A Student’s Guide to
Interpreting SPSS Output
for Basic Analyses
These slides give examples of SPSS output with notes about
interpretation. All analyses were conducted using the Family
Exchanges Study, Wave 1 (target dataset)
1
from ICPSR. The
slides were originally created for Intro to Statistics students
(undergrads) and are meant for teaching purposes only
2
. For
more information about the data or variables, please see:
http://dx.doi.org/10.3886/ICPSR36360.v2
3
1
Fingerman, Karen. Family Exchanges Study Wave 1. ICPSR36360-v2. Ann Arbor, MI: Inter-university Consortium for Political
and Social Research [distributor], 2016-04-14. http://doi.org/10.3886/ICPSR36360.v2
2
The text used for the course was The Essentials of Statistics: A Tool for Social Research (Healey, 2013).
3
Some variables have been recoded so that higher numbers mean more of what is being measured. In those cases, an “r” is
appended to the original variable name.
Frequency Distributions
Frequencies show how many people fall into each answer category on a given question
(variable) and what percentage of the sample that represents.
Total number of people in the
survey sample
Number of people with valid (non-
missing) answers to the question
Percent of the total sample who
answered that “child1” was married
Number of people who responded
that “child1” was married
Percent of those with non-missing
data on this question who answered
that “child1” was married
Cumulative percent adds the
percent of people answering in one
category to the total of those in all
categories with lower values. It is
only meaningful for variables
measured at the ordinal or
interval/ratio level.
Crosstabulation Tables
“Crosstabs” are frequency distributions for two variables together. The counts show how
many people in category one of the first variable are also in category one of the second and
so on.
Number of people who answered that
their children are biologically related to
them and that their children need less
help than others their age.
Marginal: Total number of people who
answered that their children need more
help than others their age.
Marginal: Total number of people who
answered that all of their children are
biologically related to them.
Marginal: Total number of people who
had valid data on both D34r and A1A.
Crosstabulat ion Tables (Column %)
Crosstabs can be examined using either row or column percentages and the interpretation
differs depending on which are used. The rule of thumb is to percentage on your
independent variable.
Marginal: Percent of sample who feel that
their children need less help than others
their age.
Percent of the sample whose children are
all biologically related to them that said
their children need less help than others
their age.
Marginal: 100% here tells you that you’ve
percentaged on columns.
Can you interpret this number?
(17.6% of those whose children were not
all biologically related to them felt their
children needed more help than others
their age.)
Crosstabulation Table (Row %)
Percent of those who said their children
need more help than others their age
whose children are all biologically
related to them.
Marginal: Percent of the sample whose
children are biologically related to them.
Marginal: 100% here shows that you are
using row percentages.
Can you interpret this number?
(17.4% of those who said their children
need about the same amount of help as
others their age had children who were
not all biologically related to themselves.)
Chi Square (
)
Based on crosstabs,
2
is used to test the relationship between two nominal or ordinal
variables (or one of each). It compares the actual (observed) numbers in a cell to the
number we would expect to see if the variables were unrelated in the population.
Actual count (f
o
).
Count expected (f
e
) if variables were unrelated
in pop.
f
e
=
value (obtained) =
This could be
compared to a critical value (with the degrees
of freedom) but the significance here tells you
that there appears to be a relationship between
the perceived amount of help needed and
whether the children are related to the R.
Degrees of freedom = ((# rows – 1)(# col – 1))
The
2
test is sensitive to small expected
counts, it is less reliable if f
e
is < 5 for multiple
cells.
Independent-samples T-test
A test to compare the means of two groups on a quantitative (at least ordinal, ideally
interval/ratio) dependent variable. Acomputed variable (dmean_m) exists in this dataset
that is the average amount of support R offers mother across all domains (range 1-8); that
will be the dependent variable.
The actual mean and standard
deviation of dmean_m for each of the
groups.
The number of females and males who
have non-missing data for dmean_m.
T-value for the difference between 4.7233 and
4.3656. t
=
where
=
+
.
(Note, as long as the sig. of F is ≥ .05, use the
“equal variances assumed” row, for reasons
beyond the scope of these slides).
Significance (p) level for t-statistic. If p ≤ .05, the
two groups have statistically significantly different
means. Here, females provide more support to
their mothers, on average, than do males.
Confidence interval for the difference between
the two means. If CI contains 0, the difference
will not be statistically significant.
Paired-samples T-test
Like the independent-samples t-test, this compares two means to see if they are significantly
different, but now it is comparing the average of same people’s scores on two different
variables. Often used to compare pre- and post-test scores, time 1 and time 2 scores, or, as in
this case, the differences between the average amount of help Rs give to their mothers
versus to their fathers.
Mean amount of support R provided to mothers.
Mean amount of support R provided to fathers.
Number of cases with valid data for both
variables. SPSS also gives the correlation
between the two dependent variables, that was
left off here for space.
The difference between the average amount of
support provided to mothers and fathers and
accompanying standard deviation.
T-statistic for the difference between the two
means and the significance. In this sample,
respondents provide significantly more support
to their mothers than to their fathers.
Oneway ANOVA
Another test for comparing means, the oneway ANOVA is used when the independent
variable has three or more categories. You would typically report the F-ratio (and sig.) and
use the means to describe the groups.
Number of cases in each group of the
independent variable.
Average amount of support provided (and
standard deviation) by those with incomes
< $10k.
Confidence Interval: the range within which
you can be 95% certain that the group’s
mean falls for the population.
Average amount of support provided by all
521 people with valid data on dmean_m.
F-statistic (and associated p-value) test the null hypothesis that all groups have the same
mean in the population. A significant F means that at least one group is different than the
others. Small within groups variance and large between groups produces a higher F –value: F=
. Here we see that at least one group’s mean amount of support is
significantly different than the others. Additional (post-hoc) tests can be run to determine
which groups are significantly different than each other.
Variability of means between the groups:
Mean Square Between =
where SSB=
(
)
2
and dfb = k-1. (k is number
of groups;
is # people in a given group;
is mean for that group)
Variability within each group: Within
Groups Mean Square =
where SSW =
SST-SSB and dfw = N-k.
Total Sum of Squares (SST) = SSB + SSW
Correlation
Pearson’s r: measures the strength and direction of association between two quantitative
variables. Matrices are symmetric on the diagonal.
Correlation coefficient (r) tells how strong
the relationship is and in what direction.
Range is -1 to 1, with absolute values closer
to 1 indicating stronger relationships. Here,
the frequency of visits is moderately related
to the amount of emotional support given to
the mother; more visits correlate with more
frequent emotional support.
r =
(
)(
)
[
(
)
(
)
Significance (p) tells how certain you can be
that the relationship displayed is not due to
chance. Typically look for this to be ≤ .05.
N=sample size: this number may be different
in each cell if missing cases are excluded
pairwise rather than listwise.
Can you interpret this number? (How
often mother forgets to ask about R’s life is
negatively, albeit weakly, related to the
amount of emotional support R gives mother
– If mother forgets to ask a great deal of the
time, she gets less emotional support from
R.)
Bivariate Regression (model statistics)
Examines the relationship between a single independent (“cause”) variable and a
dependent (outcome) variable. While it’s good to look at all numbers, the ones you
typically interpret/report are those boxes marked with an * (true for all following slides).
Regression line: = + .
Coefficient of determination (R
): the amount of
variance in satisfaction with help given to mother that
is explained by how often the R saw mother. R
2
= (TSS
– SSE)/ TSS. *
Multiple correlation (R): in bivariate regression, same
as standardized coefficient
Independent (predictor) and Dependent variables.
F-value (and associated p-value) tells whether model is
statistically significant. Here we can say that the
relationship between frequency of visits with ones
mother and satisfaction with help given is significant;
it is unlikely we would get an F this large by chance. *
Residual sum of squared errors (or Sum of Squared
Errors, SSE) =
( )
.
Total Sum of Squares (TSS) =
( )
.
Bivariate Regression (coefficients)
Standardized coefficient (β): influence of x on
y in “standard units.”
Confidence Interval – the slope +/- (critical
t-value * std. error) – shows that you can
be 95% confident that the slope in the
population falls within this range. If range
contains 0, variable does not have an
effect on y.
Standard error of the estimate: divide the
slope by this to get the t-value. *
Y-intercept (a): value of y when x is 0.
Slope (b): how much y chances for
each unit increase in x. Here, for every
additional “bump up” in frequency of
visits, satisfaction with the amount of
help given to mother increases by
.157. *
T-statistic (and associated p-value) tells
whether the individual variable has a
significant effect on dependent variable. *
The regression equation (for this model
would be = 1.770 + .157().
Multiple Regression (OLS: model statistics)
Used to find effects of multiple independent variables (predictors) on a dependent variable.
Provides information about the independent variables as a group as well as individually.
Regression line:
= a +
+
+
…
R= multiple correlation. The association between the group of
independent variables and the dependent variable. Ranges from 0-1.
R
2
and Adjusted R
2
– how much of the variance in satisfaction with
amount of help R provided mother is explained by the combination of
independent variables in the model. Also called “coefficient of
determination.” R
2
= (TSS – SSE)/ TSS. Adjusted R
2
compensates for the
effect that adding any variable to a model will raise the R
2
to some
degree. About 14% of the variance in satisfaction is accounted for by
financial and emotional support, seeing mother in person, and whether
mother makes demands on R. *
Residual sum of squared errors (or Sum of Squared Errors, SSE) =
( )
.
Total Sum of Squares (TSS) =
( )
.
F-value and associated p-value tells whether model is statistically
significant (chance that at least one slope is not zero in the population).
This combination of independent variables significantly predicts
satisfaction with amount of help R gives mother. *
Df2 = # of cases – (# of independent variables +1)
Df1 = # of independent variables
Multiple Regression (coefficients)
Y-intercept (a): value of y when all Xs are 0.
Slope (b): how much satisfaction changes for
each increase in frequency of visiting mother. *
Standard error of the estimate: divide the
slope by this to get the t-value. *
Zero-order correlation = Pearson’s r. Bivariate
relationship between frequency of visits and
R’s satisfaction w/ help R’s given to mother.
Partial correlation: bivariate relationship
between frequency of visits and R’s
satisfaction, controlling for demands mother
makes and emotional/financial support given.
T-statistic (and associated p-value) tells whether the
individual variable has a significant effect on
dependent variable, controlling for the other
independent variables. *
Regression Equation: = 1.477 + .130 b3ar + .195(m_6) .009 d1r + .033(d22r)
Standardized coefficient (β): influence of x on y in
“standard units.” Can use this coefficient to
compare the strength of the relationships of each
of the independent variables to the dependent.
Largest β = strongest relationship, so here
frequency of visits has the strongest relationship to
satisfaction, all else constant.
OLS with Dummy Variables
Using a categorical variable broken into dichotomies
(e.g., race recoded into white, black, other with each
coded 1 if R fits that category and 0 if not) as
predictors. In this case, the amount of help R
perceives his/her adult child to need was recoded
into 1 = “more than others” and 0 = “less or about
the same as others.” If the concept is represented
by multiple dummy variables, leave one out as the
comparison group (otherwise there will be perfect
multicollinearity in the model).
Slope is interpreted as the amount of difference between the “0”
group and the “1” group. Here, those who perceive their
children needing more help than their peers are .321 more
satisfied with the amount of help they give their mother than
those whose children require less help, controlling for frequency
of visiting mother, demands made, and emotional and financial
support provided (and the difference is statistically significant).
Dummy variable. Children who need less or the same amount of
help as their peers is the reference category (0).
