Theory Construction

This assignment is based on the following theorizing handouts:

An example of how to do this assignment can be found here. Unlike the other quizzes in this class, this assignment can be done individually or in teams.

The assignment begins with a simple observation (given at the bottom of the
page below) about the world which we
will assume is true. **Do not attempt to change the observation nor test it**.
Just take it as a fact, even if you don't believe it. Then you have to generate
two
theories that account for (i.e., explain) that fact. To do this, think of the fact as the outcome of
an unknown process. Your job is to deduce that process. Focus on the process (also called
"mechanism").

Note that the theory is not about the observation. If the observation were "Amy
gained a lot of weight during Freshman year", the theory would be about how people
gain weight -- it is not really about Amy. For example, the explanation might be "**When
people are lonely, they eat. When they eat, they gain weight**. ** Many freshman are
away from all their friends and family for the first time, so they get lonely**." It is
really the first part which is the theory. See how that part is completely general. It's
only at the end that we relate the observation (that freshman gain weight) to the theory
(how people in general gain weight). The observation falls out as an implication of the
theory. Because of the general way that the theory is stated, it easily generates other
implications that can be used for testing. For example, freshmen who live in far away
states or countries should gain more weight, on average, than those that grew up in Newton
(because they don't have their friends and family with them). Or, freshmen who make
friends easily will gain less weight (because they won't be lonely for long).

Then, generate at least two implications for each theory. Implications always have the form: "If the theory is true, that would mean that we should see X occurring." Implications can be used to test a theory, because if we don't see X occurring, the theory is wrong.

Finally, devise a minimal set of empirically testable critical questions (they can be based on the same implications generated in part 2) that, taken together, will allow us to tell which theory is supported.

Theory 1 | Theory 2 | |
---|---|---|

Question 1 | Yes or No | Yes or No |

Question 2 | Yes or No | Yes or No |

Question 3 | Yes or No | Yes or No |

Question 4 | Yes or No | Yes or No |

Where it says "Question #" you write a short version of the implication in the form of a question. For each cell in the table (where it says "Yes or No") you put a "yes" if the empirical results are consistent with the theory in that column and a "no" if not. Don't put "Yes or No" -- pick one or the other. To figure out which to put for a given implication and a given theory, ask yourself: If this theory is absolutely true, would it mean that this (the implication) would be happen? For instance, one of your questions might be, "Will football players appear to be dumb off-season as well as on-season?" Then you look at each theory and think 'what does this theory predict about this? Should the football players continue to appear dumb during the off season?'.

The initial observation that you must explain is:

- BC is suffering grade inflation --- i.e., the average grade for
__all__students is increasing a little bit each year.

**Step 1**. State at least two theories that would explain this fact.
The theory should be expressed in general terms (don't limit to BC). Focus on the underlying mechanism.

**Step 2**. List at least two implications (besides the original
observation) of each theory which can be used
to test the theory. An implication is an expectation or prediction generated by the
theory. It must follow logically from the theory.

**Step 3**. Build a table, as described above, that relates each
implication to each theory. Make sure that, taken together, the implications distinguish
between all the theories. In other words, no two columns of the table can have the same
pattern of Yes's and No's. If the set of four implications generated in Step 2 do not
distinguish between the theories, add more until they do.

To see an example of how to do this homework, click here.

Copyright ©1996 Stephen P. Borgatti | Revised: October 03, 2000 | Home Page |