Abstract
The multilevel data analysis, a conceptually and technically appealing statistical procedure. is becoming increasingly popular among researchers who focus on evaluation of educational programs. student achievement and assessment of growth in achievement (Sheltzer. 1995). Although application of multilevel model in econometrics, biometrics. and sociological research is not novel. its application to educational data is a very recent development. To date. there is only a small but grovving number of studies utilizing this model in both developing and developed countries (Aitkin & Zuzovsky. 1991. 1992: Bernstein. 1990: Goldstein. 1989: Lockheed & Longford. 1989. 1991: Mislevy & Bock. 1989;Raudenbush&Bryk. 1989. 1992: Riddel. 1989). Since the first publication of Coleman Report in 1966. tlıere has been an increasing research focus on examining thc detenninants of student achievement. Research fındings regarding sources of variation in student achicvement among schools have been inconelusive and contradictory in many cases. However. thesc fındings have been utilized in development and dcsıgn of educational policics and programs for improving student achicvement in pııblic schools Most studies use regression based analysis methods in examining thc relationship bcrvveen achievement outeomes and various SES and school related variables. As noted by Madaus and others (1980) existence of interrelations among explanatory variables can result in biased estimation of regression coeffıcients. Policy implications drawn from these studies regarding \vhat constitutes an effective school may be misleading.
Keywords
Abstract
The multilevel data analysis, a conceptually and technically appealing statistical procedure. is becoming increasingly popular among researchers who focus on evaluation of educational programs. student achievement and assessment of growth in achievement (Sheltzer. 1995). Although application of multilevel model in econometrics, biometrics. and sociological research is not novel. its application to educational data is a very recent development. To date. there is only a small but grovving number of studies utilizing this model in both developing and developed countries (Aitkin & Zuzovsky. 1991. 1992: Bernstein. 1990: Goldstein. 1989: Lockheed & Longford. 1989. 1991: Mislevy & Bock. 1989;Raudenbush&Bryk. 1989. 1992: Riddel. 1989). Since the first publication of Coleman Report in 1966. tlıere has been an increasing research focus on examining thc detenninants of student achievement. Research fındings regarding sources of variation in student achicvement among schools have been inconelusive and contradictory in many cases. However. thesc fındings have been utilized in development and dcsıgn of educational policics and programs for improving student achicvement in pııblic schools Most studies use regression based analysis methods in examining thc relationship bcrvveen achievement outeomes and various SES and school related variables. As noted by Madaus and others (1980) existence of interrelations among explanatory variables can result in biased estimation of regression coeffıcients. Policy implications drawn from these studies regarding \vhat constitutes an effective school may be misleading.
Keywords
Details
| Primary Language | Turkish |
|---|---|
| Journal Section | Articles |
| Authors | |
| Publication Date | August 1, 1996 |
| Published in Issue | Year 1996 Volume: 7 Issue: 7 |