FInal Flashcards
Six threats to internal validity
Maturation History Regression to the mean Attrition Testing Instrumentation
Three other threats to internal validity
Observer bias
Demand characteristics
Placebo effect
Neither the experimenters nor the participants knows who is in what group.
Double-blind study
Same as double-blind, plus
The control group receives a placebo, such as a sugar pill, fake therapy, etc.
Double-blind placebo-control study
No significant difference between the levels of the independent variable
null effect
Why might you get a null effect?
- Because there really is no effect!
- The different levels of the independent variable weren’t different enough.
- There was too much unsystematic variability.
Why might the difference between groups be too small?
- Manipulation was not STRONG enough.
- Measure was not SENSITIVE enough.
- Ceiling or floor effects.
everyone is maxed out on the measure
Ceiling effect
- The questions were so hard that people got most of them, even the easy-to-read ones, wrong.
- The anagrams were so hard that a red test cover couldn’t make performance worse.
Floor effect
How might unsystematic variability lead to null results?
- Measurement error
- Individual differences
- Situation noise
is the likelihood that a study will yield a statistically significant result when the IV really has an effect
Power
People adapt to the environment, the conditions on their own. Children get better at walking; trees grow taller; depressed people get better over time; campers get used to the camp and aren’t so wild. Time just passes and things change
Maturation
Something happens to most of the people in the experiment. Weather got colder for the “using electricity” folks. Maybe the campers started a swimming program that tired them out. This is a historical or external event. It isn’t that time went by and things got better (like above). Something happened.
History
The statistical likelihood that things even out. An extreme score at the pretest usually
leads to a regression to the mean on the post score.
Regression to the mean
A reduction in participant numbers that occurs when people drop out before the end of the study
Attrition
The change in the participants as a result of taking the test—they may become better because of practice, or they may be fatigued or bored
Testing
When a researcher uses two different measures, but they don’t capture the same concept, or capture it at different levels. Example: observers change their observation criteria over time, a researcher uses different forms of a test at pretest and posttest and they’re not equivalent forms. The instrument changes from time 1 to time 2
Instrumentation
researchers’ expectations influence their interpretation of the results—or even influence the outcome of the study.
Observer bias
participants guess what the study is supposed to be about and change their behavior in the expected direction.
Demand characteristics
people receive a treatment and really improve - but only because they believe they are receiving a valid treatment.
Placebo effect
A replication study in which researchers repeat the original study as closely as possible to see whether the original effect shows up in the newly collected data.
direct replication
Pertaining to a study whose results have been obtained again when the study was repeated.
replication
A replication study in which researchers replicate their original study but add variables or conditions that test additional questions
replication-plus-extension
A replication study in which researchers examine the same research question (the same conceptual variables) but use different procedures for operationalizing the variables.
conceptual replication
the idea that a meta-analysis might be overestimating the true size of an effect because null effects, or even opposite effects, have not been included in the collection process.
file drawer problem
The extent to which a laboratory experiment is designed so that participants experience authentic emotions, motivations, and behaviors.
experimental realism
Help demonstrate how seemingly basic human processes can work differently in different cultural contexts. And how when theories are tested only on WEIRD people they may not represent everyone.
cultural psychologist
Western, educated, industrialized, rich, and democratic
WEIRD samples
A third variable that changes the relationship between two variables.
moderators
– when the influence of one independent variable changes (is moderated by) another independent variable
Interaction
An experiment with more than one independent variable is said to have a …..
factorial design
– an statistically significant effect of each independent variable separately, ignoring the other independent variable.
Main effect
This language applies mostly to ANOVA-type analyses.
factorial design Note 1:
Interactions can be statistically significant even if main effects are not.
factorial design Note 2:
A significant interaction overrides any significant main effects. The interaction is always more important!
factorial design Note 3:
Types of Factorial Designs
- Between-subjects (AKA independent groups)
- Within-subjects (AKA within-groups)
- Mixed Design
Why Use Factorial designs?
TO TEST LIMITS:
-Factorial designs can explore whether an effect generalizes across groups (external validity)
- Do children and adults both like clowns?
- Does makeup improve the perceived likability of people of different ethnic backgrounds equally?
And TO TEST THEORIES:
-If a theory predicts that an effect will be present for one group or situation but not another, then a factorial design can test this.
- Theory: makeup makes people appear more attractive, and attractive people are more likable.
- -People who are already attractive should benefit less from wearing makeup.
-Theory: clowns make people happy because they wear bright colors, and the colors make people happy.
2X4
8 cells
2X4 independent groups factorial design, 2X4 within subjects factorial design–what changes between the two studies?
- Within=same subjects tested for both variables
- Independent=different subjects but only tested in one variable
What’s the difference between a 2x4 independent design and a within subjects design?
Independent groups: twice as many participants
For a 2x4 factorial design, there are 2 main effects
-one main effect for each variable
A 2x2x2 design would have 3 main effects
there are 3 independent variables
Between Subjects (independent groups) design
separate groups, participants are only exposed to one level of the independent variable (you are given coke OR pepsi and asked to rate it) then your results are compared with those of someone else, random assignment is essential
- Post test
- Pre-test/post-test
Within group (within subjects) design
each participant experiences all levels of independent variable, and each person is compared with themselves (they are their own control) -> no need for random assignment
- concurrent measures- dependent variable is measured once, after exposure to all levels of independent variable (usually a preference measure, after trying coke and pepsi, which do you like more?)
- repeated measures-dependent variable is measured after exposure to each level (try pepsi, rate it from 1-10, try coke, then rate it also)
Weaknesses in Between subject & Within Groups Design-
Order effects (the order of which level of IV they are exposed to effects outcome, because you tried Coke first, you prefer it more) that is why you have to counterbalance (randomize the order)
- Theory
* Hypothesis
- is a set of statements that describes general principles about how variables relate to one another.
- is a way of stating the specific outcome the researcher expects to observe if the theory is accurate.
Open-ended -
Forced-choice -
Likert scale -
o allow respondents to answer any way they like – forces the respondent to answer with more then yes or no.
o in which people give their opinion by picking the best of two or more options.
o people are presented with a statement and are asked to use a rating scale to indicate their degree of agreement. When such a scale contains more than one item and is anchored by the terms strongly agree, agree, neither agree nor disagree, disagree, and strongly disagree.
Applied research
is a form of systematic inquiry involving the practical application of science - for a specific, often state-, business-, or client-driven purpose.
Understand the difference between positive, negative or zero association
+ = as one variable increases, the other variable also increases
- = as one variable increases, the other variable decreases
0 = no correlation between variables
Meta-analysis -
Longitudinal Study –
- Some review journal articles provide results of many studies and give a number that summarizes the magnitude, or the effect size, of a relationship
- Long-term study to see how a variable changes. (A benefit is that you are using the same participants, so able to factor out more variables)
Why do we use PsychInfo?
All Psychological Information
- Justice
- Beneficence
- Respect for Persons
- how participants are selected and whether the selected participants represent the population benefiting from the experiment.
- Cost benefit analysis for: Participants and society as a whole
- Informed consent – protecting venerable groups of people
Was an infamous study conducted between 1932 and 1972 by the U.S. Public Health Service to study the natural progression of untreated syphilis in rural African American men who thought they were receiving free health care from the U.S. government.
Tuskegee Study
Know about operational
when testing hypotheses with empirical research. Turning concept of interest into a measured or manipulated variable
Causal Claim
a. Temporal precedence (A comes before B in time)
b. Covariance (association between A and B)
c. Internal Validity (there are not other possible causes for B except A)
Association Claims
Association Claims – At least two variables – No causation is implied
What’s the difference between ordinal scales, interval scales, categorical measurements, and self-report measurements?
- Categorical- same as nominal (Male or Female)
- Ordinal- In order (like in a race)
- Interval- has no absolute zero, equal distance between two units (ratio does have an absolute zero)
face validity -
does it look like it’s measuring what it’s supposed to
construct validity -
is your experiment as a whole measuring the construct that you actually want/purport to measure?- “The momma of all validities”
Content -
do the items on your test measure the desired construct, or a different one?
Criterion validity -
how well one variable can predict another-How well does your driver’s test measure your ability to drive?
convergent -
Your measure has similar outcomes of other measures of the same construct
When do we use scatterplots and when do we use bar graphs and when do we use line graphs?
oScatter Plots -for 2 continuous variables – correlation test
oHistograms-For continuous data (income) & Interval Data
oBar Graphs- For nominal & ordinal data (gender)
Why do we use random assignment?
To prevent selection effects
o Mediator
o Moderator
is a mechanism that explains how one variable influences another. (Variable A influences Variable B because of the mediator)…like Christ.
when the relationship between two variables changes depending on the level of another variable, that other variable is…
Bivariate
normal correlation between 2 measured variables, does establish covariance, but not temporal precedence or internal validity
Longitudinal designs
Measure the same variables in the same people at two different points in time. Is a multivariate design because you have to measure each variable twice, shows covariance and temporal precedence, but does not control for confound.
Experimental designs
you can make a causal statement because of the design (that’s what it’s for) control, random assignment, variable manipulation, control for confounds.
Autocorrelation-
Cross sectional-
Cross-lag-
- correlation of 1 variable with itself, in 2 different time points (HORIZONTAL)
- correlation of 2 variables in 1 time point (VERTICAL)
- correlation of variable A at time 1, with variable B at time 2 (or vice versa)
IV:
DV:
- variable that stands alone and isn’t changed by the other variables you are trying to measure.
- something that depends on other factors.
occurs in an experiment when the KINDS of PARTICIPANTS in one level of the independent variable are systematically different from those in the other
Selection effect
in which some form of CONTAMINATION, carries over from one condition to the next.
Carryover effects,
when EXSPOSURE to one level of the independent variable influences responses to the next level of the independent variable.
Order effects
matching when assigning small numbers of participants to groups (you would measure participants on something like IQ, and then take a group (like highest 3) and split them randomly into control group and experimental groups
Matched-groups
