Department of Economics

Misclassification of the Dependent Variable in Binary Choice Models: Evidence from Five Latin American Countries

Department of Economics
University of Delaware
Working Paper #2007-05

Misclassification of the Dependent Variable in Binary Choice Models:  Evidence from Five Latin American Countries

Evangelos M. Falaris


Misclassification of the dependent variable in binary choice models can result in inconsistency of the parameter estimates.  I estimate probit models that treat misclassification probabilities as estimable parameters for three labor market outcomes: formal sector employment, pension contribution and job change.  I use Living Standards Measurement Study data from Nicaragua, Peru, Brazil, Guatemala, and Panama.  I find that there is significant misclassification in eleven of the sixteen cases that I investigate.  If misclassification is present, but is ignored, estimates of the probit parameters and their standard errors are biased toward zero.  In most cases, predicted probabilities of the outcomes are significantly affected by misclassification of the dependent variable.  Even a moderate degree of misclassification can have substantial effects on the estimated parameters and on many of the predictions.

JEL Classification: C81, C25, O17, J26, J62

Key Words: Data Quality; Misclassification; Formal Sector; Pension Contributor; Job Change; Nicaragua; Peru; Brazil; Guatemala; Panama

The Lerner College is home to a state-of-the-art financial trading facility, the $1.5 million student-led Blue Hen Investment Club, a student-managed restaurant and hotel, a high-technology development center of a global bank and a start-up experience for students with new business ideas.

Undergraduate scholarships, made possible through the generosity of alumni and friends, enable us to support promising scholars. Here, Delaware Gov. Jack Markell and his mother, Leni Markell, join recipients of the William Markell Scholarship (center), which benefits students in Accounting and MIS.

The Institute for Financial Services Analytics is a collaboration between the Lerner College, UD's College of Engineering and JPMorgan Chase, and hosts events for academics and the business community designed to address consumer analytics and industry applications.

All Lerner College departments offer discovery learning experiences and emphasize data-based analytics to enrich the student experience. Here, students at Vita Nova, our award-winning restaurant run through HRIM, joined professional chefs and winemakers in hosting a 2014 Mid-Atlantic Wine and Food Festival event.

Lerner College faculty - like Meryl Gardner, whose research on foods and moods was recently published in the Journal of Consumer Psychology - are thought leaders who command attention from influential business audiences, economists and policy makers.

  • Alfred Lerner College of Business & Economics
    303 Alfred Lerner Hall  •  Newark, DE 19716  •  USA  •  ©