Stat 10, UCLA
Chapter 6 Solutions

    1. The tape may have stretched or perhaps moved.
    2. Cloth
    3. Yes, as the tape stretches or the cloth moves.

    1. False: chance errors are sometimes positive, sometimes negative. Bias pushes one way; if you, for example, consistently sample the largest rabbits in a population, then you will always over-estimate the size of the rabbits in the true population.
    2. False: same reason.
    3. True.

  3. About 0.03 inches.
    1. No, the students did not work independently of one another. Students 2 and 10 both got exactly the same answers, and both of these students recorded their answers to the wrong decimal place (2 places instead of 3). The other students seem fine.
    2. Tell your friend that these are all measurements on the same thing, and yet they differ so much from one another. No two people who worked independently got the same answer!!!

Special Review Exercises

S6.
  1. (600+650)/2 = 625.
  2. It would be more than 125 because the distributions are different, so when they are combined there will be more variability.

S8.
It would be too low, because the curve is lower than the histogram in that range.

S10.
False. The data are cross-sectional. One possible explanation is that when those 70 year olds were young, people were forced to write with their right hand, now they are not. To decide what happens to people as they age, you need a longitudinal study.

S13.
  1. If the drivers were older, they might just have higher rates heart disease because of age. (Older people have higher rates of heart disease).
  2. Any effects of exercise might take a while to show up.
  3. Drivers are more likely to be similar to conductors with respect to age, education, income, socioeconomic background, ethnicity, etc.
  4. This is an observational study and there are several plausible confounding factors. For example, drivers might have been heavier to begin with.
  5. See if the drivers were heavier (had larger uniforms) than the conductors at the time of hire. If both groups were the same, initial weight is not the confounding factor. If drivers were heavier to begin with, the argument that exercise matters is much weaker. Note: looking at size after 10 years does not help us to decide.

http://www.stat.ucla.edu/~dinov