Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
127 Cards in this Set
- Front
- Back
Qualitative methods
|
based on what people do, say, write, etc. Used to interpret aspects of peoples lives (interview survey)
|
|
quantitative methods
|
methods/analysis using the statistical summary of data. used to explain or predict causality/outcomes
|
|
inductive reasoning
|
start with data collection and little or no explanation and develop explanations from the data. Interpretive approach; phenomenological; used in qual; to understand feelings, experiences, and beliefs
|
|
deductive reasoning
|
start with theoretical concept, and gather data to either prove or disprove theory. Hypothesis testing; used in quant; positivist/scientific approach
|
|
Positivism
|
A true external reality exists and it can be known through objective observation.
|
|
Post-Positivism
|
A true external reality exists, but it cannot be known objectively because of bias.
|
|
Anti-Positivist
|
No true external reality exists.
|
|
Explanatory Research
|
Quantitative method mostly
|
|
Exploratory Research
|
Qualitative method mostly
|
|
Descriptive Research
|
Used to describe a group of people (mostly qual)
|
|
Evaluative Research
|
A particular subset of Explanatory Research mostly.
|
|
Generalization
|
Don’t over generalize so that we are protected from flaws in logic and reasoning.
|
|
Representativeness:
|
Using information that is true from a small group to represent the whole group, which might not be true for the larger mass. Make sure your data is representative of the larger population (the ratio of sub group is same as larger pop.)
|
|
Research Hypothesis:
|
Assuming there is a relationship between variables.
|
|
Evaluative Research
|
A particular subset of Explanatory Research mostly.
|
|
Null Hypothesis:
|
Assuming there is no relationship between variables.
|
|
Generalization
|
Don’t over generalize so that we are protected from flaws in logic and reasoning.
|
|
Representativeness:
|
Using information that is true from a small group to represent the whole group, which might not be true for the larger mass. Make sure your data is representative of the larger population (the ratio of sub group is same as larger pop.)
|
|
Sufficient Cause
|
Can a factor in and of itself be enough to cause an outcome? (i.e.: don’t take test= sufficient to fail the test)
|
|
Research Hypothesis:
|
Assuming there is a relationship between variables.
|
|
Necessary Cause
|
A cause that must be present in part to cause an outcome (being a woman is a necessary cause of getting pregnant)
|
|
Null Hypothesis:
|
Assuming there is no relationship between variables.
|
|
Spuriousness
|
(relationship) a true unexplored causal factor does exist.
|
|
Paradigm/ Theoretical Perspective:
|
Over arching, generalized way of seeing the world
|
|
Sufficient Cause
|
Can a factor in and of itself be enough to cause an outcome? (i.e.: don’t take test= sufficient to fail the test)
|
|
Structural Functionism
|
The way in which interrelated parts of society work together to assure continuity of the whole. (1920-1960, Macro view, Durkheim, T. Parsons) Concept: Anomie: (Durkheim) A state of normlessness and disconnection from the whole.
|
|
Necessary Cause
|
A cause that must be present in part to cause an outcome (being a woman is a necessary cause of getting pregnant)
|
|
Spuriousness
|
(relationship) a true unexplored causal factor does exist.
|
|
Paradigm/ Theoretical Perspective:
|
Over arching, generalized way of seeing the world
|
|
Structural Functionism
|
The way in which interrelated parts of society work together to assure continuity of the whole. (1920-1960, Macro view, Durkheim, T. Parsons) Concept: Anomie: (Durkheim) A state of normlessness and disconnection from the whole.
|
|
Marxism/ Conflict Paradigm
|
In society groups are in conflicts over valued resources and change is caused by these conflicts. (1960-1970, Macro view, Karl Marx) Concept: Alienation: (Marx) Condition of workers in an industrial system; alienated from product they made, from nature/environment, from fellow workers.
|
|
(Symbolic) Interactionism
|
(minority paradigm) We make meaning in our world through the use of words and symbols. (Minute interaction studies/ Micro, G.H. Mead, Blumer) Concept: Roll Distancing: (Gothman) through expressing distain for activity they are distancing themselves from someone they don’t want to be considered similar to.
|
|
Feminism
|
Gender inequalities are central to life, and we should pay attention to that. (1960-1970, this paradigm can intersect with any of the other paradigms) Concept: Emotional Labor: (Hoschild) The work people do in managing your own and other’s emotions.
|
|
Post-Modernism
|
There is no one Truth that exists across place and time, and our world is characterized by fragmentation, simulation and hybridity. (1980’s) Concept: Transmigrant: Someone who doesn’t more from one country to another, but maintains active economic and political connections with home country and new country.
|
|
Concept
|
Generalized statement about a class of phenomena empirical generalization), a word or phrase that describes something that we all know what it means. (i.e.: chair=stool.) Not theories because they don’t explain why/ how, jut are neat summarizations of terms.
|
|
Conceptualization (Conceptual Definition)
|
Process where we elaborate the precise definition of our variable/concepts (make sure we have a precise definition)
|
|
Operationalization (operational Definition)
|
The process of deciding how to measure the absence/presence/ or degree of presence of a variable.
|
|
Distribution/ Normal Distribution
|
has a distribution where there is symmetrical variation around the mean with no skewedness.
|
|
Descriptive Statistics
|
Statistics that summarize characteristics of a sample of data
|
|
Inferential Statistics
|
The body of statistical computations relevant to making generalizations from findings based on a sample to the larger population.
|
|
Measures of Central Tendency
|
Mean, Median, Mode
|
|
Mode
|
Most commonly occurring result.
|
|
Median
|
Score in the middle, used for ordinal data, equal interval data
|
|
Mean
|
Add all the scores; divide by the number of observations. Use with attitude scores.
|
|
Measures of Dispersion
|
Range, Standard Deviation.
|
|
Range
|
ex: age 21-28. Doesn’t work if there are outliers. Interquartile Range=
25-75% (no extreme cases included) |
|
Standard Deviation
|
A calculation of the average amount of variability in a set of scores/ observations. The average distance from the mean in a set of scores.
|
|
Skewness (positive and negative):
|
Positive: most results are above the mean; the graph is weighted towards the right
Negative: most results are below the mean; the graph is weighted towards the left
|
|
Z-Scores
|
Tells us exactly what % of population falls above and below a specific value on an observation.
|
|
Central Limit Theorem
|
when you sample repeatedly, it will approximate the natural distribution.
|
|
Type I Error
|
Claiming there is a significant difference when there isn’t actually one.
|
|
Type II Error
|
Failing to claim a difference where there is one.
|
|
Theory
|
set of interrelated statements that seek to explain how or why events happen
|
|
Theorizing
|
explaining underlying causes or relationships between observed phenomena and intuiting concepts and making logical, systematic explanations from them
|
|
Concept
|
(empirical generalization) generalized statement about a class of phenomena; a word or phrase that we all have a generalized idea of
|
|
Conceptual Ordering
|
(typology) organizing date into categories because of their characteristics; precursor to theorizing
|
|
Independent variables
|
variables that help to determine what you are studying; social properties; demographic
|
|
Dependent variables
|
the variable that you are measuring, its outcome is dependent on the independent variables
|
|
Control variables
|
a variable that is the same for all subjects, to keep one aspect consistent
|
|
Indicators
|
the subunits which we can combine to measure the degree of presence of a certain variable (1.e. satisfaction)
|
|
levels of analysis
|
micro- individual
maso- groups or network macro- worldwide, large institutions |
|
unit of analysis
|
who you're actually studying
|
|
unit of observation
|
from whom/what did I get my data from
|
|
Categorical variable
|
made of groups that define a person, not numerical; i.e. gender
|
|
Continuous variable
|
can be measured low to high, implies a rank; associated with number
|
|
Validity
|
a valid result represents the truth accurately
|
|
Composite Measures
|
Index: several indicators that add up to a total score to imply something
Likert: used to express attitude, usually appear in series, presents a scale for a participant to choose where they fall Scale: point system of answering, different answers are allocated certain amount of points |
|
Correlation:
|
Direct/positive: high scores are associated with high scores and low scores are associated with low scores
Indirect/negative: those who are high on one variable are low on the other |
|
Levels of Measurement
|
nominal, ordinal, interval, ratio
|
|
nominal
|
identity (being able to name or list into categories); least amount of info
|
|
Ordinal
|
identity, magnitude (can be rank ordered, but can’t know exact distance)
|
|
Interval
|
identity, magnitude, equal unit size(no zero but can know the distance)
|
|
Ratio
|
identity, magnitude, equal unit size, true zero point (has a real zero and can be said in a ratio); most info
|
|
How to make appropriate measurement categories:
|
They categories must be exhaustive (every possible answer is accounted for) and they must be mutually exclusive (Bad: 2-3, 3-4 Good: 2-3, 3.1-4)
|
|
How to collapse more precise measures into less precise measure:
|
Any more complex data can be collapsed into less precise data but not vice versa, for example, age, if initially recorded at the ratio level, could be collapsed into the interval or ordinal level.
|
|
Methods for checking reliability
|
-one of the most often used and obvious ways of establishing reliability is to repeat the same test on a second occasion, known as test/retest reliability.
-if the test is reliable, we expect the two scores for each individual to be similar, and thus the resulting correlation coefficient will be high (close to 1.00) -inter-rater reliability- rate to see if questions are interpreted as intended |
|
Hot to calculate and interpret a Z score:
|
A Z score tells exactly what percent of population falls above and below a specific value on an observation (how many standard deviations above or below the mean). (Result – mean)/SD
|
|
Chi-Square Tests of proportion
|
shows if the results of one variable are statistically significant
-tells if there’s a statistically significant difference between the expected and observed -one nominal variable -ex: number of votes cast for each candidate |
|
Chi-square test of independent means
|
shows if variables are related
- tells if there’s a statistically significant difference between the expected and observed - two nominal variables each with two or more categories -ex: have every student report their gpa and race, and collapse gpa into categories; voting preferences for men and women |
|
T-test of independent means
|
etermines if the difference in the mean between two groups in a population is statistically significant
-2 variables: one dichotomous categorical variable (nominal, two groups), one continuous variable (take a mean) - ex: students report race and gpa: white or non-white, take mean of gpa |
|
Paired sample T-test:
|
test/retest method (same population) tells if the change in mean is statistically significant
-ex: testing the success of a training program |
|
One way Anova
|
determines if the difference in the mean between three or more groups in a population is statistically significant
- two variables: independent variable- categorical with three or more categories; dependent variable is continuous (often likert/attitude/frequency variable) - ex: amongst four groups is there a difference in the total satisfaction score (combined from several indicatiors) |
|
Facotrial Anova
|
determines if the difference in the mean between three or more groups in a population is statistically significant when influenced by more than one independent variable (shows if there is an interaction effect)
-three variables: two independent categorical variables with two or more categories, one dependent scale variable - ex: among immigrants, do gender and martial status affect the mount of money send to relatives |
|
Linear regression
|
-used for prediction and description
-explanatory variable- I.V., predictatory variable- D.V - simple: one I.V., (can be dichotomous or scale), one D.V. (scale) -multiple: many 1.V.s (can be dichotomous or scale), 1 D.V. (scale) -finds ordinary least squares regression (OLS) which estimates the line of best fit -allows you to see the difference between the expected values and the observed values - ex: age at first contraception= 9.96(intercept value)-.26(foster care)-(.55)(black) - often uses dummy variables (0 or 1) |
|
Logistic regression:
|
-helps us to determine what variables have an effect on an outcome/no outcome D.V.
-tells us if the difference in chance of having an outcome or no outcome for different groups is statistically significant -at least two variables: one D.V. (dichotomous/dummy), one or more independent variables (dummy or scale) -logit coefficient- tells how likely - positive logit: more likely than (e^logit>1) - negative logit: less likely than (e^logit<1) |
|
How to calculate expected values for a contingency table or cross-tabulation:
|
-Always use percents
-Calculate down, compare across -to calculate expected values: (row total x column total)/total total |
|
When relations may be curvilinear or loglinear instead of linear:
|
Curvilinear: instead of linear correlation, the data yields a simple regression equation that produces a parabola (for example, the number or crimes committed by age)
Loglinear: the data is represented by a logarithmic equation that produces a logarithmic curve (for example, number of cigarettes smoked by smokers throughout their life) |
|
Non-Probability
|
no known chance of selection into the sample
|
|
Paired sample T-test:
|
test/retest method (same population) tells if the change in mean is statistically significant
-ex: testing the success of a training program |
|
convenience
|
get information from people you have easy access to
|
|
One way Anova
|
determines if the difference in the mean between three or more groups in a population is statistically significant
- two variables: independent variable- categorical with three or more categories; dependent variable is continuous (often likert/attitude/frequency variable) - ex: amongst four groups is there a difference in the total satisfaction score (combined from several indicatiors) |
|
snowball
|
request referal from respondant to further respondants (often used with clandestine/marginal groups/rare groups)
|
|
Facotrial Anova
|
determines if the difference in the mean between three or more groups in a population is statistically significant when influenced by more than one independent variable (shows if there is an interaction effect)
-three variables: two independent categorical variables with two or more categories, one dependent scale variable - ex: among immigrants, do gender and martial status affect the mount of money send to relatives |
|
Linear regression
|
-used for prediction and description
-explanatory variable- I.V., predictatory variable- D.V - simple: one I.V., (can be dichotomous or scale), one D.V. (scale) -multiple: many 1.V.s (can be dichotomous or scale), 1 D.V. (scale) -finds ordinary least squares regression (OLS) which estimates the line of best fit -allows you to see the difference between the expected values and the observed values - ex: age at first contraception= 9.96(intercept value)-.26(foster care)-(.55)(black) - often uses dummy variables (0 or 1) |
|
Logistic regression:
|
-helps us to determine what variables have an effect on an outcome/no outcome D.V.
-tells us if the difference in chance of having an outcome or no outcome for different groups is statistically significant -at least two variables: one D.V. (dichotomous/dummy), one or more independent variables (dummy or scale) -logit coefficient- tells how likely - positive logit: more likely than (e^logit>1) - negative logit: less likely than (e^logit<1) |
|
How to calculate expected values for a contingency table or cross-tabulation:
|
-Always use percents
-Calculate down, compare across -to calculate expected values: (row total x column total)/total total |
|
When relations may be curvilinear or loglinear instead of linear:
|
Curvilinear: instead of linear correlation, the data yields a simple regression equation that produces a parabola (for example, the number or crimes committed by age)
Loglinear: the data is represented by a logarithmic equation that produces a logarithmic curve (for example, number of cigarettes smoked by smokers throughout their life) |
|
Non-Probability
|
no known chance of selection into the sample
|
|
convenience
|
get information from people you have easy access to
|
|
snowball
|
request referal from respondant to further respondants (often used with clandestine/marginal groups/rare groups)
|
|
quota
|
need to get a certain number from each group
|
|
Probability
|
each pesron/entity has a known chance of selection into sample
|
|
Simple Random
|
generate random numbers and survey those individuals, everyone has equal chance
|
|
Systematic Random
|
start with one random number, add the same interval repeatedly
|
|
Stratified random
|
groups clumped together based on some identifiable chatacteristic, then do systematic random
-assures you get resprentative percent of each group |
|
Multi-stage sampling (complex)
|
do simple random or systematic random in several steps from larger to smaller groups
|
|
Behavorial
|
is often or do you do x
|
|
affective
|
emotional/intellectual; how do you feel about x?
|
|
social property/demographic
|
race, class, gender (usually I.V.)
|
|
Ordinal
|
identity, magnitude (can be rank ordered, but can’t know exact distance)
|
|
Interval
|
identity, magnitude, equal unit size(no zero but can know the distance)
|
|
Ratio
|
identity, magnitude, equal unit size, true zero point (has a real zero and can be said in a ratio); most info
|
|
How to make appropriate measurement categories:
|
They categories must be exhaustive (every possible answer is accounted for) and they must be mutually exclusive (Bad: 2-3, 3-4 Good: 2-3, 3.1-4)
|
|
How to collapse more precise measures into less precise measure:
|
Any more complex data can be collapsed into less precise data but not vice versa, for example, age, if initially recorded at the ratio level, could be collapsed into the interval or ordinal level.
|
|
Methods for checking reliability
|
-one of the most often used and obvious ways of establishing reliability is to repeat the same test on a second occasion, known as test/retest reliability.
-if the test is reliable, we expect the two scores for each individual to be similar, and thus the resulting correlation coefficient will be high (close to 1.00) -inter-rater reliability- rate to see if questions are interpreted as intended |
|
Hot to calculate and interpret a Z score:
|
A Z score tells exactly what percent of population falls above and below a specific value on an observation (how many standard deviations above or below the mean). (Result – mean)/SD
|
|
Chi-Square Tests of proportion
|
shows if the results of one variable are statistically significant
-tells if there’s a statistically significant difference between the expected and observed -one nominal variable -ex: number of votes cast for each candidate |
|
Chi-square test of independent means
|
shows if variables are related
- tells if there’s a statistically significant difference between the expected and observed - two nominal variables each with two or more categories -ex: have every student report their gpa and race, and collapse gpa into categories; voting preferences for men and women |
|
T-test of independent means
|
etermines if the difference in the mean between two groups in a population is statistically significant
-2 variables: one dichotomous categorical variable (nominal, two groups), one continuous variable (take a mean) - ex: students report race and gpa: white or non-white, take mean of gpa |
|
Major Challenges to Ethics:
|
1.Stanford Prison Experiment: Beasement in Stanford converted to mock prison. Subjects assigned roles, got too into roles, got harmed, didn’t understand risks.
2.Humphrey’s Tearoom Trade: Sociologist posed as “watch queen” of gay sexual encounters in pubilc restrooms. Recorded license plates, went to homes disguised and administered surveys. 3.Milgram’s Obedience Study: Subjects instructed to give electric shocks to “learners” when they answered incorrectly in memory tests. ⅔ participants gave lethal voltage. 4.Tuskegee Syphilis Study: Researchers made black males believe they were mreceiving treatment for syphilis when they were not, even once it became available. Went on for 40 years. |
|
Ethical Ground Rules:
|
-Do no harm (physical or psychological)
-Have fully informed consent -Do not invade personal privacy -Do not deceive participants -Do not betray a participant’s confidentiality |
|
How to Calculate Expected Values in 2 way contingency tables:
|
Fe= Row Total x Column Total
Total Total |
|
Conditions of Causality:
|
-Correlation: A change in 1 variable is associated w/ a change in another
-Time Order: Supposed cause must come before effect -Non-Spuriousness: Must establish that change in 1 variable is not caused by some as-yet-unidentified variable |
|
When to choose which method (quan or qual)
|
Ideally determined by the research question. More cases is more likely quantitative, less cases is more likely qualitative.
|
|
Scientific Method:
|
State a problem, suggest hypothesis, gather data, draw conclusions.
|
|
Common Logical Errors
|
-Non-systematic observation
-Over reliance on others’ answers -Overgeneralization/stereotyping -Jumping to conclusions, esp. about cause & effect -Seeing just our own interests -Influence of ego, tradition and authority |
|
Statistical Significance
|
Significant at 2 standard deviations above or below the mean (Upper and lower most 2.5%)
|