-Transform
items that need to be reversed scored

Transform…Recode…

-Run
frequencies for each item

Analyze…Descriptive
Statistics…Frequencies…Move variables into Variable(s) window…OK

Look
for impossible values

Look
at response distributions

Analyze…Descriptive
Statistics…Frequencies…Move variables into Variable(s) window…select Statistics…check
off Mean, Median, Standard Deviation, and Skewness…Continue…OK
[make sure you uncheck “Display frequency tables”]

Look
for skewed items

Look
for items with lots or very little variability

*Analysis
1**:
After you access or enter your data file, do the following:*

Click
on:

q__A____nalyze;__

q__D____ata
Reduction;__

q__F____actor.__

Then,
select variables for your factor analysis as usual.

At
the Factor Analysis box:

qClick
on __D__escriptives;

qClick
in the box next to __C__oefficients;

qClick
in the box next to __K__MO and Bartlett’s test of sphericity;

qClick
on Continue.

At
the Factor Analysis box:

qClick
on __E__xtraction;

qSelect
Principal components in the __M__ethod box (__¯____);__

qClick
in the box next to __S__cree Plot;

qChange
Ma__x__imum Iterations for Convergence to 50;

qClick
on Continue.

At
the Factor Analysis box:

qClick
on Ro__t__ation;

qConfirm
that “no rotation” is selected;

qClick
on Continue.

At
the Factor Analysis box:

qClick
on __O__ptions;

qClick
in the box next to __S__orted by size;

qClick
on Continue

Run
the first analysis.Determine the
number of factors with the number of eigenvalues
greater than one.Select two or more
other model dimensionalities to consider in the search for the appropriate
number of factors.

*Analysis
2**:
Rerun analysis (1) above with the same settings except for the following
changes:*

qIn
the Descriptives box, deselect “KMO and Bartletts
test.”

qIn
the Extraction box, select Maximum likelihood in the __M__ethod box
(__¯____),
set the “number of factors” equal to
the # you have decided to test for, and deselect “Scree
plot.”__

*Analysis
3**:
Rerun analysis (2) above with the same settings except for the following
changes:*

qIn
the Extraction box, set the “number of factors” equal to an alternate #
of factors that you wish to examine. (Results from analyses 1 through 3
help us to decide how many factors we want)

*Analysis
4**:
Rerun analysis (3) above with the same settings except for the following
changes:*

qIn
the Descriptives box, deselect “Initial
solution” and “Coefficients.”

qIn
the Extraction box, set the “number of factors” equal to the # of factors
you have decided on and deselect “unrotated factor
solution.”

qIn
the Rotation box, select “Varimax” if you think
the factors are not correlated; select "Direct Oblimin" if you think the
factors are correlated.

**Item
analysis for each factor**

Inter
item correlations

Item
criterion correlations

Alpha
if item removed

Over
all coefficient alpha

Analyze…Scale…Reliability
Analysis…move Variables into the Items window…click on the Statistics button…check…Scale
if item deleted…Inter-item Correlations…continue…OK

-You
will also need descriptive statistics for your factor/composite scores.