MODULE 6

DESCRIPTIVE STATISTICS IN SPSS

1. to determine frequency counts and percentages using

SPSS
2. to compute for the measures of central tendency of a set
of scores using SPSS
3. to compute for the measure of variability of a set of a
scores using SPSS

Introduction

Computing the measures of central tendency (mean, median and the

mode), and measures of variability (range, quartiles, decile percentiles, standard
deviation, variance and coefficient of variation) using long methods of calculations
are tedious and prone to error especially involving large data sets (i.e. hundreds
or even thousands of data). These long calculations may be one of the reasons
why statistics as a course seems to be complicated and difficult. Thus, students
tend to dislike the subject.

With technological advancement and the invention of computers, programs

have been developed to help people overcome difficulties in long calculations, and
more importantly arriving at more accurate results. In research and related
endeavors, the Statistical Package for Social Sciences (or SPSS) is one of the
many calculation programs developed by the International Business Machines,
Corporation (or IBM) to aid researchers in analyzing large data sets easily.
Specifically, this module utilizes the SPSS version 21. Some researchers prefer
to analyze data sets which are encoded in the SPSS program (as SPSS file), but
for this module, analysis involves data in excel files (as imported files from excel
to SPSS).
FREQUENCY COUNTS AND PERCENTAGES IN SPSS

In most cases, aside from analyzing the data to answer the main objective
of the study (i.e. testing the hypothesis that there is no significant difference on the
test anxiety of male and female students), we usually start in determining how
many of the respondents (from a large data) belongs to a category in a study
variable. For instance, of the 1000 respondents, “how many are females ?, without
actually counting it manually. Or maybe we are interested on determining, “what
percent of the students answered strongly agree on one of the test anxiety items
in your questionnaire, again without actually having to count it.

To illustrate this, we will use the excel file. Folder name: DATA SET FOR
LECTURES, File name: DATA SET (descriptive statistics), a portion of which
is show below:
 The sample excel file contains data gathered from a sample of 90 students,
who were asked relevant information such as their gender, type of school, how
much they like schooling, and their scores in English test, Math test and Science
test.

The categorical variables were dummy coded as:

a) Gender (1-male, 2-females)

b) Type of school (1-public, 2-private)
c) How much do you like schooling in general
( 5- very much, 4- much, 3- neutral, 2- not much, 1- not at all).

Meanwhile, English test, Math test and Science test are the actual raw
scores obtained in the test. Using the data, respondent 1 is a male (coded as 1),
enrolled in a private school (coded as 2), and answered neutral in terms of how
much he like schooling (coded as 3). He got the score of 65, 88 and 76 in the
English test, Math test, and Science test, respectively.

How to import excel files to the SPSS program

Before we analyze the descriptive statistics of the study variables (and any other
statistical analyses), it is important to learn how to import an excel file to the SPSS
program.
Step 1. Here
Openare the
your steps.version 21. Close the initial dialog box to show a
SPSS,
blank SPSS Data Editor. (Note that earlier versions can be used, but
some of the features may be different)
Step 2. Click File, Open, Data.

Step 3. Look in the data where it is saved (i.e. , desktop, flash drive or
local disc D). The file folder’s name is DATA SET FOR LECTURES.
Specify files of type: Excel (*xls, *xlsx, *xlsm). Click DATA SET
(descriptive statistics)xls.
.
 Step 4. Click Open, then OK. (If the excel file to be used contains multiple spread
sheets, select the spread sheet to be analyzed before clicking OK).

We have successfully imported the excel file to the SPSS program, and is
now ready for analysis. There are so many features of the SPSS program that are
very interesting to learn, but we will only tackle the ones that we need in as far as
our objectives are concerned.
HOW TO DETERMINE FREQUENCY COUNTS AND PERCENTAGES OF
CATEGORICAL DATA (NOMINAL AND ORDINAL) USING SPSS.

Step 1. With the excel file already imported to the SPSS, Click Analyze,
Descriptive Statistics, Frequencies.

Step 2. Choose the variables to be analyzed and put them inside the Variable(s)
by using the arrow pointing to the right. You can choose the variables one at
a time
. or simultaneously, and you can use the arrow pointing to the left if you
like to change or replace the variable). Click Statistics, then Continue. You can
also Click Charts and specify whether you like a bar graph or a histogram for
graphical representation.
 Step 3. Click OK, and the result of the analysis is shown below.

The output shows that the Valid N for all the three variables (Gender, Type
of School and How much do you like schooling in general) is 90, and 0 Missing
data. In other words, the data set is complete. The frequency table shows that out
of 90 respondents, 42 were males (coded as 1), and that is 46.7%. There are 48
females (coded as 2) comprising 53.3%. Similarly, 38 or 42.2% are enrolled in
public schools (coded as 1), while the remaining 52 or 57.8% are enrolled in private
schools (coded as 2). In terms of how much they like schooling, 6 (6.7%)
responded “very much” (coded as 4), while 38 (42.2%) were neutral.
With SPSS, you can practically count all categorical variables (nominal and
ordinal) simultaneously and easily even for a very large data set (i.e. n=2, 000).
CROSS TABULATIONS IN SPSS

It is also possible to determine the frequency counts and percentages of

samples for a specific variable across the levels of another variable/s.

Example 1. We might be interested to know how many male students are enrolled
in private school, or how many females are enrolled in public school.

Step 1. With the excel file open, Click Analyze, Descriptive Statistics, Crosstabs
.

Step 2. Put one of the variables in the Row(s) box, and the other variable in the
Column(s) box.
Step 3. Click Cells, and Row and Column Percentage (to express frequency
counts as percentages.)

Step 4. Click OK, and the result of the analysis is shown below.
 The result of cross tabulation shows that out of the 42 males (Gender 1) ,
20 or 47.6 are enrolled in public school (Type of School 1) , while 22 or 52.4 % are
enrolled in private school (Type of School 2) . We can also describe it in terms of
type of school. Out of the 38 students who are enrolled in public school, 18 or 47.4
are females while there are 20 or 52.6% males. Practically, you can also perform
cross tabulations easily to any categorical variables, even for large data sets.

Example 2. How many students enrolled in public school responded “ neutral” in

the question, “How much do you like schooling in general?

Perform the same procedure. Put type of school in the Row(s) and the How much
do you like schooling in Column(s) box. The result is:

The cross tabulation shows that out of the 38 students enrolled in public
school (School 1) , 14 or 36.8% responded “neutral”, while 24 out of 52 (46.2%)
from private schools (School 2) responded “neutral”. Moreover, of the 6 students
who responded “ very much”, 5 or 83.3% were from private school.
HOW TO OBTAIN MEAN, MEDIAN, MODE, RANGE, STANDARD DEVIATION
AND VARIANCE OF CONTINUOUS DATA (INTERVAL OR RATIO) USING
SPSS.

Quick Review !!

Measures of Central Tendency

1. Mean is the arithmetic average of the set of scores
2. Median is the middle score separating the top half from the bottom
half of the scores
3. Mode is the most frequently occuring score in the distribution

Measures of Variability
1. Range is the score distance between the highest and the lowest value
2. Standard deviation is the average distance of all the scores from the
Mean
3. Variance is the squared deviation.

Step 1. With the excel file open, Click Analyze, Descriptive Statistics,
Frequencies.
Select the variables to be analyzed and put them in the Variable(s) box.
Step 2. Click Statistics. Check all statistics that you want to compute.
(Aside from measures of central tendency and dispersion, measures of
distribution like skewness can also be calculated)
.

Step 3. Click Continue, then OK and the result is shown below.

The result shows the following:

1. Valid N for English test, Math test and Science test is 90, the data is complete
2. The Mean for English test = 76.22, Math test = 79.16, and Science test =
74.60
3. The Median for English test = 76, Math test = 78, and Science test = 76
4. The Mode for English test = 76, Math test = 83, and Science test = 76. The
English test is polymodal (multiple modes).
5. The Highest score for English test = 90, the Lowest score = 56, Range = 34
The Highest score for English test = 88, the Lowest score = 66, Range = 22
The Highest score for English test = 89, the Lowest score = 56, Range = 33
6. The Standard deviation (s) English test = 7.857, variance (s2) = 61.725
The Standard deviation (s) Math test = 6.135, variance (s2) = 37.638
The Standard deviation (s) Science test = 8.544, variance (s2) = 73.007

7. We can compute for the coefficient of variations of the three scores:

Review: Coefficient of variation is a measure that compares the variability of two

or more sets of data.
𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛
CV = x 100%
𝑀𝑒𝑎𝑛

7.857
a. CV (English test) = x 100% = 10.31%
76.22

6.135
b. CV (Math test) = x 100% = 7.75%
79.16

8.544
c. CV (Science test) = x 100% = 11.45%
74.60

8. Skewness

The normal distribution, represented by the normal curve represents

symmetry and that the measures of central tendency (mean, median and mode)
are the same. However, if these three—mean, median and mode are affected by
lack of symmetry, skewness in the data occurs.
 Typically, for standardized test the curve very closely approximates a
normal distribution. However if the distribution is positively skewed, most of the
scores piled up at the lower end and there are just few high scores. For a
negatively skewed distribution it is just the opposite- most of the scores are high
with few low scores..

Negatively Skewed Normal (no skew) Positively Skewed

Negative Direction Normal curve represents Positive Direction

perfectly symmetrical
distribution.

In the example, all scores are negatively skewed (English = -.257, Math =-
.800 and Science .108). It means that more scores are high and few are low,
especially in Math. If the data is perfectly normal/symmetrical, skewness is zero,
but it is almost impossible in real life situations. In research, normality is assumed
if skewness is -1.0<x<1.0. Thus, the three scores still meet the assumption of
normality.
HOW TO GET THE DESCRIPTIVE STATISTICS OF GROUPS WITHIN A
VARIABLE

Example 1: In our previous result, the average score for the English test is 76.22.
What are the descriptive statistics (i.e. means, and standard deviation in English
test of males and females separately?
Step 1. With the excel file open, Click Analyze, Compare Means, Means.

Step 2. Put English test in the Dependent List and Gender in the Independent List
Step 3. Click Options. The Default Cell Statistics are Mean, Number of Cases and
Standard Deviation. You can include other statistics such as Median, Minimum,
Maximum, Variance, Kurtosis and Skewness using the arrow key.

Step 4. Click Continue, then OK and the result is shown below.

 The output shows the descriptive statistics of Gender 1 (male) and Gender
2 (female) in the English test. The mean score of the 42 males is 76.24 with the
standard deviation of 7.77. The median score is 78, the highest score is 89 and
the lowest is 56. Similarly, the mean score of the 48 females is 76.21 with the
standard deviation of 8.02. The output also shows that the scores of the males
are more skewed (negatively) than the score of the females. The overall mean of
the 90 students in the English test is 76.22, the same value we obtained in the
earlier example.

Example 2. What are the descriptive statistics in Math scores according to type of
school?

Perform the same procedure. Put Math test in the Dependent List and Type of
School in the Independent List. The result is:
 The table shows that the two groups have the same minimum (66) and
maximum (88) scores. The overall mean score of the 90 students in 79.16, the
same value we obtained in the earlier example. The output shows that the mean
score in the Math test of the 38 students enrolled in public school(Type of School
1) is 78.47 with a standard deviation of 6.26 while the mean score of the 52
students enrolled private school(Type of School 2) is 79.65 with a standard
deviation of 6.06. With these means scores and standard deviations in Math test,
can we say that students enrolled in private school performed better than those
enrolled in public school? This question will be discussed thoroughly in the
succeeding module (i.e. mean comparisons).

HOW TO GET THE DESCRIPTIVE STATISTICS OF A VARIABLE BASED ON

CROSS TABULATIONS

In the earlier examples, we obtain the frequency counts and percentages of

cross tabulations (i.e. gender and type of school). Suppose we are interested to
determine the relevant statistics (i.e. means and standard deviations) of these
cross tabulations in a given variable.

Example 1. In the previous example, the mean score of male students in the
English test is 76.24, and SD is 7.77. What are the descriptive statistics of male
students who are enrolled in public school only? In the cross tabulation performed
earlier, 20 or 47.6 % of the 90 respondents belong to this group (male-public).
 Step 1. With the excel file open, Click Analyze, Compare Means, Means

Step 2. Put English test in the Dependent List and Gender in the Independent List.
Then Click Next. Put Type of School in the Independent List. Note: You can continue
adding another independent variable to determine the cross tabulations and means
( i.e. What is the mean score in English of male students enrolled in public school
(next) who responded “neutral”?)
Step 3. Click OK and we have the result.

The output shows that the mean score of all male students (n= 42) is 76.42
and SD is 7.77, the same values we obtained earlier. The mean score of 20 males
(Gender 1) who are enrolled in public school (Type of School 1) is 77.55 and SD
is 7.65. Similarly, the mean score of the 30 females (Gender 2) who are enrolled
in private school (Type of School 2) is 75.47 and SD is 8.92.
Example 2. What are the descriptive statistics in Math test based on the cross
tabulation of gender and type of school?

Perform the same procedure. The Dependent Variable is Math test. The result is:

The output shows that the mean score in Math test of male students
enrolled in public school (n=20) is 79.70 and SD is 4.68. Similarly, the mean score
of female students enrolled in private school (n=30) is 80.37 and SD is 5.55.