Mixed Topic
What does the E stands for in IDEAS?
Examine psychometric properties
Which type of validity do the items fully represent the construct that they may aim to measure?
Content
This type of factor analysis assumes the presence of a latent variable that cannot be directly measured using a single variable.
EFA (Exploratory Factor Analysis assumes the presence of a latent variable that cannot be directly measured using a single variable)
What type of reliability involves the scale being split into two and the scores for each half are correlated
Split-half reliability
In psychology, do we usually measure direct or latent constructs?
Latent (ex. sadness, motivation)
Which type of validity does the test measure the concept that it intends to measure?
construct
This type of factor analysis is used to create component variables from linear combination of a set of variables.
PCA (Principal Component Analysis - used to create component variables from linear combination of a set of variables)
What is the skewness if most scores are high and pile up to the right?
Negatively Skewed Distribution
The following are examples of indirect construct measurement of Psychometrics except:
a. happiness b. height c. volume d. weight
a. happiness
Which type of validity does content appear to suitable
face validity
This type of factor analysis makes assumption based on based on theory what items will comprise a factor
CFA (Confirmatory Factor Analysis makes assumption based on theory what items will comprise a factor)
The most common measure of scale reliability
Cronbach’s alpha (α)
True or False - Factor Analysis can be used with ordinal and nominal data
False
Which type of validity do the results correlate with other measures/ constructs
criterion
The following are required to do factor analysis except:
a. continuous data
b. have adequate sample size
c. adequate difference from each other
d. adequate correlation between items
c. adequate difference
True or False - even if subs scale reliabilities are good, you cannot use them as separate variables
False -
When sub scale reliabilities are good, you can use them as separate variables, if not, consider using the entire scale instead of subscale as a variable.
What is the process of deciding how many factors to keep?
Factor extraction
This clusters of variables that correlate highly with each other that may measure a latent variable
factor analysis
This is used to discriminate between factors such that variables are loaded maximally to only 1 factor
factor rotation
According to Klein (1999) what is the suggested cut-off for cognitive tests?
.80 for cognitive tests