Kesalahan Pengukuran Dalam Koefisien Regresi Linear: Perbandingan Antara Penggunaan Raw Score, Factor Score, Dan Structural Equation Modeling

Hasniar A. Radde

Abstract


Abstract

This research's objective is to determine effect of measurement error to linear regression coefficient value by comparing use of raw score, factor score, and Structural Equation Modeling. Measurement error can be determined by reliability coefficient value. This research simulates Monte Carlo study which varies reliability values. Data is generated using parallel measurement with variations of reliability value, consisting 40 items and 500 respondents, replication was conducted 50 times, and regression coefficient value's set in 0.8. this study used Mplus to generate the data in desired characteristic. This study used Confirmatory Factor Analysis and SEM analysis method. Regression coefficients resulted on the scores then was compared, which value is the closest to 0.8. The result indicates that attenuation occurs on all the model of raw score. Whereas on factor score, attenuation only occurs on model with IV's reliability 0.5 to DV's reliability 0.5 ; 0.7 ; 0.9. On SEM analysis, attenuation doesn't occur on all model. 

Abstrak

Penelitian ini bertujuan untuk melihat pengaruh kesalahan pengukuran terhadap nilai koefesien regresi linear dengan membandingkan penggunaan raw score, factor score, dan Structural Equation Modeling. Kesalahan pengukuran dilihat dari nilai koefesien reliabilitas. Penelitian ini merupakan studi simulasi Monte Carlo dengan memvariasikan nilai reliabilitas. Data yang dibangkitkan mengikuti pengukuran paralel dengan variasi nilai reliabilitas, terdiri atas 40 item dan 500 responden, replikasi sebanyak 50 kali, dan nilai koefesien regresi ditetapkan sebesar 0,8. Penelitian ini menggunakan software MPlus untuk membangkitkan data sesuai karakteristik yang diinginkan. Penelitian ini menggunakan metode analisis faktor konfirmatorik dan SEM. Koefesien regresi yang dihasilkan pada ketigas jenis skor kemudian dibandingkan, nilai mana yang paling mendekati 0,8. Hasil penelitian menunjukkan bahwa atenuasi terjadi pada seluruh model untuk raw score. Sedangkan pada skor faktor, atenuasi hanya terjadi pada model dengan reliabilitas independent variable 0,5 terhadap dependent variable reliabilitas 0,5 ; 0,7 ; 0,9. Sedangkan pada analisis SEM, tidak terjadi atenuasi untuk keseluruhan model.

Keywords


Kesalahan Pengukuran, Koefesien Regresi Linear, Skor Mentah, Skor Faktor, Stuctural Equation Modeling

References


DAFTAR PUSTAKA

Allen, M.J., Yen, W.M. (1979). Introduction to measurement theory. California : Brooks/Cole Publishing Company.

Brown, R.L. (1989). Congeneric modeling of reliability using censored variables. Applied Psychological Measurement. 13 (2), 151-159.

Field, Andy. (2006). Discovering statistics using SPSS. California : Sage Publication.

Guilford, J.P., Fruchter, B. (1981). Fundamental statistics in psychology and education. Singapore : McGraw-Hill Book Company.

Joreskog, KG & Sorbon. (1996). Lisrel 8 user’s reference guide. Chicago : Scientific Softwere International Inc.

Miller, M.B. (1995). Coeffecient alpha: A basic introduction from the perspective of Classical Test Theory and Structural Equation Modeling. Structural Equation Modeling. 2 (3), 255-273.

Muthen, Linda K., Muthen, Bengt O. (2012). MPlus user’s guide. Los Angeles : www. statmodel.com. Di akses pada tanggal 21 Oktober 2013.

Muthen, Linda K., Muthen, Bengt O. (2002). How to use a monte carlo study to decide on sample size and determine power. www.statmodel.com. Di akses pada tanggal 21 Oktober 2013.

Pedhazur, EJ. 1997. Multiple regression in behavioral research. USA : Thomas Learning Inc.

Raykov, T. (1997). Estimation of composite reliability for congeneric measures. Applied Psychological Measurement. 21 (2), 173-184.

Socan, G. (2000). Assessment of reliability when test items are not essentially τ-equivalent. Ljubljana : Development in Survey Methodology.

Umar, J. (2012). Mengenal lebih dekat konsep reliabilitas. Jurnal Pengukuran Psikologi dan Pendidikan Indonesia. 2 (2), 126-140.

Umar, J. (2012). Peran pengukuran dan analisis statistika dalam penelitian psikologi. Jurnal Pengukuran Psikologi dan Pendidikan Indonesia. 1 (1), 47-55.

Wijanto, SH. (2008). Structural equation modeling. Yogyakarta : Percetakan Graha Ilmu.


Full Text: PDF

DOI: 10.15408/jp3i.v4i3.9306

Refbacks

  • There are currently no refbacks.


Copyright (c) 2015 Hasniar A. Radde

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.