Homework 1 Solutions

Ch. 2: 
 
4.(a) They were controlling for possible confounding factors. 
 (See Freedman, p 13). 
 
  (b) This is the wrong conclusion to draw.  Ex-smokers are a 
self-selected group, and many people give up smoking because they
are sick.  Thus, the group of recent ex-smokers may include a higher
proportion of sick people than the group of current smokers.  


11. (a) treatment group - inmates completing the "boot camp"; 
        control group - inmates who didn't choose "boot camp" and
inmates who didn't finish "boot camp"
    (b) observational study - Inmates volunteered to join the "boot 
camp"; the researchers did not assign groups to treatment or control. 
    (c) The data do not show that "boot camp"; reduces the recidivism
rate.  It could be that the inmates who volunteered for the "boot 
camp"; were the inmates who were the most motivated to reform
themselves.  This difference in the intentions and motivations of the 
control and treatment groups may explain the correlation between the 
"boot camp"; experience and lower recidivism rates. 
 


Ch. 3: 

7. Histogram (i) shows distribution of age at death from natural
causes.  Not as many young people as old die from natural causes
(disease, cancer, etc.), so we we know the histogram with the longer
left tail is the match for this.  We would expect most of the young
people who die would die from trauma, so associating histogram (ii)
with these deaths makes sense.


12. Not necessarily.  It could be that temperatures rarely reach very
high temperatures, so even if a riot occured every day that we
experienced such high temperatures, the histogram would still look
this way.  The investigators should have looked at the number of riots
in each temperature range divided by the number of days in each
temperature range.


 
Additional problems: 
 
1. (a) This was a correlational (observational) design.  The
investigator did not assign the subjects in the treatment or control
group.  He just watched what happens and found the correlation.   
  (b) The lead exposure (degree of blood lead) is the independent 
variable.  It is quantitative and continuous. 
  (c) The dependent variables are I.Q. and achievement scores.  Both 
are quantitative and continuous. 
  (d)  An entire community may have in common certain things which
are absent from other communities.  Possible confounding variables
might include hazardous factors in the environment, such as water they
drank, or the material of the apartments or houses in which they live,
and so on.