methods and stats - exam 1

normative, empirical, combination, rhetorical

types of assertions

normative

type of assertion that discusses "what should be”

• “Healthcare is a human right.”

empirical

type of assertion that discusses “what is”

• fact/evidence based

• “Universal healthcare is associated with a decrease in maternal mortality rate”

combination

type of assertion that combines any 2 other types of assertions

rhetorical

type of assertion that discusses one’s belief as though it is fact

deductive

type of reasoning that is top-down - looks at things from a larger, broader, theoretical perspective, then works down with assumptions

deductive

deductive or inductive

• Ponies are horses.

• Stormy is a pony.

• Therefore, Stormy is also a horse.

inductive

type of reasoning that is bottom-up - generalizations that can lead to logical fallacies/false conclusions

• small observation turns into larger, broader, often wrong assumption

• however, learning from examples/observations (plural) can help build towards a larger theory

inductive

deductive or inductive

• This cat is black.

• That cat is black.

• Therefore, all cats is black.

empirical research

research based on actual, objective observations of a phenomenon - scientific knowledge of a political observation

• political theory is normative, but political science is _____

verification, falsifiability, non-normative, transmissible, cumulative, explanatory, parsimonious

elements of empiricism (7)

verification

element of empiricism

• ability of methods to be replicated to verify that research was not made up

falsifiability

element of empiricism

• hypothesis must be debatable/disproved or refuted

  • “The sky is blue” is not this

  • “Universal healthcare is associated with decreased MMR” is this

non-normative

element of empiricism

• something you can measure - what is (not what ought to be)

• “Healthcare is a human right” is the opposite of this

transmissible

element of empiricism

• applicable in other circumstances and can do the research to prove it

cumulative

element of empiricism

• based on previous research

• research does not exist in a vacuum - must be informed

explanatory

element of empiricism

• explains an event - X in relation to Y

parsimonious

element of empiricism

• not too specific to 1 thing/situation - can be broadly applied

theory in empiricism

body of statements that systemizes knowledge of and explains phenomena

• 2 aspects

  • leads to specific, testable predictions

  • the more observations that support these predictions, the more the theory is confirmed

1. ask a question

2. research existing sources

3. formulate a hypothesis

4. design and conduct a study

5. draw conclusions

6. report results

scientific method steps

hypothesis

theory-based statement about the relationship we expect to observe - “if-then”

• observable implication of a theory

null hypothesis

theory-based statement about 2 variables having no relationship - testing against this in a study

• we reject it if we find a statistically significant relationship

qualitative

type of research method that includes

• analysis of large # of observations

• enough observations to use statistical methods or descriptive methods

• more generalizable

• often tests a theory

• mostly deductive reasoning

quantitative

type of research method that includes

• analysis of small number of observations

• observations can lead to rich descriptions

• may test a hypothesis or general theories

• inductive and deductive reasoning

  • does not lead to a logical fallacy

interviews, focus groups, ethnography, grounded theory

types of qualitative research methods

interviews

type of qualitative research method - a conversation to answer a research question

• structured v. unstructured v. semi-structured

focus groups

type of qualitative research method - groups of people discussing a topic to figure out reactions to that topic

• campaign ad, product, etc.

ethnography

type of qualitative research method in which a research immerses themselves within a community, group, area, etc.

• fieldwork to observe individual/societal functions

• can be conducted on the internet

grounded theory

type of qualitative research method - getting as much info as possible to formulate a theory

topic to research to question to data to hypothesis

steps to formulating a hypothesis

surveys, survey experiments, secondary analysis

types of quantitative research methods

case studies, discourse/content analysis

research methods utilized in quantitative and qualitative

literature review

a brief, thorough recap of research on a specific topic so far that highlights recurring themes, findings, and gaps within previous research

• done before conducting research to establish a foundation, recognize different perspectives, and show that the work is distinct

correlation

association between variables - relationship/connection between 2 concepts

• positive, negative, no, spurious

causation

association between 2 variables in which 1 causes change in another

• 4 conditions

  • correlation

  • logical time ordering

  • mechanism

  • non-spuriousness

correlation, logical time ordering, mechanism, non-spuriousness

4 conditions of causation

logical time ordering

1 of 4 conditions of causation - X must precede Y

mechanism

1 of 4 conditions of causation - what is responsible for the causal relationship b/w X and Y

non-spuriousness

1 of 4 conditions of causation - another variable is NOT affecting the relationship

1. annotated bib

2. new doc - sort AB into thematic sections

3. write 1 thematic section at a time

4. write a short intro to lit review

5. write a short conclusion that justifies research

steps for writing a lit review

general, parsimonious, falsifiability

3 conditions of a theory

general

1/3 conditions of a theory - applicable to more than 1 instance

• variety of operationalizations and geological contexts

  • other circumstances rather than 1 specific scenario

parsimonious

1/3 conditions of a theory - the simplest condition that works

• needs to be simple so the theory can be generalized

• complexity makes it harder to observe distinct relationships between concepts

falsifiability

1/3 conditions of a theory - can be proven wrong

• can be challenged

unit of analysis, value of independent variable, value of dependent variable

anatomy of hypothesis (if-then)

unit of analysis

what a researcher is specifically analyzing; entity to which the concept applies

• the individual

questions

topic to question to hypothesis - conceptual to concrete

operationalize

defining a concept in measurable terms - variation

• finding out how to measure a variable so it can be analyzed

measurement

systematic observation and representation by scores and numbers - values for variables

• nominal, ordinal, interval, ratio

nominal, ordinal, interval, ratio

levels of measurement

nominal

discrete type of measurement - categories that are exhaustive and mutually exclusive

• most basic type

• measure for race, ethnicity, religion, etc. - no inherent order

ordinal

discrete type of measurement - values are mutually exclusive and can be ordered; inherent order to categories (scale)

• place in a race (first to last)

• less to more, always to never scales

• not evenly spaced

interval

concrete type of measurement - can assign different values, order them, and attach a number with equal distance between value

• zero (0) has no real meaning as absence

  • zero degrees does not equal absence of temperature

ratio

concrete type of measurement - anything in dollars or percentages

• zero (0) as the absence of something

• fulfills mathematical properties

• allows comparison

index

multivariable measure (summary)

• freedom house merges different concepts to measure democracy

  • maximalist: too many measures makes it harder to examine relationship between X and Y if Y is embedded into the index

  • minimalist: missing context

reliability

gives the same reading after repeated use

• test-retest method

• inter-coder reliability

test-retest method

method to measure reliability - results are consistent across different conditions/points in time

inter-coder reliability

measure of reliability (nominal) - trained in the use of a measure and given examples to code; more agreement amongst coders increases reliability

validity

operationalization or variable measures what should be measured

• internal: sample is unbiased - cause/effect study is accurate

  • scientific experiments have high levels of internal validity

• external: sample applies to real world (other situations)