RMP Exam 2
Experiment
a study where there are only independent variables
Conditions/levels
the groups formed by the independent variables manipulation
Treatment group
Level of independent variable that has the key active variable
Control group
The group that remains unaffected by manipulation
Placebo group
Group that remains unaffected by manipulation but they think they are
Constant
Something that does not vary in the study
Covariation
The two variables must go together
Temporal precedence
The predictor variable must come before the dependent variable
Third variable criterion
The ability to eliminate other possible alternative explanations for the results
Random assignment
Confounds→ a possible alternative explanation for the results but it varies with the manipulation
Confounds ex. Brian vs. Karen
Manipulation Check
Designed to see how well the manipulation worked
Pilot study
Just tests a new manipulation to decide whether it is appropriate or not
Main effect
The effect of a predictor variable on a DV and IV ignoring other predictor variables (each predictor variable has its own main effect)
Interaction
The effect of one predictor variable on the DV which depends on the level of another predictor variable (two predictors together produce an interaction)
Correlation
a statistical analysis which tests the association between two variable (pearsons’ r)
THIS IS ABT STATISTICAL ANALYSIS
Correlational study
a study which all predictors are measured
THIS IS ABT METHODOLOGICAL APPROACH
Statistical significance
The likelihood of getting our data by chance ASSUMING there is no effect in our population
Effect size
the strength of magnitude of a relationship
Larger effect sizes are often but not always more important
In general, larger effects are more likely to be statistically significant
Restriction of range→ having a limited range on one variable can make relationships look weaker than they are
Internal validity
The strength of the claim that the predictor variable (vs. something else) is causing change in the dependent variable
Covariation
Two variables must be related
Temporal precedence
The predictor variable must come before the dependent variable
Third variable criterion
The ability to eliminate other possible alternative explanations for the results
Matched Design/Matched Groups
participants who are similar to some important measure are grouped into sets and then members of each set are randomly assigned to different conditions
potential disadvantages are…
Pretest/Posttest Design
participants are tested on the key dependent variable twice, once before and after manipulation
potential disadvantages are…
Within-Groups
each participant experiences all levels of the independent variable
potential disadvantages are…
Order Effects
one experience or level of the IV can effect the others
Demand Effects
cues that lead participants to guess what the study’s purpose
Counter Balancing
presenting levels of IV to participants in different orders to control order effects
Quasi Experiment
one measured and one manipulated predictor variable
statistics
math procedures to interpret information
parameter
a numerical value we use to describe a population- does not vary
A statistic
a numerical value used when we describe a sample
sampling error
differences that exist from one sample to another, natural
descriptive statistics
numerical and graphical ways to analyze data including mean and mode, used to simplify and organize data
graphs, correlation and effect size, z-score percentile
central tendency
inferential statistics
uses sample info to make estimates, approx, conclusions and predictions about population
Null hypothesis significance testing, Type I and Type II error, Power, Types of inferential tests and when to conduct them
central tendency
descriptive statistics that provide an average, typical, or representative score that defines the center of a distribution
mean, median, mode
variability
Describes the distribution
Estimates distance we expect between scores
Provides a sense of how well an individual score will represent the population
How much error to expect if you’re using a sample to represent the population
variance
average deviation ² distance from the mean
standard deviation
average distance or deviance score from the mean
P<0.05
The data is statistically significant
P>0.05
the data is not statistically significant
Type I error
false positive
Type II error
false negative
Power
the probability that the test will correctly reject the false null
Inferential tests
chi squared
Correlational
independent t test
ANOVA
factorial ANOVA
replication
conducting the study again and getting the same results
replication plus extension
has new variables
conceptual replication
has same conceptual variables but new operational
meta-analysis
quantitative review of effect
floor effects
data is clustered at the bottom of the scale
ceiling effects
data is clustered at the top of the scale
noise
too much variability associated with predictors