selecting variables you will use for the 3 major test categories based on the correct levels of measurement for variables that are required for each test

1. You will want to start by selecting variables you will use for the 3 major test categories based on the
correct levels of measurement for variables that are required for each test
a. T-test: IV = nominal/ordinal with only 2 groups/levels; DV = interval/ratio
b. ANOVA: IV= nominal/ordinal with 3 or more groups; DV = interval/ratio
c. Correlation: both variables must be measured at the interval/ratio level
i. Hint: can’t use nominal variables for correlations
d. Linear regression = both variables must be measured at interval/ratio level
i. Because you didn’t learn ‘dummy coding’ you will just use I/R level
2. Go to ‘variable’ view in the dataset and look at the ‘measure’ column to identify level of measurement
for each variable. 3 options:
a. Nominal
b. Ordinal
c. Scale (which is the label for interval and ratio variables)
3. Make a list of the variables you want to examine for each pair of tests for the 3 major categories. So
final list should look something like:
a. T-test #1: IV: DV:
b. T-test #2: IV: DV:
c. ANOVA #1: IV: DV:
d. ANOVA #2: IV: DV:
e. Corr #1: Var 1: Var 2: (remember there is no IV/DV for r tests)
f. Corr #2: Var 1: Var 2:
g. Lin Reg #1: IV: DV:
h. Lin Reg #2 IV: DV:
i. **You can experiment and run lots of these tests beforehand but ultimately you will report on
a list like this.
4. Once you have decided on the variables you will use for your tests, you can then go and run your
DESCRIPTIVES tests on those variables that you use.
a. You can use some variables for multiple tests as long as they are the correct level of
measurement. So, for example you might use the same variable as the DV for a t-test, for one
of the anovas, and use it for a correlation test since it is an interval or ratio level variable.
b. If you are unsure, email me your list of variables and I will give you feedback before you
do your write up
5. Descriptives write up
a. Once you run the descriptives on the variables you use for the 4 major tests you will write up
your descriptives paragraph, which will look like a ‘sample’ description in the method section
of a published study. This is like what you did for Lab 1.
b. For each variable you need to report the correct central tendency measure based on its level
of measurement and then also include a few variability statistics that you think will help
describe the sample.
i. For example, if you used AGE as a variable (ratio level) you would report the mean or
median (depending on skewness), and then maybe include the range
ii. FOR NOMINAL VARIABLES: Always include the mode and give the %s for
each group/category in the variables when reporting central tendency
6. A note about running tests in SPSS:
Finding Your Variables
a. SPSS will run a test with whatever variables you include but it won’t tell you if they are not
the correct level of measurement.
i. This is why we need to know this info ahead of time