We examine the gift exchange hypothesis on both the quantity and quality of output using a hybrid field-laboratory labor market experiment. We recruited participants to enter survey data for a well-known charitable organization. Workers were paid either a high or low wage. We find that although the total number of surveys entered did not vary with the wage, high wage workers made fewer errors and entered more surveys after controlling for errors. We further find that for low costs associated with errors, offering the low wage maximizes profits, but for higher costs paying the higher "gift exchange" wage maximizes profits.