Transcript Q5.pptx

You cannot be serious?!?
Question #5
Intent of Question
The primary goals of this question were to
assess a student’s ability to
(1) calculate proportions from a two-way table;
and
(2) interpret a confidence interval for the
difference between two proportions.
Question #5
In 2006, tennis introduced a challenge system in which a player can
challenge a decision as to whether a tennis ball was correctly called
‘in’ or ‘out’ by an official. In the 2015 U.S. Open Tennis
Championship there were a total of 850 player challenges between
both the Men’s and Women’s Singles matches. These challenges
were determined either to be correct (meaning the player was
correct and the umpire was incorrect) or incorrect (meaning the
player was incorrect and the umpire was correct). These 850
challenges are summarized in the table below.
Player Challenge
Men’s
Women’s
153
72
394
231
Correct
Incorrect
Part (a)
Part (a)
Calculate the proportion of all
challenges that were
determined to be correct.
Part (a) solution
The proportion of all
challenges that were
determined to be correct is
153+72
225
=
≈ 0.265.
850
850
Part (a) Scoring
Part (a) is scored as follows:
Essentially correct (E) if the response correctly calculates the
proportion with work shown.
Partially correct (P) if the response provides the correct answer with
no work shown.
Incorrect (I) if the response does not meet the criteria for E or P.
Note: Answers reported as fractions rather than decimals are
acceptable and can be considered to have work shown.
Question #5 continued
(b)Using these data as a representative sample of
all player challenges in professional tennis, a 95%
confidence interval for the difference in the
proportion of men’s and women’s challenges that
are determined to be correct (men – women) was
found to be 0.04 ± 0.06. All conditions for inference
were met. Does this confidence interval provide
convincing statistical evidence that there is a
difference in the effectiveness with which men and
women use the challenge system? Justify your
answer.
Part (b) solution
No. The confidence interval is (−0.02, 0.10), which
includes the value of zero. Therefore, it is plausible
that the proportion of all men’s challenges that will
be determined to be correct is equal to the
proportion of all women’s challenges that will be
determined to be correct, and the confidence
interval does not provide convincing statistical
evidence that there is a difference in the
effectiveness with which men and women use the
challenge system.
Part (b) scoring
Part (b) is scored as follows:
Essentially correct (E) if the response states that because the interval contains 0, it does not provide
convincing statistical evidence that there is a difference in the effectiveness with which men and
women use the challenge system.
Partially correct (P) if the response indicates that it is necessary to check whether the value of 0 is in the
computed interval, but there are errors in implementation. Examples of errors include:
The response notes that 0 is within the interval but does not draw a conclusion.
The response has an arithmetic error in the computation of the endpoints of the interval but provides a
correct conclusion with justification that is consistent with the computed interval.
Incorrect (I) if the response does not recognize how to use the confidence interval to determine
whether there is a difference in men and women;
OR
if the response states that the interval shows that the difference in proportions is equal to 0;
OR
if the response notes that 0 is within the interval and concludes that there is a difference in proportions;
OR
if the response otherwise does not meet the criteria for E or P.
Question #5 continued
A scientific study has
demonstrated that the best time
for a player to use a challenge is
when the tennis ball was called
‘out’ but the player believes that
the ball was actually ‘in’ because
when objects travel faster than
the human eye the umpire is left
to fill the gap with their own
perception. The table below
shows the Women’s Singles
challenges based on if the tennis
ball was called ‘out’ or ‘in.’
Player Challenge
Umpire Called
Ball Out
Umpire Called
Ball In
61
11
111
120
Correct
Incorrect
Part (c)
Part (c) solution
Calculate the proportion of
correct player challenges when
the umpire called the ball ‘out’
and the proportion of correct
player challenges when the
umpire called the ball ‘in.’
The proportion of correct
player challenges when the
umpire called the ball ‘out’ is
61
61
=
≈ 0.355.
Out proportion:
In proportion:
61+111
172
The proportion of correct
player challenges when the
umpire called the ball ‘in’ is
11
11
=
≈ 0.084.
11+120
131
Part (c) scoring
Part (c) is scored as follows:
Essentially correct (E) if the response correctly performs both
calculations with work shown.
Partially correct (P) if the response correctly performs one of the two
calculations with work shown;
OR
if the response provides both correct answers with no work shown.
Incorrect (I) if the response does not meet the criteria for E or P.
Question #5 continued
Part (d)
Part (d) Solution
Using these data as a representative
sample of all player challenges in
professional tennis, a 95%
confidence interval for the difference
in the proportion of correct
challenges when the umpire calls the
ball ‘out’ and when the umpire calls
the ball ‘in’ (out – in) was found to
be 0.27 ± 0.09. All conditions for
inference were met. Does this
confidence interval provide statistical
evidence to support the scientific
findings detailed above? Justify your
answer.
Yes. The confidence interval is
(0.18, 0.36). Because this interval is
entirely above zero, this suggests
that the proportion of correct
challenges is higher when the umpire
calls the ball ‘out’ than when the
umpire calls the ball ‘in,’ and the
confidence interval does provide
statistical evidence to support the
theory that the best time for a player
to use a challenge is when the tennis
ball was called ‘out’ but the player
believes that the ball was actually
‘in.’
Scoring part (d)
Part (d) is scored as follows:
Essentially correct (E) if the response concludes that there is statistical evidence to support the scientific
findings because the interval is entirely above 0.
Partially correct (P) if the response indicates that it is necessary to determine whether the interval is
entirely above 0 but there are errors in implementation;
OR
if the response notes that 0 is not in the interval and concludes that there is statistical evidence of a
difference in proportions, but does not address the direction of the relationship between the two
proportions;
OR
if the response notes that the interval is entirely above 0 but confuses the order of subtraction and
argues there is statistical evidence that the better time to challenge is when the ball is called 'in.'
Incorrect (I) if no argument is made based on whether or not 0 is in the interval;
OR
if the response otherwise does not meet the criteria for E or P.
Scoring
Each essentially correct (E) part counts as 1 point. Each partially correct (P)
part counts as ½ point.
4
Complete Response
3
Substantial Response
2
Developing Response
1
Minimal Response
If a response is between two scores (for example, 2½ points), score down.