1. unbiased observation

2. description and explanation 

3. constructing of causal theories 

1. Empirical verification 

2. Falsifiability 

3. Explanatory 

4. Prediction

5. Parsimonious 

1. Transmissible: methods are explicit so others can analyze and repeat 

2. Cumulative: techniques and findings are based on previous studies 

3. Replicable: ability to repeat 

4. Generalizable: ability to apply to many cases

1. Empirical: What is 

2. Normative: What should be 

3. Rhetorical: My belief is a fact

general to specific 

true premise = true conclusion 

Specific to general 

true premise = likely conclusion 

•         Actions: human behavior done for a reason

•         Interpretation: one must understand the way individuals see their world to understand their behavior

•         Social Facts: values and institutions that have a subjective existence to the people living in a particular culture 

1. authoritative

2. mystical 

3. rational 

-information percieved as fact because it is recorded by "qualified sources"

-these facts are typically unquestioned by the majority

-Knowledge obtained from logic and reasoning 

-derives from syllogisms, proofs, formal models 

-rational knowledge may not be an accurte statement of reality 

ex: if i can balance my checkbook so can the U.S Govnt.

Components of Research

(5)

•         Research Question: Asks why or how a political phenomena behaves.

•         Literature Review: A synthesis of prior research and knowledge.

•         Theory: A body of statements that explain phenomena.

•         Hypotheses: The terms to be tested through the collection and analysis of empirical data.

•         Empirical Analysis: Tests hypotheses and answers research question.

it is the gap your research hopes to fill

Organizing Literature by: concepts, methods, data and findings

1. System

2. Dyadic

3. State 

4. Individual 

-how the international system influences state behavior 

·         The assemblage of units, objects, or parts united by some form of regular interaction

Constrains behavior: This include states, IGOs, MNCs, etc.

Strengths: inclusive and parsimonious

Weaknesses: deterministic, lack of detail, and black box the state 

-focus on relationship between two states 

·         Two States: Shared features or interests

·         Shared Regime Type: Two democracies very unlikely to fight.

·         Other Comparisons: Religion, Culture, Power, Alliances, etc.

Strength: Parsimonious& State no longer a “black box”

 Weakness: Deterministic, Not Inclusive – leaves out third parties

-focus on particular state characteristics

·         State/Societal Level: State-Specific Characteristics

·         Examples: Regime Type, Geography, Revolutionary/Post-Revolutionary States

Strength: Permits Differentiation

 Weakness: Distortion & Excluded Outside Forces Reliance on Perception 

-looking inside the state at individuals and groups
-Focus on sub-state actors, emphasizes the decision-making process.
-Cognition: How knowledge is acquired
-Perceptions: How we view circumstances.

-Examples: Public Opinion,  Ethnic Groups, and Nationalism
-Strength: Focus on Actual Actors & Great Amount of Detail
-Weakness: Predictions Less Clear, Not Inclusive & Distortion

1. X & Y covary

2. change in x preceeds change in Y

3. covariation is not a coincidence (spurious)

null hypothesis 

Ho states there is no relationship between the variables in question and any relationship is due to a stat error

A hypothesis has support if the null is rejected

Definitions of concepts should be:

(4)

1. clear  2. accurate 3. precise 4. informative 

examples: democracy, political participation, economic development

-deciding what empirical observations should be made to measure the occurences of a concept

-provides specific terms on concept measurement 

Ex: Democracy, conceptually is a system governed by the people

Operational, presence of elections in an area 

The extent to which an experiement or test yields the same results on repeated results 

*consistent results = high reliability 

inaccurate measures lead to erroneous conclusions 

poor measurements inhibit the ability to observe the actual relationship between variables 

-the amount of precison associated with a variable

- the mathametical properties of the variable 

-describes variables that indicate a difference between categories

-very limited percision. cannot be used with many stats tools

ex: which school did you attend?

-indicates a difference in categories and ranks them in order and has a little more precision 

-can be used with more stat tools 

ex: what level of school did you achieve?

-includes meaningful intervals between values of the variable but does not use a meaningful zero 

-more precise, can be used with most stat tools

ex: waht did you score on the SAT?

-includes meaningful intervals between valyes as well as a meaningful zero

-precise and can be used with most stat tools 

ex: what is your batting average?

7 points on a scale: strongly agree- strongly disagree

only select questions in the calculation of the score

-how far along the spectrum you are willing to go

-respondents agree with each of the lower-ranked answers if they agree with a higher ranked answer

covariation: the cause and effect vary together 

time orer: the cause precedes the effect in time

elimination of alternative explanations to isolate causation to one factor

-the procedure demonstrates a true cause-and-effect relationship

-the results are not produced by spurious factors 

-potential issues include selection bias, experiemental mortality, and demand characteristics

-the extent to which the results of a study can be generalized across populations, time, and settings

-the population and stimuli should correspond with the actors and enviornment in the political world 

1. establish an experiemental group & control group

2. random assignment of individuals to each group

3. measurement of a DV before and after the treatement with a pre-test & post-test 

4. controlled administration of the treatment 

5. controlled enviornment of the experiement

1. presence of a single group

2. lack control over the assignment of subjects

3. lack control over the application of the IV

4. inability to measure the DV before & after exposure to the IV occurs

-measurements of variables at different points in time

-models change across time and examine time order

-certain events can alter the influence of a variable

-simplified representation of reality express through a verbal, mathematic, or symbolic system

-rules or logic govern connections between the parts of the model used for making decisions 

-primitive term: undefined in a model and its meaning is taken for granted 

-any well defined set of units of analysis

-determined largely by the RQ and should be consitent throughout the research project

-describe characteristics of the population 

-most research uses samples b/c resources & feasibility precludes the use of a population 

-creates a sample that is identical to the population in all  characteristics except size

-draw concluions about a population based on sample statistics 

-on average sample statistics will equal the value of populatin parameter 

-any single sample statistic may not equal the value of population parameter 

1. validity of measurements 

2. effect of phenomena being measured

3. population covered by a data collection method

4. resources and the cost of a method 

5. avaliability of data

-provides depth of understanding 

-establishes theoretical plausability 

-generation of new theories/hypotheses

-more appropriate for complex phenomena 

-may be the only strategy possible

-small sample size and lack of generalizability 

-difficult to control for alternative explanations

-selection bias and possibly less objective 

-difficult to replicate

-people behave naturally and behavior can be observed over time

-increased accuracy and completeness absent in documents and surveys 

direct: field studies

-a lab setting allows control over the enviornment including a more rigourous experimental design

-observation may be easier and more convenient to record and preserve 

indirect: physical traces of behavior

-created by the depostion and accumulation of materials 

-erosion & accreation measures may be biased 

-certain traces more likely to survive because the materials are more durable

Ethical concerns 

(5)

1. invasion of privacy

2. stress during the research interaction 

3. disclosure of sensitive information to the researcher

4. negative reprucussions from associateing with the researcher because of the researcher's sponsors, nationality, or outsider status 

5. research involving human subjects must be submitted to an IRB

1. Milgram Experiement (1961): shocking people

2. Stanford Prison Experiement (1971): prison roles

Concept to remember in ethics:

(5)

-allow access to subjects who are difficult or impossible to research through direct personal contact

-records are often avaliable for analysis over time

-can use a larger sample size than with interview or direct observation 

-less expensive because record keeping costs are borne by record keepers 

-selective survival

-large gaps may exist in many archives 

-contet may be biased through incomplete, inaccurate, or falsified records 

-some written records are unavaliable to researchers 

-records may lack a standard format 

-cost in both time and money 

-accessibility of records-locating episodic records can be time consuming 

-covers an extensive period of time

-organizations decide how to keep records 

-sometimes difficult to identify record-keeping practices 

-raw data is not always avaliable

content analysis procedures 

(4)

1. select materials relevant to the research subject and then sample the actual material to be analyzed from that sampling frame

2. define the categories of content that are going to be measured- the topics of interest within the content 

3. choose the recording unit

4. decide on the numerical values that will be used to code each category in each unit

-2 or more analysts use the same procedures and definitions and agree on the contet categories 

*serves as a check on bias

-collect information through a questionare 

-forma and structured 

-survey insturments

-direct and personal communication 

-lacks formality and structure

1. cost

2. completion rates

3. sample-population congruence 

4. questionare length

5. data-processing issues

costs of a survey based on 

(6)

1. length of the questionare 

2. geographical dispersion of the sample

3. callback procedures 

4. responder selection rules 

5. avaliability of trained staff

6. method-specific costs 

-failure to maximize the representativeness of the sample risk bias 

-a low response rate limits the ability to make statistical inferences about the population 

-those who fail to participate in the survey share common characteristics 

-conslusions made with the data could be incorrect 

-gather sufficient information without extensive length

-keep respondents attention to help achieve the best possible response quality 

-substantial time and effort are required to process data 

-surveys using complex designs or a large number of questions wil require more time and effort 

-processing costs are a major reason for the adoption of internet and telephone surveys

-high response rates 

-flexibility/control in questioning 

-observation (personal intervewing)

-supervision of interviews (telephone)

-expensive 

-lack of anonymity 

-characteristics of interviewer may introduce bias: mail and telephone

low cost 

avoid biases

-low response rates 

-cannot probe 

-cannot clarify 

-cannot be sure who completes questionare

question wording 

question type 

question order

motivate respondent with friendly administrative style

avoid biasing responses through cues 

probe for additional information or clarification 

-interviews feed respondents false and information under the guide of asking a question 

-conducted by campaigns and political organizations

close-ended questions

(5)

-provide answer choices 

-easy to code and analyze 

-respondents more likely to answer these questions with respect to sensitive topics 

-may force respondents to choose pre-coded answers that do not match their own position 

-may include inappropriate choices leading to choice of "other"

Open-ended questions

(4)

1. do not provide answer choices 

2. allow respondents to more fully explain answers 

3. create the opportunity for researchers to find answer choices they had not anticipated 

4. questions are time consuming & difficult to analyze

questions that ask to agree or disagree with a single substantive statement

*example: do you agree or disagree that the government should see to it that every person has a job with a good standard of living?

question that offers two substantive choices 

*example: should the government see to it that every person has a job with a good standard of living or should it let each person get ahead on their own?

-the order in which you place your survey questions can cause the error of a saliency effect 

-specific mention of an issue in a survey may cause a respondent to associate the issue in connection with a later question 

-respondents may strive to appear consistent in their answers 

-an answer to a question may be constrained by a previous response 

-sorts respondents into subgroups 

-subgroups respond to certain sets of questions

-asking individuals a series of questions and recording respondents 

-particularly useful when studying political elites 

explicit purpose & direction 

prepared topics & questions

-a table that shows the number of observations having each value of a variable 

-lists a variable's possible values & how frequently they occur

computed using all values

describes the average value

not resistant to outliers 

outliers will pull the mean up or down, sometime significantly

removes a percentage of high and low values

reduces influence of outliers

describes the middle value in an ordered set of values

resistant to outliers

indicates the most common observation 

count the number of times you observe each value 

mode is resitant to outliers 

by definition, the mode cannot be an outlier

-indicate how the other values vary around the typical value

1. range

2. interquartile range 

3. variance 

4.standard deviation

divide data into 4 quartiles

describes the midde 50% of observations

-the bulk of observations lie in the center 

-half the observations lie above and below the mean 

-mean, median, and mode have the same value

skewed left

few observations on the low end of the scale

skewed right 

few observations on the high end of the scale

bar chart 

pie chart 

histogram 

time series plot: general trends over time

positive: high value in one means high values in another

negative: high values in one means low values in another

rule 1: no association 

rule 2: score on related variable is known

Linear regression formula and its parts 

Y=Bo + B1x + E

E (Y): expected value of the DV

Bo= expected value of Y when X = 0 

B1= expected change in Y for a unit change in X

X= value of the IV

E-error term in linear regresison 

R= correlation coefficent 

perfect positive association: 1 

perfect negative association: -1

no relationship: 0

assumes Y is continuous 

provides the best fit line for squared distances 

total SS= regression SS + residual SS 

R2= ReggSS/TSS*identifies the proportion of explained variance 

survey failure 

small sample size 

representation (car registries during depression)

Multivariate Analysis

• Relationships with more 

than two variables.

• Including more than two 

variables allows you to 

control for a spurious 

relationship.

• A spurious relationship is 

one in which the association 

between two variables is 

caused by a third.

Polling data

Bivariate: % of people who voted for obama or mccain 

Multivariate: looks at % of women, men, minorities, etc. 

a coefficient indicates how 

much and in what direction the DV changes with 

a one-unit increase in the IV, controlling for all 

other variables in the model.

Each B indicates the relationship of X and Y after all 

other IVs have been controlled.

A logistic regression is a better choice for a 

dichotomous DV.

• Coefficients in a change when each independent 

variable is set at a different value.

The coefficients only indicate the direction of the 

relationship.

• To assess the magnitude of a relationship you must 

calculate the predicted probability or odds ratio.