Psychology 241
Lab Assignment: Statistics (t-tests)
Instructions
This lab exercise will introduce you to the use of SPSS for data analysis
with quantitative dependent variables. Follow the instructions below to perform
the analyses, and as you do so, fill in the accompanying Answer Sheet. Turn in your answer
sheet at the end of the lab session to receive credit for the lab.
The data set you will use for this assignment is available from the
following link. Right-click on it and save it to your computer's
desktop. You will start SPSS and then import the data file to SPSS.
DATA FILE: lab3-data-ttests.txt
The data file contains 40 rows and 5 columns of numbers, plus one row of
variable labels in the first row. These are hypothetical (made up) data for an
experiment on balance. In the hypothetical experiment, male and female
participants balanced on one foot for as long as they could, first on the right
foot and then on the left. Participants assigned to one condition had to count
backwards while balancing. Particpants in the other
condition did not have to count backwards. The experimenter measured how long
(in minutes) the particpant could balance on one
foot. The data file is described by the following coding guide:
Column variable possible values meaning
1 ID 1-40 subject number
2 count 1, 2 1 = counting backwards; 2 = not counting
3 sex 1, 2 1 = male, 2 = female
4 right >= 0 number of minutes balanced on right foot
5 left >= 0 number of minutes balanced on left foot
Note regarding importing data files: SPSS can also import Excel
spreadsheets. You may want to use a spreadsheet as your data file rather than
creating a text file like the one used in this lab exercise. Either way, it is
convenient to use the first line of your data file to list the variable names.
If your data file is a text (.txt) file, SPSS gets fussy if you have more than
one space between the variable names. Excel files seem to import without
problems, however. Here is the path I would suggest for creating a data file
for your project:
personals ads -> content analysis
protocol -> data summary sheet -> coding guide -> Excel spreadsheet
(data file) -> import to SPSS
In other words, use the content analysis protocol to assign values to the
variables in your data summary sheet based on the content of the ads, use the
coding guide to tell you what values to put in the spreadsheet, use the
spreadsheet as your data file, and import it to SPSS for analysis.
Assignment for this lab session:
1) After saving the data file to your local hard drive, import it to SPSS as
a data file by following these steps:
- Start SPSS from the Start
Menu
- If the opening dialog
appears asking "What would you like to do? " select Open an
existing data source or Open another type of file and click OK.
If not, go to the File menu and select Open -> data
- In the dialog box that
appears, for Files of type, select Text (*.txt)
. Then find the file
"lab4-data.txt," click it, and click Open. A text import
wizard should then appear.
- In the text import wizard, give
the following answers on each screen:
- Does your text file match
a predefined format? Click No. Click Next.
- How are your variables
arranged? Click Delimited.
- Are variable names
included at the top of your file? Click Yes. Click Next.
- Click Next
for each succeeding page, then click Finish. You should now see the
data file in the "Data Editor" window of SPSS.
2) Create a frequency table of all of the variables in the data file, in
order to check for invalid data. Use the following steps:
- Analyze -> Descriptive
Statistics -> Frequencies (A dialog box labeled
"Frequencies" should then appear.
- Select all 5 variables in
the left window and move them to the Variable(s): window using the
arrow key in the center of the dialog box
- Display frequency tables
should be checked. If it is, click OK.
- The output window should
then appear. The first chart should be labeled "Statistics."
Check the number of valid and missing cases. There should be 40 valid
cases for each variable and no missing cases. Answer question 2a on the
Answer Sheet.
- Below the first table, you
should see a table for each of the variables. The first column in each
table contains the values that variable takes in the data set. Look
through the list of values and check to see if there are any invalid
values (values that are not in the range specified in the coding guide).
Answer question 2b.
- The second column of each
table gives the number of times each value occurred for that variable
(frequency). For the variables "count" and "sex" there
should be an equal number of each condition (the values "1" and
"2" should occur an equal number of times). Check the tables to
see whether they do. If the frequencies are unequal, this would mean that
there was an error in either our coding of the data or in our design of
the experiment. Answer question 2c.
3) Use stem plots (stem and leaf plots) to check for outliers (outlying data
points) for each of the variables "right" and "left." SPSS
will identify any points it decides are outliers and label them as "Extremes"
in the stem plot. Follow these steps to produce stem plots:
- Analyze -> Descriptive
Statistics -> Explore
- Move the variables
"right" and "left" to the Dependent List.
- Under Display, select Plots.
- On the lower right of the
"Explore" dialog box, click the rectangular Plots button.
- In the "Explore:
Plots" dialog box that appears, under Boxplots
select None. Under Descriptive
select Stem-and-leaf. Click Continue.
- Click OK.
- Scroll down the SPSS Output
Viewer window to see the stem plot for each of the two variables.
- If there are any outliers,
they will be labeled as "Extremes." Answer question 3 on the
Answer Sheet.
4) Create a table of means comparing males vs. females on each of the
variables "right" and "left." Use the following procedure:
- analyze -> Tables
-> General Tables
- Move the variables
"right" and "left" into the Rows box.
"Right" should be above "left" in the resulting list
in the box.
- Click on right and
then under Selected Variable click Is summarized. Then
do the same for left.
- Move "count" and
"sex" into the Columns box. "Count" should be
above "sex" in the box.
- Click on sex. Click Nest.
- Click OK. You should
see a table with the first two rows containing labels for the levels of
"count" (1 and 2, in the first row of the heading for the table)
and for "sex" (1, 2, 1, 2, in the second row of the table
heading). In the body of the table, you should see two rows of means: one
for "right" (the top row) and one for "left" (the
bottom row).
- Note that the table has no
variable labels. You must write down the variables you use as you create
the table so that you will be able to label the rows and columns
correctly. Answer question 4.
- The table in the output
window can be cut-and-pasted into a word processing document. You may want
to practice doing that now so that you will know how when you do your
final project.
5) Test the hypothesis that counting backwards makes it harder to balance on
the right foot. To do this, you will need to perform and independent samples
t-test. An independent samples test is used when each subject was in only one
condition - when you have a separate control group and experimental group. In a
repeated measures (or within-subjects design) on the other hand, each subject
participates in both the control condition and the experimental condition. In
that case, you would need to use a paired-samples t-test. Follow these steps to
perform an independent samples t-test:
- Analyze -> Compare
Means -> Independent-Samples T Test
- Move the variable
"right" into the Test Variable(s) box.
- Move the variable
"count" into the Grouping Variable box.
- Click Define Groups
- For Group 1 enter 1
in the box; for Group 2 enter 2. Click Continue.
- Click OK.
- In the output window, you
should see the title "T-Test" followed by two tables. The first
table gives the means for each condition (among other things). The second
table gives the results of the t-test. The fourth column (labeled
"t") gives the value of t, the fifth column gives the
degrees of freedom (labeled "df"), and
the sixth column gives the p value (labeled "Sig.
(2-tailed)").
- The results of a significant
t-test are reported like this: t (df)
= x.xx, p < .05. (note that df is the number
representing the degrees of freedom, and x.xx is
the t score to the hundredths place.)
- The results of a
non-significant t-test (when p > .05) are reported like this: t (df) = x.xx,
n. s.
- Answer questions 5a and 5b.
6) The variables "right" and "left" actually measure the
same thing (time balanced on one foot) in two different conditions (left vs.
right foot). Test the hypothesis that which foot
someone is standing on makes a difference in how long they can balance on one
foot. Because the DV in this hypothesis (time balanced) is measured twice for
each participant, we call this a repeated measure (or within-subjects) design.
Another way to tell that it is a within-subjects design is because each subject
gets all levels of the IV (in this case, the IV is "which
foot" and the two levels are "right" and "left"). For
a within-subjects design, you must use a Paired Samples t-test to compare the
means of the two conditions. Follow these steps to test the hypothesis with a
Paired Samples t-test:
- Analyze -> Compare
Means -> Paired-Samples T Test
- Move "right" and
"left" into the Paired Variables box
- Click OK.
- Answer question 6.
7) Is there any relationship between how long someone balances on the right
foot and how long they balance on the left foot? In other words, are the
variables "left" and "right" correlated? Follow these steps
to calculate the correlation coefficient for these two variables:
- Analyze -> Correlate
-> Bivariate
- Move the variables
"right" and "left" into the Variables box.
- Select Pearson under
Correlation Coefficients, then click
OK.
- Find the "Pearson
Correlation" between "right" and "left" in the
table in the SPSS output.
- Answer questions 7a, 7b,
and 7c.
8) Create a scatterplot for the variables
"right" and "left."
- Graphs -> Scatter
- Simple, Define
- Y Axis = left
- X Axis = right
- OK
- Can you see evidence of a
correlation in the scatter plot? Can you see any outliers?
- Answer question 8