22 february from theory to evidence. hypothesis vs. research question when no expected direction to...
TRANSCRIPT
22 February
From Theory to Evidence
Hypothesis vs. Research QuestionWhen no expected direction to relationship.Exploratory research, not deductive
Elements in HypothesesVariables
characteristics of the units we are studyingTheoretical populationUnits of Analysis
What the variables are characteristics of: individuals or a larger unit [e.g. states, countries]
Direction of relationship
ConceptsDemocracy, power, representation,
justice, equality, trust, social capitalMany interesting concepts are abstract
and lack a completely precise meaningOne strategy is to defining concepts by
being explicitExpress a concept in a form that is
observable and includes some notion or degree of differentiation
This process is called operationalization
What is a variable?Interesting questions in social science
center on concepts that involve variation and how changes in one phenomenon help to explain variation in another
Process of Operationalizing ConceptsDemocracy could be defined as having
certain attributesIe. Number of parties
Political Engagement as another exampleForms of political participation
MeasurementAre these measures any good?ReliabilityValidity
Types of VariablesDependent
Phenomenon trying to explainIndependent
The variable you are using to explain variation in the phenomenon
Dependent independentThe value of the dependent variable
depends on the value of the independent variable
Units of AnalysisAlong with proposing a relationship, a
hypothesis also specifies the types or levels of political actor to which the hypothesis is thought to applyIndividuals, groups, cities, countries
Problems with mixing different levels of analysis
Ecological fallacyUsing information that shows a relationship
for groups to infer that there is the same relationship for individuals
Direction/Nature of RelationshipPositive – as one variable increases the
other also increasesNegative – as one variable increases the
other decreasesAs age increases, probability of voting
increases (+)As political cynicism increases, probability of
voting decreases. (-)
Pro
b o
f V
oti
ng
Pro
b o
f V
oti
ng
Age
Cynicism
Problems in Formulating HypothesesVague or trivial predictionVague relationshipValue statementDeals with a case and is not generalizable
[theoretical population should not be one]
Problem Hypotheses?"Ronald Reagan was a stronger president than
Jimmy Carter." "The number of convicts on death row is greater in
Texas than in Utah." "Republican electoral success is related to how the
economy is doing." "Americans are generally satisfied with their
congressional representative." "People who are dissatisfied with the government
tend to exhibit political cynicism." "More legislation was passed by Congress in 1934
than in 1954." "Corruption among world leaders is related to
average annual rainfall in their country's capital." "Political participation is high among residents of
affluent suburbs."
Variables, units of analysis and direction of relationship:"The greater the inequality in the ownership of land in
countries, the greater the civil strife." "Women are more likely to oppose capital punishment than
men." "Local television news stories are more likely to be about
crime than network news stories." "The proportion of the vote a party receives determines the
proportion of seats it receives in the legislature." "Daughters tend to acquire their interest in politics from their
mothers." "The greater the number of highway patrol officers per capita
in a state, the fewer the number of highway fatalities." “Interest groups that spend the most on professional lobbyists
receive the greatest financial rewards from government programs."
"The size of a government agency influences the decision-making authority of its director."
"Lawyers who attended highly rated law schools are less likely to run for elective office than those who attended less prestigious law schools."
Types of RelationshipsNull hypothesisCorrelation/associationCausalSpurious
Spurious Relationshipa relationship in which two variables that
are not causally linked appear to be so because a third variables in influencing both of them
Spurious Relationship
Fire damage in $# of fire trucks responding to call
+
Intensity of fire
+ +
Spurious Relationship
%women electedParty quotas +
Political culture
+ +
Correlation/association vs. causationCo-variationTimeConsistent with other evidenceRule out rival explanations