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;
93 Cards in this Set
- Front
- Back
|
Factorial Design
|
-research design that involves all combinations of at least two values of two or more IV's
-simplest is 2x2 where have two variables each with two levels; it gives rise to 4 combinations |
|
|
Why would you want to examine the effect of two or more variables in a single experiment?
|
-efficiency
-may have more than one alternative hypothesis to rule out -reveal interactions among variables |
|
|
Why avoid having too many factors and levels?
|
-increasing complexity requires more time to conduct experiment
-considering a large number of interactions taxes the mind -usually use 2 or 3 factors with 2 to 6 levels |
|
|
Main Effect
|
-in a factorial experiment, the effect of one IV averaged over all levels of another IV
-does not mean the principal effect of a variable -main effects can be misleading when interactions are present |
|
|
Column means and row means
|
-column means is averaging the columns of the table
-row means is averaging the rows of the table |
|
|
Interaction
|
-when the effect of one IV depends on the level of another IV
-whenever there is an interaction in the data, the main effects cannot be interpreted w/o discussing the interactions -can have interaction if one or the other IV has no main effect or even if neither IV has a main effect |
|
|
Graphical representation of interactions
|
-if graphical representation of a factorial experiment shows curves that are not parallel, there is an interaction bwt the variables
|
|
|
Antagonistic Interaction
|
-interaction in which the two IV's tend to reverse each other's effect
-effect of A equals effect of B in table -graph, main effect line is horizontial |
|
|
Synergistic Interaction
|
-interaction in which the two IV's reinforce each other's effect
-graph will show steeper slope of line relating the DV to A when B is larger and vice versa |
|
|
Ceiling-Effect Interaction
|
-interaction in which one variable has a smaller effect when paired with higher levels of a second variable
-graph shows all lines in close proximity |
|
|
Types of factorial experiments
|
-within subjects
-between subjects (more common) -mixed where have at least one within-subjects variabel and at least one between-subjects variable |
|
|
2X2 factorial design table
|
factor A
Factor B A1 A2 B1 A1B1 A2B1 B2 A1B2 A2B2 |
|
|
Within-in subjects factorial experiment
|
-each subject experiences each condition
-each subject (s) will be seen in each condition |
|
|
Between-subjects factorial experiment
|
-seperate groups experience each condition
-s1-8 group A1B1 s9-16 group A1B2 s17-24 group A2B1 s25-32 group A2B2 |
|
|
Mixed facotrial experiment
|
-one variable is with-in subjects and the other is between-subjects
-subjects either experience B1 once with A1 and also A2 or B2 once with A1 and also A2 |
|
|
Advantages of within-subjects design
|
-whenever possible to present each condition to every subject within-subject should be considered as long as order and sequence effects aren't expected
-requires fewest subjects to achieve a particular degree of power, the mixed desing is the next fewest, and between-design the most |
|
|
Control in within-subjects factorial experiments
|
-controll for order and sequence effects over both variables
-if have multiple presentation of each condition situation becomes very complex -conditions of variable B: block randomized -conditions of variable A: reversed for half of the subjects |
|
|
Gustav Fechner
|
-worked extensively on individual participants
-psychophysics and methods to measure sensory thresholds |
|
|
Hermann Ebbinghaus
|
-experimental work on memory
-used himself as participant |
|
|
Wilhelm Wundt
|
-introspectionism
-measured various psychological and behavioral responses in individual participants |
|
|
Why early researchers used single participant designs
|
-didn't have statistical methods so solved problems of reliability and validity by having extensive observations and frequent replication of results
|
|
|
What frequent assumption did researchers using single-subject designs traditionally make?
|
-individual participants are essentially equivalent and that one should study additional participants only to make sure the original participant was not abnormal
|
|
|
B.F. Skinner
|
-behaviorism, operant conditioning
-disdained use of statistics |
|
|
What does the single-participant tradition assume
|
-most variability in the participant's behavior is imposed by the situation and therefore can be removed by careful attention to experimental control
|
|
|
What does the group research tradition assume?
|
-variability is intrinsic and should be statistically controlled and analyzed
|
|
|
Advantage of single-subject design
|
-people or animals act as own control
-avoids possibility that the average picture is a distortion of the behavior of the individual participants which is a potential problem whenever data are averaged over many participants -not distracted by effect of a minor variable -spend time reducing variability so effect of variable is maximized instead of spending time testing more participants |
|
|
Clinical significance
|
-practical importance of a result
-expt that employs large group of participants will likely discover that an IV has an effect even if the effect is a minor one but still can have little clinical significance even if have statistical significance |
|
|
Power
|
-probabilit that a statistical test will find a significant difference when a difference actually exists in the population
|
|
|
Power
|
-power depends on size of difference that exists in population and size of the sample drawn from the population
|
|
|
How can researcher increase probability of finding a significant result?
|
-increase the size of the effect or
-increase the size of the sample (but this decreases variability of the data) -first tactic is favored by single-subject researchers |
|
|
Ehtical and Practical problems avioded with single-subject design
|
-if testing efficacy of treatment that benefits participants problem arises with those that aren't treated
-soln is to treat all participants but to evaluate them from a single-subject standpoint -use single-subject if cannot locate enough participants to constitute a group to study |
|
|
Single-subject offers flexibility
|
-modify expt on spot if not going right way with subject (not responding to reinforcer that's worked on previous subjects)
-switch conditions immediately if you think a large change in subjects behavior is due to outside force and see if behavior changes correspondingly |
|
|
Disadvantages of single-participant approach
|
-some effects are small relative to the amount of variability in the situation (may be impossible to control the other sources of variability sufficiently to observe the experimental effect in one participant)
-statistical procedures for analyzing single-subject data are not well developed -some experimental effects are by definition bewteen-subject effects (ie can't have participant simultaneously be taught same material by 2 different methods) |
|
|
How do you compare outcomes when only have one participant?
|
-compare the behavior that occurs before and after the introduction of the experimental manipulation
-measure later behavior against baseline |
|
|
Baseline
|
-the measure of behavior before treatment that establishes a reference pt for evaluation the effect of treatment
-need long enough time to obtain this |
|
|
Purpose of the baseline
|
-measure the current level of behavior
-predicts what the behavior would be in the future if no treatment was administered |
|
|
When can evaluation of a treatments effect occur?
|
-only if a baseline measurement shows the behavior to be either remaining at the same level or changing in the direction that is oppisite to the predicted treatment effect
|
|
|
AB (or comparison) Design
|
-single participant research design that consists of a baseline followed by a treatment
-baseline: A -treatment: B -difficult is will not know whether other variables that may have coincidentally changed at the same time that the treatment was administered actually produced the change in behavior |
|
|
ABA Design
|
-includes baseline period, a treatment period, and a subsequent withdrawal of treatment
-strengthen argument that treatment is cause of the change |
|
|
Problems with ABA design
|
-effect of the manipulation may not be fully reversible
-may want to leave the participants in the new condition rather than return them to their original state |
|
|
ABAB Design
|
-an ABA design w/ treatment repeated after the withdrawal phase
-aka replication design -another oppurtunity to evaluate effect of treatment and its reliability -repeated presentation and withdrawal of a varibale can produce strong evidence for the validity of the IV effect |
|
|
How should you change variables in a single-participant research design?
|
-change only one thing at a time
-if 2 variables are changed simultaneously, it's impossible to decide whether change in behavior was caused by one or the other or the two together |
|
|
Alternating treatments design
|
-single participant design that allows the comparison of 2 different IVs
-meets requirement that only one variable is changed at a time by alternating treatmenst across sessions -similiar to ABAB design w/ proviso that more than one type of treatment is administered during the first nonbaseline condition |
|
|
Multiple-baseline design
|
-introduces experimental manipulation at different times for different behaviors to see if behavior change coincides w/ manipulation
-demonstrates that the manipulation caused the behavior change -espcially useful if expected behavior change is irreversible |
|
|
what can the experimental baselines be in a mutliple-baseline design?
|
-different behaviors in the same individual
-same behavior is different ppl -same behavior in same person by in different behavior settings |
|
|
Changing-criterion design
|
-introduces successively more stringent criteria for reinforcement to see if behavior change coincides w/ the changing criteria
-after baseline measurement reward is given for meeting lax criterion of behavior, after behavior stabilizes at that level criterion can be raised until behavior stabilizes again and so forth |
|
|
What can changing-criterion design show?
|
-if the behavior begins to change after each change in the criterion, then the conclusion that the reward is the cause of the improvement is rather convincing
|
|
|
When is changing-criterion design useful?
|
-behavior change is irreversible
-when a return to the initial baseline is not desirable |
|
|
Quasi Experiment
|
-research procedure in which the scientist must select subjects for different conditions from preexisting groups
-don't control assignment of subjects to conditions |
|
|
How do quasi and true experiments differ?
|
-true: assign subjects to conditions, manipulat variables
-quasi: select subjects for the different conditions from previouslt existing groups, observe categories of subjects |
|
|
Example of assignment in quasi experiment
|
-gender
-cannot create groups of males and females but instead select members from preexisting groups |
|
|
Subject variable
|
-IV in a quasi experiment if it is a characteristic of the subjects on which they have selected such as gender
|
|
|
Ex Post Facto
|
-after the fact
-another name for quasi experiments b/c the experiment is conducted after the groups have been formed |
|
|
What additional problems arise from presenting some IV to two preexisting groups and measuring the effect of the variable on their behavior?
|
-don't know whether difference in behavior was caused by difference between the groups or by the IV
|
|
|
True Experiment and Control
|
-greatest degree of control in ruling out alternative hypotheses or alternative IV as cause of difference bewteen groups
-most powerful control for confounding b/c eliminate other possible IVs by random assignment of subjects to conditions |
|
|
Features of Quasi experiments
|
-can have one experimental variable and one quasi-experimental variable
-often data must be collected at particular time or not at all -may want to do experiment certain ways but cannot b/c of practical reasons -enough compromise of experimental control takes place |
|
|
Validity and Quasi Experiments
|
-uncontrolled and confounded variables reduces internal validity but not rendered invalid
-can have higher external validty if quasi expt studies subjects or settings that are more appropriate to the question than a true expt could |
|
|
Randomization and Internal Validity
|
-randomization permits greatest degree of control
-inability to randomly allocate subjects to groups reduces internal validity |
|
|
Choosing experiment type
|
1.true expt
2.quasi expt 3.nonexperimental method -only use other when believe gain in validity will be worth the loss of control |
|
|
nonequivalent-contol-group design
|
-research design having both an experimental and a control group wherein subjects are not randomly assigned to groups
-most typicall quasi -problem: how to compare results bwt experimental and control groups when they were not equivalent to begin with |
|
|
nonequivalent-control-group design with pretest and posttest
|
-typical quasi
-since no random allocation, no good reason to believe groups were equivalent before experimental manipulation -interpretability depends on whether the pattern of the results obtained can be accounted for by possible differences in the groups or by something else in the expt |
|
|
nonequivalent-control-group design: good patterns
|
-two groups showed same performance on pretest
-experimental group improved on posttest but control group did not change -can compare b/c behavior was same in the beginning -also, experimental was lower on pretest but higher on posttest (see effectiveness of manipulation) |
|
|
nonequivalent-control-group design: uninterpertable patterns
|
-experimental group improved but control group did not and experimental group was superior to control from start
-experimental did better on pretest, both improved on posttest but experimental showed twice the improvement (proportional improvement) |
|
|
Designs without control groups
|
-use design that allows same group to be compared over time if no control group can be obtained that can be considered comparable enough to be useful
|
|
|
Interrupted-Time series design
|
-allows same group to be compared over time by considering the trend of the data before and after experimental manipulation (don't just compare average)
-ideal situation is flat and stable baseling before change followed by either abrupt change to new level or gradual change to new level -similiar in design and interpertation to many single-subject designs and nonexperimental methods |
|
|
Repeated-Treatment design
|
-a treatment is withdrawn and then presented a second time
-attempt to improve validity by presenting treatment more than once -limitation: treatment must be one that can be withdrawn w/o causing complication in anaylsis of data -used in single-subjet designs -reversal in trend bwt posttest 1 and pretest 2 when treatment is withdrawn is desirable to rule out possibility of continious change regardless of treatment |
|
|
Developmental Psychology
|
-involves time and nonequivalent control groups
|
|
|
Cross-Sectional Study
|
-in developmental research, a study that test different age groups at the same time
-advantage: all age groups can be tested at the same general time -disadvantage: age is confounded with date of birth b/c ppl at different ages were all born at different times -since every group was born in a different year, cohort effects are likely |
|
|
Cohort
|
a group that has something in common, such as age
|
|
|
Longitudinal Study
|
-in developmental research, a study that test individuals in a single cohort over the course of time
-avoid cohort effects b/c pppl have the same birth dates |
|
|
Problems of Longitudinal Study
|
-practical problem that the research has to wait years to complete the study as the cohort ages
-theoretical problem is it confound age with time of testing meaning have secular trends |
|
|
Secular Trends
|
-a change that is taking place in the general population over time
-can influence a study -ie technology |
|
|
Cross-Sequential Design
|
-design used to help seperate developmental, cohort, and secular effects
-tests individuals from two or more cohorts at two or more times -able to see the effect of time lag |
|
|
Reading Cross-Sequential Design
|
-longitudinal trends: horizontal across the cohorts
-cross sectional effects: vertical, look at ppl tested -time lag effect: diagonal -cross sectional and longitudinal come up with opposite results |
|
|
Time-Lag Effect
|
in a cross-sequential design, the effect resulting from comparing subjects of the same age at different times
|
|
|
Program Evaluation
|
-a set of techniques for determining the effectiveness of a social service program
-intended to determine some factual question not theoritical concerns -true expts, archival research, and emphasis on quasi and nonexperimental methods |
|
|
Why not use same techniques to evaluate programs as employed in evaluating companies?
|
-agencies usually don't operate in competitive market
-product is difficult to evaluate -profit or loss is often irrelevant to program evaluation |
|
|
Sources of Resistance to Program Evaluation
|
-political dimension so evaluation involves many ppl and affects many ppl
-fear program will be terminated (usually modified not terminated) -fear of losing control of the program -fear that information will be abused -fear that the wrong measure will be used (use qualitative and quantitative measures) -belief evaluation is ptless -hopes that are too high (compare new program to some standard program of treatment) |
|
|
Steps in Planning an Evaluation
|
1. Identify Stakeholders
-involve all 2.Arrange Preliminary Meetings 3.Decide Whether an evaluation should be done 4.Examine Literature 5.Determine Methodology 6.Present a written proposal |
|
|
Stakeholders
|
ppl in an organization who stand to gain or lose by any change in it
|
|
|
Preliminary Meetings
|
-meet with players
-sponsors need to be convinced evaluation is good use of thier money -program personnel need to be made comfortable with the idea of being evaluated -type of evaluation can differ among ppl -understanding of when evaluation is needed |
|
|
Summative Evaluation
|
evaluation of the quality of a project often after it is completed
-sponsor may want it |
|
|
Formative Evaluation
|
an evaluation of ways to improve a project while it is ongoing
-project personell may want it |
|
|
Deciding to proceed in an evaluation
|
-is program soundly based in theory
-evaluation concerns the effectiveness of the application of the theory not the validity of the theory underlying the program |
|
|
Methodology
|
-What is it exactly you want to evaluation
-methodology depends on this |
|
|
Science as Conservative: Sources of Bias
|
-older scientists who may be slower to change
-dependence on financial support which is subject to political pressure -type of research that gets proposed and funded is subject to political pressure -status quo control where a group of powerful ppl find in necessary to control the behavior of others |
|
|
Science as Liberal: Sources of Bias
|
-by definition its changing so many find it a change directed against established institution
-search to truth leads to answers that are not palatable to society -dilemma bwt science and values |
|
|
Limitations of Science: Essential or Theoretical
|
-agnostic to concerns of values and God
-objectively observe phenomenia -culturally relative -incomplete |
|
|
Limitation of Science: Practical
|
-oppertunistic nature (problems easier to study or financial support are studied more)
-costs money -complexity of problems (psychological and experiments are usu liminted to 2-3 variables) |
|
|
Chaos Theory
|
-used to understand more complicated psychological interactions
-psychological theories address simple S-R phenomenia not phenomenaw/ complex feedback from R's to S's -this limitation is being addressed using non-linear dynamic systems theory |
|
|
Responsibilities of Scientist
|
-work/research should benefit society
-should foster the ideal of free speech (propse unpopular problems and ideas) -educate public about findings -"watch dogs" for society b/c have additional training and skills |
|
|
Fraud
|
-honesty is essential to science
-aren't many safeguards to insure honesty so left up to integrity of individual |
|
|
Why does fraud occur so infrequently?
|
-major deterrent is that important experiments will be replicated
-replication is ultimate test for reality of finding and thus ultimate deterrant of fraud |
