Chapter 2

#7 (a) this is observational, confounding may be a problem (b) rates of cervical cancer go up with age, women of different marital statuses have different patterns of sexual activity and they are exposed to different kinds of risk; this is the same for education. These are potential confounders. (c) Pill users are more sexually active than non-pill users and have more partners, this may be making cancer higher among pill users. (d) no.

#9 (a) false (b) true (c) false -- this is why we experiment. People who eat lots of fruits and vegetables are different from the rest of us in many other ways. Some other aspect of their lives may be protecting them. Of course, it is possible that the observational study is correct.

Chapter 3

#1 66 inches and 72 inches

#4 (a) 25%, (b) 99%, (c) 140-150mm, (d) 135-140mm (e) about 5 x 2.1 = 10.5% (f) 102-103 mm (g) guess between 115 and 120mm

#7 (i) natural causes (ii) trauma. Reasoning - youn gpeople tend to die of accidents and murders etc. Old people usually die of disease.

#8 (a) true but (b) and (c) flase. The figure does not adjust for the different widths of the class intervals and is misleading.

Chapter 4

#1 Average = 50, SD =5, (b) 48, 50 and 50 are within 0.5 SD of average, 48, 50, 50, 54, 57 are within 1.5 SD

#6 (a) (i) 60, (ii) 50, (iii) 40

(b) (i) median > average (ii) median = average (iii) median < average

(c) 15

(d) False. They have the same SD.

Page 428

#1. No, teachers might be tempted to put disadvantaged kids in treatment.

#3. a) yes, b) (ii) see section 2.5

#5 (iii)

Page 568 #4 this is not a histogram, if you fix it, the pattern goes away. It just shows what you know, most students are in their late teens and early twenties...

#6 The difference between the 90th and 50th is bigger because this distribution has a long right handed tail.