Political Science 5 Lecture

1
Political Science 5Lecture 7, 2/17/04

• Homework 2 is posted
• Midterm two weeks from today study guide will be
  posted soon. It will cover text Chapters 1-6,
  and all material and readings for Lectures 1-9.
  Review will be Thursday next week.

2
Review Characteristics of a Good Hypothesis

• It must state an empirical relationship between
  an independent and dependent variable
• Relationship can be positive or negative
• Generality
• Plausible
• Directional
• Consistent with data
• Testable
• Add should be derived from a theory

3
What about hypotheses that contain nominal
variables?

• Example IVreligion, DVsupport for tax cuts
• Example IVplace of residence, DVparty
  affiliation
• Example IVwealth, DVparty affiliation

4
Hypotheses with Antecedent or Intervening
(control) Variables

• Question
• A political researcher expects that higher income
  leads to higher political trust. At the same
  time, the researcher expects that part of the
  relationship between income and political trust
  might be due to the effects of education on both
  income and political trust. What is the
  hypothesis in this situation?

5
Hypotheses with Antecedent or Intervening
(control) Variables

• Answer
• There is a positive relationship between income
  and political trust while controlling for
  education.
• If all variables involved are at least ordinal,
  then we can simply add the phrase while
  controlling for and specify the control variable.

6
Hypotheses with Antecedent or Intervening
(control) Variables

• Question
• Suppose you are doing research that investigates
  the relationship between age and support for
  income equality. You deduce, however, that this
  relationship will also depend on whether the past
  system of government was communist or not. How
  would we state a hypothesis that takes into
  account the control variable?

7
Hypotheses with Antecedent or Intervening
(control) Variables

• Answer
• In countries with a non-communist legacy, there
  is a negative relationship between age and the
  level of support for income equality.
• In countries with a communist legacy, there is a
  positive relationship between age and the level
  of support for income equality.
• If the control variable is nominal, then we can
  specify the kind of relationship we expect to
  find among different categories of the control
  variable.

8
The Null Hypothesis

• A null hypothesis states that there is no
  relationship between the variables in a
  hypothesis.
• In statistical analysis, we always test the null
  hypothesis and not the actual hypothesis.
• This is because, as we learned in Lecture 6, we
  cannot prove causality.
• Instead, we must focus on disproof.
• If our research leads us to reject the null
  hypothesis, we have decided that there is some
  kind of relationship between the independent and
  dependent variable.
• But we need further investigation to find the
  true nature of the relationship--by bringing in
  more control variables, for instance.
• We have supported the theory but not proven it.

9
Example of a Null Hypothesis

• Answer
• Males are more supportive of sexual equality than
  females are or females are more supportive of
  sexual equality than males are.
• Null There is no relationship between support
  for sexual equality and gender.

10
Survey Research

• Very common method of data-gathering.
• Scarcely a day goes by when we are not told the
  results of one poll or another.
• Surveys depend on questionnaires. Well-designed
  questionnaires must
• Have well-worded questions
• Open- or closed-ended questions?
• Place questions in the appropriate order

11
Survey Research Question Wording to Avoid

• Double-barreled questions
• Do you agree with the statement that the Soviet
  Union is definitely ahead of the United States in
  the arms race and that the United States should
  increase defense spending?
• How does a person who believes the Soviets are
  superior in military capacity but who does not
  want an increase in defense spending answer this
  question?
• Researcher does not know whether the respondent
  meant for his answer to apply to both components
  or whether one component was given precedence
  over another.

12
Survey Research Question Wording to Avoid

• Ambiguous Question
• What is your income?
• Why is this ambiguous?
• Do you prefer Brand A or Brand B
• Why is this one ambiguous?
• Carter/Ford debate poll were they good, bad, or
  indifferent?

13
Survey Research Question Wording to Avoid

• Leading Question
• Dont you think that global warming is a serious
  environmental problem?
• Encourages a particular response because the
  question indicates that the researcher expects
  it.
• Choice of words may also lead respondents.
• Research has shown that people are more willing
  to help the needy than those on welfare.
• Linking personalities or institutions to issues
  can affect responses
• What is wrong with the following wording
• It is generally better to have a man at the head
  of a department composed of both men and women
  employees.

14
Survey Research Question Wording to Avoid

• Avoid technical words, slang, and unusual
  vocabulary.
• Questions including words with several meanings
  will result in ambiguous answers
• How much bread do you have?
• Answer depends on the measure one uses.
• Be aware of cross-cultural differences.
• Should look at wording of questions in existing,
  reputable surveys.

15
Survey Research Open-ended vs. closed-ended
questions

• Closed-ended questions provide respondents with a
  list of responses from which to choose.
• Alternatively, closed-ended questions can provide
  multiple choices for the respondent to accept or
  reject.
• Open-ended questions do not provide respondents
  with any answers from which to choose.
• Is there anything in particular about Mr. Kerry
  that might make you want to vote for him?

16
Survey Research Open-ended vs. closed-ended
questions

• Closed-ended questions advantages and
  disadvantages
• Advantages
• Easy to answer and takes little time
• Answers can be precoded (assigned a number) and
  easily transferred to a computer
• Answers are easy to compare
• Easier to elicit responses to sensitive questions
• Disadvantages
• may not be accurate--forces people to accept
  categories, or puts too many people into other
  category
• Wording of response choices may influence
  responses (example)
• Single-sided vs. two-sided questions (example)

17
Survey Research Open-ended vs. closed-ended
questions

• Open-ended questions advantages and
  disadvantages
• Advantages
• Not forced to choose between categories
• Lets respondent have the say, let him tell the
  researcher what he means, and not vice-versa
• Disadvantages
• Respondent may say too much or too little
• Coding problems!
• Interpretations of answers may vary
• Too much variance in response
• Expensive and time-consuming

18
Survey Research Question Order

• First questions should break the ice
• Questions on personal or sensitive topics left
  towards the end
• Avoid a series of questions that are likely to
  elicit the same response
• One question can affect another
• If you ask about crime early on, and then ask
  about problems in their community, likely to get
  crime as an answer
• Or, redundancy or consistency problems
• Dangers of filter questions