School Libraries DO Increase Student Achievement
For decades researchers have tried to link school librarians with increases in student achievement. However, prior research generated very little convincing evidence of an impact - until now.
Michael Radlick and Joette Stefl-Mabry, educational researchers, presented results of their large-scale initial pilot research about school libraries at the recent American Educational Research Association conference in Chicago (April, 2015). Their paper "Finally - Convincing Evidence for the Impact of School Librarians" focused on answering this critical question about school capacity and resources: is there evidence that a full‐time (or more) certified school librarian (SL) has a positive impact on student achievement? Their study was the initial step in a much larger on-going statewide school library research project. With over two prior decades of library research that used perceptual data or had small samples of schools using purely descriptive or correlational techniques, prior research provided, at best, weak causal evidence for the impact of school librarians on student achievement.
In contrast, their study was framed within the context of a strong statistical analysis technique called structural equation modeling (SEM), and the study reflected a very large sample size while controlling for a number of other variables including prior year student achievement. Specifically, the causal models in this study examined the SL’s impact on building level student English and math achievement, while taking account for prior academic achievement, a wide range of student demographic variables (gender, special education status, limited‐English proficiency, minority status, free/reduced lunch participation), and building level characteristics (size, discipline climate, district accountability status, and overall district financial capacity). Data for the study came from a variety of New York State Education Department sources including the annual Basic Educational Data System survey of schools, which includes a range of data collected each year about school libraries.
Their longitudinal, between‐schools design studied all of New York State’s 2,245 public schools (which excluded schools in New York City) that had students in grades 3 thru 8. Of those 2,245 schools, there were 1,511 (67.3%) that had a full time or more school librarian, while there were 743 schools (32.7%) that did not have at least a full time SL. The student achievement outcome measures used in the study were the New York State Education Department’s annual state assessments transformed into the Department’s school performance index measure for ELA or math. SEM path model were used to examine the New York State ELA and math Common Core Performance Index for 2012-2013 (and also the change in ELA and Math Performance Index in the school from 2011‐2012 to 2012-‐13) while controlling for a wide range of demographic and school characteristics.
Based on the results from the structural equation models, which controlled for many other variables, school librarians (SLs) are shown to have a statistically significant impact on student achievement in English Language Arts (both the 2012-2013 ELA Performance Index and the Change in ELA Performance Index from 2011-2012 to 2012-2013), but not on student achievement in math. This impact was significant, even after taking controlling for student demographic factors (gender, disability status, limited English status, minority status, and poverty), school factors (NCLB status, size of school, high resource-need status, and numbers of disciplinary incidents reflective of school climate), and prior academic performance in the building in both ELA and math. At the outset the researchers had hypothesized that the impact of school librarians would be on English language arts, but not math achievement, and in fact the SEM results confirm this impact both for the 2012-2013 ELA Performance Index and the Change in ELA Performance Index from 2011-2012 to 2012-2013. The interpretation was that school librarians have a greater impact on literacy related achievement than mathematics because of their training and focus on information and digital literacy.
While the school librarian effect on school ELA academic performance is statistically significant, the effect size is relatively small and, not unexpectedly, explained a relatively small part of the variance in total academic achievement as compared with some of the other variables in the models. The path coefficients from the SEM models show that there are other factors that have a much greater impact on the outcome measures in both ELA and math, such as prior academic performance and poverty - which had the highest effects in the model. The relative size of the School Librarian Effect is not surprising, given that the school librarian is a school-level resource in effect spread across the entire school building, and given that it is a school-level outcome measure they used. In contrast, a regular classroom teacher’s effect is estimated by other researchers to account for 7% to 21% of the variance in student achievement. However, each individual teacher is only impacting a small group of students in the school (i.e. a single classroom of students) whereas the effect of the school librarian is at the building level.
Both researchers were selected to attend the American Association of School Librarians (AASL) national research summit in April 2014 as two of 50 leading school library and educational researchers to investigate causal phenomena in SL instruction, resources and services. Based on the study results, the authors were recently awarded a three year Institute of Museums and Library Services (IMLS) Research Grant to expand their pilot research to model New York City, as well as across other school years and other school levels. The IMLS research grant will also allow investigation of other school library factors and resources that may impact student achievement.
Joette Stefl-Mabry, Ph.D. University of Albany, SUNY
N.B.: Dr. Stefl‐Mabry is the contact person for this study and for the IMLS Research Grant.