Reading...
Play button
Play button
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;
188 Cards in this Set
- Front
- Back
|
What are two requirements for an experimental design?
|
1. manipulation of variables
2. random assignment of subjects to condition |
|
|
What are the four basic experimental designs?
|
1. posttest only
2. pretest/posttest 3. repeated measures design (most complicated) 4. matched pairs design |
|
|
What is a posttest design?
|
you only test subjects once
participants randomly assigned- a. empty cage -aa. testosterone b.female in cage - bb. testosterone |
|
|
What are limitations of a posttest design?
|
1. potential variability between groups-only fixed by random assignment when sample size is large enough
2. rats in hormone study = hard to get large sample size 3. testosterone levels can range from person to person |
|
|
What are sources of variability in testosterone?
|
1. pulsability of testosterone (happens every 90 min)
2. circadian rhythms 3.individual diff. |
|
|
Describe the pretest-posttest design
|
1. test subjects twice: once before any manipulation and once after manipulation
2. allows for comparison of how scores on measure change as a result of the manipulation: can control for individual diff. in response |
|
|
What are advantages of a pretest-posttest design?
|
1. even in cases of small sample size, can check to see groups are approx. equal on variables of interest
2. can directly measure changes w/in an individual, not just group. experimental manipulation might have diff. effects depending on initial value of measure (ceiling/floor effects) 3. can draw inferences about manipulation despite mortatlity |
|
|
What are disadvantages of a pretest-posttest design?
|
1. if measure is expensive = increase in cost per subject, esp. with hormone studies
2. testing same measure a 2nd time can affect how subjects respond =practice/carryover effects. and requires more rigorous set of controls. solution= don't give any feedback from first test. give diff. order of qs. or give time between 2 tests |
|
|
What is mortality?
|
differential dropout of subjects between conditions. people who dropout of a study and skew results
|
|
|
What is the Solomon 4 group design?
|
half the subjects get only posttest and the other half get pretest and posttest.
it allows for evaluation of potential carryover effects: if there is no impact of the pretest, the posttest scores will be the same for the two control groups and experimental groups |
|
|
What is a repeated measures design?
|
1.All subjects are exposed to all conditions.
2. combines pretest-posttest logic with strong controls for individual diff. btwn groups. 3. need to worry about order of progression (which you are randomly assigned to) |
|
|
What are limitations of a repeated measures design?
|
1. potential complications in setting up order in which subjects complete measures
2. problem of order effects: simple with 2 condition but can be complicated with 2+ 3. requires manipulations that can be repeated on multiple occasions. |
|
|
What is a matched pairs design?
|
1. subjects are matched on a characteristic related to the dependent variable and then randomly assigned to one condition.
2.technique to ensure that experimental groups are equal before being subject to the experimental manipulation 3. used with small sample sizes (matched on gender and IQ to ensure distribution is about equal btwn conditions, don't rely on random assignment to make equal, you do it yourself) |
|
|
What do repeated measures and matched pair designs both do?
|
seek to reduce effects of individual difference in groups
|
|
|
What are advantages of a repeated measures design?
|
Since they compare each individual to themselves, it decrease effects of subject variability and makes effect of individual variables easier to detect
|
|
|
What are disadvantages of a repeated measures design?
|
order effects
|
|
|
What are advantages of a matched pairs design?
|
similar to repeated measures without having problem of order effect. decreased by subject variability
|
|
|
What are disadvantages in a matched pairs design?
|
1.cost of id'ing/measuring matching variable
2. procedure is worthless if matching variable isn't related to dependent measure 3. may be unnecessary with large enough sample size |
|
|
How to pick btwn the 4 simple designs?
|
1. posttest only= most basic, simple to run
2. pretest-posttest= if worried that individual might respond very differently to manipulation, if interested directly in change produced by manipulation 3. repeated measures= controls for individual differences btwn subjects 4. matched pairs= controls for individual differences with small samples, measures that can't be used in repeated measures |
|
|
How can you analyze results for the 4 experimental designs?
|
1. posttest= group means
2. pretest-posttest= difference scores btwn groups 3. repeated measures= difference scores btwn groups 4. matched pairs= group means |
|
|
What are two techniques to analyze results?
|
1. t-tests=comparing 2 groups
2. ANOVA (Analysis of Variance)= comparing 2+ groups |
|
|
What is a null hypothesis?
|
population means are equal. any differences btwn sample means are due to chance.
|
|
|
What is a research hypothesis?
|
population means are not equal. differences due to real differences btwn populations. result of manipulation
|
|
|
What is a T-test?
|
a stat. that computes likelihood of obtaining the observed results (the observed difference btwn. the two populations) if the null hypothesis is true.
p-value= is the probability that you see pattern of resultsif the pop is really the same t=diff. btwn groups (means) / normal variability w/in groups |
|
|
What happens if t is large/small in a t-test?
|
large= difference btwn groups is much bigger than the normal variability w/in groups
2 groups are sig. diff. from each other. p=small small= difference btwn groups is much smaller than the normal variability w/in groups 2 groups aren't sig. diff.p=large |
|
|
What is a F-test?
|
Stat. that computes likelihood of obtaining the observed results if the null hypothesis is true. its an extension of the t-test. computes ratio of systematic variance to error variance.
F=large=variance btwn groups is big, therefore, at least one of the groups is diff. F=small=variance is much smaller, means from group to group not diff. doesn't examine means exactly but looks at variance |
|
|
What is a Type 1 error?
|
rejecting the null hypothesis when it is in fact correct.
this is a worse error to make |
|
|
What is a Type 2 error?
|
accepting the null when it's false.
|
|
|
What is gullability? In which error do you find this?
|
accepting an effect when there's insufficient evidence
Type 1 error |
|
|
What is blindness? Which error is this?
|
blindness to a relationship, discarding an effect thats real
type 2 error |
|
|
What is confidence and power?
|
confidence=accurately concluding no effect
power=accurately concluding effect |
|
|
What is alpha?
|
the p-value at which we decide to reject the null hypothesis:
as alpha increases, the prob. of a type 1 error increases and the prob. of a type 2 error decreases |
|
|
Other errors
|
as sample size increases, prob. of type 2 error decreases
|
|
|
When not to use t-tests and ANOVA
|
when you need to test the relationship of continuous variables with other techniques like correlation and regression
|
|
|
What are the three order effects?
|
practice, fatigue, and contrast
|
|
|
What is a practice effect?
|
scores change as subjects gain experience with the measure. usually increase with increased exposure
|
|
|
What is a fatigue effect?
|
scores change as subjects repeatedly complete the measure. usually decrease with greater exposure
|
|
|
What are contrast effects?
|
scores might change depending on which items were completed most recently. similar to assimilation/ contrast effects for survey data
the experience of one condition produces an opposite effect on the experience of subsequent condition |
|
|
What is complete counterbalancing?
|
present the different conditions in every possible order. 2 conditions = 2 orders 4 conditions = 24 orders (4!)
large number of orders=requires more subjects, sufficient sample size in each order |
|
|
What are randomized blocks?
|
presents conditions in random order. each subject individually experiences random order
|
|
|
What is a Latin Square?
|
matrix of conditions that minimizes effects of order effects.
1. each condition appears at each ordinal position = should reduce practice fatigue effects 2. each condition precedes and follows each condition one time = should reduce/cancel contrast effects |
|
|
What are expectancy effects?
|
subjects expectations about the nature of the experiment. also experiment's expectations, if these influence how subjects are treated.
if these vary with the manipulation, they present potential confounding variables that can interfere with the ability to make inferences from the experiement |
|
|
What are demand characteristics?
|
subjects may alter their responses to be consistent with the hypothesis they think is being tested . need to disguise nature of q's
|
|
|
What are placebo effects?
|
effects of the expectation of having received a treatment. drug trials with placebo control group where subject gets no active med.
|
|
|
What is a factor?
|
another term for an independent variable
|
|
|
What is a level?
|
different conditions within a factor
|
|
|
What are 2 reasons for one-factor designs with 2 levels?
|
1. may want multiple control groups/ interested in multiple conditions
2. need more than 2 levels to detect curvilinear/ complex relationships btwn variables |
|
|
What is a full factorial design?
|
involves 1+ independent variable/factor
all levels of each ind. variable are combined with levels of the other ind. variables each combo produces a cell in which diff. combos of the ind. variables are present |
|
|
What are some factorial design notations?
|
3x4= 2 factors: factor 1= 3 levels. factor 2= 4 levels =12 cells total
2x3x5 factors=3 factor 1=2 levels factor 2=3 levels factor 3=5 levels cells =30 |
|
|
What are main effects?
|
are the effects of each independent variable, considered separately.
compares the effects of levels within a factor avg. across all levels of the other factors number of main effects possible= number of factors |
|
|
What is an interaction effect?
|
when the effect of one independent variable depends on the level of another independent variable
can't fully understand the influence of one independent variable without reference to another independent variable. if lines are parallel=no interaction effect. lines must not be parallel to have interaction effect |
|
|
What is external validity?
|
extent to which results of a study would be found with different subjects/setting
|
|
|
What are simple main effects?
|
examine the effects of one factor at one level of another factor.
|
|
|
How do you conduct a mixed designs experiment?
|
Use:
1. repeated measures design 2. latin square |
|
|
What are mixed designs?
|
1. have at least one w/in subjects factor and one btwn subjects factor
|
|
|
What are within subjects?
|
the same individuals appear in all levels of the factor
|
|
|
What are between subjects?
|
different individuals appear in each levels of the factor
|
|
|
What are three-way interactions?
|
test whether 2 way interactions differ based on levels of a third factor. number of interactions depend on total number of factors in experiment.
|
|
|
Explain developmental research
|
1. focuses on how responses change over time.
2. theories about how attitudes might change with age. much research focuses on kids and old ppl |
|
|
What is cross-sectional research?
|
subjects of different ages are studied at one point in time. need to gather data from a number of diff. age ranges to be able to make inferences. btwn subjects design.
assume that subjects at each age represent what a single individual would experience across course of his life |
|
|
What are disadvantages of cross-sectional research?
|
age and cohort are always confounded.
cannot differentitate developmental changes from cohort effects |
|
|
What is a cohort effect?
|
groups of ppl who were born around the same time have experienced common cultural, tech., and env. changes at the same age.
|
|
|
What are advantages of cross-sectional research?
|
Data can be gathered quickly, data from various ages without having to wait for subjects to age. common in child research where cohort effects aren't strong because kids are so young
|
|
|
What is longitudinal research?
|
measure same subjects on dependent variables at different ages
can document within person, age related changes that cross sectional designs cant may reveal cohort effects when compared to results from cross sectional studies if scores same over time = no cohort effect. if scores diff over time = cohort effect |
|
|
What are limitations of longitudinal designs?
|
1. selective attrition/mortality= has rate of dropouts. healthier and better educated ppl are more likely to stay in design over time = not random dropouts.
selective attrition makes generalizations of age trends more difficult. problematic for older adults 2. time of measurement effects= age related trend may differ by cohort = cohort/time interaction |
|
|
What are sequential designs?
|
mix of longitudinal and cross sectional designs
allows for testing both developmental and cohort effects minimizes limitations of each design. most difficult no cohort=same responses |
|
|
What are pilot studies?
|
trial runs with small number of subjects
training/testing for strength of manipulation. few data pts. random assignment can get messed up because of small sample size |
|
|
What are manipulation checks?
|
assessing the construct validity of the independent variable manipulation ( are you manipulating the intended construct?)
often performed with surveys at the end of the study to not give away hypothesis and produce expectancy effects success= eliminates alternative explanations for failed predictions |
|
|
What is a straightforward manipulation?
|
no deception, simply present different types of stimuli across conditions and measure their effects
|
|
|
What are staged manipulations?
|
create an artificial environment as a means of ensuring a strong manipulation of independent variables: uses deception and confederates
|
|
|
What is deception in a staged manipulation?
|
misleading subjects as to intent of study/consequences of a particular course of action
|
|
|
What is a confederate in a staged manipulation?
|
research assistants who pose as other subjects and play other roles
more confederates=more conformity |
|
|
What is a quasi-experimental design?
|
a design that lacks all of the control features of an experiment. extraneous variables are not controlled.
inability to randomly assign participants or failure of r.a. reduces confidence in internal validity |
|
|
What are the most common types of quasi-exp. designs?
|
1. one group pretest/posttest design
2. non-equivalent control group design posttest, pretest/posttest 3. time series analysis 4. reversal design (ABA) |
|
|
What is a one group pretest/ posttest design?
|
same as a pretest/posttest but no random assignment to multiple conditions
no control condition, only experimental with no control condition, you can't rule out confounding effects |
|
|
What are problems with a one group pretest/ posttest design?
|
1. history effects
2. maturation effects 3. testing effects 4. instrument decay 5. regression towards the mean |
|
|
What are history effects?
|
any event that happens btwn pre and posttest is confounded with the manipulation
|
|
|
What are maturation effects?
|
systematic changes in individuals over time confounded with the manipulation.
big problem with child research |
|
|
What are testing effects?
|
any effects of the pretest on the posttest. aka practice effects, etc
|
|
|
What is instrument decay?
|
posttest measurement of the dependent variable has changed since the pretest measurement.
observers change how easily they observe certain behaviors |
|
|
What is regression towards the mean?
|
problematic when subjects are chosen because they score high or low on some measure
good performance one week is likely to be assc. with more avg. performance the next week |
|
|
What is a non-equivalent control group design?
|
want to avoid problems assc. with one group pretest/posttest by adding a control group that is systematically chosen. still no random assignment. control and experimental group are different=non-equivalent
|
|
|
What is a time series design?
|
like a pretest/posttest but you record multiple measures, before and after manipulation
if there's one extreme value, lets you observe more stable patterns across timepoints. should decrease issues with regression towards the mean requires measures that can be taken repeatedly |
|
|
What is an interrupted time series?
|
time series data collected at very long intervals. (good with tracking relationships that might only appear over a long amt of time)
time stretches allow for number of intervening/confounding events like history effects, etc. |
|
|
What are reversal designs?
|
subjects complete a pre-manipulation test a post-manipulation test and another post-manipulation test
baseline-posttest-baseline (ABA) has good evidence of manipulation |
|
|
What are limitations of reversal designs?
|
some effects of treatments are difficult to reverse
reversal designs most useful for seeing short term effects of treatment |
|
|
What are multiple baseline designs?
|
introducing treatment at different times for different subjects, behaviors, and situations. can rule out maturation effect
|
|
|
Explain research ethics
|
Science strives to be objective: Repeatable conclusions based on concrete observations
ethics is unavoidably subjective to some degree, based on agreed standards, conventions and social norms. unavoidable component of intuition/judgement needs methods to buffer against potential costs of subjectivity |
|
|
What is the Institutional Review Board (IRB)?
|
all studies to be conducted on humans must receive human subjects approval. review committees
similar things are used for animals, hazardous materials, etc |
|
|
What are human subjects review?
|
only when benefits outweigh the cost, it's approved.
benefits: knowledge to be gained , basic science, practical applications. cost:duress (physical, mental) inflicted on subjects |
|
|
What is informed consent?
|
subjects must be told in reasonable detail what will be expected of them/what they'll be exposed to in the study
no need to state hypothesis |
|
|
What is deception?
|
informed consent not truly possible in such cases, chances of lying possible for aims of study to be achieved
|
|
|
What is debriefing?
|
if deception occurs, after the experiment is over, subjects will get debriefed and discover the purpose of the study
explains the true nature of the study, and why it was necessary to deceive the subject. to make sure the subject isn't experiencing stress about the experiment |
|
|
Which experiment is most responsible for the IRB?
|
Zimbardo's prison experiment
|
|
|
What is data integrity?
|
results of studies are also subject to ethical questions, specifically when they could be trusted. could be wrong due to:
idiosyncrasies in data (type1/2 error) mistakes in data analysis: worst, hardest to catch, lying about the data. wrong way to do analyses, incompetence in carrying out analysis deliberate misrep. in data |
|
|
What are data integrity safegaurds?
|
1. peer-review: should catch mistakes. can catch fraud
2.replication: can ID idiosyncrasies, can catch fraud |
|
|
What is external validity?
|
measures the extent to which research findings can be generalized to other subjects, settings, etc
even if a study has strong internal validity, would you get the same results with diff. subjects? |
|
|
What is internal validity?
|
You can conclude that one variable caused another.
|
|
|
describe how one can replicate based on external validity?
|
when replicated in a different setting, with different subjects, and with different operational definitions.
|
|
|
What is exact replication?
|
goal is to closely as possble reproduce exact circumstances of an original study.
useful in checking for idiosyncrasies in results |
|
|
What is conceptual replication?
|
goal is to test the same hypothesis with different measures.
distinct results with different procedures suggest alterations to theory |
|
|
What is a meta-analyses?
|
collapse of results of all relevant experiments and subject to common statistical analysis.
helps evaluate external validity |
|
|
What is the file drawer problem?
|
positive results are often published, negative results aren't
over-rep. of positive results, under-rep of negative results. |
|
|
correlation is not...
|
causation
|
|
|
What is representative heuristic?
|
Over-representation because it looks more common. assume an answer because it is more common
|
|
|
What is availability heuristic?
|
difficulty to look beyond personal experience. what your friends tell you, what you hear on the news. you look at your personal experiences because it so available, you override other data
|
|
|
What is base rate fallacy?
|
underweighting stat. averages and overweighting personal preference
|
|
|
What is gamblers fallacy?
|
see relationships btwn events that don't really exist. see dependence where events are independent
|
|
|
What is empirical evidence?
|
based on observations
non-empirical=1 set of data empirical=lots of types of data |
|
|
What is objective evidence?
|
observations allows others to get the same results. data is gathered and lets ideas be evaluated
|
|
|
What is systematic evidence?
|
observations allow for casual inferences
|
|
|
What is falsifiability?
|
a good theory capable of being false when tested using scientific methods...speaks directly if idea is correct or not
|
|
|
What is intuition?
|
it is not evidence, it's subjective and unsystematic. cannot always be accurate
|
|
|
What is change blindness?
|
ppl aren't good at noticing small changes
|
|
|
What are the steps of creating an experiment?
|
theory-hypothesis-predictions-data/observations-conclusion
|
|
|
What is a theory?
|
ideas about why things should work in a particular manner. generate/explain sets of hypotheses
|
|
|
What is a hypothesis?
|
specific idea about how a particular phenomenon works. a statement about the world that might be true. 2+ variables that are related
|
|
|
What is a prediction?
|
hypotheses translated into concepts used in the study
|
|
|
What is data?
|
experimental results that address predictions
|
|
|
What is a conclusion?
|
Whether results confirm obsv. and how they speak to the hypotheses
|
|
|
What is meta-theory?
|
generates many theories with diff. research
|
|
|
What is evolutionary psychology?
|
meta-theoretical position that the human brain contains a collection of specialized proccessing mechanisms designed by natural selection to address specific problems encountered by our ancestors over the course of human evolution: food choice, parenting, navigation
|
|
|
What are the goals of psychological research?
|
1. to describe behavior
2.to predict behavior 3.to determine the causes of behavior 4.to explain behavior |
|
|
What is a variable?
|
event, situation, behavior with at least 2 values
|
|
|
What is an operational definition of variables?
|
a definition of a variable in terms of how it's measured or used to manipulate the concept. puts concepts in public realm where they can be tested and improved
|
|
|
What are different types of linear relationships and what do they predict?
|
how one variable affects another.
positive linear, negative linear, no linear relationship, and curvilinear. |
|
|
What is predicting behavior?
|
showing a relationship (correlation) btwn 2 variables that don't show cause and effect. leads to problems making causal statements (does a cause b?) counfounding 3rd variables that might cause a and b
|
|
|
What is a clarifying explanation?
|
how we get to the explanation of the 2 variables. give more details.
|
|
|
What is an alternative explanation?
|
a doesn't cause b. something else causes b. no original relationship
|
|
|
Why are correlation studies helpful?
|
even though they cant directly prove causation, they can be useful evidence in building other causation cases
|
|
|
What is the Wechster Adult Intelligence scale?
|
1st assement of IQ. long verbal subtest and performance subtest
|
|
|
Who was Alfred Binet?
|
he simplified the wechster adult intelligence test with questions and a mean of 100 (normal distribution)
|
|
|
What is reliability?
|
the degree to which a measure is consistent
|
|
|
What is test-retest reliability?
|
scores on a measure in month a should correlate with the scores on a measure in month b
|
|
|
What are the correlation numbers?
|
r=.8+=positive correlation
r=0=no correlation r=-.7=negative correlation |
|
|
What is internal consistency reliability?
|
if you divide the items on a test in half, both halves should have the same score.
|
|
|
What types of construct validity are there?
|
face, criterion-oriented, convergent, predictive, concurrent, discriminant.
|
|
|
What is face validity?
|
measure appears to measure what it is intended to measure. involves judgment, intuition and is subjective. not sufficient to conclude from. reflects the meaning of the construct being measured
|
|
|
What is criterion-oriented validity?
|
measure is related to theoretically similar factors.
|
|
|
What is convergent validity?
|
if scores of 1 construct are related to the scores of the same/similar construct. 2 diff IQ tests should give similar results.
|
|
|
What is predictive validity?
|
research uses a measure to predict some future behavior. good IQ score leads to good grades in school.
|
|
|
What is concurrent validity?
|
does the measure correlate with current behavior? do diff. groups of ppl differ on the measure in expected ways? phsyics ppl smarter than psych ppl
|
|
|
What is discriminant validity?
|
does the measure asses the intended construct better than other constructs?
|
|
|
What is the independent and dependent variable in experimental research?
|
Independent=manipulated variables
Dependent= outcome variables |
|
|
What is the independent and dependent variable in non-experimental research?
|
Independent=the cause
Dependent= the affect |
|
|
What is random assignment?
|
assign randomly so other factors won't matter (like weight, size, gender) only if sample size is big enough
|
|
|
What happens in experimental research with the manipulated variable and the control variable?
|
only the manipulated variable changes between the two groups. the control variable doesn't affect experiments
|
|
|
What is top-down, bottom-up justification?
|
top-down=need of theory and why 1 causes the other
bottom-up=put data together then find out what it means |
|
|
What is automaticity?
|
to what extent is social behavior driven by influences we are unaware of and are thus outside of conscious control
|
|
|
What are nonexperimental methods, advantages and disadvantages?
|
methods=measure variable but don't manipulate them
advatages=can observe ppl in real-life situation, can sample large # of ppl (rep. survey) disadvantages=difficult to demonstrate cause/effect relationships |
|
|
What are experimental methods, advantages and disadvantages?
|
methods=must manipulate a variable and randomly assign to condition
advantages=can demonstrate causal relationships disadvantages= lab settings are often unrealistic. may not generalize to the real world |
|
|
What is the goal of naturalistic observations?
|
to provide a complete/accurate picta of behaviors, events, rather than to test hypotheses formed before the study. analyze observations and form hypothesis post-hoc.
|
|
|
What are issues with naturalistic observations?
|
1. participation= involved with subject of study v. not involved
2. reactivity=reacting to your presence. you can conceal and hide yourself or habitutation- around animal so much they don't notice you 3.lack of operational definitions=without pre-exisiting theories, how to create operational definitions and ensure objectivity |
|
|
1.What is the low ball technique?
2.What is the reciprocity technique? 3.what is the commitment technique? |
1. offer low price, then offer upgrades that increase price after commitment to purchase was made
2. give something for free (create obligation) more likely to buy something 3.get commitment to 1 package then change the package |
|
|
1. What is the authority technique?
2. What is the the liking technique? 3. What is the scarcity technique? |
1. appear as authority and encourage purchase of a particular package
2. customer likes salesman, have salesman offer advice 3.the rarer an object, more valued assc. with it |
|
|
What is the goal of systematic observations?
|
careful observations of 1+ behaviors in a particular setting.
interested in specific behaviors. quantifiable, measurable. hypothesis driven. |
|
|
What are methods of systematic observations?
|
require coding system for behavior. operational definitions of those behaviors to be measured.
sampling-what segment of time should be recorded? |
|
|
What is inter-rater reliability?
|
how much 2 ppl agree on the data
|
|
|
Limitations on systematic observations?
|
reactivity is still a problem
results depend entirely on decisions about coding system. |
|
|
What is a coding system?
|
a set of rules used to categorize observations
|
|
|
What is the Costly-Signaling Theory?
|
signals can be regarded as more reliable/ honest if they are more costly, expensive and hard to fake. more you trust more likely to allow diddling
|
|
|
What is the goal of case studies?
|
to provide detailed descriptions of the behavior of an individual, usually in rare circumstances
most are qualitative v. quantitative=exist as proof of concept |
|
|
What are limitations of case studies?
|
conclusions drawn are often post-hoc
difficult to determine causalty |
|
|
What is the goal of archival research?
|
use existing data to answer novel research questions. no further data collection
|
|
|
What are the methods of archival research?
|
data mining= stat analyses, survey archives, written records
content analysis-coding systems that can be used to quantify data |
|
|
What are limitations of archival research?
|
records can be difficult to obtain
data may not speak directly to hypothesis difficult to verify accuracy |
|
|
How should survey research questions be phrased properly?
|
1. questions need to be clear and concise
2.questions should be directly relavent to the hypothesis 3.avoid loaded/leading questions 4.avoid double barreled questions (2 qs in1, cant phrase out which is the driving q) |
|
|
Whats the halo effect?
|
avoid associating a position with a person so that responses may reflect feelings about the person rather than the position. might be inclinded to agree/disagree based on feelings about the person not the position
|
|
|
Explain open-ended questions
|
1. answer in own words
advantages= can discover unexpected answers, don't restrict responses disadvantages= costly time consuming to code. respondents may not think of answers of interest to investigator |
|
|
Explain closed ended questions
|
limit response alternatives from a list
advantages=easier to code and ensure cover intended options disadvantages=may be putting words into respondents mouths. may have given other responses not on list must make sure options are mutually exclusive make sure options are exhausted (all possibilities are covered) |
|
|
What are the effects of question order?
|
valence of accessible information in previous questions affects how subsequent questions are answered.
other q's can prime specific things coming to mind |
|
|
What is the assimilation effect?
|
when positive info currently in mind produces more positive evaluation of a target (same direction)
|
|
|
What is the contrast effect?
|
when positive info in mind (about past events) produces a negative evaluation of a target (opposite direction)
|
|
|
what are constructing attitudes?
|
no reminder v. reminder. how they impact and don't impact judgement
|
|
|
what are response sets?
|
ya-saying (tend to agree with all)
nay-saying (tend to disagree with all) possible solution=reverse the direction of some items |
|
|
what is social desirability bias?
|
tendancy to respond in anyway that makes subject look good
|
|
|
Whats the bogus pipeline?
|
ppl attached to fake lie detector. ppl tend to report higher #s when they think they're being tested
|
|
|
Whats survey cognition?
|
ppl construct answers to surveys on the spot using any relevant info (wording qs, response options, env. variables)
proper interpretation of surveys requires knowledge of the context |
|
|
What is true random assignment?
|
all subjects have an equal probability of being assigned to any condition in the study
|
|
|
What is random sampling?
|
subjects are chosen so that every member of the population of interest has an equal chance of being selected to participate in the survey
|
|
|
What is selection bias?
|
by taking names from phone books, survey systematically excluded poor voters. results when sampling frame doesn't equal population
|
|
|
What is a sampling frame?
|
actual population of individuals from which the sample is drawn
|
|
|
What is a non-response bias?
|
non-respondents may have diff. preferences that from respondents. greater response rates reduce the odds of biased samples.
|
|
|
What is a population?
What is a sample? |
population=the group of all individuals of interest (political polls=eligible voters)
sample=some subset of the population. purpose is to estimate the response of the total population without having to survey everyone |
|
|
What is the response rate?
|
perecent of ppl in sample who complete the survey
|
|
|
1. What is simple random sampling?
2. What is stratified random sampling? 3. What is cluster sampling? |
1. each member of the population has an equal chance of being included
2. population is divided into strata followed by random sampling from each stratum 3.clusters of individuals are identified then clusters are randomly selected. all individuals within a selected cluster are sampled. 1-3=probability sampling=draw stronger inferences |
|
|
1. What is convenience sampling?
2. What is quota sampling? |
1. subjects are chosen in a haphazard way based on whoever is available.
2. sample is chosen to reflect the numerical composition of various subgroups in a pop. 1-2=nonprobabililty sampling |
|
|
explain margin of error and confidence intervals
|
a representative random sample can estimate the true pop. mean within a margin of error or confidence interval
math formula that determines confidence interval relies on the size of the sample, not size of the overall pop. |
