Midterm 1 Flashcards
Writing Longhand produces what results
Better results
Epistemological Orientation
The beliefs about how research should be conducted and influence research. ex. scientific approach or more sensitive to social world
2 Types
- Postivist
- Interpretivist
Ontological Orientation
Considers the nature of social phenomena ex. is it inert and beyond our influence or are they a product of social interaction
2 types
- Objective
- Constructionist
Positivist
Advocates scientific method
Knowing something or not
Generate hypothesis that can be tested
-epistemological
Interpretivist
Social world can’t be treated the same as the natural world (it moves)
-epistemological
Realism / realist
The natural and social sciences can and should apply the same kinds of approach to the collections of data
Objectivist
Social Phenomena confront us as external facts that are beyond our reach or influence
ex. an organization that exerts pressure on its members
- Does not matter who you are, only environment
- Ontological
Constructionist
Social factors and their meaning are constantly being achieved by actors
- Matters who you are
- Ontological
Measurement Validity
Whether a measure that is devised of a concept reflects the concept being measured
Primarily quantitative (quantitative)
-face validity can be a part of it
Internal Validity
Causality, does A cause B
- independent and dependent variable
- How confident are we that it does
External Validity
Can the study be generalized beyond the specific research context
Ecological Validity
Does the study represent the real world
-stanley-trustgame
The money buys happiness study which measures how amount of money contributes to mental health only to a point in people has the strongest type of what kind of validity
Measurement validity
Quantitative research
typically numbers based
deductive-theory to confirmation
Qualitative research
typically observation or words based
inductive- observation to theory
Reliability
Can you get the same result multiple times
ex. how much variations were there in the telephone interviews and will we get the same answers in the evening
Replicability
Can someone else achieve this result
-different than reliability
Validity
Do findings mean what they say (4 types)
Face Validity
Intuitively, does it measure what it say it does ex. using gpa for intelligence
Concurrent Validity
Does it correlate with another measure in another study
-does absenteeism from work relate to job satisfaction?
Predictive Validity
Does it correctly predict some future event
Research Design
Framework for the collection and analysis of data 5 types -experimental -cross sectional -longitudinal -case study -comparative
Research Strategy
Your ideas about how theory and research and linked.
Your idea about what counts as knowledge (epistemological) and how things exist in the real world (ontological)
Research Method
The technique you use for collecting and analyzing data
Experimental study
High control of setting manipulates independent variable(s) control and test groups always trumps other designs Strong: internal validity Weak: ecological validity and external validity ex amodio & devine study (research design)
Cross-sectional study
Multiple cases (because variation) at a single point in time and establishes quantifiable data
-patterns of association
-typically a survey
Strong: External validity (generalizable)
Weak: Internal Validity
(research design)
Longitudinal study
When a sample is surveyed at least one extra time on a further occasion
panel study- random selection with data collected and not necessarily associated with a group
cohort study- data pulled from a cohort (group of people) randomly selected or otherwise multiple times
needs to be the same group to be longitudinal
Internal validity weaker, but time might improve this
(research design)
Case Study
Detailed analysis of a single case -the case is the unit of analysis critical case-chosen to test theory extreme case-chosendue to atypicality revelatory case- a lucky opportunity to study something longitudinal case- can be studied over time ex case study of "genie" who was locked away until she was 13 -case study trumps longitudinal Strong: internal strongish: ecological weak: external (research design)
Comparative Study
two case studies or two cross sectional studies compared ex 2 hosptials, 2 individuals, 2 couples Strong: external validity Weak: internal validity (research design)
In the money buys happiness study, where emotional well being and evaluation of life were studied, what was the independent variable? and what were the results?
income was the independent variable
Happiness increased to 75k and then leveled off
evaluation of life went up
has measurement validity
Unit of analysis
what is being studied ex. group, town, arrests
Main steps in quantitative research
(just remember there are 4-5)
theory, hypothesis, research design, analysis/conclusion
Likert Scale
Common technique in surveys about attitudes, perceptions, and behaviors
- 5 to 7 point scale
- must involve statements, statements related to some object, items should be interrelated, usually some reverse coded
Implicit prejudice in the white/black face response
participants responding faster or slower to negative or positive words following black faces
trustworthiness scale: 1-9 scale on neutral male faces black and white
Criticisms of quantitative research
Fails to distinguish people/ social institutions form the world of nature
The idea that the measurement process interpreted consistently by different people is flawed
analyzing can become disconnected
Sample population
The universe of units from which a sample is selected
sample
the segement of the population selected for study
Sampling frame
what you will draw on to obtain the sample ex. list of all registered voters
two types of probability sampling
probability sampling and non probability sample
sampling bias and when does it happen
a sample that does not represent the population well–some members are more likely to be selected than others
- non probability sampling method
- sampling frame is inadequate
- systematic non-response within sample
Simple random sampling
equal chance ex. hat drawing
systematic sampling
sampling frame
-take every 3rd person
(random)
stratified random sampling
equal chance, but first sorted into groups
inclusion is random within criteria
and you can set up proportions ex. every 5th republican
(random)
multistage clustering sample
random at each level, but there are multiple levels
-look for tree looking thing
(random)
convenience sample
whoever you can get
non random
snowball sample
you start with one and add
-when the population is not clear
(non random)
Quota sample
you take what you can until you reach a quota
non random
Open questions vs closed questions
open- write in answer
closed- categories
principles of ordering
general to specific
broader question followed by components
demographic questions late
-delicate and open questions should not be first
order bias
order of questions specificity will guide understanding of overall experience
can put “overall experience” at the beginning so small details don’t create bias
consistency in regards to order effects
1 person should have the same outlook over multiple questions
ex. same political beliefs over multiple questions or environmentally friendly over multiple questions
contrast in regards to order effects
once more than one question is asked, the first effects the latter
ex. more people favor civil unions after being asked about gay marriage
additive or positive in regards to order effects
when the answer to the second question goes up or the response is higher in response to the first
negative or subtractive in regards to order effects
when the second response goes down or decreases
Good and bad of open questions
Good: allows for respondents to answer (so unusual responses)
No answer suggested
Generates fixed answers
Bad: time consuming
can be unreliable to code
more respondent effort
variability in recording can compromise validity
Good and bad of closed questions
Good: easy to process answers easy to compare answers meaning of questions clarified less interviewer variability Bad: less spontaneity difficult to be exhaustive interpretation may be different
Presuppositions in question design
things that are assumed
ex. how often do you goto the movies? assumes you goto the movies
Question polarity in question design
a question can tilt it negatively
ex in the evening do you eat (any) sugar?
Social preference questions often work in some areas but others they dont
ex do you have sex with several different partners?
how many sexual partners do you currently have?
Agreement bias in question design example
To what extent do you think obama has done the best job he could have during 2010? more likely to agree
To what extent do you think obama has done the worst job in history?
Interval/ratio variable
variable where the distances across the categories are identical
ex. minutes
Ordinal variable
variables who’s categories can be rank ordered, but the distances are not equal across the range
ex 1-2 years 3-4 years 5-6 years 6+ years
Categorical or nominal variables
variables who’s categories cannot be ranked
ex race and ethnicity
Dichotomous variables
variables containing data that only have 2 categories
What is univariate analysis and what is the most popular tool to visualize it?
the analysis of one variable at a time
frequency table
ex. number of teenagers that drink 83%
Interval data and measure of central tendency in univariate analysis
mean-average
median-mid point
mode-most frequent value
What is bivariate analysis
analysis of two variables at a time to uncover whether the two variables are related
-relational
methods of bivariate analysis
scatter plot (correlation line from trending points) works well with interval data (shows correlation) contingency tables-work well with ordinal/dichotomous/categorical data
what is used as a confidence to see if an association is real?
p-value - probability by chance
p
what are statistical tests in variable analysis
tests assess probability that association between two variables is real
which test depends on variable type
-the relationship between ind. and depend. because we don’t know causality
difference between association and causality
causality- Independent variable must actually lead to dependent variable
association hypothesis- IV and DV are associated with each other
-surveys aren’t great at supporting causality
Structured observation
A method for systematically observing the behavior of individuals in terms of a schedule of categories
advantage: behavior observed directly can be more valid
disadvantage: more finite access to behavior
Incident observation
wait for something to happen and then observe what happens
strategy for systematic observation
Across time observation
across some period of time
ex. code for 20 minutes everything
At particular times observation
code every x seconds
What is an example of a threat to validity (guinea pig)
guinea pig effect
aware of study so you create a better impression
express socially preferred attitudes