While adverse selection is an important theoretical explanation for credit rationing it is difficult to quantify empirically. Many studies measure the elasticity of credit demand of existing or previous borrowers as opposed to the population at large; other studies use cross-sectional approaches that may confound borrower risk with other factors. We circumvent both issues by surveying a representative sample of microenterprises in urban Uganda and by measuring their responses to multiple hypothetical contract offers, varying in interest rates and collateral requirements. The two seminal theories on selection provide contradicting predictions following a change in the contractual terms. Under adverse selection, a lower interest rate or a lower collateral obligation should increase take up among less risky borrowers. By contrast, advantageous selection implies that take up should increase among the riskier borrowers. We test these two predictions by examining if firm owners respond to changes in the interest rate or the collateral requirement and whether higher take up varies by firms' risk type. We find support for the presence of adverse selection as contracts with lower interest rates or lower collateral obligations increase hypothetical demand - especially for less risky firms. Our results imply that changes to the standard loan product available to microenterprises may have substantial effects on credit demand.