Lab Assignment 2: Correlations with SPSS!!!
Generally, a "correlation" is an analysis computed to examine the relationship between two variables. The most common kind of correlation is the Pearson Product correlation which examines linear relationships between two continuous variables (e.g., years of education and income).
When SPSS computes a correlation, it ultimately provides two important statistics.
First, it provides r which is a "correlation coefficient." This number varies between 1 and -1. Numbers near 1 indicate a strong positive correlation (e.g., the correlation between height and weight). Numbers near -1 indicate a strong negative correlation (e.g., the correlation between years of education and years of prison time). Numbers near 0 indicate weak, negligible relationships (e.g., the correlation between SAT scores and number of rubber bands owned).
Second, SPSS will provide you with a p value. This statistic refers to the probability that the obtained correlation (r) is due to chance alone. Traditionally, psychologists will say that a correlation is "significantly different from chance" if p is less than .05. Importantly, the p that is provided is for a "two-tailed significance test." This kind of test is for when you are looking for a relationship, but you have no hypothesis regarding the specific direction predicted." If you have a specific direction that is predicted for r ahead of time, then you are to divide the p value by 2. This new value is the p value for you.
Example: Suppose you hypothesize that age is negatively correlated with the number of phobias (i.e., specific irrational fears) people have; older people are predicted to have fewer phobias. To test this zany hypothesis, you measure (a) people's age in years and (b) the number of specific phobias people report. You collect data from 5 people. Your data would look about like:
Age Phobias
30.00 25.00
80.00 3.00
50.00 15.00
25.00 25.00
75.00 7.00
To see if age and number of phobias are, indeed, negatively correlated, you would need to conduct a correlation like so:
1. Click on "Analyze" on toolbar.
2. Click on "correlate"
3. drag to "bivariate"
4. variables would be "age" and "phobias"
5. Click paste
6. Go to the .sps file and highlight the relevant commands.
7. Click on "run."
SPSS will give you output that looks something like this:
- - Correlation Coefficients - -
AGE PHOBIAS
AGE 1.0000 -.9941
( 5) ( 5)
P= . P= .001
PHOBIAS -.9941 1.0000
( 5) ( 5)
P= .001 P= .
(Coefficient / (Cases) / 2-tailed Significance)
" . " is printed if a coefficient cannot be computed
Notice that your prediction is clearly a one-tailed hypothesis. Thus, divide p (.001) by 2 to come up with .0005 (whoa!). This number refers to the probability that your correlation was due to chance alone. 5 out of 10000 is pretty low odds!
Your report of the data would sound like:
A strong and significant negative correlation was observed between age and number of phobias (r(5) = -.99, p < .05). As people age, they seem to have significantly fewer phobias.
Note that the number "5" in the above parenthetical expression refers to the N, or the number of cases on which the correlation is based.
_______________________________________________________________
Assignment:
Part ONE: Correlations among variables in the made up data.
Find two continuous variables in the data that you think may be related to one another. Compute the correlation between the two variables. Report the correlation and briefly discuss what the result implies about the relationship between these two variables conceptually.
Hand in:
A. a one-page summary including the hypothesis, a brief description of the variables of interest, a report of the correlation, and implications regarding the nature of the relationship between these two variables.
B. a printout of the .spo (output file).
_____________________________________________________________
Part TWO: Correlations among variables in a naturalistic college setting.
For this part of the assignment, you will be collecting data by observing college students in a naturalistic setting. You may want to collect data from people walking the halls of humanities or checking out books in the library.
Before you collect data, think of two variables that you would be able to measure that would lend themselves to a correlational analysis. For instance, you could go the library and measure the correlation between the amount of time it takes someone to walk down a particular hallway and the number of books a person is carrying (fascinating data!).
Get in groups of 3 or 4.
Once you have come up with two CONTINUOUS variables that you could measure in such a setting, collect data from no less than 10 people (unobtrusively). Next, enter the data into an SPSS .sav file. Compute the correlation between the two variables. Report the correlation and briefly discuss what the result implies about the relationship between these two variables conceptually.
The Pearson r correlation is designed for examining relationships among continuous variables (e.g., age) not categorical variables (e.g., biological sex). Thus, note that you are required to collect data from continuous variables only for this assignment.
Hand in:
A. a one-page summary including the hypothesis, a brief description of the variables of interest, a report of the correlation, and implications regarding the nature of the relationship between these two variables.
B. a printout of the .spo (output file).