Think about the relationship between and independent variable and a dependent variable that I propose and do the following: Suggest THREE control variables that you would need for each example (do NOT tell me how you would measure them) Each time you suggest one, explain WHY you think it is necessary to control for it.In the above example you would say, “Older people are more likely to be Republican and your partisanship is likely to affect whether you vote for Trump.” I uploaded the document for the assignment that will explain much more. It is a controlling assignment for research in politics class.
controlling_homework_.docx
Unformatted Attachment Preview
Name: ________________________________
Controlling Homework Exercise
UPDATED DUE DATE WED. MARCH 6: hard copy in class (beginning of class time)
In quantitative research, we’re trying to make claims that an independent variable (X) is affecting or
causing the dependent variable (Y) and that this relationship is a pattern across many units being
analyzed. But showing causality requires running an experiment and sometimes we don’t run them
for a number of reasons. What if someone claimed that old people were more likely to vote for
Donald Trump? Is age really driving people to vote for Donald Trump? Maybe older people just also
happen to be more Republican than younger voters and it’s their partisanship that’s affecting whether
they vote for Trump- not their age. In this scenario, when we collect data, we would need to collect
information about each person’s partisanship if we want to see if age and partisanship have effects
on voting for Trump that are SEPARATE FROM ONE ANOTHER. Put another way, does age
have its own effect on voting for Trump that is unique from partisanship? Or is age masking the real
driver: partisanship. This is the essence of “controlling”- finding alternative explanations for what
could be happening and making sure we measure/collect those variables so we can figure out which
one (or both, or neither) has its own effect on the DV.
There are TONS of variables that we can potentially control for in an observational study but we
can never control for everything with this design. So, we need to select a handful of variables that we
think are the MOST important alternative explanations and would have the largest effect on the
dependent variable.
***Assignment***
In the exercise below, I want you to think about the relationship between and independent variable
and a dependent variable that I propose and do the following:
•
Suggest THREE control variables that you would need for each example (do NOT tell me
how you would measure them)
•
Each time you suggest one, explain WHY you think it is necessary to control for it.
o In the above example you would say, “Older people are more likely to be Republican
and your partisanship is likely to affect whether you vote for Trump.”
THE MOST COMMON MISTAKE:
• You need to make sure that the controls you’re proposing make sense given the unit of
analysis. Ask yourself: what is the UNIT being analyzed (individual, town, county, country,
event etc.)? You need to control for characteristics that fit THAT unit and not another unit.
If I say “Better looking candidates get more TV coverage”, the unit is CANDIDATES (who
are individuals). You need to tell me features that candidates have (such as their age or
whether they are incumbents etc.), not features and traits of something else. THIS IS
IMPORTANT.
1. Relationship: Individuals who go to concerts are more likely to vote. Concert
attendance predicts whether people will vote or not.
Name 3 variables you should control for in this relationship and explain why it makes sense
to control for them. I’ll give you the first of the three for this example:
a. We need to control for someone’s INCOME. People who go to concerts can afford
the ticket and therefore may have higher incomes. Maybe people with higher
incomes are more likely to vote and it’s the income, not the concert attendance
affecting the DV. IF YOU MIMIC THE LOGIC OF THIS EXAMPLE FOR
YOUR ANSWERS YOU WILL DO FINE ON THIS ASSIGNMENT.
2. Relationship: Individuals who have met an immigrant are more likely to support
increased level of immigration. Direct contact with immigrants affects immigration
opinion.
Name 3 variables you should control for in this relationship and explain why it makes sense
to control for them.
3. Relationship: Candidates that get more press coverage get more votes. The amount
of coverage a candidate gets affects the candidate’s vote share.
Name 3 variables you should control for in this relationship and explain why it makes sense
to control for them. BE SURE YOU UNDERSTAND THE UNIT OF ANALYSIS FOR
THIS ONE.
4. Relationship: Countries that have Starbucks in them are more likely to have a free
press. Presence of Starbucks affects country press freedom level. Name 3 variables you
should control for in this relationship and explain why it makes sense to control for them.
5. Come up with your own example of a relationship between any Independent
Variable and Dependent variable. You cannot use one we’ve been doing in class or a
slightly altered version of one of the above examples. Tell me:
a. State the relationship
b. Which is the DV and which is the IV in this relationship? You MUST specify
this for the rest to make sense
c. What is the unit of analysis?
d. Name 3 things you would need to control for in this relationship and why
STOP. BEFORE YOU TURN THIS IN, ASK YOURSELF:
DID I MIMIC THE LOGIC OF THE EXAMPLE QUESTION FOR #1 IN MY ANSWERS?
…
Our essay writing service fulfills every request with the highest level of urgency.
attachment
