Calculating Statistics using Excel

This document was put together following the third edition of Quantitative Research Methods for Communication:
A Hands-On Approach by Wrench, Thomas-Maddox, Richmond, and McCroskey. In the 2nd and 3rd editions, we
discussed calculating statistics using Microsoft Excel. Starting in the 4th edition, we’ve moved more information to
the online website (https://oup-arc.com/wrench) including numerous walk-through videos for running the statistics
discussed in our book. We wanted to create this special addendum to include the step-by-step walk throughs of
using Excel on its own to calculate statistics (as was seen in the 2nd and 3rd editions). Starting in the 4th edition,
we’ve recommended using Statistician for Excel (http://statsaddin.com/) because it makes calculating statistics in
Excel easier. However, as this document demonstrates, it’s still very possible to calculate a wide range of statistics
using Excel on its own.

EXCEL AND CRONBACH’S ALPHA

Unfortunately, there is not an easy way to calculate Cronbach’s alpha in Excel. As such, you must use a variety of
formulas for the calculations of alpha. Thankfully, a psychometrician at the University of Leicester in the United
Kingdom named Professor Glenn Fulcher has created an easy-to-use Excel template to conduct an alpha reliability
(http://languagetesting.info/statistics/excel.html/). When you download the alpha reliability file, it will open with
all of the calculations in place, but there will be no actual data within the file yet. The Excel file is built for up to
200 scale items and 3,005 individual cases (Figure Excel.1). Note that there are two tabs to this Excel file. The one
currently selected in Figure Excel.1 is where the data will be pasted. The arrow in Figure Excel.1 is pointing to the
tab where you can see the alpha reliability after you have pasted the data.

Figure Excel.2
Select Data in Recoded Dataset
Figure Excel.1
Fulcher’s Excel File To paste the data, find the data you want to include in the alpha reliability
test in the Recoded Dataset (Figure Excel.2). Note that you do not select
the variable names at the top of each column. Then go to copy function and click “copy” (note the pointer in
Figure Excel.2). Once you have copied the data, you can then go to the Alpha Reliability Excel File and paste that
1
copied data into the file. Select the square located at B2 and paste the data (Figure Excel.3). Once you have pasted
the data, just click on the “Results” worksheet at the bottom of the screen and you can view the alpha reliability
calculated (Figure Excel.4). Please note that it is using a slightly different formula for calculating alpha reliability
than the one used by SPSS, which has resulted in a slightly elevated alpha reliability. To learn more about the specific
calculations utilized in this alpha reliability test, check out Fulcher’s discussion on conducting alpha reliability in
Fulcher and Davidson (2007) or Fulcher (2010).

Figure Excel.3 Figure 8.4

Pasted Scale Items into Excel Excel Alpha Reliability Results

EXCEL AND FREQUENCY DISTRIBUTIONS

At this point, we will examine how to calculate
a frequency chart using Excel as well. Given
that Microsoft Excel is readily available to most
students, we will include information on using
this program to calculate statistics. Although
Excel may be a convenient program for analysis,
keep in mind that the primary purpose for the
program is not statistical analysis. Thus, the
process for computing even basic statistics
Figure Excel.5
often involves more steps than in other software
Find and Select
programs, and the options for post hoc analyses
are often more limited than in programs such as SPSS.
Before you begin using Excel, consider the following caveats. First, missing data must be entered as a blank
cell in Excel. The program is unable to read any other format for missing data. If you are importing a database that
automatically fills an empty cell with data, utilize the “Find and Select” feature to convert all missing data (see Figure
Excel.5). The imported “Recoded Dataset” file may include several cells with “#NULL!” to represent missing
2
 values. To convert the data, click on the “Replace” tab at the
top of the screen, type the value that appears in the cells for
missing data in the “Find What” box, and leave the “Replace
with” box blank. Click on “Replace All” to convert all missing
data to empty cells (Figure Excel.6).
Second, the steps for calculating statistics may differ
depending on the version of Microsoft Excel that you are
using (e.g., Excel 2007 vs. Excel 2010). In addition, the data
Figure Excel.6
analysis process for statistical computation varies depending on
Replacing #Null!

whether you are using a Mac versus

a PC for data analysis. Finally, Excel
does not come preinstalled with the
data analysis tools that are required
to complete statistical analyses.
However, this is easily remedied
with a free download. To determine
whether your computer has the data
analysis tool kit installed, click on
“Help” and type “Data Analysis.”
If you are using a PC, the Data
Analysis ToolPak is available from Figure Excel.7
the Microsoft website (http://www. Excel Data Menu
office.microsoft.com/). The Mac
platform will require you to download a program called “StatPlus”(http://www.analystsoft.com/). In this text, we
will describe the procedures for calculating statistics using the PC version of Excel 2010.
Once you have
downloaded the data analysis
package to Excel, you will see a
tab on the top of the spreadsheet
screen titled “Data.” Clicking
will provide access to an area
labeled “Analysis.” This area will
enable you to calculate a variety
of basic statistics in Excel (Figure
Excel.7).
To calculate frequencies
in Excel, open the data file titled
“Recoded Dataset” from the
textbook site. Remember—if
the dataset contains any missing
data, you will need to utilize the
“Find and Select” function to
Figure Excel.8 convert any data that are used
Pivot Table to represent missing values into
blank cells. To be able to easily
view your data and output, it is recommended that you select and work with one column or variable at a time. To
compute the frequency counts for each value, Excel provides the PivotTable and PivotChart options. Begin by
highlighting the “Sex” column in the spreadsheet and clicking on the “Insert” tab at the top of the screen. Click
3
 on “PivotTable” at the left side
of the screen, and the range of
values that is highlighted should
appear in the box. Check the
box for “New Worksheet” and
select “OK” (Figure Excel.8). A
new sheet will open, and you will
be asked to indicate the fields
that you wish to add to your
report. Drag and drop “Age”
from the list into both the “Row
Labels” and the “Values” boxes.
A frequency table with counts
for each age reported will appear
on the screen. If you would like
to add a histogram, select the
“PivotChart” option and follow
Figure Excel.9
the same process by dragging
Frequency Distribution in Excel
“Sex” into the “Axis Field” and

“Values” boxes. A frequency table

and histogram will appear on the new sheet (Figure Excel.9).

Excel and Descriptive

Statistics
To calculate the descriptive
statistics for the CA scores for
participant, open the “Recoded
Dataset” file. Select the ”Big CA”
field and highlight all values in
the column. Click on the “Data”
tab at the top of the sheet, select
“Data Analysis” from the upper
right side of the screen, and
choose “Descriptive Statistics”
from the menu of options.
(Figure Excel.10) The range of
highlighted values should appear
in the “Input Range” box. Select
“New Worksheet Ply” to create a
Figure Excel.10
new sheet for the output. Select
Excel’s Descriptive Statistics Dialogue Box
the “Summary Statistics” option
and indicate the “Confidence
Level for the Mean” in the box provided (Figure Excel.11). Click “OK” and the output will appear on a new sheet
(Figure Excel.12).

4
Figure Excel.11
Means in Excel

Figure Excel.12
Descriptive Statistics Calculated by Excel

5
EXCEL AND CHI-SQUARES
To calculate the chi-square using Excel, select the Excel file titled “Chi-square” from the textbook’s website. Click
the “Insert” file tab at the top of
the screen and choose the “Pivot-
Table” function (Figure Excel.13).
The data for both the “sex” and
the “class” variables should be
highlighted. The range for the data
will appear in the “Table/Range”
box. Select “New Worksheet” from
the menu to create a new page for
the output (Figure Excel.14). Drop
and drag “sex” to the section titled
“Row Labels” and drag “class” to
the box titled “Column Labels.”
The sum of scores for “sex” should
automatically appear in the “Values”
section (Figure Excel.15). Click on
the “X” to close out the dialogue Figure Excel.13
box. The output table includes the Pivot Table
total number of responses for each
category we identified (e.g., males in
the freshman class). Be careful, Excel
often does a great job of bringing
in the counts for the first row but
not for the second row. As such, you
may find yourself with a percentage
value and not a count value. Look at
Figure Excel.16. You will note that
the first number in the second row is
14 when it should have been

7. To change this, right click on the

first number and scroll down to
“Summarize Values By.” The first
one, which will be checked, is Sum.
You want to uncheck Sum and check
Count (Figure Excel.17). This will
correct that row to counts instead of
summed totals.
Computing the chi-square
Figure Excel.14
statistic involves a couple of steps.
New Pivot Table
First, copy all of the information
from the PivotTable. Click on an
empty cell below the table, right click, choose “Paste Special,” and choose the “Paste Values” option. This will
duplicate the table directly below the original PivotTable (Figure Excel.18). Doing so will enable you to keep the
data separate while you compute the expected values versus the observed values. In the duplicate table of values
that you created, highlight and delete the values for each of the cells. Be careful not to delete any of the totals from
6
 the table.
To compute the expected values into the table,
click on the first empty cell and type the formula
“=column total value * row total value/grand total value”
(Figure Excel.19). Repeat these steps for each empty cell
in the table until you have computed the expected values
for each cell (Figure Excel.20). Finally, select an empty cell
below the two tables and click. In the cell, type the Excel
formula for a chi-square analysis, which is =CHITEST
(actual values, expected values). To enter the range of
values, simply highlight the block of expected values from
the top table and click Enter (Figure Excel.21). Complete
the same steps for the expected values in the table below.
Click “Enter,” and the probability for the chi-square value
will appear in the designated cell (Figure Excel.2).
Please note that the actual chi-square statistic
is not actually calculated by Microsoft Excel. Instead,
Figure Excel.15 Excel calculates your p value, so you would still need to
Pivot Table Field List Dialogue Box calculate the chi-square statistic by hand. Now that we
have examined the chi-square test itself, let’s see how a
researcher would write up this finding using APA style in a results section of an article.

Figure Excel.16
Summed Total Problem

7
 Figure Excel.17
Summarize Values By

Figure Excel.18 Figure Excel.19

Paste as Values Computing Expected Values

8
 Figure Excel.20
Computed Expected Values

Figure Excel.21
Entering Actual Frequencies Figure Excel.22
Chi-square Test Results

9
EXCEL AND t TESTS
To compare the means for independent samples in
Excel, first open the “t-Test” file from the textbook’s
website. Once the file is opened, you will see data for
the two dress groups. As we explained earlier, Excel
uses the numeric representations of 1 and 2 for each
of these groups (Class A and Class B). Click on the
“Data” tab at the top of the spreadsheet, select “Data
Analysis” from the right side of the screen, and choose
the “t-Test: Two Sample Assuming Equal Variances”
option from the menu (Figure Excel.23). To fill in the
data ranges under the Input category, click the box
labeled “Variable 1 Range” and highlight the values in
cells B2 through B6 because this range corresponds
with those in the first group. Next, click in the box
titled “Variable 2 Range” and highlight the values in
Figure Excel.23
cells B7 through B11 because these scores correspond
t Test: Two Sample Assuming Equal Variances
with those in the second group. Click the “Labels” box
and the
“New Worksheet Ply” box. A default of 0.05 is set for alpha, but
this can be adjusted if desired. Click “OK” and the data output will
appear in a new screen (Figure Excel.24). In addition, the z-Test
option is provided in the menu of options.
Please note that Excel does not provide you with the
Levene test for equality of variances, so there is no way to ascertain
whether this assumption of the t test is violated. You can repeat
the test and run it where you assume that equal variances has been
violated, but without Levene’s test for equality of variances, it is
hard to know whether the assumption has been violated.

APA WRITE-UP (EXCEL)

A paired two sample for means t test was conducted to determine
whether an instructor’s dress in a classroom (fashionable [M
= 90.20, SD = 4.82] vs. out-of-date dress [M = 72.80, SD =
7.73]) affected the scores students received on their midterm
examinations. Overall, no significant difference was noted between
Figure Excel.24 the two groups, t(8) = 2.31, p = 0.06.
Computed t Test in Excel

EXCEL AND ONE-WAY ANOVAS

To compute the one-way ANOVA in Excel, begin by locating the “One-Way ANOVA” spreadsheet on the
textbook’s website. After opening the file, you will see two columns of data. The “color” column represents the
three room colors (1 = white, 2 = yellow, and 3 = blue). The column labeled “time” represents the number of
minutes required to complete a 10-item quiz. First, you must separate out the time it took to take the test into three
distinct columns. Excel needs the three groups separated into distinct columns (Figure Excel.25). Next, click on the
“Data” file tab at the top of the spreadsheet, select “Data Analysis” from the right side of the screen, and choose
“Anova: Single Factor” from the drop-down menu (Figure Excel.26). Highlight all the data included in cells D1
10
through F5. The data range will appear in the box labeled “Input Range.” Be sure the “Columns,” “Labels in First
Row,” and “New Worksheet Ply” options are all selected (Figure 18.15). Excel sets the Alpha default at 0.05, but this
can be adjusted if you choose to do so. Click “OK” to proceed to the output worksheet (Figure Excel.27).

Figure Excel.25
Select ANOVA: Single Factor

Figure Excel.26
“ANOVA: Single Factor” Dialogue Box

Figure Excel.27
ANOVA Summary Table

11
EXCEL AND PEARSON PRODUCT-MOMENT CORRELATIONS
To calculate a correlation using Excel, open the file titled
“correlation” in the Excel data file on the textbook’s website. In
this file, data for two variables will be listed—“CA” and “hrchange.”
“CA” represents the CA scores for each respondent, and
“hrchange” reflects the change in heart rate when the respondent
is asked to give an impromptu speech. The process for calculating
the correlation is similar to other Excel statistical analyses. Click on
the “Data” tab at the top of the spreadsheet, select “Data Analysis,”
and select “Correlation” from the menu. Click on “OK” and you
will be prompted to enter the values in the “Input Range” box.
Click on the box and highlight all of the data in cells A1 through
B21—this includes the variable labels included in the first row of
the spreadsheet. Check the options for “Labels in First Row” and
“New Worksheet Ply” before clicking “OK” (Figure Excel.28). The
data output will appear in a new sheet (Figure Excel.28). With the
Excel file, you will have to look at the critical value table in (SEE Figure Excel.28
BOOK) to determine whether the test is statistically significant. Excel Correlation Dialogue Box
Again, we have bolded the significant correlation on the
results to make it stand out more easily. Excel correlation tables can become somewhat confusing because they are
not placed in a nice chart format like in SPSS. Although you may not
notice this problem here, in a few minutes when we look at a larger
example you will see it more clearly. Again, to make these tables easier
to understand, we recommend finding the diagonal lines of 1.0 and
marking a line through them with a highlighter to make it easier to view.

Figure Excel.29
Excel Correlation Output

EXCEL AND THE SIMPLE LINEAR REGRESSION

To compute a regression in Excel, begin by opening the “Regression” data file on the textbook’s website. In this
spreadsheet, you will see two variables, “CA” and “hrchange”. These are the same variables used in our calculation
of correlations from the last chapter. Recall that “CA” represents an individual’s scores on the PRCA-24, and
“hrchange” reflects the change in an individual’s heart rate while giving an impromptu speech. To begin, select
the “Data” tab at the top of the spreadsheet, click on “Data Analysis” at the right side of the screen, and select
“Regression” from the drop-down menu (Figure Excel.30). Next, click on the “Input Y Range” box and highlight
the data in cells A1 through A21. Select “Input X Range” and highlight the data in B1 through B21. Excel sets a
default confidence interval of 95 percent for the regression statistic. Click on “New Worksheet Ply” and “Labels”
before selecting “OK” (Figure Excel.31). The output data will appear in a new sheet (Figure Excel.32).

12
Figure Excel.30
Select Regression from the Data Analysis Dialogue Box

Figure Excel.31
Regression Dialogue Box

Figure Excel.32
Excel Regression Output
13
14