GVPT100 Midterm ReviewI. Lecture 1a. Conceptsi. Concept: an idea or mental construct that represents phenomena in the realworldii. In empirical political science, concept refers to facts, not values. A concept is an observable and measureable phenomenon. iii. Primary goals of political science research are to describe concepts and analyze the relationship between them. (How does X affect Y?)iv. Examples1. Democracy2. War 3. Partisanship b. Concrete questioni. question expressed using tangible properties and can be answered empiricallyii. Need to transform concepts into concrete terms iii. Example1. Are women more liberal than men?c. How do we measure concepts?i. Step One: Conceptual definition1. Must decide what is important to the concept. What are the essential elements? What is important to the concept?2. Three potential problems:a. Might think of empirical attributes that refer to a completely different conceptb. May include terms that are not measurablec. Empirical properties may represent different dimensions of a conceptd. Multidimensional concept = has two or more distinct groups of empirical characteristicsi. Example: Liberalism- economic (support for gov’t funded healthcare, aid to poor, funding education); social (gay/lesbian rights, abortion, legalization of marijuana)3. Unit of Analysis: the entity to which the concept appliesa. Examples: Individual, Household, Municipality, Congressional District, Country, International systemii. Step Two: Operational definition1. The operational definition explains exactly how the concept will bemeasured empirically a. How are you going to measure the concept in practice?b. Is the measure valid and reliable?2. Examplesa. What about war? How do we know a war when we see one?iii. Misc. Describing Concepts1. Make a list of empirical attributes or concrete properties2. Refine list to exclude any attributes that do not vary3. Refine list to exclude any attributes that are not measurable4. Refine list to exclude any attributes that measure a different dimension of the conceptiv. Examplev. Format/Example1. The concept of [concept label] is defined as the extent to which [unit of analysis] exhibit the characteristics of [measure].2. The concept of economic liberalism is defined as the extent to which individuals exhibit the characteristics of supporting government spending for social programs.d. Ecological Fallacyi. Aggregate-level phenomenon is used to make inferences at the individual level1. Example—average citizen well-being based on country’s level of development2. Example—your neighborhood’s average level of wealth after Bill Gates moved inII. Lecture 2a. Judging operational definitions: To determine if an operational definition (or “operationalization”) is a good measure, we ask two questions:i. Reliability1. If we used the same measurement instrument repeatedly, would weget the same results?2. If not, its not as good of a measure, because the results we get are partly a function of when we do the testii. Determining Reliability 1. Test-retest method: investigator applies the measure once and then applies it again to same units of analysisa. If measure is reliable, then two results should be same or similar2. Split half method- internal consistency approach a. Measurement obtained from half scale’s items should be same as other half (items reliably measuring same concept, two sets of scores should be same)b. Ex. Ten government spending questionsiii. Validity1. Even if it is reliable, a measurement can be wrong2. Validity addresses whether the operational definition actually measures the concept that we are interested in iv. Determining Validity1. Face validity: consider meaning; “on the face of it” 2. Construct validity: analyze empirics to compare measure to relatedconcepts3. Exp. GRE and grade point average in Graduate Schoolb. Measurement Errori. Systematic measurement error: introduces consistent, chronic distortion into empirical measurements1. Durable2. Inherent in instrument3. Examplea. Hawthorne effect: Knowing you are being studied.b. Test anxiety ii. Random measurement error: “random;” introduces chaotic distortion into measurement process, producing inconsistent readings of concept1. Not durable2. Not related to instrument3. How to stopa. Random AssignmentIII. Lecture 3a. Variablesi. Third step in describing conceptsii. Variables are “empirical measurements of characteristics” iii. Variables can take on different values—they can varyiv. There are three main levels of measurement for variables—1. Nominal—categories2. Ordinal—the categories are ordered3. Interval—the values of the variable communicate actual differencesb. Nominal Variablesi. Also called “categorical variables,” if they have numbers the numbers just indicate categoriesii. Categories with no rank – no category is “better” or “more of” or “greater”than another; number assigned is arbitraryiii. Examples—1. Region names—North America, Latin America, Africa, etc.2. State names3. Party labels—Republican, Democrat, Greenc. Ordinal Variablesi. Can be ranked categories1. Examples—low, medium, highii. Can assign numbers1. Examples—Countries are poor (1), middle income (2), rich (3)iii. Numbers show order, but we cannot say that 2 is twice 1, etc.iv. Make up many of the questions in survey research.v. Gauge approval or disapproval of government policies1. Immigration reform2. Legalization of Marijuana3. Same Sex Marriage4. Government-run Healthcared. Interval Variablesi. Can use math on numbersii. Distance between numbers is meaningfulintervalordinalnominaliii. Example1. Average household incomea. A household making $100,000 is twice as wealthy as a household making $50,000e. Dichotomous Variablesi. An additional kind of nominal variables– dichotomous, or “dummy” variablesii. Can only take on two valuesiii. Examples:1. Dead/alive2. In war/not in war3. Republican/non-Republicanf. Indexi. Additive combination of ordinal variables, each of which is coded identicallyii. Also called summative scale or ordinal scaleiii. Allows for more reliable measure of a characteristicg. Likert Scalei. 5-7 value ordinals that capture strength and direction of agreement.h. Levels of Measurementi. It is often possible to measure the same concept at multiple levels of measurement1. Example:a. How does level of education affect wealth?b. Measuring level of education:i. Dichotomous: high school graduate/non-high