Statistically Differentiating Between Interaction and Non - Linearity in Multiple Regression Analysis : A Monte Carlo Investigation of A Recommended Strategy  

Oleh:

Kromrey, Jeffrey D. ; Foster-Johnson, Lynn

Jenis: Article from Journal - ilmiah internasional
Dalam koleksi: Educational and Psychological Measurement vol. 59 no. 3 (1999), page 392-413.
Topik: STATISTICAL; statistics; tests; testing
Fulltext: 392.pdf (101.19KB)

Ketersediaan

Perpustakaan Pusat (Semanggi) Nomor Panggil: EE30.3 Non-tandon: 1 (dapat dipinjam: 0) Tandon: tidak ada

Lihat Detail Induk

Isi artikel

The procedure recommended by lubinski and humphreys for differentiating between moderated and non linear regression models is shown to evidence statistical problems that are charateristic of stepwise procedures. A monte carlo study was conducted in which random samples were generated from populations with known non linear or non additive forms. Factors manipulated in the study included the correlation between regressors, regressor reliability, population R (2), effect size of the non linear or non additive components and sample size. Type I error was not adequately controlled under any condition examined and statistical power was poor in many conditions. Results were interpreted in terms of researchers' need to differentiate between exploratory (stepwise regression) and confirmatory (theory testing) aspects of inquiry.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali