Statistics Detective Review Lab
Recall that an underlined phrase means that it is a link off of the
web page itself. The lab can be accessed from
http://www.cs.appstate.edu/~sjg/class/1010/wc/stats/detectives03.html
Use the clues and Excel calculations in order
to eliminate people in the class, until
you are down to one person.
To mark off eliminated people on the computer, click on the row number
(in grey) and then click on B on the formatting toolbar
(if you don't see this then make sure it is checked under View/toolbar/
formatting)
Here are some Excel commands to help you:
The average of data in column G, with numerical entries
in rows 2 thru 46, would be
=average(g2:g46)
The median of the data in column g would be
=median(g2:g46)
The lo of the data (the 0th quartile) in column g would be
=quartile(g2:g46,0)
The first quartile, q1, of the data in column g would be
=quartile(g2:g46,1)
and so on for the other numbers in the "5 number summary".
Use the
clues below and effective reasoning skills in order
to identify the "odd person out" in our
class data set.
Thue suspect...
- hits the 1st or 3rd quartile (q1 or q3)
on the "head" for the number of competitors of
the family bathroom (compute these and write them down here)
- on this histogram of the
number of seconds that breath was held,
falls within the most common class range ("a" less than or
equal to breath held less than "a" + 10)
- falls within the middle 50% of the class on
this boxplot of the
distance from home
- pulls
this regression line for "does height predict armspan" down
since the actual armspan value corresponding to the actual height is
below the
predicted value for the actual height given by plugging into
the linear regression
equation.
- hurts the statistics that show that, on average,
women perform more poorly than
men on the timed test, but confirms the statement
that (even though it was not true for our class),
on average, women do just as well on the untimed test.
(Recall from a previous lab that
the female average for our class on
the untimed test was 1.5, while the male average
was 1.89, and the female average for the timed test was 1.67, while the
male average was 2.)
- has an exact golden mean of (1+sqrt(5))/2 embedded in the
forearm to hand
ratio when the numbers are rounded to one decimal value.
Compute the golden mean rounded to one decimal value.
In Excel, compute the forearm to hand ratios of the remaining
suspects and use the clue to eliminate people.
How many additional people have been eliminated?
How many total people have been eliminated?
Your Approved Suspect
Once you have your guess, call me over and show it to me, and have
your suspect approved.
Wait until your guess has been approved BEFORE you write it down here:
Gender______
Height______
Armspan______
Foot Length______
Head Circ______
Hand______
Forearm______
Smoker______
Class_________________
Siblings______
Home_______
Major_______________________
Minute______
RealWds______
NonsenseWds______
Breath______
Timed MRT______
Untimed MRT______
Verification of the Clues
Explain why each
clue is satisfied by your approved suspect.
There is NO need to explain whether other classmates do or do not
satisfy the clues - you only need to address your approved suspect.
Give detail
that conveys deep understanding of the statistical representations that are
involved.
This means that you will have to include Excel computations in your
explanations and that you should explain calculations
and graphs in depth to show how you can see that your suspect
satisfies the related clue.
Your lab grade for today
will be based on the depth, quality and clarity of your responses.
This is also a good review so you should look up any statistical
representations that you do not understand.
Skim through the major writing assignment directions at the end of this
lab before you answer the following.
I will explain in depth why my suspect
- is not one of the values dragging the mean above the median on this bargraph of the number of nonsense words remembered
(compute the mean and median and write them down here)
- hits the 1st or 3rd quartile (q1 or q3)
on the "head" for the number of competitors of
the family bathroom (compute these and write them down here)
- on this histogram of the
number of seconds that breath was held,
falls within the most common class range ("a" less than or
equal to breath held less than "a" + 10)
- falls within the middle 50% of the class on
this boxplot of the
distance from home
- pulls
this regression line for "does height predict armspan" down
since the actual armspan value corresponding to the actual height is
below the
predicted value for the actual height given by plugging into
the linear regression
equation.
- hurts the statistics that show that, on average,
women perform more poorly than
men on the timed test, but confirms the statement
that (even though it was not true for our class),
on average, women do just as well on the untimed test.
(Recall from a previous lab that
the female average for our class on
the untimed test was 1.5, while the male average
was 1.89, and the female average for the timed test was 1.67, while the
male average was 2.)
- has a golden mean of (1+sqrt(5))/2 embedded in the
forearm to hand
ratio when rounded to one decimal value.
Major Writing Assignment for Statistics - A Persuasive Report
Turn in the above sections (filled in) by the end of lab.
This will be returned to you with
feedback.
Write a persuasive report, convincing the reader that have found
the "suspect" who satisfies all of the clues. The first
draft of your writing assignment,
worth 20% of this writing assignment's grade,
is due Friday April 29th. You may work by
yourself, or with at most one other person. Groups will be self-evaluated,
and the evaluations will be taken into account for your final course average.
Your draft should have a cover page, a table of contents,
the report, references and this packet
(attached as an appendix). The references will be
Greenwald, Sarah J, Math 1010 labs and private communication, 2001.
Math 1010 class data, collected on 2/18/2002.
Microsoft Excel software.
and any other references that you use such as textbooks,
help from other people, ...
Each of the clues
should be a separatly labeled
section in your report, with each section beginning
by telling the reader what you will be presenting. You must also
explain our data collection methods (see Math 1010
Data Collection Sheet), and the relevant statistics and methods
that provide a convincing argument that this clue has been satisfied
by your "suspect". Your explanation needs to convey deep understanding
of the statistical methods involved.
In addition, your report
needs to satisfy the writing checklist,
which means (among other things) that it needs to have an introduction
that introduces the major themes by summarizing what will be done in the paper.
Place the 4 graphs (accessible as links from this lab on the web - the
bar graph, linear regression, histogram, and boxplot) into your
paper into the relevant sections. To satisfy the checklist points
regarding graphs, each graph needs to be placed in the text
after the first time it is mentioned, and each graph must be fully explained.
You report is aimed at someone who does not know statistics.
This means that you must explain in depth each statistical method
(including a summary of how they are calculated and what they mean) and
each graph (including what the graph shows on the x and y axes, how
the data shown on the graph is obtained from the class data,
how to read the graph, ...) used in your analysis.