Langston, Research Methods, Sample Exam 3 Answered

This is an old version of the exam.  Study first, use the sample exam as a check on comprehension.  Remember, even though the exam is open notes, failure to study will lead to failure.

Page 1
Instructions:  All of the questions will be based on the following description of an experiment.  Read the entire description carefully before trying to answer the questions. The point breakdown is not even.

Note well:  Answer the questions on page 2 and have your answers checked before doing anything else.  Try not to spend more than 10 minutes on page 2.  You may have the answers to questions 1 and 2 for the price of all of the points.  If you're stumped, it's worth a few points to have a chance to get the rest correct.

My research question is:  What is the relationship between divorce, television violence, and aggression?  I have two hypotheses culled from the literature.  The D-I-V-O-R-C-E hypothesis is that children whose parents divorce will be more aggressive than children whose parents stay together.  The TV hypothesis is that children whose parents divorce will be more aggressive if those children are exposed to aggressive television programming.
To test between these hypotheses, two variables are manipulated.  Parents of half of the kids in the sample are assigned to divorce (yes, assigned).  The parents of the other half of the kids stay together.  If the D-I-V-O-R-C-E hypothesis is correct, then kids of divorced parents should be more aggressive than kids of non-divorced parents.
The second variable is TV violence.  One third of the kids watch nothing but Barney and Friends (the show has no violence, but it might induce violence for other reasons which we'll overlook).  One third of the kids watch Bugs Bunny etc. cartoons.  They see violence, but it's cartoon violence.  The final third watch violent movies.  Amount of TV violence should have no effect if the D-I-V-O-R-C-E hypothesis is correct.  However, if the TV hypothesis is correct, the kids whose parents divorce who also watch violent TV should become more aggressive.  The other kids (no divorce, or divorce without violence) should not become aggressive.
To measure aggression, we put the kids in a frustrating situation, and measure the number of times they act with aggression towards another kid in the room.

Page 2
USE VARIABLE AND HYPOTHESIS NAMES CONSISTENTLY.  WHATEVER IS VARIABLE 1 ON THE FIRST PAGE SHOULD BE VARIABLE 1 FROM THEN ON, ETC.

Based on this experiment, answer the following:
1.  Two hypotheses can be derived from the above description.  What are they? (2 pts. each)

Hypothesis 1 =  Name of hypothesis is:  ________D-I-V-O-R-C-E__________________________
Describe it briefly.  Divorce causes aggression

Hypothesis 2 =  Name of hypothesis is:  ________TV__________________________
Describe it briefly.  Divorce + TV violence causes aggression

2.  The experiment that was designed uses a _2_ X _3_ factorial design.  (7 pts.)
Variable 1 is:  Marital status

Its levels are:
a.  Divorced

b.  Not divorced

Variable 2 is:  TV violence

Its levels are:
a.  No violence

b.  Cartoon violence

c.  Real violence

3.  I would use a between subjects / within subjects design (circle one).
Justify your response with one reason for choosing this design type.  (5 pts.)

Divorce would carry over, violence would probably carry over.

Page 3
4.  Hypothesis 1 predictions:  (30 pts. total)  Name of hypothesis 1:  ____D-I-V-O-R-C-E_________

a.  Main Effect for variable 1:  yes  /  no  (circle one)  (2 pts.)

b.  Order the names of the conditions from the one with the largest expected mean to the one with the smallest expected mean.  (3 pts.)

Divorced > Not

c.  Why do you predict this (based on the description on page one)?  (4 pts.)

"If the D-I-V-O-R-C-E hypothesis is correct, then kids of divorced parents should be more aggressive than kids of non-divorced parents"

d.  Main Effect for variable 2:  yes  /  no  (circle one)  (2 pts.)

e.  Order the names of the conditions from the one with the largest expected mean to the one with the smallest expected mean.  (3 pts.)

None = Cartoon = Real

f.  Why do you predict this (based on the description on page one)?  (4 pts.)

"Amount of TV violence should have no effect if the D-I-V-O-R-C-E hypothesis is correct"

g.  Interaction:  yes  /  no  (circle one)  (2 pts.)

h.  Graph the interaction predicted by hypothesis 1.  Label all of the parts.  (6 pts.)

i.  Why is this interaction (or lack of interaction) being predicted (based on the description on page one)?  (4 pts.)

"If the D-I-V-O-R-C-E hypothesis is correct, then kids of divorced parents should be more aggressive than kids of non-divorced parents" and "Amount of TV violence should have no effect if the D-I-V-O-R-C-E hypothesis is correct"

in other words, the effect of one variable is not different at different levels of the other variable

Page 4
5.  Hypothesis 2 predictions:  (30 pts. total)  Name of hypothesis 2:  _____TV________

a.  Main Effect for variable 1:  yes  /  no  (circle one)  (2 pts.)

b.  Order the names of the conditions from the one with the largest expected mean to the one with the smallest expected mean.  (3 pts.)

Divorced > Not

c.  Why do you predict this (based on the description on page one)?  (4 pts.)

"kids whose parents divorce who also watch violent TV should become more aggressive"

d.  Main Effect for variable 2:  yes  /  no  (circle one)  (2 pts.)

e.  Order the names of the conditions from the one with the largest expected mean to the one with the smallest expected mean.  (3 pts.)

Real > Cartoon > None

f.  Why do you predict this based on the description on page one)?  (4 pts.)

"kids whose parents divorce who also watch violent TV should become more aggressive"

g.  Interaction:  yes  /  no  (circle one)  (2 pts.)

h.  Graph the interaction predicted by hypothesis 2.  Label all of the parts.  (6 pts.)

i.  Why is this interaction (or lack of interaction) being predicted (based on the description on page one)?  (4 pts.)

"kids whose parents divorce who also watch violent TV should become more aggressive.  The other kids (no divorce, or divorce without violence) should not become aggressive"

In other words, the effect of one variable is different at different levels of the other variable

Page 5
6.  Analysis:  Here are the data from the experiment (label everything):  (10 pts. total)

a.  Fill in the table of means.  Label everything.  (5 pts.)

b.  Make a graph of the data.  Label everything.  (5 pts.)

Page 6
7.  Based on number 6:  (14 pts. total)
a.  Have you found a main effect for variable 1:  yes  /  no  (circle one)  (2 pts.)

b.  Describe the effect for variable 1:  (2 pts.)

Divorced > Not

c.  Have you found a main effect for variable 2:  yes  /  no  (circle one)  (2 pts.)

d.  Describe the effect for variable 2:  (2 pts.)

Real > Cartoon > None

a.  Have you found an interaction:  yes  /  no  (circle one)  (2 pts.)

b.  Describe the interaction:  (4 pts.)

For children of divorce, there is an increase in aggression when you go from no violence to real violence on TV.  For chuldren whose parents are not divorced, there is no change in aggression due to violence

BONUS:
a.  Which hypothesis is supported by your data?  (2 pts.)

TV

b.  Why is that hypothesis supported?  (3 pts.)

There are two main effects and the predicted interaction

Research Methods Sample Exam 3 Answered
Will Langston