Carrying out a principle component analysis using varimax rotation
1 of 14
Descriptive Statistics
The case to variable ratio is _/ _:1, greater/less than our min of 10:1, have sufficient/too few cases for the analysis
2 of 14
Correlation matrix
The determinant of _ is GREATER than 0.00001, therefore there is no evidence of multicollinearity or singularity (less than is evidence)
3 of 14
KMO
The KMO test of sampling adequacy (_) is less than 0.7 therefore factor analysis may be appropriately undertaken on the data
4 of 14
Bartlett's test of sphericity
... is significant (p_) therefore the data correlation matrix is sufficiently different to the identity matrix so we have sufficient significant correlations to proceed with the analysis
5 of 14
Anti-image matrices
Only needed in KMO is below 0.7, drop variables with the lowest value until it raises
6 of 14
Communalities
The commonalities are all greater than our minimum requirement of 0.2 therefore the factor solution account for an acceptable amount of the variance of each variable
7 of 14
Total variance explained
Using the kaiser (K1) criteria suggests we retain _ factors, accounting for _% of the total variance
8 of 14
Scree Plot
The scree plot suggest an elbow at _, in indicating a (one less) factor solution, dis/agreeing with the K1 criteria
9 of 14
Monte Carlo
(Table: random and data eigenvalue) Using parallel analysis suggests a _ factor solution, dis/agreeing with the K1 criteria and scree plot
10 of 14
Component matrix
ignore
11 of 14
Rotated component matrix
Factor one can be called _. Write each variable and its value. Repeat for each factor.
12 of 14
Component transformation matrix
ignore
13 of 14
T-test
(to interpret factor scores in relation to another variable) report levene's and compare the means
14 of 14
Other cards in this set
Card 2
Front
Descriptive Statistics
Back
The case to variable ratio is _/ _:1, greater/less than our min of 10:1, have sufficient/too few cases for the analysis
Comments
No comments have yet been made