Measurement of an Attitude
This assignment asks you to use computers and employ either SPSS
Excel to explore evidence for reliability of measures. We will give you a
data set from which to work.
Here is a summary of the study from which these data were
Researchers attempted to determine if there really is anything to the
notion that fallacious reasoning is less persuasive than sound argument
(some wags claim that people are basically irrational [except for them,
of course] and that sound reasoning really does not matter in the
persuasion process). So a group of researchers took a speech and
developed examples of fallacies to represent informal fallacies in the
family of non sequitur (literally "not in sequence"). Two examples of
each fallacy form were used. In contrast, a message that deleted the
fallacious arguments was included. The messages were presented on
two topics to control for topic-subject interaction.
The hypothesis to be tested was:H1: Subjects exposed to persuasive messages including non sequitur
Subjects were presented with the messages and then asked
complete a set of four seven-point semantic differential-type scales:
wise-foolish; good-bad; positive-negative; beneficial-harmful. These
scales are presumed to measure attitudes toward the topic.
Identify the level of measurement for the measure of
change (you may wish to write your answer to this question your own
or you may wish to use the sheet at the bottom of this web page).
To determine whether the attitude measure is reliable, it is
to examine the evidence for it. To do so, we will compute coefficient
alpha. Analysis of data is slightly different depending on whether you
use SPSS or Excel to analyze your data. Instructions for each are
If you use SPSS to analyze your data, follow these
|If you have not already done so, go to the Data
Sets location and|
download the file attitude.sav under the list of SPSS Files.
|Start your SPSS program. From the "File" menu load
the data file|
"ATTITUDE.SAV" from the location where the you saved it after
|Click on the "Analyze" menu, and on subsequent menus
appear click on "Scale . . .," and "Reliability Analysis." In the
"Reliability Analysis" dialog box, click on the names of the
variables that compose the attitude scale: wise, good,
positive, benefici. These names appear in the left window
of the dialog box. Transfer each of these variables to the
window marked "Items:" by clicking on the arrow key in the
middle of the dialog box.
When you have completed this process the dialog box should
look as follows.
|Click on the "Statistics" button. The following
dialog box will|
To keep things simple,
we will ask for essential elements of
reliability analysis only. Check the boxes for all three
"Descriptives" as in the above box.
|Click on "Continue" and "OK" in the subsequent
dialog box. Your |
data analysis results will appear as output and you may save
these results as a file or simply interpret them directly.
What is the coefficient
alpha for the attitude measure
(you may wish to write your answer to this question your own
or you may wish to use the sheet at the bottom of this
Once reliable items have been isolated, it is
a simple task to add
the reliable items together to make a composite scale. In SPSS
this step is completed by using the "Compute" from the
If you use Excel to analyze your data, follow these
If you have not already done so, go to the Data Sets location and
|Start your Excel program. From the
"File" menu load the data file|
"attitude.xls" from the location where the you saved it after downloading.
NOTE: In Excel, there is no preprogrammed reliability subroutine.
Thus, you must use your own formula. There is more than one way
to use Excel functions, but we will use a rather direct method in this
illustration. In our example, the following steps are taken.
|The formula for coefficient alpha requires
that one have a variable|
that it is the sum of all the items on the measure. Thus, you
should add the scales that are found in columns E through H on the
spreadsheet. This step is accomplished by moving the cursor to the
nearest empty column (column X in this case) and clicking on the
first row to highlight the cell. In this cell, type "attitude," the label
for the composite scale. Highlight the second row and enter "=."
The formula field above the spreadsheet will be activated. Simply
enter "E2+F2+G2+H2" to identify the columns that are to be added.
Press "Enter" when done.
To copy this cell down the column,
click on the cell and then select
"Copy" from the "Edit" menu. When the cell is highlighted in this
fashion, hold down the left mouse button and drag down the column
until all critical rows are highlighted (through row 53 in this case).
|The formula for coefficient alpha is:|
To compute coefficient alpha, we must prepare a formula. For
our data, the following formular applies:
= ((4)/3) * (1-((VAR(E2:E53))+(VAR(F2:F53))+(VAR(G2:G53))
Note: Though the formula takes more than one line on this page,
it is entered as one long formula. Do not forget to start the
formula with an equals sign.
|To compute coefficient alpha, we need only
find a blank cell (for|
instance, cell Y2) and click on it to make it active. Then, the above
formula may be inserted.
What is the coefficient alpha for
the attitude measure (you may
Once reliable items have been isolated, it
is a simple task to add the
Measurement Comments for Computer Assignment
Go to "Measurement of an Attitude" found on the book
website for Chapter 4.
What is the level of measurement for the dependent measure? ___________________________
Of what items is the measure composed?
What is coefficient alpha for the scales identified as the dependent measure? ________________