36-201 INTRODUCTION TO STATISTICAL REASONING

Assignment #8  

                Due:  In Lab, Thurs Mar 11 and Fri Mar 12.

Be sure to write your name and SECTION on your paper.  

STAPLE, DO NOT PAPERCLIP OR FOLD/TEAR, YOUR PAGES TOGETHER.

1. Use a WWW Browser (e.g. Netscape) to look at the following polling
report from the Pew Memorial Trusts.

	http://www.people-press.org/tech98sum.htm

Answer the following questions.  A sentence or two at most is needed for
each answer.

(a) The first table of results, near the top of the page, states that
while only 45% of users who started using the interenet over a year ago were
female, 52% of users who started using the internet in the past year
were female.

    (i) Compute margins of errors for each of these percents (note that
    the sample sizes are different!), and use these margins of error to
    fill in the blanks:
      * The percent of females among users who started using the internet
        over a year ago is likely between ___ and ___.
      * The percent of females among users who started using the internet
        less than a year ago is likely between ___ and ___.

    (ii) Was there a statistically significant increase in the fraction 
    of females among users who were new to the internet a year or more
    ago, to the fraction of females among users new to the internet
    in the last year?  (Hint: do your answers for part (i) overlap? Why
    does this help answer the question?)

(b) Now follow the link at the bottom of the page called "The
    Questionnaire and Overall Breakdowns". 

    (i) The second paragraph lists several sample sizes, and a different
    margin of error for each sample size.  Explain to someone who knows
    no statistics how the margin of error changes as the sample size
    changes, and why it changes in that way.

    (ii) Reading on, did the poll takers try to compensate for some possible 
    biases in their sample?  Generally speaking, how did they do this?

(c) The same web page lists the actual questions the poll takers asked
    in their interviews.  Browse through the questions and find a
    question that would be a violation of your privacy, if you were one
    of the people polled and your answers were not kept confidential.
    Explain why keeping that information private or anonymous would be
    important.  [ Obviously, I am not asking that you reveal this
    specific information on the hw.  Just explain why revealing that type 
    of information might matter to a survey respondent. ]

2. Moore, pp. 41-42, problem 1.33.
 
3. Moore, p. 42, problem 1.34 and problem 1.38 [ the answers may be
very different! ]

4. Moore, p. 413, problems 7.3 and 7.6.

5. Moore, p. 414, problem 7.9

6. Moore, pp. 416--417, problem 7.17