Empirical Education has extensive experience with district-publisher partnerships, which typically take the form of a publisher needing to measure the effectiveness of a new product and seeking districts to serve as test sites. Alternatively, a district choosing to try a publisher’s new product may seek to obtain it at zero or reduced cost in exchange for serving as a test site. In either case, using MeasureResults^® to plan, design, set up, and conduct a study can lower costs for all parties involved.

In these partnerships, MeasureResults users can include the publisher’s research staff, curriculum specialists, district and campus liaisons, or others needing proof of their product’s effectiveness in school settings.

Based on the questions of interest and on the available or anticipated data, the web-based MeasureResults system guides users in designing and setting up a study that matches local needs and resources. After an initial meeting with an Empirical Education researcher, users set up their research study online via a series of prompts. Design choices may be retroactive (when data have already been collected) or prospective (when users can control data collection going forward). If one or more research settings are already available, their responses help to determine whether a study is feasible there. If study sites are still to be recruited, MeasureResults can help the user determine the number of subjects and the nature of the setting(s) that will be needed, and thus to choose the most cost-effective and useful approach to the study. At each step of this process, Empirical Education’s research design and data analysis experts are available to assist.

When users have questions that are best answered by including qualitative data, our SurveyCenter^® is available for collecting and uploading responses to surveys, implementation logs, and similar instruments without placing extra demands on district or school staff.

When a study is complete, MeasureResults automatically generates a brief, clear report written in language that is accessible to a variety of audiences.

Please click a topic to learn more.

Types of Studies

• Interrupted Time Series

  This design examines the average achievement over the prior two or three school years for each grade in order to establish a baseline for that grade. The current year achievement for each grade is then compared to the baseline to determine the extent of the change from previous years. This design calls for two or three years of prior testing at that grade level using the same test; if the state changed the test at anytime during the study’s duration, however, the comparison cannot be made. Such studies can be very useful as preliminary investigations and for providing customers with evidence of local effectiveness of the program.

• Quasi-experimental Studies

  These are conducted in a district that has partially implemented the intervention. Data for at least one school year are used. By matching schools, teachers, and classrooms that are using the intervention with very similar schools, teachers, and classrooms, we can provide a rough estimate the difference that it made. We use a variety of matching-based approaches, including propensity score analyses and differences-in-differences estimation, as well as more rudimentary methods involving multilevel regression.

• Randomized Experiments

  These are also called “randomized control trials” or RCTs. We conduct these to help district decision-makers determine what difference a new intervention will make before they invest in it. We usually work with between 30 to 40 teachers who volunteer to implement the new product and are then randomly assigned to either pilot the intervention or to continue to use the existing school program. Random assignment is recognized throughout the scientific community as the only way to produce unbiased results. These trials are often run for a school year but in some cases, a semester-long intervention can be studied.

• Longitudinal Studies

  These are randomized experiments that extend into a second or third year following the same teachers and often the same students. These studies are important where teacher and student experience with the intervention is expected to result in continuing gains over time. Longitudinal experiments are structured so that results are available after the first year. Often federal grants require a longitudinal study, especially when the grant period is two or three years. Our designs also accommodate control groups that enter treatment prior to the end of the study period.

• Sample Reports

Study Processes

• Experimental Design and Evaluation Methodology (32K)
  We can use our expertise in educational settings to help you understand and connect with current and potential customers.
• Research Operations (55K)
  In response to your customers' needs and interests, we can conceive, design, and carry out research projects using your products as intended.
• Analysis, Reporting, and Consultation (63K)
  Our analyses are appropriate to the study, and our deliverables match the level and interests of the audience you specify.
• Tools and Infrastructure (80K)
  Our dedicated engineering staff creates tools to refine our processes, which enables us to work quickly, accurately, and efficiently with you.
• Research Partners, Sites, and Content Areas (35K)
  We can help you test your products' effectiveness across multiple and diverse sites.
• Download our complete capabilities statement here. (180K)

ARRA Funding

Empirical Education can help strengthen any proposal you may be working on with your customers. For example, we can gather evidence from quasi-experiments or surveys, or we can build a scientific feedback component into an implementation. We can do this as an effectiveness test for products already in use or as part of a pilot test for newly implemented products. Click here to learn more.