Research Methods Authors & Concepts

Feldman & Luna

• Topic: Good Literature Review

• Components:

  • Justification of their study

  • Synthesis of the State of the Art on their topic

Gladwell

• Topic: Accuracy

• Components:

  • Conceptual Clarity: what is a “good” vs. “bad” college?

  • Validity: how do the elements that US News measures signify a “good” vs. “bad” college?

  • Reliability: if examined again, would the same measurements be reliably used to signify the concept of the “best” college?

Munck & Verkuilen

• Topic: Precision

• Components:

  • Conceptual Clarity: how is democracy conceptualized?

  • Validity: how well do each of the measures selected represent democracy?

  • Precision: does the aggregate data collected by Freedom House contain enough detail and nuance to accurately reflect measures of democracy across the world?

Levitt

• Topic: Causal Hypotheses

• Components:

  • Causal Factors: demonstrates direct relationships between the four factors he attributes with causing falling crime rates

  • Spuriousness: claims that the other factors he mentions are spurious

  • Correlation: is careful to identify causation (in his four factors) vs. correlation with the other factors he mentions

Howard

• Topic: Spuriousness

• Components:

  • Confounding Variable: a confounding variable that Rugy was not aware of was present in her research (presence of state capitals)

  • Spuriousness: Rugy did not recognize a confounding factor and thus that the relationship she studied was spurious

  • Endogeneity: identifying which variable came first (which is the cause and which is the effect)

Taub & Fisher

• Topic: Spuriousness and Causal Mechanisms

• Components:

  • Correlation: there were strong correlations between Facebook usage and increased anti-immigrant attacks

  • Causal Mechanisms: there were numerous causal mechanisms embedded in the Facebook usage independent variable

Kellestadt & Whitten

• Topic: Deterministic vs. Probabilistic Relationships

• Components:

  • Deterministic Relationship: when the IV occurs, the DV will occur with certainty

  • Probabilistic Relationship: a cause is associated with an effect but it is not a certainty that the effect will happen

Banzhaf et. al

• Topic: Correlation vs. Causation

• Components:

  • Causal Mechanisms: Banzhaf establishes the mechanisms by which race and pollution exposure are connected

  • Proving Causation: Banzhaf et al. had to prove that the factors have causal relationships with the effects 

  • Confounding Variable: socio-economic status was a variable closely intertwined with both the IV and the DV

McDermott

• Topic: Experimental Studies

• Components:

  • Experimental Study: a study in a controlled environment where the variables can be manipulated to prove a causal relationship

  • Internal Validity: being able to prove that the IV has a causal relationship with the DV

  • External Validity: being able to generalize the findings of a study to a broader population

  • Laboratory Experiments: the gold standard for experimental studies, because all variables can be controlled

Iyengar et. al

• Topic: Experimental Study in Social Science

• Components:

  • Controlled Variables: certain factors (such as the way that the subjects were exposed to the stimulus) were controlled for to eliminate any confounding factors

  • Internal Validity: there was only one factor that fluctuated significantly

  • External Validity: the study was generalizable to the broader population of adults who regularly tune into the news

  • Random Assignment: subjects were randomly assigned to either the control or test group

Butler & Broockman

• Topic: Field Study (Experimental Study)

• Components:

  • Random Assignment: politicians were randomly assigned to test aliases

  • Sample Size: the sample size in the study was significant and thus added legitimacy to the findings

  • Controlled Variables: as best as possible, other variables were controlled for to establish a causal relationship

Bocar et. al

• Topic: Observational Study

• Components:

  • Recognizing the Failing to Control for Certain Factors: 

    • No randomization

    • No recognition of the difference in policing experience due to race

  • Randomization: Bocar addresses the problem by:

    • Randomizing police distribution

    • Creating similar policing conditions (ie. time of day)

    • Taking a large sample size

Gerring

• Topic: Case Study

• Components:

  • Units of Measurement:

    • Population

    • Sample

    • Unit

    • Case

    • Variable

    • Value

  • Types of Case Study Variation:

    • Cross-Case

      • Most-Similar

      • Most-Different

    • Longitudinal

    • Within-Case

Gonzalez

• Topic: Case Study

• Components:

  • Covariation: Gonzalez includes the full spectrum of covariation in order to determine the causal factors of police reform in her case studies

  • Multiple types of Variation: this helps to rule out spuriousness

  • Reverse Causation: establishes the causal direction of the relationship 

  • Causal Mechanisms: Gonzalez fleshes out the logic of her findings to establish credible causal mechanisms

Geddes

• Topic: Selection Bias

• Components:

  • Selecting on the Dependent Variable: choosing cases based on similarities in the dependent variable which will produce invalid results

  • Equifinality: there are multiple paths to arrive a the same outcome or dependent variable

    • Necessary Condition: necessary but not enough on its own

    • Sufficient Condition: if present, is enough to produce the outcome

  • Single Case Study Designs:

    • Typical

    • Deviant

    • Extreme

George & Bennett

• Topic: Focused, Systematic Comparison

• Components:

  • Process Tracing: the process by which sequences of events are analyzed to determine the sequence of causal events

  • Tools Involved in Process Tracing:

    • Causal Analysis

    • Sequence Analysis

    • Testing for Necessity and Sufficiency

    • Pattern Matching and Congruence Testing

Mahoney

• Topic: Process Tracing

• Components:

  • Tools Involved in Process Tracing:

    • Causal Analysis

    • Sequence Analysis

    • Testing for Necessity and Sufficiency

    • Pattern Matching and Congruence Testing

  • Hoop Tests: tests that provide the minimum evidentiary support for a finding

  • Smoking Gun Tests: a piece of evidence that leads to strong support for a finding

  • INUS Condition: by itself is not sufficient, but is part of a sufficient condition

Pollock & Edwards

• Topic: Randomly Selecting Cases for Surveys

• Components:

  • Sampling Frames

  • Types of Sampling Methods:

    • Simple Random Sampling

    • Stratified Sampling

    • Cluster Sampling

    • Systematic Sampling

  • Weighting

  • Selection Bias

Feldman & Mendez

• Topic: Nonresponse Bias

• Components:

  • Nonresponse Bias: certain demographics did not answer follow-up surveys, and thus, the longitudinal study results were skewed

Enten

• Topic: Bias and Error

• Components:

  • Nonresponse Bias: people do not answer, and thus the results are skewed

  • Desirability Bias: people will answer with what they think is the most socially desirable answer

  • Margin of Error: surveys represent approximations so there is a margin of error in either direction