Exploring Missing Data Substitions for Binomial Response Student Test Data

• Activity Number: 186
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Social Statistics Section
• Author(s): Christopher Gonwa-Reeves*+
• Keywords: Binomial Data ; Mean Substition ; Model Comparison ; Missing Data ; Imputation ; Education Statistics

Abstract

As part of a four year study to test the effectiveness of a teaching technique called the Science Writing Heuristic (SWH) on 3rd-5th grade students, 48 school buildings accross Iowa have been split between treatment and control groups, where the treatment schools have adopted SWH and the control schools are teaching science as previously. To evaluate the effectiveness of SWH scores on the Iowa Test of Basic Skills (ITBS) and the Cornell Critically Thinking Test (CCTT) were collected. The CCTT was given to 5th graders and the being and end of the school year. For each student the answer they gave for each of the questions was recorded (including missing). Mean substitutions (whether by student, sub-set of students, question, sub-set of questions) are common for this type of missing data. For this study the above four mean substitutions as well as another substitution method are applied to CCTT and then compared by fitting various models to see how they deviate.