You are on page 1of 3

a) Since both types of corn are planted on each of the 33 plots of land, we cannot assume the yields of each

type are
independent. There may be subtle, unknown sources of variation that impact how well corn grows from plot to plot.
It would be reasonable to assume that an increased yield in regular corn on one plot is likely to suggest an increased
yield in GM corn from that same plot.

b) It would be more likely to obtain a yield of 123 or more bushels per acre from a plot of GM corn. This is because the
median of the GM corn yield is 127.5, more than 123. This suggests that more than half of the plots produced a yield
greater than 123. On the other hand, the median for the regular corn yields was 119.4, less than 123, suggesting
that fewer than half of the plots produced a yield greater than 123.

c) The distribution of differences appears to be skewed to the right, with a high outlier plot that produced a difference
of approximately 20 bushels per acre. It is centered at around 3 bushels per acre, and half of the plots produced a
difference in yield from approximately -1 (1 bushel per acre higher for the regular corn) to 6 (6 bushels per acre for
the GM corn)

d) We are conducting inference on the mean difference in yield, and our sample suggests this difference may not be
symmetric. As a result, we need a larger sample size in order to “overcome” this lack of normality in the hypothetical
population of bushel yields.

e) With a p-value of 0.0003, less than any reasonable alpha level, the students can reject the null hypothesis. They can
conclude that there is a difference in yield between GM and regular corn in plots similar to those used in the study.
(In fact, we can be 95% confident that the average yield is 2.25 to 6.82 bushels per acre greater for GM corn than for
regular corn.)

SCORING – Scoring is divided into four sections. Section 1 is composed of part a, Section 2 is composed of part b,
Section 3 is composed of parts c and d, and Section 4 is composed of part e

Section 1 is scored as follows:

Essentially correct (E) if the response includes the following two components:

 The response identifies the planting of both types of corn on each plot of land as the reason for the lack of
independence between yields, AND
 Suggests a reasonable explanation for why yields of corn growing on the same plot of land cannot be considered
independent

Partially correct (P) if the response identifies the sharing of plots as the reason for a lack of independence, but does not
offer a reason why OR discusses how yields of corn from the same plot may be related, but does not explicitly tie this to
the question of independence

Incorrect (I) if the response does not meet the specifications for an E or P
Section 2 is scored as follows

Essentially correct (E) if the response includes the following three components:

 The response states that GM corn is more likely to produce a yield of 123 bushels or more, AND
 Gives a statistical explanation for why GM corn is more likely (using mean or median), AND
 Gives a statistical explanation for why regular corn is less likely to produce such a yield

Partially correct (P) if only 2 of the three components are included.

Incorrect (I) if the response does not meet the specifications for an E or P

Section 3 is scored as follows:

Essentially correct (E) if:

 The response in part c includes reasonable comments on four or five the following five components
o Shape (skewed right)
o Outlier (at least one) or gap (one yield around 20, next highest around 15)
o Center (approximately 3 bushels per acre)
o Variability, by noting that yields vary from about -2 to 20, or that the central 50% of yields vary from -1
to 6.
o Context: bushels per acre
 The response in part d contains a reasonable explanation for the necessity for n > 30

Partially correct (P) if the response to part c includes only three of the needed components AND the response in part d
contains a reasonable explanation for the necessity for n > 30 OR the response to part c includes four or five of the
needed components AND the response to part d does not contain a reasonable explanation for the necessity for n > 30

Incorrect (I) if the response does not meet the specifications for an E or P

Section 4 is scored as follows:

Essentially correct (E) if the response includes the following three components:

 Makes an explicit statement about the size of the p-value, either by comparing it to an explicit or implied
(“reasonable”) significance/alpha level, or by generally stated the p-value is small
 Makes the correct decision based on the size of the p-value (“Reject the null hypothesis”)
 Correctly states the conclusion in context

Partially correct (P) if only 2 of the three components are included.

Incorrect (I) if the response does not meet the specifications for an E or P
Each essentially correct (E) step counts as 1 point. Each partially correct (P) step 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 2 ½ points), use a holistic approach to decide whether to score up or
down, depending on the overall strength of the response and communication.

You might also like