Research Methods Flashcards
The Research Cycle
1.Figure out what matters to you (values, context, state of field)
2.Formulate a testable theory (tentative explanation of a phenomenon that usually specifies cause[s] + effect[s])
•Choices governed by social context/funding
The Research Cycle
- Review existing literature for inspiration + prevent duplication, we want to add to knowledge
- Select method(s) appropriate for the problem
- Collect data
The Research Cycle
- Treat subjects ethically to ensure no harm, anonymity, etc.
•aware of risks, respect of privacy - Analyze data
The Research Cycle
- Report results so they can be evaluated, rejected/accepted, applied
Conference, journal, book: we can’t contribute if we keep knowledge a secret – science is a social institution
•rule is attack: everybody else is obliged to systematically find flaws in research
•encourages researchers to do better next time
•add to stock of knowledge + formulate policies
•give others opportunity to criticize
Operationalization
establishing criteria for assigning numerical value or qualitative values to variables
reliable
consistent results
valid
measurement procedure measures what it is supposed to
Causality
measurement of causes and their effects
generalizable
research findings hold in many contexts
Validity, reliability, and generalizability may be explained by drawing an analogy between measuring a variable and firing at a bull’s-eye
case 1, measures far apart (not reliable) + far from bull’s-eye (not valid)
case 2, shots close (reliable) but far from bull’s-eye (not valid)
Validity, reliability, and generalizability may be explained by drawing an analogy between measuring a variable and firing at a bull’s-eye
case 3, shots close to bull’s-eye (valid) + close to each other (reliable).
case 4, use a second target - shots close to each other (reliable) + close to the bull’s-eye (valid) again.
Because measures were valid + reliable in both cases 3 + 4, we conclude that our results are generalizable.
participant observation
observe social setting from outsider’s POV + take part in activities of subjects
•analytically, systematically, from the outside observe
•participation only: not able to observe objectively, lack of outside information
•only by regularly stepping apart do we get the chance to see what subjects are blind to
•only outside: we don’t understand the culture or why, we don’t know what it means
•we have to move back and forth
Participant Observation: Advantages
Promotes understanding meaning of subjects’ actions
Good for exploratory research (theory construction)
•Often engaged in process of theory construction – mechanism for causal theory
Participant Observation: Disadvantages
High potential for reactivity
Bad for theory testing (establishing cause and effect)
Low reliability: subjective – might be different if someone else did it
•Researchers right there – subjects acting in way to please researchers
•Gotta gain confidence + trust – important if subjects are very different from you
experiment
carefully controlled artificial situation that allows researchers to isolate hypothesized cause + measure effect precisely
•Remove some of the problems from observation
Experiments
1 Recruit subjects: need ppl, advertise
2 Randomly assign them: randomization removes bias
3 Experimental group: group exposed to independent variable (hypothesized cause)
3 Control group: group not exposed
Experiments
4 Measure dependent variable in experimental + control
5 Expose experimental group to independent variable
5 Do not expose control group to independent variable
6 Re-measure dependent variable in experimental + control to determine the degree to which it has changed
•if any change there can only be 1 cause
•we have to do this over and over again to ensure both groups are alike
Experiments: Advantages
Very high reliability
Very good for theory testing (establishing cause and effect)
•Measure things precisely consistently
Experiments: Disadvantages
Do not promote understanding of motives + culture Not useful for exploratory research (theory construction) Low validity (in sociology): not real life setting
Surveys
form containing questions mailed to the respondent + returned to the researcher through the mail system, not very effective
face-to-face interview survey: questions presented by the interviewer during a meeting
Surveys
telephone survey: questions presented by the interviewer over the phone, frequent
•Self select – must be highly motivated, not representative
Surveys
closed-ended question provides respondent with a list of permitted answers
open-ended question allows respondents to answer in their own words.
Sampling
sample is part of a group.
voluntary response sample: group of people who chose themselves in response to a general appeal – not scientific
Sampling
convenience sample: ppl easiest to reach – results not reliable, don’t like this
representative sample: group chosen randomly so characteristics closely match those of the population of interest
Sampling
sampling frame: list of all ppl in the pop of interest.
randomizing method: way of ensuring every person in sampling frame has chance of being selected
•1000-1500 – gonna be pretty accurate
•because of unknown biases/bad luck – it might be outside standard error
Sampling Error
•conventionally tolerate 5% chance that sample is different than population
Support for party A is 48%. Support for party B is 50%
measures are accurate ± 2.5%, with margin of error measures fall within overlapping margins of error, conclude measured diff in pop not statistically significant
Sampling Error
48%-55% measures fall outside margins of error, conclude measured diff in pop statistically significant.
•probability sampling permit us to generalize within known margins of error
Surveys: Advantages
Good reliability + validity
Good for theory testing (ease of quantification)
•because it can be repeated
•allow us to see relationship between causes + effects
Surveys: Disadvantages
Not great for exploratory research
contingency table
cross-classification of cases by at least two variables that allows you to see how/if variables are associated
contingency table
Percent of cases in each column that fall into a category of each row variable
Independent variable in column
Dependent variable: rows
contingency table
total frequency (n): Number of cases in each column
total percent: Percent of cases in each column
percentage difference: Measure of strength of association
•bigger the percentage, the stronger the association
•no difference means there is no effect
•strength of association determined by percentage difference
Testing an Association for Spuriousness (= “phoniness”)
(1) believe there is a causal relationship between TV viewing and aggressiveness
(2) By controlling for gender we can see whether gender has created a spurious association between TV viewing and aggressiveness
Testing an Association for Spuriousness (= “phoniness”)
•engaging in statistical manipulation to eliminate irrelevant variable
•analysis of survey data looking at relationship between hypothesized causes + effects through statistical
control factors: explain original association
•allows you to conclude original association was spurious it was actually the control factors
Positive + Negative Freedom
- Negative freedom: freedom from constraints that would otherwise prevent me from doing as I wish
- Positive freedom: capacity to act rationally, involves taking control of one’s life and realizing one’s best interest
Conclusions
- Peace in the Middle East
- Decreasing social solidarity = lower rate of altruistic suicide
- higher standards of security + well-being to most vulnerable - natural disasters
- More negative + positive freedom
- State intervention
- engaging + higher-quality schools, widely accessible childcare programs to inner-city
Spuriousness - Experimental
remove effect of extraneous variables by randomizing allocation of subjects to experimental + control groups + repeating experiments many times
Spuriousness - Nonexperimental
remove the effect of extraneous variables by controlling for variables that we expect may be irrelevant