Doing More with SPSS

Today's topics

Basic and intermediate inferential statistics  
Introduction to SPSS syntax


Graphs and Charts

Bar charts  summarize variables by groups and subgroups

Simple displays summaries by groups
Display frequencies or percents for groups [Ex: N for HSP]
Display summary statistics for a variable by groups [Ex: Mean(RDG) by HSP]
Statistical displays for a set of variables [Ex: Medians (RDG WRTG MATH)]

Cluster displays summaries by groups and sub-groups (clusters) [Ex: %'s for HSP SEX]
An elaboration of simple bar charts
Bars for "missing" can be discarded in the options menu
Can use pairwise or listwise deletion

Histograms  summarize an interval variable [Ex: SCI with normal curve]
A normal curve can be superimposed

Line charts  summarize by groups

Simple is like a simple bar chart, midpoints of bar tops are connected [Ex: Mean(RDG--CIV)]
Multiple is analogous to cluster bar charts
Can get separate lines for groups or variables
Series--transpose data switches groups with variables [Ex: Mean(RDG--CIV) by SES]

Scatterplots  show relationships between variables

Simple shows the relationship between two variables [Ex: RDG w/ WRTG by SEX]
Axes are usually interval variables
Markers can be used for separate groups
Options can fit regression lines to the total group or subgroups
Sunflowers are for marking # of hidden cases (can't be done for subgroups)

Some other charts
Pie, area, high-low, boxplots, control, time series


Editing Charts

Done in the chart editor (or in chart windows for other versions of SPSS)
Double-click on chart to open it in chart editor
Most changes are made by editing the chart, highlighting, and selecting from menus
Depending on the type of chart being edited, different characteristics can be changed, including:

       Axis scaling and ticks
       Number of intervals
       Interpolation, fitting lines, reference line
       Labels, text styles, and fonts
       Titles, footnotes, annotation, and legends
       Fill patterns, spacing, and colors for bars
       Bar style
       Marker colors, styles, and sizes
       Line styles and colors
       Swapping axes
       Transposing data

Sizing charts
Saving charts

Save as SPSS output  (.spo)
Export as a jpeg, pict, png, tiff, or bmp (not SPSS files)
Cut and paste to another application


Editing and Saving Output

Output Viewer

Click on tree in left sends you to a table on right
Double-click on table to get in edit mode
   (or go to Edit menu and open/edit SPSS Pivot Table Object)
Double-click on cell you wish to change
Somewhat limited editing options for tables
Objects can be cut and pasted to other applications

File Menu

Tables can be saved as SPSS output files (.spo)
They can be exported as HTML or text files


Basic Inferential Statistics

Correlate: analyze relationships between variables [Ex: RDG--CIV CONCPT MOT]
Pearson vs. Spearman
Significance: 1 or 2 tailed p-values
listwise or pairwise deletion
Graph: Scatterplot
Crosstabs: analyze differences in proportions [Ex: RACE by SES]
Chi-squared {Sum of ((O-E)2/E)}
phi and Cramer's V
Cells:  observed; expected; row, column, total percents; residuals: unstandardized
Graph: Cluster bar chart with frequencies
Compare means: analyze differences in means

Paired-samples t-test (dependent t-test) ) [Ex: RDG with WRTG & MATH ]
Two variables
Confidence interval
Graph: Bar chart with means

Independent samples t-test [Ex: RDG & MATH by SEX]
Define groups
Graph: Bar chart with means


Intermediate Inferential Statistics-ANOVA

One-way factorial ANOVA

Compare means: One-way ANOVA[Ex: LOCUS by RACE, with Bonferroni post-hoc tests]
Post hoc tests show pairwise comparisons while controlling for the number of tests being conducted
Graph: Bar chart with means

General Linear Model: Univariate [Ex: LOCUS by RACE]
Differences between group means (like an extension of independent t-test)
You must specify the range for the independent variable 
Graph: Bar chart with means

Repeated measures ANOVA
General Linear Model: Repeated measures [Ex: RDG WRTG MATH SCI]
Differences among means for variables (like an extension of dependent t-test)
Often used for analyzing differences over time
Specify the factor name and number of levels
Define: click on variable names
Contrasts: only limited contrasts are available through menus (seldom useful)
Graph: Line chart with means


Intermediate Inferential Statistics-Regression

Multiple Regression: forming models for prediction from multiple predictor variables
Outcome variable is usually an interval variable
Predictor variables are usually interval variables, but can include dichotomous variables

Linear regression forms a linear composite of the predictor variables in such a manner that the correlation between the composite of predictors and the outcome variable is maximized


Using SPSS Syntax

Why use syntax?
It can save you time if you are running similar procedures repeatedly, or at different times
SPSS syntax can be saved as a file and retrieved for future use
Some SPSS operations are not available through menus
Downside: reading manuals

Open a syntax window
Preferences: open a syntax window at startup and save preferences
Paste commands to a syntax window
Open a syntax file

Pasting syntax vs. writing
Pasting is easy, writing is not
Writing gives you options unavailable from menus, but:
    You have to know SPSS syntax
    You will frequently refer to SPSS manuals
    Errors are commmon and not always easy to solve
Executing syntax
Highlight the commands you want executed, then click the right arrow button on the menu bar

Saving syntax

Syntax window-- File-- save or save as