The Importance and Dilemma of Publishing Studies That Do Not Produce Positive Results

December 18, 2017

(1) An Evaluation of a Learner Response System (2) The Effects of Financial Incentives on Standardized Testing (3) Do Teacher Observations Make Any Difference to Student Performance?

Commentary: This piece reports on three examples of studies of practices that did not produce positive results and highlights the issue of publication bias in educational research. There are powerful contingencies that shape the publication process in ways that do not always work in the best interest of science. For example, promotion and tenure committees do not give the same weight to published replication studies. Also, journals generally do not publish studies that show no effect resulting in the “file drawer problem”. The only exception to this rule is if a study shows that a widely accepted intervention is not effective. Studies that show no effect may be very experimentally rigorous but because they did not show an experimental effect the studies are relegated to the researchers file drawer. These contingencies result in a publication bias for original research that demonstrates a positive effect. This can result in efforts to systematically review the evidence for an intervention over-estimating its effectiveness. Publishing in peer-reviewed journals is a critical component needed to safeguard the quality of research but these biases reflect potential publication biases. Replication is a fundamental cornerstone of science.  Replication studies demonstrate the robustness of a finding. The biases against publishing non-results is a bit more complicated. Some studies that report non-results are unimportant. For example, demonstrating that a car will not run if gas is put in the tires is unimportant. The only important demonstration is one that shows a positive relation between where the gas was put in the car and the car actually running. Other studies reporting non-results are important because they show that a variable that has been experimentally demonstrated to have an impact on student behavior does not have that effect in a replication study or under a particular set of conditions.

News Summary:

  • An Evaluation of a Learner Response System: A Learner Response System (LRS) is a classroom feedback tool that is becoming increasing popular. LRS is the practice of teachers and pupils using electronic handheld devices to provide immediate feedback during lessons. Given that feedback has been found to be a powerful tool in learning, it is not surprising that LRS are being adopted. The important question remains, do LRS increase student performance. This study tests a Learner Response System using Promethean handsets to assess whether it improves student outcomes. The study found no evidence that math and reading were improved using the system for 2 years.

 

  • The Effects of Financial Incentives on Standardized Testing: Standardized testing has increasingly been used to hold educators accountable. Incentives are often offered as a way to improve student test performance. This study examines the impact incentives for students, parents and tutors on standardized test results. The researchers provided incentives on specially designed tests that measure the same skills as the official state standardized tests; however, performance on the official tests was not incentivized. This study finds substantial improvement for performance when there were incentives on the results did not generalize to the official test. This calls into question how to effectively use incentives so they will actually produce desired outcomes.

 

  • Do Teacher Observations Make Any Difference to Student Performance? Research strongly suggests that feedback obtained through direct observations of performance can be a powerful tool for improving teacher’s skills. This study examines a peer teacher observation method used in England. The study found no evidence that Teacher Observation improved student language and math scores.

Citation:

(1) Education Endowment Foundation (2017). Learner Response System. Education Endowment Foundation. Retrieved https://educationendowmentfoundation.org.uk/projects-and-evaluation/projects/learner-response-system/.

(2) John A. List, Jeffrey A Livingston and Susanne Neckermann. “Do Students Show What They Know on Standardized Tests?” working papers (2016) Available at: http://works.bepress.com/jeffrey_livingston/19/

(3) Education Endowment Foundation (2017). Teacher Observation. Education Endowment Foundation. Retrieved https://educationendowmentfoundation.org.uk/projects-and-evaluation/projects/teacher-observation/.

Link:

(1) https://educationendowmentfoundation.org.uk/projects-and-evaluation/projects/learner-response-system/

(2) https://works.bepress.com/jeffrey_livingston/19/

(3) https://educationendowmentfoundation.org.uk/projects-and-evaluation/projects/teacher-observation/