|
Table 1': Mean Responses of Many Participants
(3 = Probably A, 1 = Maybe A, -1 = Maybe B, -3 = Probably B. Individual mean computed first for each cell, then mean across 71 individuals is tabulated here.) |
|||
| Disease A | Disease B | Blocking | Overshadowing |
| 2.92 SS/df=0.078 N=71 |
-2.96 SS/df=0.041 N=71 |
-0.31 SS/df=0.44 N=71 |
0.51 SS/df=0.51 N=71 |
For both diseases, accuracy during the test phase was extremely good.
There was a blocking effect, on average, with people tending to prefer disease B for the blocking cases. That is, the mean of -0.31 is significantly less than zero, t(70)=3.95, p<.01.
There were notable individual differences, however. Below is a frequency distribution of individual means for the blocking cases. It can be seen that many people showed no blocking effect, and a few even the opposite of a blocking effect. But a significant number of people did exhibit blocking.
Cum
Value Frequency Percent Percent
-3.00 1 1.4 1.4
-2.00 1 1.4 2.8
-1.50 2 2.8 5.6
-1.00 11 15.5 21.1
-.50 18 25.4 46.5
.00 29 40.8 87.3
.50 6 8.5 95.8
1.00 3 4.2 100.0
------- -------
Total 71 100.0
|
-3.00 |** 1
|
-2.00 |** 1
|
-1.50 |*** 2
|
-1.00 |****************** 11
|
-.50 |****************************** 18
|
.00 |************************************************ 29
|
.50 |********** 6
|
1.00 |***** 3
|
+---------+---------+---------+---------+---------+
0 6 12 18 24 30
Frequency
|
There was an overshadowing effect, on average, with people tending to prefer disease A for the overshadowing cases. That is, the mean of 0.51 is significantly greater than zero, t(70)=6.04, p<.01.
There were notable individual differences, however. Below is a frequency distribution of individual means for the overshadowing cases. It can be seen that many people showed no overshadowing effect, and a few even the opposite of an overshadowing effect. But a significant number of people did exhibit overshadowing.
Cum
Value Frequency Percent Percent
-.50 5 7.0 7.0
.00 29 40.8 47.9
.50 11 15.5 63.4
1.00 18 25.4 88.7
1.50 4 5.6 94.4
2.00 2 2.8 97.2
2.50 1 1.4 98.6
3.00 1 1.4 100.0
------- -------
Total 71 100.0
|
-.50 |******** 5
|
.00 |************************************************ 29
|
.50 |****************** 11
|
1.00 |****************************** 18
|
1.50 |******* 4
|
2.00 |*** 2
|
2.50 |** 1
|
3.00 |** 1
|
+---------+---------+---------+---------+---------+
0 6 12 18 24 30
Frequency
|
It turns out that there was a significant correlation of blocking with overshadowing. That is, if a person showed overshadowing, then he or she tended also to show blocking. The correlation was highly significant, with r = -.612. A scatterplot of 71 individuals is shown below. Notice that many people (21) showed zero blocking and zero overshadowing, but many other people showed both, with covarying magnitude.
Blocking vs. Overshadowing
++----+----+----+----+----+----+----+----+----+----+----+----+----++
1+ 1 1 1 +
| |
| |
| |
| |
.5+ 1 4 1 +
| |
| |
| |
| |
0+ 2 (21) 3 2 1 +
| |
| |
B | |
l | |
o -.5+ 1 2 7 8 +
c | |
k | |
i | |
n | |
g -1+ 1 1 6 3 +
| |
M | |
e | |
a | |
n -1.5+ 1 1 +
| |
| |
| |
| |
-2+ 1 +
| |
| |
| |
| |
-2.5+ +
| |
| |
| |
| |
-3+ 1 +
++----+----+----+----+----+----+----+----+----+----+----+----+----++
-2 -1 0 1 2 3 4
Overshadowing Mean
|