An investigation of the effects of missing data technique selection on student performance results

Abstract

allowed for a narrowing of focus to useful information. The Need The components of standards-based reform, standards, assessments, and accountability, are the same three components of what has been called the "new educational accountabUity" (Elmore, Abelman, & Fuhrman, 1996). In order to make a "complete" accountability system, Newman, King, and Rigdon (1997) add a fourth component: a body that reviews data, judges against standards, and hands out sanctions and rewards. These systems are increasingly focusing on educational outcome measures, with a primary emphasis on smdent performance (Ebnore, Abelman, & Fuhrman, 1996; Meyer, 1997). With demands on smdent performance data increasing, it is not unreasonable to invest some time addressing the quality of the data. An area of concern in regard to the quality of data in statistical analyses, especially when results are tied to major decisions or policy, is that of missing data (Henry, 1995). Social and behavioral science studies "frequently suffer"" from missing data (Little & Schenker, 1995). The existence of incomplete data in educational research and evaluation is an "undeniable Research Question 7-Posttest by SES Research Question 7 asked whether smdent posttest level of mastery scores on criterion-referenced reading tests would differ between highand low-socioeconomic (SES) students. The research hypothesis addressing this question stated that there would be no significant difference in posttest level of mastery score between high-and low-SES students. Independent t-tests were used to test this hypothesis.