Salah Ukur dan Variabel Tak Terukur dalam Penelitian Ekonomi

Economist growing interest in individual data and nonmonctcr objects has led the concern on measurement errors and unobservable variables. Inaccurate data collection gives more pressure to the study on measurement errors and unobservable variables. Unfortunately, econometrics - the main empirical tools economists use — used to limit its analysis on observable variables and variables measure with little or no errors. Econometrics is convcntioanlly an extension of regression analysis to the problems commonly encountered by economists. The problems of measurement errors and unobservable variables arc surely incompatible with the conventional econometrics. Worse, the lack of sufficient tools sometimes brings economists to ignore the measurement errors by assuming that the errors will be accomodated by the errors in equation (residual). Such a treatment is shown here to produce bias in the estimation of the equation coefficients. A wrong choice of proxies for unobservable variables will also produce biased estimations.