ANSWERS TO PRACTICE QUESTIONS

PART ONE

1. No, there is no control for comparison. Hangovers always go away some time.

2. A control group, given a placebo, would help.

3. If the y-axis is percent per thousand dollars of income, the heights of the bars should be:

$0-$10000	2
$10000-$25000	2
$25000-$50000	1
$50000-$100000	0.5

4. SD = 15.2. Misleading because of outlier.

5. 16%

6. Whether she's correct or not is arguable. It depends on whether the errors in jellybean counts are like draws from a box, and in particular, whether counts are independent.

PART TWO

1. 0.916

2. Still 0.916

3. Predicted weight = 1.5 * height - 3.7

4. Prediction = 6.05 lb, new SD = 0.927, z = 0.49. Probability = 31.2%.

5. No: 2 inches would be extrapolating (the prediction would be negative).

6. Small sample size is the main thing to worry about.

PART THREE

1. 2/9 or 22%

2. 7/9 or 78%

3. 1/2

4. 1/2

5. 14/9 or 1.56

6. 1/10

PART FOUR

1. (b), the law of averages.

2. (a), the law of averages.

3. 56

4. 4.96

5. 21% (answer may differ slightly if you use a continuity correction)

6. Response bias: minimise by allowing anonymous responses. Non-response bias: minimise by interviewing in person. (other answers are possible.)

7. EV = 153, SE = 8.66

8. 34.3% to 45.7%

9. 26.4 to 33.6

10. The distribution of songs downloaded by students is likely to be highly right-skewed.