- Our Research
- Working With Us
- Our Data
- Our Team
- About Us
FAQ - Using the Data for Research
I know that I want to conduct longitudinal research, but I don’t know exactly which data I need yet. Can I get a copy of the database so I can see what is available?
- While there is no specific cap on data file size, our intention in granting access to the GRD is to extract well defined samples for defined projects. Our agreements with our clients who participate in the GRD state that we may share their data with researchers who are using the data for specific research purposes, so to that end, we will work with you to define a data specification that is appropriate for your research. We do not grant access to the entire database.
Is it reasonable to compare growth across groups that start at different places on the vertically-aligned scale?
- No. In our own work, we always adjust growth comparisons to account for differences in starting position on the scale. Thus the growth norms we use with schools show growth for students not only at each grade, but also at each starting score within a grade. When we employ HLM methodologies, we include starting score in the model. And when we create Virtual Comparison Groups, each member of a control group starts with the same score as a study group member.
What types of research questions is the Kingsbury Center interested in?
- Our website can give you some good examples of the types of research we engage in. Please go to About Our Research to learn more about the areas of education research we undertake. Also be sure to check out our reports page.
How do external researchers access the data? Do researchers log in to your data warehouse, or do you extract data and send it to the researchers?
- Eligible researchers will be provided access to their dataset (created per the specifications that are developed with Kingsbury Center staff) via our secure ftp website.
Do the reading/math scales guard against ceiling effects?
- Yes - but all tests ultimately have some ceiling. Because the reading and mathematics tests are adaptive, they have much lower standard errors of measure at the upper end of the achievement distribution than fixed form assessments. Over the past several years, we have significantly increased the number of items at the upper end of the measurement scale which has helped reduce ceiling effect artifacts. In reading, the test measures quite accurately through a college entrance level of achievement. Students in the upper five percent of our norms in ninth and ten percent of our norms in tenth grades might be underestimated some because of a ceiling. In mathematics, high performing students generally move from a general math assessment to end-of-course assessments in mathematics (Algebra I, Algebra II, Geometry, Integrated Math 1 and 2), usually in eighth grade. End-of-course tests are reported on the same scale as the general math assessment and have considerably more range. We work with researchers on ways to interpret data across these tests.