__Sample Items for the Final__

__Psychology 303, Fall 2002__

**(20) For the following examples, please list the IV(s) and
the DV(s), and select the best description of the design employed.
All examples are drawn from journal article abstracts. In some cases
the abstract has been altered slightly to simplify or clarify.**

A visual preference procedure was used to examine 18-, 6-, and 4-month-old
infants’ sensitivity to phrase structure in music. Infants listened
to sections of Mozart minuets that were divided into segments that did
and that did not correspond to the phrase structure of the music.
Infants in all age groups listened significantly longer to the segments
that corresponded to the phrase structure of the music. There was
no difference among the age groups, suggesting that protracted musical
experience may not be necessary to perceive phrase structure in music.
Krumhansl & Jusczyk (1990) Psychological Science, 1, 70-73.

IV(s):

DV(s):

DESIGN (Selected from below): ________

“Work expands so as to fill the time available for its completion,”
Parkinson’s law, is an explanation classic that has survived without an
artifact-free demonstration at the individual level. To evaluate
Pakinson’s law, undergraduate subjects expected to judge four sets of photos
of faces with reference to a subjective criteria. The experimental
subjects, who were told that the fourth set was canceled before they began
work on the third set, dallied on the third set; that is, as compared with
control, they prolonged work. The cancellation-dalliance effect was
re-obtained to two exact replications. The generalizability of the
effect and explanations for it are discussed.
Brannon, Hershberger, & Brock (1999). Psychonomic Bulletin and
Review, 6, 148-156.

IV(s):

DV(s):

DESIGN (Selected from below): ________

STATISTIC (Selected from below): ________

Designs

a. Basic two group between-subjects experiment

b. Basic within-subjects with two levels

c. Basic status variable with two groups

d. One multiple level between-subjects IV

e. One multiple level within-subjects IV

f. One multiple level status variable

g. factorial design: 2 between-subjects variables

h. factorial design: 2 status variables

i. factorial design: 2 within-subjects variables

j. factorial design: 1 status; 1 between

k. factorial design: 1 between; 1 within

l. factorial design: 1 status; 1 within

**(10) For the following source table, please fill in the missing
information.**

A 3x3 between-subjects ANOVA on number of words recalled revealed
a main effect of number of rehearsals, F (2, 81) = 15.02, p < .001,
MSE = 15.20, an effect of learning condition, F (2, 81) = 6.65, p <
.001, MSE = 15.20, and no interaction, F (4, 81) = 1.67, p > .05, MSE =
15.20.

__Source
SS Df
MS
F-ratio
Probability__

Between

Rehearsals

Learning Condition

Interaction

R x L

Error

Total

**(10) For the following one-way within-subjects source table
do two things. First, please fill in the missing information. Second,
calculate the F-ratio and find the associated probability if the data had
been analyzed using a one-way between-subjects ANOVA (do this in the space
below the source table).**

__Source
SS Df
MS
F-ratio
Probability__

Within

Social Setting
4.45

Error
24.99 38

Subjects
19

Total
334.67

If the same data was analyzed using a one-way between-subjects ANOVA used:

F-ratio:

p:

**(10) Describe the post-hoc comparisons needed in a 2x3 factorial
design when a two-way between-subjects ANOVA finds no main effect for the
first factor, a main effect for the second, and an interaction. Explain
why these comparisons are needed.**

**(20) For the following set of means, please draw a graph of the means.
In drawing your graph, please label the x and y-axes carefully. Then
predict whether a 2 x 2 between-subjects ANOVA would find main effects
(for each factor) and whether the interaction would be significant.
If you suspect that the ANOVA would give you a main effect that would be
an artifact of an interaction, please state that as well. (Keep in
mind that your experiment has reasonable power -- i.e., enough subjects
-- 20 per cell, and moderate error variance to discover effects if they
are meaningful.)**

Post-event information

Delay interval
Misleading
Consistent

5 m
3.85 (SD = 1.89) 7.75 (SD
= 1.94)

10 m
4.13 (SD = 1.43) 4.45 (SD
= 1.61)

DV: Confidence in the correct answer on a scale from 1 (low confidence)
to 9 (high confidence)

** (10) Define an artifact of the interaction. Comment
on how this will impact the interpretation of an ANOVA finding of a main
effect of an independent variable.**

** (20) This is a 2 page question. What do you know
about this study based on the following results paragraphs? Be sure
to mention basic research question, IV and levels, DV(s), design, statistics
employed, number of subjects, and findings. Answer these on the next page!**

Performance on the first main test questionnaire is summarized in Table
1. This gives the numbers of subjects and means scores out of 16
for each group (UGM1 – first year math student who received training in
reasoning; UGM2 – second year math students who received training the previous
year; BSM – boys studying math in a high school; PGMT – graduate students
learning to teach math; UGA – undergraduates not taking math classes),
each form of the questionnaire (W – words only; ID – inessential pictures
with the words; ED – Essential pictures), and for each of the 5 x 3 cells.
It also gives the overall mean score.

Analysis of variance was applied to the 118 individual scores
classified by group and form of questionnaire. This showed a significant
difference in performance between the groups (F (4, 111) = 5.10, p <
.001). . Scheffe’s comparisons showed that the UGM1 group,
which received recent explicit training in reasoning, though not in problems
of this type, performed significantly better than all others in this study.
The mean score for the UGM1 group was 14.2 and for the other groups 11.0.
The mean scores for the W, ID, and ED questionnaires were 11.9, 11.4, and
10.9, respectively. This difference between questionnaires was not
significant (F (2, 111) = 1.03). Makovits’ findings says nothing
about scores on the ID, but suggests that the mean score of the ED should
be significantly better than that on the W. An independent samples
t-test on the subjects who took the W and ED forms gave no more than weak
support (t (84) = 1.31, p < .1; one-tailed).

Table 1. Mean (standard deviation) performance scores of the five groups on the first main questionnaire (W, ID, or ED).

W
ID
ED

Group
N
n Mean
n Mean
n Mean
Group Mean

UGM1
19
9 14.1 (2.6)
6 14.8 (1.7) 4 13.3 (3.1)
14.2 (2.5)

UGM2
23 11
11.3 (3.1)
6 11.8 (2.7) 6 10.3 (3.1)
11.2 (3.1)

BSM
38 19
12.0 (2.6) 10 9.8
(2.1) 9 9.8 (3.0)
10.9 (2.8)

PGMT
24
11 11.5 (3.2)
6 11.3 (2.8) 7 11.7 (3.9)
11.5 (3.3)

UGA
14
7 10.1 (2.0)
4 10.0 (2.0) 3 10.7 (2.5)
10.2 (2.1)

Total Number
118
57
32
29

Overall Mean
11.9 (3.0)
11.4 (2.9)
10.9 (3.4) 11.5 (3.1)

From: Nelson & Hannan (2002). Applied Cognitive Psychology,
16, 157-170.

Basic Research Question:

IV(s):

Levels of IV(s)

DV(s):

Design:

Stats Employed:

Number of Subjects:

Findings: