Specifying a Research Question. Searching for “What causes y to vary” Facts are interesting but research into facts is usually limiting. Usually they are a building block in a more significant research activity. Questions. Causal. Factual.
What percentage of registered voters cast a ballot in 2008?
How much trash did citizens of Austin and Houston recycle last year?
What did Obama do to respond to the Arab Spring in Tunisia, Egypt & Libya?
“Rules” for Hypotheses
1. Identify 2 variables and how they are (plausibly) related
2. Clear, concise, unambiguous
3. Communicate the population of interest
4. Be as general as possible
5. Should not include value statements
6. Should be empirical (about the real world)
7. Should be testable
1. Null (A statement of no relationship. Typically used in hypothesis testing as a baseline to be disproven.)
2. Correlative (variables are related but the causal status of each is unknown)
As X increases, Y increases (or decreases)
(Category A of X) tends to show more of Y than (Category B of X).
(Category A of X) tends to show more of (Category C of Y) than (Category B of X).
(Category A of X) tends to show more of (Category C of Y) while (Category B of X) tends to show (Category D of Y).
As (education) increases, (tolerance) increases.
(Poor citizens) tend to show more (alienation from the political system) than (wealthy citizens).
(Older Americans) are more likely to (vote) than (younger Americans).
(Older Americans) tend to (vote) while (younger Americans) tend to (abstain from voting).
1. Statements with only one variable.
Citizens are alienated from the political system.
More tolerant people are less prejudiced than less tolerant people.
2. Unclear relationships.
Urban families have low incomes.
Urbanization is related to literacy.
3. Assertions lack generality.
a. Avoid Predictions
e.g. More Incumbents will be elected this Fall than non-incumbents.
b. Avoid Personal names
e.g., Abraham Lincoln was a better President than Grover Cleveland.
4. Value judgments.
e.g., Conservatives are better than liberals.
“Cause” = “Producing”
If X is a cause of Y, then a change in X “produces” or “forces” (rather than is simply “followed by”) a change in Y.
If X causes Y, then a change in X produces a change in Y. This doesn’t mean a change is Y produces a change in X.
Two Types of Arrows: