PENGARUH METODE DAN UKURAN SAMPEL TERHADAP VARIANSI SKOR HASIL PENYETARAAN

Abstract

Penelitian ini bertujuan untuk memperoleh informasi perbedaan variansi skor hasil penyetaraan (equating) metode linear dan metode eki-persentil untuk ukuran sampel 200, 400, dan 800 pada Ujian Akhir Sekolah Berstandar Nasional (UASBN). Metode yang digunakan adalah simulasi de-ngan variabel metode penyetaraan dan banyaknya responden. Data pene-litian berupa respons peserta UASBN SD/MI tahun pelajaran 2008/2009 mata pelajaran IPA di Jakarta Timur, yang ditentukan menggunakan teknik penarikan sampel acak dengan pengembalian. Hipotesis diuji menggunakan uji kesamaan variansi. Hasil penelitian dengan α = 0,05 menunjukkan: (1) variansi skor penyetaraan metode ekipersentil (σ2ekp200) tidak berbeda de-ngan variansi skor penyetaraan metode linear (σ2lin200) untuk ukuran sampel 200, (2) variansi skor penyetaraan metode ekipersentil (σ2ekp400) tidak ber-beda dengan variansi skor penyetaraan metode linear (σ2lin400) untuk ukuran sampel 400, dan (3) variansi skor penyetaraan metode ekipersentil (σ2ekp800) berbeda dengan variansi skor penyetaraan metode linear (σ2lin800) untuk ukuran sampel 800. Kata kunci: variansi skor, equating, metode ekipersentil, metode linear ______________________________________________________________ THE IMPACT OF METHODS AND SAMPLE SIZE TO THE SCORE VARIANCE OF EQUATING RESULT Abstract This study was aimed to obtain information on the difference of score variance as a result of equating linear method and equipercentile method for the sample size of 200, 400, and 800 in the Final Examinition of National Standardized Schools. The method used was a simulation of variables equating method and the number of respondents. The population are examinees from the 2008/2009 elementary school final examination for science class in East Jakarta. Random sampling with replacement technique was used. The hypotheses were tested using similarity variance. The results with α = 0,05 showed that: (1) the equated score variance from equiper-centile method (σ2ekp200) was not different from the equated score variance from linear method (σ2lin200) for the sample size of 200, (2) the equated score variance from equipercentile method (σ2ekp400) was not different from the equated score variance from linear method (σ2lin400) for the sample size of 400, and (3) the equated score variance from equipercentile method (σ2ekp800) was different from the equated score variance from linear method (σ2lin800) for the sample size of 800. Keywords: score variance, equating, equipercentile method, linear method