Suppose you are an employer for a large firm that hires only candidates with undergraduate college degrees. You are interested in measuring the effect of an undergraduate degree at a top-tier university (vs. an undergraduate degree elsewhere) on first-year performance with your firm. You have data for several years on new hires' degree institutions and first-year performances. However, you believe the data-generating process for first-year performance looks like: Performance i = + 1TopDegreei + 2Extroverti + Ui.

Here, Top Degree is a variable that equals one if the new hire got a degree from a top-tier university, and Extrovert is a measure how extroverted the new hire is; it is believed greater extroversion leads to greater performance for this particular job. You worry that Extrovert may then be a confounding factor in a regression of Performance on TopDegree. How might you use a proxy variable to address this concern? (LO4)