In a study, researchers compared the effectiveness of two software programs (A and B) which identify different types of land cover by the pixel values in satellite imagery. In fact, there is no difference in the effectiveness of the software programs. On average, each program correctly identifies land cover type for 80% of pixels.

(a) Four satellite images were analysed by each software program. What is the probability that program A correctly identifies a higher percentage of pixels than program B for each of the four images?

(b) If you observed this result (and you didn't know that their effectiveness is equal), would you think that program A is more effective than program B in correctly identifying land cover types?