2. Population
The specific group that is being studied is
soufflés.
A soufflé is a lightly baked
cake made with egg yolks
and beaten egg whites
combined with various
other ingredients and
served as a savory main
dish or sweetened as a
dessert. They puff up when
cooked, but can go flat if
not prepared correctly.
3. Experiment Facts
Experimental units: Soufflés
Explanatory variable: The oven door being opened or kept closed
(categorical)
Response variable: Whether or not the soufflé falls (categorical)
Blinding: It is not necessary to use blinding in this experiment,
because nothing will change in the analysis if it is known
that the
door was opened.
Experimental design: Matched pairs, because with varied soufflé
recipes, it will be easier and more accurate to find a difference
between soufflés of the exact same recipe.
4. Experiment Design
200
soufflés
(Two from
each recipe)
e
On
Soufflés
divided
into pairs by
recipe
m
fro
e
hp
ac
air
After randomly selecting 100 soufflé recipes, two
soufflés of each recipe will be made. These pairs
will be divided, with one soufflé being cooked with
the oven door being opened during the bake time
and the other, the control, will be left alone. No
placebo will be needed. The results will be
collected to see if the soufflés receiving the first
treatment fell more often than the soufflés
receiving the second treatment. The experiment
will be conducted by statisticians, but the soufflés
will be prepared by professional cooks to insure
proper preparation.
fro
re
su
Oven opened
during cooking
Randomi
zation!!!
On
e
Co
ll e
ct
m
ea
ch
pa
ir
Compare
t
lec
l
Co
Oven closed
during cooking
l ts
lts
su
re
5. Soufflés rise because they are filled with air bubbles.
This occurs because of the eggs that are used to make
them, and the soufflé will fall when the air bubbles
are popped. Since all soufflés must contain eggs, it
would not make sense to block them by their types
(dessert, entrée, spinach, banana, etc.).
Blocking
For this experiment, one half of each soufflé pair will
receive the treatment of the door being opened, and
the other half will be left alone as the control group.
No other blocking will be used.
6. Method of Randomization
Two hundred soufflé recipes will
be randomly selected from a list.
The list will contain all soufflé
recipes able to be located
through popular recipe
databases. The recipes will be
numbered on the list, and a
computerized random number
generator will select one hundred
recipes.
Various cookbooks and websites
contain a plethora of soufflé recipes,
and individual cooks may have their
own variation. The number of
recipes available makes it difficult to
know how many soufflé recipes
there really are, so it is difficult to
know how to represent this
population.
The matched-pairs design will
help to show clear results in the
experiment. It will be easier to
tell whether the oven door being
opened contributes to the
soufflés’ falling, because both
soufflés in each pair will have the
same composition. This will help
to eliminate lurking variables.
The members of each pair will be
assigned treatment by flipping a
7. Conclusion
There are a few possible
concerns that I have about
this study. One is that there
may be lurking variables to
why the soufflés would fall
flat, such as the skill of the
person preparing them. This
could be prevented by hiring
a team of experienced cooks
to prepare the soufflés.
Another is that the recipes
chosen are not
representative of all soufflés.
This is hopefully minimized
by randomization.
To analyze the data gathered in this
experiment, I would need to
compare the matched pairs and look
at the difference that the treatments
made in the final result. What I
would be looking for would be
whether the soufflés that had the
oven door opened fell more often
than those that did not. I would
analyze the results to see if there
was a statistically significant
relationship between opening the
door and having the soufflé fall. If
there were to be a relationship, it
could be concluded that opening the
oven door causes a soufflé to fall.