Using Data from the West Virginia Growth Model
Student growth percentiles and the growth-to-standard calculation allow West Virginia to provide a comprehensive overview of student academic progress and to associate that progress with longer term academic goals like achieving mastery of the state’s academic achievement standard or achieving college- and career-readiness. The model allows the state to summarize the progress of students across various groupings of interest to further inform investigations about best practices and decisions about where critical resources are needed to increase the rate of academic progress for at risk students. The following are some key areas where data from the West Virginia Growth Model can inform practice:
School-level growth data such as the median growth percentile and percentages of students who achieve probabilistically low and high levels of growth and those who achieve growth-to-standard will be useful in the process of determining the relative need for and impact of various school improvement interventions. Figure 1 below illustrates how school classifications can be informed by the inclusion of both status and growth measures of academic progress. For example, schools that fall into Quadrant 4 would likely be thought to require intensive school improvement efforts under a model that only considers student performance. However, with the inclusion of the growth component, it is clear that, while these schools have demonstrated low performance (i.e., low levels of student attainment) for the current year, students within the school are exhibiting high levels of academic growth and could be on track to attain desired performance levels (e.g., mastery) within a reasonable amount of time. As such, less intensive school improvement efforts would likely be required in these schools. On the other hand, ignoring student growth, schools that fall into Quadrant 2 appear to not require any intervention at all under the status model since enough students achieved mastery for the current year. However, when student growth data are added, it is found that students in these schools might be in danger of declining below proficiency if there is not some level of intervention.
Perhaps most promising is the utility of the proposed growth model in identifying schools that fall into Quadrant 1 or Quadrant 3. These are schools that either legitimately require intensive school improvement efforts (i.e., Quadrant 3) or are exhibiting performance and/or growth to the extent that they may be potential sites of interest where the state can discover best practices (i.e., Quadrant 1 and 4). The central learning from this figure is that a status model alone offers a very limited picture of school performance. Such a model considers schools in Quadrants 3 and 4 as a single category of low performing and schools in Quadrants 1 and 2 as high performing and does not provide enough information to differentiate supports.
Figure 1: Status vs. Growth and School Classifications
Figure Adapted from Patz (2008). Growth and Value Added Modeling Approaches in K-12 Accountability Systems.
Student growth data aggregated to the school and classroom level provides one useful measure of the amount of student growth demonstrated within the school and classroom. Such school and classroom summaries can be used as part of a larger evaluation effort to understand teacher and school quality that can be incorporated into a robust and multifaceted annual evaluation system. As noted above, the median growth percentile can be calculated for nearly any group of interest and interpreted as a summary statistic quantifying the average student growth for that group. School-level median SGPs are being used as one of the many factors included in West Virginia’s new Educator Evaluation Pilot.
School-wide estimates of the percentage of students exhibiting levels of growth adequate to achieve or maintain mastery (e.g., growth-to-standard) will also be used to improve the utility and fairness of West Virginia’s accountability system. With broad stakeholder involvement, the WVDE has developed an ESEA Flexibility Request which includes a detailed proposal for incorporating student growth data into the accountability system following the 2012-13 school year.
Informing Instructional Decisions
WVDE also understands that student growth data will be greatly useful for teachers as they decide upon a range of instructional interventions for their students. Specifically, growth data will provide teachers with individualized information regarding which specific students are progressing toward proficiency and which require more intensive intervention in order to meet proficiency.
It is important to note that the growth model does not have the ability to inform teachers as to the specific content that requires remediation. However, the information provided by the model will paint a clear picture of how each individual student’s academic growth compares with similar students across West Virginia and will yield a better understanding regarding how much growth is necessary for each student to maintain or reach proficiency. As such, the West Virginia Growth Model will promote appropriate instructional intervention for all students, not just those who are considered below proficient via the status model.
More details about how teachers and principals can access student growth data are included in the “Reports & Resources” section of this Website.
Program Evaluation and Research
Data from the West Virginia Growth Model will provide unprecedented value to WVDE with respect to systemic research initiatives and program evaluation. The individual student-level data generated via the model will offer an entirely new window into the academic growth of students who are impacted by a variety of intervention programs provided by the department and via external vendors. This information has great potential to result in better informed policy decisions about which programs should receive continued support and which programs should be abandoned in pursuit of more effective alternatives.
Informing StakeholdersThe West Virginia Growth Model will make statewide assessment data more accessible and useful to a wide variety of stakeholders in the public educational system including parents, students, teachers and school administrators. This is because the student growth percentiles calculated by the model quantify individual student growth within a norm-referenced framework that is easily understood by the public. For example, any parent can immediately understand what it means for her/his child to have achieved at the 75th or 50th percentile of growth. However, the concepts of a “proficiency level” or a “scale score” are nebulous and, as such, their utility is very limited for parents and other stakeholders. Beyond student growth percentiles, “growth-to-standard” calculations will make the criterion-referenced goals and objectives set by the state much more accessible to parents and other stakeholders by very simply illustrating the level of sustained growth that is necessary to keep students at proficiency or to ensure they meet proficiency within a predetermined period of time. This information will be greatly valuable to educators as they select instructional techniques and for parents as they provide the necessary support for their children.