectors (i.e., child care, prekindergarten, Head Start) in a way that reduces the duplication of multiple assessment processes. In preparation for evaluation, consider the benchmarks that are being set and how to document their achievement, including coordination of standards using data from other assessment processes, such as accreditation, Head Start performance standards, and prekindergarten standards assessment.
Implementing an Evaluation
QRIS evaluation is essential for supporting continuous system improvement. Evaluation results can inform four activities that shape how the QRIS evolves:
- Identifying implementation successes and challenges. At any stage of QRIS implementation, but particularly when the QRIS is newly launched or has undergone a major revision, evaluation can reveal which activities are working well and which activities need attention. Findings from focus groups or surveys with the implementation team or with providers participating in the QRIS add context and depth to administrative data. For example, administrative data can be used to track provider enrollment in the QRIS and to see how enrollment patterns differ across regions of the state. Additional data collection, such as surveys with eligible providers, can provide insights into the motivations and experiences of providers that underlie the patterns observed in the administrative data. Evaluation results can inform the development or revisions of marketing materials, implementation partner communication protocols, roles and responsibilities of staff, the content of training sessions for staff, and the development of orientation materials for providers.
- Examining the effectiveness of new and existing activities in the QRIS. Marketing, recruitment, distribution of financial assistance, provision of technical assistance, and assignment of program ratings are QRIS activities that require significant investments of staff and financial resources. Evaluation is a critical tool for learning about the effectiveness of QRIS activities and identifying whether and how different activities are contributing to intended outcomes.
- Documenting outcomes for stakeholders. Stakeholders for QRIS expand beyond state agencies and implementation partners and include providers, legislators, parents, and business and community leaders. These stakeholders are eager for information about QRIS outcomes. It is important to set clear expectations for outcomes that align with the stage of QRIS implementation. For example, early in implementation, realistic outcomes include program enrollment and engagement in quality improvement activities. Realistic outcomes at later phases of implementation include increased awareness of the QRIS among the public, greater density of program participation, increases in program quality, and provision of quality at the highest levels of the QRIS.
- Engaging in short- and long-term planning. Evaluation results can inform immediate adjustments to the QRIS and support development of plans for the future. For example, an implementation evaluation typically produces results that can be acted on right away to address challenges or to expand activities that are working well. Evaluation results also can be used to set long-term goals for outcomes, such as quality improvement. Results may support projection of the expected pace of improvement among programs, which can help with planning for technical assistance staffing and distribution of financial incentives to participating programs over 5 years or longer.
Using a QRIS logic model provides the guiding framework for evaluation efforts and development of an evaluation plan. The Quality Rating and Improvement System (QRIS) Evaluation Toolkit (Lugo-Gil, Sattar, Ross, Boller, Tout, & Kirby, 2011) includes a chapter that serves as a workbook for logic model development. The key steps to developing a logic model include the following: (1) describing the context and environment for the QRIS and articulating the QRIS goals; (2) identifying the inputs and the resources needed to support the work; (3) outlining the implementation activities; (4) indicating the outputs that can be tracked; (5) articulating short-, mid-, and long-term outcomes; and (6) linking expected outcomes with activities to identify any gaps or unrealistic expectations about the impact of the QRIS. The “Initial Design” section of this guide includes information about the Massachusetts QRIS logic model.
When embarking on logic model development, it is important to convene a group of stakeholders to inform the process. The logic model should reflect connections to other systems (e.g., licensing, professional development) and serve as a platform for identifying and leveraging implementation resources and cross-system evaluation opportunities.
Once the logic model is complete, it can be used to develop an evaluation plan. An evaluation plan contains the following: research questions (with a designation of their priority levels); the data elements needed to address the research questions; preferred timing for each research question; whether the data are currently available or need to be collected; an estimate of the cost for each type of research question; a note about whether the evaluation can be conducted internally or whether an external evaluator should be identified; and strategies for disseminating results.
It is important to designate a staff person within the QRIS implementation team to be the coordinator and facilitator of the work. Evaluation planning will be challenging to launch and manage if it is not assigned as an explicit work activity. If possible, it will also be important to engage an experienced evaluator to help guide the process of evaluation planning.
In the same way that logic model development will benefit from stakeholder participation, it is helpful to invite community stakeholders to be part of the evaluation plan development. Key stakeholders for evaluation planning include state agency partners, local or national funders, university partners, or other research partners who can bring new ideas, resources, and even evaluation capacity to the process.
The evaluation plan and the logic model can be viewed as living documents that should be revisited on a regular basis to ensure that they still reflect QRIS priorities and features of the system. Both documents can help guide planning if funding opportunities become available or if opportunities arise to evaluate other parts of the early learning system that offer a platform to add research questions related to the QRIS.
Developing an Evaluation Plan
Because a QRIS serves as a systemic structure with activities to support multiple goals related to program quality, children’s development, and provision of information to parents and caregivers, there are many possible research questions to address through QRIS evaluation. A key planning task is to identify the research questions that will be most beneficial for informing system improvement. Research questions could be developed to understand implementation and outcomes for each of the primary QRIS activities, for example: program recruitment, technical assistance, program ratings, financial incentives (including tiered reimbursement), consumer education/dissemination of ratings, and system access and equity. Research questions may be prioritized to reflect the areas of the QRIS requiring the largest investment of resources, areas of particular concern in QRIS functioning, or areas required by a funder.
A critical aspect of planning and selecting research questions is making sure they align with the QRIS stage of implementation. Exhibit 1 provides general guidance about matching topics with the QRIS stage.
Exhibit 1. Matching Evaluation Topics to the Stage of QRIS Implementation
Sample QRIS Research Questions
The following is a set of sample research questions related to QRIS participation and ratings, program quality improvement, effectiveness of financial incentives, access to high-quality programs, validity of QRIS ratings, and use of QRIS ratings by parents and the public.
- What are the characteristics of programs enrolled in the QRIS?
- How effective are QRIS recruiting efforts with different types of providers (for example, urban versus rural, centers versus family child care programs, and programs serving a high proportion of children who receive subsidies)?
- Do the characteristics of programs that are not enrolled in the QRIS differ from those that are enrolled? For example, are there differences in key characteristics, including geography, program type, funding, or director qualifications?
- What is the distribution of program sites across quality levels?
- What are the differences in program characteristics at each rating level?
- What are the characteristics of children who have access to high-quality programs?
- Which providers are improving and what resources are used for improvement?
- What is the quality of the program’s learning environment as measured by an independent measure of quality?
- Are observed quality scores improving over time among programs in the QRIS? How is this related to quality improvement investments?
- What are the perceptions of non-enrolled early care and education (ECE) programs on QRIS?
- What is the reach of technical assistance in the state?
- What are the characteristics of teachers and family child care providers who receive onsite technical assistance?
- Does the stability of the early care and education workforce increase over time?
- How does the system’s training and professional development impact child outcomes of interest? Provider outcomes of interest?
- Do parents know and use the QRIS to make ECE decisions?
- What are parent perceptions of program services and quality?
- How are website visitors using information to search for ECE programs (e.g., star ratings, distance, search terms)?
The Illinois Early Learning Council (Data, Research and Evaluation Committee) Research Agenda (2015) is an example of a research plan (not limited to QRIS). This group uses an overarching frame for its plan in which it asks, “What information would cause us to behave differently in policy and practice in ways that would likely lead to better outcomes for young children?”
Ideally, evaluation should be planned for as soon as a QRIS is designed so that evaluation can address each of the four purposes outlined in the “Evaluation Purpose” section. In addition, planning early can be efficient, especially when data collection and system activities are planned with the evaluation in mind. For example, a needs assessment conducted when planning the QRIS could also serve as baseline data that could be used to chart progress over time. In addition, QRIS data collection protocols for assigning ratings and technical assistance case management data will be better suited for evaluation if data elements needed to address high-priority evaluation questions are identified in advance. Planning for evaluation at the outset does not necessarily mean that an evaluation should be launched immediately but rather that the building blocks are in place for evaluation when the timing and resources are appropriate.
The timing and focus of evaluation should be matched to the stage of QRIS implementation (see Approaches to Implementation section). Some research questions may benefit from an annual study while others may only need to be addressed every 3 to 5 years. For example, a survey to understand provider experiences in the QRIS may be useful to launch annually, particularly early in QRIS implementation, so that adjustments can be made to the QRIS in response to the findings. In contrast, an examination of children’s development in programs at different levels of quality could be planned for a 5-year cycle to allow for system changes to be more established before investing in an expensive data collection effort. When establishing different timeframes and focal points for evaluation efforts, it is important to develop messages for stakeholders that convey the value of ongoing evaluation and how it will support system improvement.
Though planning for evaluation as part of QRIS design is ideal, it is never too late to engage in QRIS evaluation. An evaluation plan can be developed at any time during implementation. There may be some limitations in availability of data that can be used for evaluation, but these challenges can be addressed. It may be useful to work with an evaluation consultant to assess needs and capacity and to support evaluation planning once a QRIS is underway.
The data used in QRIS evaluation typically come from existing administrative data—that is, data collected for the purposes of administering the QRIS—and new data collected exclusively for the purpose of research and evaluation. The process of identifying data and developing data protocols for QRIS administration described in the first part of this section can be very useful when planning for QRIS evaluation. A comprehensive data matrix that describes the available data in the early care and education system can provide evaluation planners with information about what exists already in the system and what new data would need to be collected to support evaluation.
Exhibit 2 provides an overview of the types of data elements that may be useful for QRIS evaluation.
Exhibit 2. Possible Data Elements to Support QRIS Evaluation
When using administrative data for QRIS evaluation, it is important to be aware of potential limitations. The following questions can be asked to learn about the data:
- What is the data coverage? Does it include all ECE programs, the entire ECE workforce, all geographic areas of the state, all children, all families? Data sets are typically limited in specific ways that are relevant to evaluation. For example, it is important to know if a provider registry is voluntary and the proportion of eligible providers that are included in the data.
- Are there duplications in the data? It is useful to know if counts or frequencies calculated in the data are taking into account the fact that a program or provider, for example, may be included in the dataset more than once. Unique identification numbers are helpful for dealing with this challenge.
- What is the quality of the data? Before analyzing data, know whether procedures are in place to ensure accuracy and reliability of the data. Staff entering data, for example, should receive training and be monitored over time to ensure they are following data quality protocols.
- What is the availability of historical data? In some cases, only current data or data from a limited time period are available. The existence of archived data will determine whether it is possible to ask certain research questions that require the availability of data over time.
Child Care & Early Education Research Connections provides Working with Administrative Data (n.d.) a web page of resources organized by topic, including managing, analyzing, and linking administrative data and issues related to data confidentiality and security.
New data collection can fill gaps in data elements not covered by administrative data. When possible, consider using or modifying existing measures or surveys to facilitate comparisons and improve data quality. It is also important to consider the samples that will be tapped for data collection and the response rates of data collection efforts.
Two broad types of data can be collected: quantitative and qualitative. Quantitative data are expressed numerically and define a construct (e.g., a quality score). These data are collected through surveys, administrative data, structured observations, and direct assessments. Comparisons are made using statistical analysis. Quantitative data are useful for estimating trends, analyzing group differences, and understanding the factors that link to changes over time. Qualitative data describe a construct. Data include responses to open-ended questions that are used to describe perceptions, experiences, concerns, and recommendations for improvement. Qualitative data are usually collected through focus groups or interviews, and key themes are coded and reported. Qualitative data are useful for understanding complexity of experiences and underlying motivations.
The costs of evaluation activities vary greatly depending on the type of activity, the scope of the research questions, and whether a third-party will be contracted to do the work. It is useful to examine the relative costs of different data collection activities commonly requested for QRIS evaluation (Exhibit 3). Observations of classrooms and family child care homes and collection of child development data are relatively more expensive data to collect than survey and focus group data or use of existing administrative data.
Exhibit 3. Comparison of Relative Cost of Different Evaluation Activities
Choosing an evaluator is an issue that states address within the restrictions of their resources and the state bidding and contractual requirements. Other considerations that also influence the choice of evaluator should be incorporated in the request for proposal, including the following:
- Qualifications and experience: States look for evaluation teams with qualifications that match the task, i.e., early childhood and research qualifications and experience with QRIS research. They also look for evaluators who have experience completing the research within contract requirements.
- Credibility: Potential evaluators should be highly credible to the primary target audience. This is one of the reasons that many states use their own state universities, even though those universities may bring in national or out-of-state experts to partner on selected portions of the evaluation.
- Stability: If plans call for conducting a series of evaluations, an organization’s longevity in the field and probability of continuing in the work will be important traits to consider.
As noted, evaluation studies serve multiple purposes, including the provision of evidence-based insights into the design or implementation process, and informing funders and policymakers of the impact of the QRIS on child care programs and child outcomes. A strong communications strategy is needed to relay information..
It is important to plan a communications strategy at the beginning of each evaluation activity. Stakeholders should be involved in this planning effort. The plan should include details about which types of products will be developed and how they will be disseminated to different groups. Audiences and specific considerations for communications include the following:
- Providers: Consider multiple outreach strategies (such as videos and flyers) that use different communication techniques. Identify options for public forums, such as town hall meetings, that facilitate two-way dialogue and give providers the opportunity to ask questions about the findings. Talking points should be developed for technical assistance providers and licensors to help them communicate key messages, results, and implications for providers.
- Policymakers: Develop factsheets that provide vital information on the background of the program or initiative. Brief documents should define the problem, the intervention/approach, the results, and recommendations.
- Funders: In addition to the considerations for policymakers, include data to provide important context or rationale for the study or resulting recommendations, such as public opinion data or state or local population indicators.
- Parents: Communicate clear messages with brief details about the goals and objectives of the QRIS. Ensure that key terms such as “quality” are defined using simple, plain language.
Overall, research summaries shared publicly should use plain language, simple formatting, a question-and-answer structure (or other straightforward headings), and provide links to full technical reports and contact information.