The following exercises are designed to give you hands-on practice of the skills learned in this chapter.
Listed below are hypothetical
scores from a research study that employed three groups of
participants. The mean score in the three groups should be
compared using an analysis of variance. Using the formulas
provided on this
website or using
SPSS for Windows, compute the F-ratio for these data.
This will serve as a baseline for comparison in the following
exercises.
Group A | Group B | Group C |
104 | 91 | 129 |
121 | 107 | 103 |
93 | 99 | 117 |
73 | 73 | 95 |
109 | 86 | 114 |
(a) You are now to take the numbers above and add 10 to each
score in Group C and subtract 1 from each score in Group B.
Before you re-compute the analysis of variance with this new set
of data, ask yourself what type of variance you have changed
with this procedure (i.e., between-groups variance or
within-groups variance?). What effect should this have on the
F-ratio (i.e., should the F-ratio increase or
decrease? Will the differences between the groups be more or
less significant?)? Now, re-compute the F-ratio to see if
your ideas are correct.
(b) Now take the original data and modify the data using the
following procedure. Roll two dice, and determine the count
showing. Then flip a coin: If the coin is heads, subtract the
number showing on the dice from the score; if the coin is tails,
add the number showing to the score. Continue this process for
each and every score in all of the groups. Before you re-compute
the analysis of variance with this new set of data, ask yourself
what type of variance was affected by this manipulation (i.e.,
between-groups variance or within-groups variance?). What effect
should this manipulation have on the F-ratio? Now,
re-compute the analysis of variance with your revised data. Do
the results confirm your expectation?
(c) What do you think would happen if you took the original data
set and repeated the process described in b above, but this
time, instead of using two dice, you use a single die to
determine how many points to add or subtract from each of the
scores? Try it, and see whether your prediction is correct by
comparing the results of the analysis with the original F-ratio
and with the F-ratio computed in b above.
For each of the following designs (i) give an example, (ii)
identify the major threats to validity, and (iii) state the
limits on the conclusions that can be drawn.
a. Single-group, posttest-only design
b. Ex post facto design
c. Single-group, pretest-posttest design
d. Pretest-posttest, natural control-group design
For each of the following types of experimental design, (i) give
an example and (ii) indicate what threats to validity are
controlled.
a. Randomized, posttest-only, control-group design
b. Randomized, pretest-posttest, control-group design
c. Multilevel, completely randomized, between-subjects
design