A taxicab company wants to test two programs for improving the gasoline mileage of its drivers. 20 drivers are randomly assigned to two programs. Under program A, drivers are given a target mileage and receive modest bonuses for better performance. Under program B, drivers are allowed a maximum monthly quota of gasoline; if it runs out, a driver has to pay for extra gasoline out of pocket. All taxis used are standard models and they are given standard maintenance. After three months, each driver's mileage per gallon is calculated. The data are shown below. Assume that the data are normally distributed. Do these data provide sufficient evidence at the 10% level of significance to infer that program A yields a higher mileage than program B?