pol 2001

ratio level of measurement

categories that can be compared, have equal intervals, and a true zero, like money

interval level of measurement

categories that can be compared and have equal variables, like years

nominal level of measurement

non-ordered categories/classifications without comparison, like countries

ordinal level of measurement

are nominal measures, but can be compared and do have order, like education level

tautological reasoning

the dependent and independent variable are the same, the claim you’re making is included in the definition

inductive reasoning

general conclusions are drawn from specific observations

deductive reasoning

specific claims are formed from general observation

black-box mechanism

x and y are related but its unclear how ; the mechanism is hidden in a “black box”

normative research

aims to explain how things should be in perfect conditions

empirical research

aims to explain how things really are ; evidence based research

face validity

does an experiment “look right"? does it match up with the evidence?

convergent validity

it correlates with other measures it should relate to. e.x. show academic burnout is correlated with academic stress

discriminant validity

ensures your measure does not correlate with other measures it should not correlate with. e.x. show academic burnout is not related to academic success

ecological fallacy

generalizing a sample from one population to another. e.x. cats jump high, so dogs must jump high

mediating variable

explains why X causes Y

confounding variable

explains why X may not even be related to Y ; the argument destroyer

moderating variable

controls the strength of the independent variable

Ethics

the moral principles and professional standards that govern interactions with participants

Milgram Obedience Study (1961)

about: studied how far people will go with something they know is wrong just because the instructor told them to. 65% continued to deliver shocks at the max voltage to someone with a heart condition.

ethical issues: deception, no debriefing, lack of informed consent, no clear right to withdraw

Stanford Prison Experiment (1971)

how do assigned social roles influence behavior? found that violence, aggression, and uneven power dynamics ensued.

ethical issues: uninformed, psychological harm, no clear right to withdraw, researcher was involved in the experiment’s execution

Tuskegee Syphilis Study (1932-1972)

600 black men were told they had “bad blood” when they had syphilis so the government could find out what untreated long term effects were.

ethical issues: recruited under false pretenses, 100 died, cure was available 1947…just really fucked up

National Research Act (1974)

regulated human subjects research

Belmont Report (1979)

3 core principles of human subjects research: respect, beneficence, justice

Institutional Review Board (IRB)

required to review and approve all human subjects research

Case study designs

in-depth qualitative analysis of a single case aimed at understanding broad phenomena

strengths: allow for rich context, uncovers causal mechanisms, allows for hypothesis generation/extension, reveals outliers

limitations: limited generalizability, selection bias (picking interesting cases for a specific outcome), replication is difficult, hard to distinguish correlation from causation, hard to estimate causal effects

Idiographic case study

explains a single historical episode, with no claim to generalize beyond the case

Hypothesis-generating case study

examine 1 or more cases to develop general theories

Hypothesis-testing case study

examine 1 or more cases to test general theories

Plausibility probes

if you cannot decisively test a theory:

gauge whether a theory is viable, clarifies scope conditoions, and informs a fuller test later

Method of difference

selects cases in which the outcomes differ then look for the single factor the cases do not share. usually used to generate hypotheses not test

method of agreement

selects cases that share the same outcome then identify the causal factors they share. used to generate hypotheses not test them

Deviant cases

where researchers choose a case that does not conform to a theory or fit a normal pattern. The goal is to revise and refine theories.

• Processing tracing

an in-depth case study that unpacks and empirically traces causal mechanisms.

Interviews

purposeful, researcher-guided conversations with a participants to elicit detailed, first-person accounts

Structured interviews

fixed wording, order, and response options. high comparability, low flexibility and specificity

semi-structured interviews

open-ended questions with planned probes, order can vary and can go off script. lower comparability, higher flexibility and specificity

unstructured interviews

respondent leads. conversational, few preset questions. maximizes depth and discovery but low comparability

elite v. public interviews

elite: interview individuals that possess specialized knowledge about a political phenomenon — like a congressman about congress. elite=they’re at the top.

public: anyone else, typically to understand public opinion

Participant observation

researcher becomes regular participant in the activities of the people they’re trying to observe.

e.x. joining a local pilates class to report on conditions inside the facility

• Focus groups

a qualitative research method in which a small group of people engages in a discussion with a neutral moderator.

strengths: multiple views at once , provides rich description of conversations, reveals how people think in real time

limitations; generalizability can be a problem, group dynamics can influence the discussion (i.e. parents and children), resource intensive, ethical concerns.

• Experimental v. quasi-experimental v. observational research

experimental research: researcher manipulates the design

• Treatment/intervention

• Random sampling v. random assignment

• Treatment effect

• Between-subjects design

• Within-subjects design

• Multiple condition designs

• Threats to internal validity within experimental research

• Field experiments

• Audit experiments

• Lab experiments