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;
182 Cards in this Set
- Front
- Back
Experimental research is more ____ than other techniques
|
positivistic
|
|
where is it most widely used?
|
psychology
|
|
Commonsense language (3)
|
when you modify something in a situation, then compare an outcome to what existed without the modification
1) begin w hypothesis 2) modify something in a situation 3) compare outcomes w or without the modification |
|
3 conditions for causality
|
temporal order
association no alternative explanations * these are best met in exp. research |
|
appropriate research questions for exp (6)
|
- must be a topic in which conditions can be manipulated
- best for issues with a narrow scope - when trying to control elements that would be difficult to control in real life - limited resources and few variables - explantory - hpothesis testing |
|
random assignment (2 points)
|
a method of assigning cases to groups for the purpose of making comparisons.
- they must be exactly the same except for the variable being tested |
|
subjects
|
the cases or people used in research projects and on whom variables are measured
|
|
parts of the experiment (7)
|
1) treatment/indepdent
2) dependent 3) pretest 4) posttest 5) experimental group 6) control group 7) random assignment |
|
independent
|
treatment. what is modified. the term treatment comes from medicine, in which there is an intervener
|
|
dependent
|
things that will change in response to a treatment
|
|
pretest
|
measurement of DEPENDENT prior to intro of treatment
|
|
posttest
|
measurement of DEPENDENT after treatment
|
|
experimental group
|
grou that receives the treatment
|
|
control group
|
group that doesnt receive the treatment
|
|
deception
|
used when the research intentionally misleads sujects through written/verbal instructions, the actions of others, or aspects of the setting. (may involve confederates/stooges- people pretending to be subjects)
|
|
classical exp design (seed and cup)
|
- all other variations based off this
- random assignment, pretest, posttest, experimental group and control group pros: - random effect treatment effect - statistical analysis - can be made complex - strong internal validity cons: - low external validity (too over-controlled) |
|
pre-experimental design
|
- used when classical doesnt allow for interpretation of causal relationship.
3 types: 1) one- shot case study design 2) one group pretest-posttest design 3) static group comparison *None of these are random |
|
one-shot case study design
|
part of pre-experimental. It is one group, one treatment, and one posttest .used in exploratory
|
|
one group pretest-posttest design
|
part of pre-experimental. one group, pretest, treatment, posttest
|
|
static group comparison
|
part of pre-experimental. 2 groups, posttest, and treatment
* milgram? |
|
quasi-experiment and special designs
|
- variations of classic
- overall has less control over the independent than classical 6 types: 1) Posttest only control group 2) inturrupted time series 3) equivalent time series 4) latin square desigins 5) solomon 4-group 6) factorial designs |
|
posttest only control group (lecture also)
|
identical to static group comparison (2 groups, posttest, treatment)- with one addition: random assignment
|
|
inturrupted time series
|
one group + multiple pretests before and after
|
|
equivalent time series
|
logitudinal (ish)
one group + pretest + treatment + posttest +treatment + posttest + treatment + posttest + treatment... etc... |
|
latin square design
|
testing out different treatments in different orders
|
|
solomon 4 group (lecture also) (7)
|
- used to address issues of pretest effects (internal validity effects)
- combo of designs (ideal) - 4 groups - random assignment - 2 groups get pre test (treatment and control groups) - other 2 groups dont get pretest (treatment and control groups) - all 4 groups given same posttest and results are compared (to see if pretest had an effect) |
|
factorial designs
|
considers impact of several independent variables at the same time.
can have 2 effects: 1) main effects- each independent affected independently 2) interaction affects - categories or combos interact to create an effect beyond independence |
|
design notation
|
shorthand system for symbolizing the parts of an exp. design (see page 208)
|
|
selection bias
|
threat to internal validity. when groups in an experiment are not equivalent at the beginning
|
|
history
|
threat to internal validity. something unexpected that occurs and affects dependent during exp. more common in logitudinal studies
|
|
maturation
|
threat to internal validity. natural processes or growth, boredom, etc that occur to subjects during the experiment and affect dependent. more common in longitudinal
|
|
testing effect
|
threat to internal validity. when the pre-test itself affects the dependent. solomon 4 group helps this
|
|
instrumentation
|
a threat to reliability occurring when the dependent variable measure changes during the experiment.
|
|
mortality
|
threat to internal validity. due to subjects failing to participate through the entire exp
|
|
statistical regression
|
a problem of extreme values, or a tendency for random errors to move group results towards the average.
- some subjects may have began unusual and are unlikely to change much - or examine pre-tests for big changes |
|
diffusion of treatment/contamination
|
a threat to internal validity. occurs when the treatment "spills over" from the experimental group, and so control group modifies their behaviour because they learn of treatment
|
|
experimentor expectancy
|
threat to internal validity. indirectly communicating desired findings to subjects
|
|
what has been designed to control researcher expectancy?
|
double-blind experiment= neither the subjects nor the person who deals directly with the subjects knows the specifics of the experiment (which is control, which is treatment) (or the hypothesis)
|
|
placebo
|
a false treatment that appears to be real
|
|
external validity
|
the ability to generalize experimental findings to events and setting outside the exp.
|
|
reactivity
|
threat to external validity. arises when subjects are aware that they are in an experiment and being studied
|
|
hawthorne effect
|
a special kind of reactivity. threat to external validity. named after a famous case in which subjects reacted to the fact they were in an experiment, more than they reacted to the treatment
|
|
lab exp vs field exp on internal and external validity
|
lab= good internal validity (control), lower external validity (generalizability)
field= good extrnal validity, lower internal validity |
|
Experiments and surveys are both _____ which means...
|
reactive. the people are aware of the fact
|
|
nonreactive (and 3 examples)
|
- also called unobtrusive
- people being studied are not aware that they are part of a research project examples: -medical objective gaze (looking at body, not listening to patient) - physical traces - economic/fincancial audits |
|
Physical traces: erosion measures
|
non reactive measures of the wear or deterioration on surfaces due to the activity of people
|
|
Physical traces: accretion measures
|
nonreactive measures of the residue of the activity of people or what they leave behind (i.e. garbage)
|
|
archives
|
running records= regularly produced records that may reveal much
other records= irregular or private records that may also reveal much |
|
observation (3)
|
nonreactive
external appearances= how people may indicate social factors count behaviours= counting how many people do something time duration= how long people do something may show their attention |
|
creating nonreactive measures follows the logic of _____
|
quantitative. conceptualization--> links construct to evidence --> operationalization
|
|
content analysis
|
a technique for gathering and analyzing the content of a text.
used in a quantitative manner for nonreactive |
|
coding
|
turns aspects of content that represent variables into numbers
|
|
content analysis (nonreactive) is useful in three types of situation
|
- 1) helpful for problems involving a large volume of text (lots of newspapers)- using coding
- 2) it is helpful when a topic must be studied at a distance (people who have died, historical moments) - 3) can reveal messages in a text that are difficult to see with casual observation. |
|
coding system
|
a set of instructions or rules on how to systematically observe and record content from text.
|
|
structured observation
|
a method of watching what is happening in a social setting
is highly organized and follows sysematic rules for observing and documenting |
|
Coding systems identify four characteristics of text content
|
frequency, direction, intensity and space.
|
|
frequency
|
: counting whether or not something occurs, and if so, how often.
|
|
direction
|
: noting the direction of messages in the content along some continuum
|
|
intensity
|
the strength or power of a message in a direction
|
|
space
|
size of text. Volume allocated to it. Counting words, sentences, paragraphs, or space on a page
|
|
manifest coding
|
content analysis coding. research first develops a list of specific words, phrases, or symbols, and then finds them in a communication medium. Validity is a bit low because the word may mean different things in different situations
|
|
coding frame
|
an exhaustive list of all possible values that codes may take in content analysis
|
|
can a researcher use both manifest and latent
|
yes
|
|
intercoder reliability
|
- measuring the consistency between all of the coders you have hired
measured by the researcher with a statistical coefficient that tells the degree of consistency among coders. 1.0 means perfect agreement among coders - 80 or better is usually required |
|
what is the method for content analysis?
|
random sampling often used . Define population (i.e. all words, all sentences, etc) and the sampling frame (i.e. a magazine, a book, etc)
|
|
when is existing stats best
|
topics that involve info routinely collected by large bureaucratic organizations
|
|
social indicators movement
|
1960s. - Social indicators movement to develop indicators of social well-being
- this information could influence public policy |
|
Secondary survey data
|
- Reanalysis of previously collected survey or other data that were originally gathered by others.
- Stats can |
|
fallacy of misplaced concreteness
|
when a person uses too many digits in a quantitative measure in an attempt to create the impression that the data are accurate or the researcher is highly capable
|
|
Validity issues for existing stats/ secondary data (3)
|
- when the researchers theoretical definition does not match that of the info.
- Official stats may also be used as a proxy - the researcher lacks control over how info is collected. |
|
reliability issues for existing stats/secondary data (2)
|
- consistent measurement over time is not possible
- missing info |
|
A researchers ability to infer causality or test a theory on the basis of nonreactive data is...
|
is limited
|
|
Chapter 11
|
ANALYSIS OF QUAN DATA
|
|
Codebook
|
- Codebook= a coding system in a c format that computers can use.
|
|
precoding
|
the code categories on the questionnaire (i.e. 1-female 2-male)
|
|
in a grid
|
Row-represents a respondent subject or case
Column- represents specific variables |
|
- When data are in the computer, researchers verify data in two ways:
|
- 1) Possible code cleaning: checking the categories of all variables for impossible codes (i,e. Finding a 4 is impossible when the coding is 1-male and 2-female)
- 2) Contingency cleaning: cross-classifying two variables and looking for logically impossible variables. (e.x. occupation is crossed with education. If someone recorded never passing the 8th grade but is also recorded as being a doctor it lacks logic) |
|
- The easiest way to describe numerical data of one variable is with a
|
frequency distribution
|
|
bar and pie charts often used for..
|
discrete variables
|
|
histograms
|
upright bar graphs for interval or ratio data
|
|
mode
|
easiest to use and can be used with nominal ordinal interval or ratio. It is the most frequent number.
|
|
median
|
= middle point. 50th percentile. Can be used with ordinal , or skewed interval ratio
|
|
mean
|
arithmetic average. Most used. Only interval or ratio. Strongly affected by changes in extreme values
|
|
if there is a bell shaped normal distribution, than the three measures of central tendency...
|
equal eachother (zero variation_
|
|
variation is measured in three ways (measures of dispersion)
|
range (simpleist), percentile, and standard deviation
|
|
range
|
consistis of the largest and smallest scores. Ordinal interval ratio
|
|
percentiles
|
for one variable. indicates the percentages of cases at or below a point. ordinal, interval, ratio
|
|
standard deviation
|
= measure of dispersion most difficult to compute. However, it is the most comprehensive and widely used.
Interval ratio. Based on the mean and gives an average distance between all scores and the mean |
|
- If you add up the absolute distances between each score and the mean...
|
you get zeo
|
|
z-scores
|
determining the number of standard deviations it is above or below the mean. need the mean and sd
|
|
- Statistical relationships are based on two ideas
|
covariation and independence
|
|
covariation
|
the idea that two variable vary together, such that knowing the values in one variable provides info about values found in another variable
|
|
independence
|
opposite of covariation. It means there is no association or no relationship between variables.
|
|
- If the null hypothesis is used, the hypothesis is
|
that there is independence.
|
|
- Three techniques help researchers decide whether a relationship exists between two variables:
|
1) a scattergram, graph, or plot
2) cross-tab, or percentage table 3) measures of association, or other statistical measures. |
|
scattergram (3 aspects)
|
(interval and ratio)
- A researcher can see three aspects of a bivariate relationship in a scattergram: form, direction, and precision |
|
form
|
independence, linear, and curvilinear. Independence has no pattern, linear means a straight line can be pictured, and curvilinear means that a U or an S shape would be seen
|
|
direction
|
positive or negative*
positive is from the lower left to upper right (higher values of X go with higher values of Y etc), negative is a line from upper left to lower right (higher values on one variable go with lower values on the other). |
|
percision
|
*high or low level
the amount of spread in the points on the graph. A high level of precision occurs when the points hug the line that summarizes the relationship. A low level ovccurs when the points are widely spread around the line. |
|
contingency table
|
cross-tabulating two or more variables.
|
|
if there is no relationship in a table
|
the cell percentages look approximately equal across rows or columns.
|
|
a linear relationship looks like
|
larger percentages in the diagonals
|
|
the simplist way to see a strong relationship is..
|
to circle the highest value (stongest value) in each column or row. This rule only works if the percentages table is ordinal or interval, with the lowest variable categories beginning at the bottom left.
|
|
measures of association
|
a single number that expresses the strength and often the direction, of a relationship. It condenses info about a bivariate relationshop into a single number.
|
|
lambda
|
used for nominal. Based on a reduction in errors based on the mode and ranges between 0-indepdendance, and 1-perfect
|
|
gamma
|
used for ordinal. Based on compairing pairs of variable categories and seeing whether a case has the same rank on each. Ranges from -1 to plus 1, with 0 meaning no association
|
|
tau
|
also used for ordinal. Takes care of a few problems with gamma. Kendall’s tau. Ranges from -1 to plus 1, with zero meaning no association
|
|
rho
|
also called persons product moment correlation coefficient. Most commonly used measure of correlation. For interval or ratio. Used for the mean and standard deviation of the variables and tells us how far cases are from a relationship line on a scatterplot. -1 to plus 1, with 0 meaning no association. Linear only. If u square it, it can be used as a PRE
|
|
chi
|
two uses. Can be used as a MOA in descriptive and inferential. For ordinal or nominal. Has an upper limit of infinity, and lower limit of 0 meaning no association.
|
|
control variable
|
a `third`variable that shows whether a bivariate relationship holds up to alternative explanations (is it really spurious?). it can occur before or betwen other variables
|
|
trivariate tables
|
consist of multiple bivariate tables.
|
|
partials
|
tables that show the relationship between control variables and original variables
|
|
elaboration paradigm
|
. it describes the pattern that emerges when a control variable is introduced
|
|
replication pattern
|
the easiest to understand. When the same bivariate table is replicated because the control variable had no effect
|
|
the specification pattern
|
the next easiest. It occurs when one partial replicates the inital bivariate relationship but the other partials do not.
|
|
interpretation pattern
|
control variable intervenes so much that the bivariate relationship weakens or disappears
|
|
explanation pattern
|
looks the same as interpretation. The difference is the temporal order.
control variable comes before the independent |
|
supressor variable pattern
|
bivariate tables show independence, but a relationship appears in partials
|
|
multiple regression analysis (3)
|
- Requires interval ratio
- It controls for many alternative explanations and variables simultaneously - It s widely used in social science |
|
MRA tells the researcher 2 things:
|
- 1) the results have a measure called R squared, which tells how well a set of variables explains a dependent variable.
- Ex: an r squared of .50 means that knowing the independent and control variables improves accuracy in predicting the dependent variable by 50 % - 2) using beta, the direction and size of the effect of each variable on a dependant variable is measured - Ex: the bivariate correlation between X and Y is .75, next, the researcher statistically considers four control variables. If the beta remains at .75, than the 4 control variables have no effect. If the beta gets smaller, than the control variables had an effect |
|
statistical significance
|
means that results are not likely to be due to chance factors. It indicates the probability of finding a relationship in the sample when there is none in the population
|
|
level of significance
|
- Level of stat significance= a way of talking about the likelihood that results are due to chance factors- that is, that a relationship appears in the sample when there is none in the population. Its usually at 0.5, 0.1, or 0.001
|
|
if it is at 0.5, what does that mean (2)
|
- The odds of such results based on chance alone are 5%
- One can be 95% confident that the results are due to a real relationship in the population, not chance factors. |
|
type I error (alpha error)
|
occurs when the researcher says that a relationship exists when in fact, none does. Fasely rejecting a null hypothesis.
|
|
type II error (BETA ERROR)
|
- Type II error= occurs when a researcher says that a relationship does not exist, but in reality it does. It means falsely accepting a null hypothesis.
reudce this? - increase alpha - increase sample size |
|
which error are we worried about in sociologY?
|
type one
|
|
how are qual interviews different from conversationÉ
|
they have a set purpose
|
|
Select participants through nonprobability sampling, and which are most common
|
Snowball and purposive are used most often
|
|
selective transcription
|
They may only transcribe the parts that are relevant.
But, fully transcribed interviews are the best way to ensure that our findings are dependable and trustworthy. |
|
informants
|
people involved in a qualitative research project. They have 4 characteristics
|
|
4 characteristics of informants
|
- 1) very familiar with the culture
- 2) currently involved in the culture. Ex-members are ok, but the longer theyve been away form the culture, the more likely it is that they have reconstructed their recollections - 3) the person can spend time with the researcher - 4) nonanalytic individuals make better informants. They use native folk theory or common sense. |
|
interview guide
|
a list of questions that the researcher wants to ensure is covered during the course of the interview.
|
|
problems with reliability and validity (3)
|
- Sample sizes are small
- researchers should make all data available. - They can maximize the transferability (generalizability) of their findingins to populations beyond be selecting participants who are not entirely homogenous. (i.e. dancers from a variety of clubs and ages and martial statuses) |
|
focus groups
|
a special qual research technique in which people are informally interviewed in a group discussion setting. usually 6-12 people
|
|
moderator
|
the person who leads the focus group and asks questions to prompt group discussion.
|
|
group members should be...
|
homogeneous but not contain family or friends
|
|
advantages of focus groups (3)
|
- natural and open setting,
- feelings of empowernment, - may talk to eachother |
|
disadvantages of focus groups (5)
|
-polarization effect,
- only a few topics, -may produce fewer collective ideas, - researchers cannot reconcile differences - groupthink |
|
polarization effect
|
attitudes become more extreme after group discussion
|
|
groupthink
|
persons natural desire to avoid conflict and lean toward group consensus, even when the opinion of the group does not reflect his or her own opinions. To counteract this, use a devils advocate
|
|
devils advocate
|
a person whose role it is to argue against a dominant idea. Prevents tendencies towards groupthink. Not played by the moderator.
|
|
chapter 14
|
Historical comparative
|
|
what famous theorists used this technique
|
durkheim, marx, and weber
|
|
it is used extensively within (4)
|
social change,
political sociology, social movements, and social stratification .. good for addressing big questions |
|
positivist researchers may study..
|
historical or comparative issues, but they deny that there is a distinct H-C method.
|
|
3 dimensions of H-C
|
1) one nation, two, or many?
2) single time periods, or many years? 3) is analysis based primarily on quan or qual? |
|
- The logic and goals of H-C research are closer to those of
|
field research
|
|
6 similarities between field research and H-C
|
- 1) interpretation
- 2) subjective feelings - 3) grounded theory - 4) standpoint of others - 5) process and sequence. - 6) generalization and theory are limited. |
|
a H-C researcher reads evidence for 3 things:
|
conceptual frameworks
details empirical generalizations. |
|
history (3)
|
events of the past,
a record of the past, and a discipline that studies past |
|
histriogrpahy
|
the method of doing historical research or of gathering and analyzing historical evidence.
|
|
H-C Researchers draw on 4 types of historical evidence/data:
|
primary sources, secondary sources, running records, and recollections.
|
|
recollections
|
the words or writings of individuals about their past lives or experiences based on memory. Oral histories are often a form. This is useful for groups who do not write things down. The oral history technique began in the 30s.
|
|
problems with secondary sources (5)
|
- Too much info
- Rarely objective facts - Doesn’t know whether info left out is relevant - organize evidence as a narrative history. - Theory often remains hidden |
|
problems with primary sources (2)
|
- Only a fraction of everything written/used has survived
- Must hold back judgements and become a moral relativist |
|
1) internal criticism
2) external criticism |
1) appraises the meaning and intent of the data source (is it firsthand or secondhand)
2) appraises authenticity and ownership... is it fake? |
|
Galton`s problem
|
we've created units that are actually related to one another. (consistencies b/w doctors, nurses, and practicioners... we should instead study them as a whole) This is an important issue because cultures rarely have clear and fixed boundaries.
|
|
why isnt the nation-state a good unit of analysis
|
cultures may vary drastically within it
|
|
equivalence
|
It is the issue of making comparisons across divergent contexts, or whether a researcher, correctly reads, understands, or conceptualizes data about people from a different historical era or culture. Four subtypes: lexicon, contextual, conceptual, and measurement
|
|
Lexicon equivalence
|
language translation. finding equivalent words or phrases to express meaning in a different language
* use back translation |
|
contextual equivalence
|
the correct application of terms in different contexts. (i.e. the status of a priest across many cultures) (i.e. the meaning of attending university throughout history)
|
|
conceptual equivalence
|
the ability to properly use the same concept across divergent cultures or historical eras. Views are coloured by current life situations and experiences. ( i.e. the concept of class may vary, even though everuone has a word for it) concepts also vary over time.
|
|
measurement equivalence
|
measuring the same concept in different settings.
|
|
all data analysis is based on...
|
comparison
|
|
a qual researcher divides explainations into two categories
|
1) highly unlikely
2) plausable |
|
casing
|
bringing theory and data together
|
|
coding is two simultaenous activities;
|
data reduction
data categorization |
|
3 types of coding
|
open
axial selective |
|
open coding
|
find conceptual categories in the data. no concern for making connections
|
|
axial coding
|
look at relationships between the categories.
|
|
selective coding
|
to account for relationships, find core categories around which all other categories `fit`
|
|
analytic memo (2)
|
- made up of reflections, and ideas about coding.
- creates the link between raw data, formal theorizing and hypothesis creating |
|
Four strategies researcher use to analyze qualitative data
|
the narrative, ideal types, successive approximation, and the illustrative method
|
|
the narrative
|
detailed story of experiences in the field. lets the data speak for itself
|
|
ideal types
|
weber. compare ideal forms to empirical observation. used in two ways:
1) to contrast the impacts of context 2) as an analogy |
|
contrasting impacts of contexts
|
researchers who adopt a strongly interpretive approach may use ideal types to interpret data in a way that is senstitive to the context and cultural meanings of members.
|
|
analogies
|
- comparisons
Serves as a heuristic device (one that helps people learn or see). |
|
successive approximation
|
repeadely move back and forth between data and theory, until the gap between them shrinks and disappears. things are modified over and over to become successively more accurate
|
|
The illustrative method
|
find empirical examples in the data to support the theory. it is `filling the empty boxes`of theory with qualitative data
|
|
network analysis
|
sociograms
|
|
time allocation analysis
|
Examine duration or amount of time devoted to activities in order to examine priorities
|
|
multiple sorting procedure
|
Can be used in field research or oral history.
Its purpose is to discover how people categorize their experiences or classify items into systems of similar and different. |