The **Analyze** Menu is the work horse of SPSS. Nearly all procedures
that generate output are located on this menu. For this review, however,
we only focus on several of these hundreds of analyses. In fact, the three
procedures that follow all provide some of the same statistics.

The frequencies procedure is primarily used for discrete data (e.g., nominal
and ordinal data), although there are a number of options that are useful for
scale level data.

This option brings up a dialogue box, and we need to move the variables of interest
from the field on the left to the field on the right.

For nominal variables, for which further descriptives statistics are not
appropriate (with the exception of the mode), we can skip the **Statistics** to obtain
frequencies for each category.

For ordinal and scale variables, though, we will want to specify additional
descriptive statistics to be calculated. These can be broken down into measures of
central tendency (mean, median, mode, sum), variability (variance, standard deviation,
range, minimum, maximum), and percentiles. This last category includes quartiles
(25th, 50th, and 75th percentiles), cut-points for an arbitrary number of groups,
and any arbitrary percentile.

Most options are selected simply by clicking on the box next to each item.
For specific, arbitrary percentiles, select the option, type the desired percentile
in the field to the right, and then click on the add button below:

Which results in the desired percentile being added to the list. Note that one can always
delete or modify an entry. Also, more than one entry may be made.

These options generate the following output:

We can get many of these same statistics from the **Descriptives** item. The
options available, however, are fine-tuned to scale level variables.

We first select the desired variables from the field on the left by moving them
to the right.

Then, we click on the **Options** button to determine which statistics should be
computed.

These options, then, generate the following output.

One final method for obtaining descriptive statistics focuses on generating
statistics from multiple goups quickly and efficiently. This procedure is obtained from
the **Compare Means** item of the Analyze menu, and then the **Means** item on
the submenu.

The dialogue box requires that we select *two* variables: The dependent variable
is the one on which the statistics are computed, and the independent variable list contains
the discrete variables that characterize the different groups.

For example, if we want to compute average stress values based on pet ownership, the
dialogue box would look like the following.

And the output would look like the following.

If we want to consider more than one different group, we can add layers to the
independent variable list. For instance, we might want to compute means separately
for men and women within each pet ownership group. We start by clicking on the
**Next** button to add another layer

Doing so creates a field for the second layer, in which we specify the next grouping
variable.

Then, we select the variable to be used in the second layer, in this case, gender.

These options, then, create a full table of means and standard deviations.

- Create a data set for the following data:

Group Gender Hw1 Hw2 Hw3 expt Male 92 84 93 expt Female 77 84 85 expt Male 87 86 81 expt Female 89 90 93 expt Male 64 73 78 control Female 81 84 93 control Male 83 90 91 control Female 84 88 86 control Male 82 80 78 control Female 96 91 88 - Using the Frequencies option, find the mean, median, mode, quartiles, 95th percentile, variance, standard deviation, minimum, and maximum of Hw1, Hw2, and Hw3.
- Using the Descriptives option, find the means and standard deviations of Hw1, Hw2, and Hw3.
- Using the Compare Means -- Means procedure, find the means on Hw1, Hw2, and Hw3 for everyone, for the experimental group, for the control group, for men, for women, and for all combinations of gender and group.