Poli Sci 106 Lecture 5 Outline of Last Lecture 1. Two Cornerstones of Comparative Politics2. The Scientific Method3. Basis of Explanation: Theory4. Key Characteristics of a Theory5. Research Design6. Variables7. Correlation Between Variables8. Correlation vs. Causation9. Spurious Correlation10. Quick Review11. ExperimentsOutline of Current Lecture 1. Theory (Review)2. Review from Previous Lecture3. Theories and Models4. Hypotheses5. Formulating Hypotheses6. 5 Steps Involved in Hypothesis Testing7. "Controlling"Current Lecture - The Scientific Method II- theory - a proposition of how we think the world works, represents what we think happens in the world-based on our discussion so far, a theory is a generalization that seeks to explainrelationships between variables-it is the theory that helps us to tell a causal story, because the theory specifies the direction of the "causal arrow" when we observe a correlation-Variable 1 ---> Variable 2 (change in VI causes change in V2)-Variable 2 ---> Variable 1 (change in V2 causes change in V1)These notes represent a detailed interpretation of the professor’s lecture. GradeBuddy is best used as a supplement to your own notes, not as a substitute.- key qualities of theories - -falsifiability - necessary for a theory-parsimony - not necessary but desirable -the preference for the least complex explanation for an observation-to explain as much as possible with as little as possibleReview from Previous Lecture:- political scientists engage in "empirical analysis," the systematic analysis of facts- identify cause and effect relationships- independent, dependent, intervening variables- IV and DV correlated (either positively or negatively)- correlation does not equal explanation- correlation does not demonstrate causality- may be spurious- correlations are necessary to demonstrate causality- key characteristics of theories: falsifiability and parsimonyTheories and Models:- theory - a generalization that seeks to explain relationships between variables- model - represents and describes reality in simplified/abstract terms-derived from a theory-causal models specify causal relationships between variables-always probabilistic- Ex. Research question - Why do congressmen vote the way that they do?-Theory: politicians have two primary objectives: re-election and the policies they enact-Model: constituency + ideology = Congressmen preferences = Congressmen vote choiceHypotheses:- not the same as a theory- an observable implication of a theory- a proposition we make based on our theory and that must be empirically testable- posits a testable cause and effect relationship between IV's and DV's- we state what we expect to see in the real world based on our theory and then proceed to test this expectation against reality- "if our theory/model is correct, then we should observe in the real world that ____"- hypothesis testing: lies at the heart of the scientific method-Ex. Observation: Scrappy gets excited every time Werder Bremen scores a goal. -Research question: Why?-Theory 1: He is a devoted Werder Bremen fan.-Theory 2: He reacts to the excitement of his owner. -Hypothesis derived from theory 1: -if theory 1 is correct, then we should expect that Scrappy gets excited even when his owner leaves the room-Hypothesis derived from theory 2:-if theory 2 is correct, then we should expect that Scrappy gets excited every time his owner gets excited, no matter what the cause may be-Ex. 2 - Observation: College library is more crowded in December than September -Research question: Why?-Theory 1: Finals coming up-Theory 2: Cold outside-Hypothesis derived from theory 1: -if theory 1 is correct, then we should expect that the library is equally crowded in May-Hypothesis derived from theory 2: -if theory 2 is correct, then we should expect to observe that the people who spend time in the library are not just students-these hypotheses are testable-critical!-can determine if people in the library are studentsFormulating Hypothesis:- 2 ways - inductively and deductively- induction - -proceeds from the specific to the general-we observe specific facts then devise a generalization on that basis, which we can then test against additional evidence- deduction - -proceeds from the general to the specific -the path from existing generalization to specific cases must follow with logical necessity- inductive approach - collect data on people's vote choice and their income, observe thatsome rich people tend to vote Republican, collect more data to test to see if observation holds- deductive approach - assumption: people act based on self-interest, Republicans tend tosupport policies that support wealthy- hypothesis in both cases: positive correlation between income levels and the probabilityof voting Republican 5 Steps Involved in Hypothesis Testing:1) Defining key terms2) Identifying variables3) Specifying hypotheses4) Collecting and examining evidence - important: we look for evidence that would confirm our hypotheses as well as evidence that would contract it! Also measurement is important5) Drawing conclusions: evidence may be...a) constant with/confirms our hypothesisb) inconsistent with/contradicts/disconfirms our hypothesisc) mixedd) inconclusive"Controlling" - - basic idea: ensuring that we only compare "like to like" (and not apples to oranges)- Ex. testing children's "general knowledge"- finding: positive correlation between height and what children know- age?- solution: "controlling" for age-means no univariate analysis (height, knowledge) but multivariate analysis (height, age, knowledge)-only compare relationship between height and knowledge of children who are the sameage-no correlation between height and knowledge once we control for age-must account for all independent variables- "social network analysis" - who is connected to whom, joint membership in one or more caucuses means that two legislators are connected to one another, "network" of
View Full Document