The flashcards below were created by user
blueclues
on FreezingBlue Flashcards.

Compare multilple groups by using a ______ design
multiplegroup

More than two ___ ___ are needed to accurately compare multiple effects
data points

We need to see the full shape of the graph in order to determine the ___ ___
Functional Relationship

A manipulation's construct validity is weakened by ___ variables
Confounding

A group that recieves no treatment  not even a placebo
empty control group

a treatment that has no effect
placebo

When group means are further apart, there is greater ___
variability

ANOVA
Analysis of Variance

Allows us to compare betweengroups variance to withgroups variance
ANOVA

the ratio of the betweengroups variance to the withingroups variance
F ratio

Our environment varies in ___ and to ___ ___ a variable is present
whether and to what degree

level of a variable
data point

Finding the functional relationship requires ___ more than 2 data points
mapping

To make accurate statements regarding effects we need to know the ___ ___ of the IVs and DVs
functional relationship

Multilevel experiments means greater ___ and ___ due to ___ ___
external and construct
random selection

variables are ___ confounded
frequency

___ ___ experiments are stronger than the simple experiment
multiple group

in Multiple Group experiments, ___ validity is improved and we are able to map ___ relationships
construct
functional

Determine the ___ we will use BEFORE the type of Design
analysis

Type of analysis used will influence (3)
treatments
# of participants
hypothesis

To find a treatment effect, we look for ___
variability

difference among group means
variability

A ___ difference between means is most likely due to chance
small

A ___ difference between means is likely due to the effect of the treatment
large

Withingroup variability is NOT due to ___
Treatment

Withingroup variability is due to ___ ___
Random Error

Betweengroups variability is due to ___ and ___ ___
Treatment and Random Error

the effects of random error
withingroup variability

the effects of random error plus any treatment effects
betweengroup variability

Effects of random error (3)
individual differences
unreliability of the measure
poor standardization

tells us the extent to which chance caused individual scores to differ from each other
withingroups variability

tells us the degree to which our groups actually vary from one another
betweengroups variability

Two factors affect the extent to which group means differ
random error
treatment effect

In an ___, we see any possible effects of individual IVs and ALSO any interaction effects between the variables
ANOVA

We can ask more refined questions in a ___ ___ experiment
multiple group

factors other than the treatment
extraneous factors

We can eliminate extraneous factors by using a ______ or the ___/___ design
twogroup
pretest/posttest

If groups differ before the treatment, differences in scores could be due to ___ ___ which threatens ___ validity
selection bias
internal

By using ___, groups should have nearly identical characteristics
matching

groups are the same on the pretest, but develop at different rates
selection by maturation interaction

Average scores tend to become less extreme at the ___
posttest

tells us that extreme scores will revert back to more normal levels on retest
regression toward the mean

Groups may also end up being different due to ___  when participants drop out of a study
mortality

use the same participants in the notreatment group and the treatment group. This approach uses the ______ design
pretest/posttest

Participants may change because of ___. A change in a participant's environment has a ___ effect on their scores.
History
Systematic

when participants become better at taking our test due to practice resulting in ___ ___
testing effects

changes in the measuring scale resulting in changes in scores
instrumentation

Any change in the cause must be reflected by change in the ___
effect (outcome)

Differences in the ___ variable must be reflected by a change in the ___ variable
independent/ dependent

changes in the ___ occured before changes in the ___
treatment / activity

Experimental design where everything is identical for both treatment groups, except that one group recieves the treatment
Group scores are compared
twogroup design

8 Threats to internal validity
 groups were different to begin with
Selection

8 Threats to internal validity
 groups were destined to grow in different ways
selection by maturation interaction

8 Threats to internal validity
 extreme pretest scores tend to normalize
regression effects

8 Threats to internal validity
 participants drop out
mortality

8 Threats to internal validity
 natural growth and development appear at treatment effects
maturation

8 Threats to internal validity
 changes in environment caused changes in participants
history

8 Threats to internal validity
 practice on the pretest caused changes in posttest scores
Testing

8 Threats to internal validity
 the questionnaire changed between pre and posttest
Instrumentation

Arbitrary assignment to groups and when participants choose in which group they want to be.
These individuals are different!!!
selection bias

Measurements are only as accurate as the ___
tool we are using

Using unreliable tools produces ___ ___
random error

Group's averages tend to become less extreme during the ___
posttest

scores that are extreme will revert back to more normal levels at the retest
regression toward the mean

Design used to determine: Why people behave the way they do and How we can help them behave differently
Simple Experiment

Design used to isolate underlying causes:
Requires internal validity (cause and effect)
Simple Experiment

In a simple experiment, groups must NOT differ ___.
One group will receive treatment and the other will not.
Systematically

In a simple experiment, treatments can involve different ___ and ___ of activity
types / amounts

To ensure that the treatment is the only difference between groups of a simple experiment, we use ___ ___
random assignment

In a simple experiment, we test the experimental hypothesis against the ___ ___
null hypothesis

Hypothesis that states that the treatment WILL HAVE an effect
experimental hypothesis

Larger groups tend to be more ___, and effects will stand out better
similar

we need to know the scores of our Dependent variable to be able to calculate any ___ ___ of our Independent variable
significant effects

If a difference is ___, we can be certain, beyond a reasonable doubt, that the difference is NOT due to ___
Significant / Random Error

With a ___ ___ error, we state that a difference between our groups was Significant when it was NOT
Type 1

With a ___ ___ error, we state that a difference between our groups was NOT significant, but due to Chance, when it actually WAS significant
Type 2

To avoid Type 2 errors
Do Not set the significance level Too LOW

We need enough ___ to detect true Differences
power

Participants who are similar
homogenous

___ limits random error from disguising our treatment effect
power

Power 
Increase the ___ of our treatment effect
size

We use the ___ ___ to see whether the difference between Mean group Averages is big enough NOT to be due to random error.
ttest

