To find out what proportion of dogs are given heartworm medication, we survey 50 dog owners by telephone each night for a period of 20 nights. From the data collected, twenty different sample proportions and 95 percent confidence intervals (one for each night's sample) were computed.

1.Do all 20 of the confidence intervals have the same length?

2.Suppose you calculated 90% confidence intervals instead of 95% confidence intervals. Would they be narrower or wider? 3.Suppose you surveyed 100 dog owners by telephone each night instead of 50. Would the width of the 90% confidence intervals constructed using n=100 be narrower or wider than the 90% confidence intervals constructed using n=50 ? 4.What about 99% confidence intervals - for the same sample size, would they be narrower or wider than 95% confidence intervals ?