Chapter 12, Problem 35SE

The earth’s temperature affects seed germination, crop survival in inclement weather, and many other aspects of agricultural production. Temperature at various locations can be measured using ground-based sensors or infrared-sensing devices mounted on aircraft or space satellites. Ground-based sensoring is tedious and requires many replications to obtain accurate estimates of ground temperature. On the other hand, airplane- or satellite-mounted sensors appear to introduce a bias in temperature readings. To estimate the amount of bias, both methods were used to measure ground temperature at five locations. The readings, measured in degrees Celsius, are given in the following table.

1. a Do the data present sufficient evidence to claim a difference in average ground-temperature readings using ground- and air-based sensors?
2. b Construct a 95% confidence interval for the difference in mean ground-temperature readings using ground- and air-based sensors.
3. c We want to estimate the difference between mean temperature readings for ground- and air-based sensors to within .2^°C at the 95% confidence level. Approximately how many paired observations (measurements at different locations) are required?