The Effect of an Action's Outcome on the Evaluation of the Action's Object (pairfocus9a+b) – report.
Purpose
In previous studies (pairfocus5, pairfocus4), we tested the effect of action's outcome on the evaluation
of the action's object.
In pairfocus5, we exposed participants to four types of actions: Actions with positive outcomes
performed on positive stimuli (e.g., giving a cute animal to the participant); Actions with negative
outcomes performed on positive stimuli (e.g., taking a cute animal from the participant); Actions with
positive outcomes performed on negative stimuli (e.g., taking a nasty animal from the participant); and
actions with negative outcomes performed on negative stimuli (e.g., giving a nasty animal to the
participant). We presented the actions one at the time. After each presentation, participants evaluated
the stimulus that was the object of the action.
In pairfocus4, we exposed participants to four types of behaviors performed by four men. We used 32
behaviors. Eight described an action on a positive object with a positive outcome (e.g., David drastically
increased the profits of his company), eight described an action on a positive object with a negative
outcome (e.g., David drastically reduced the profits of his company), eight described an action on a
negative object with a negative outcome (e.g., David drastically increased the debts of his company),
and eight described an action on a negative object with a positive outcome (e.g., David drastically
reduced the debts of his company). We presented the behaviors one at the time. After each
presentation, participants evaluated the stimulus that was the object of the action.
In both studies, we found that people evaluate objects more positively if the action that was performed
on the objects had a positive outcome than if the action had a negative outcome. The link between the
object and the action's outcome led to an assimilative effect on evaluation. However, it is not clear
whether this is only a momentary effect (misattribution at the time of rating) or an effect on
evaluation that lasts longer. The current study will test if action's valence affects the evaluation of the
action's object when people evaluate the object with a delay from the learning episode.
We had two versions of this experiment: pairfocus9a and pairfocus9b that were very similar, except for
the following changes:
In pairfocus9a:
There were 32 learning trials.
There was not a constant men-roles pairing.
The evaluation of the animals was measured with a separate questionnaire using three types of
questions and a scale of 9 points.
Participants did not evaluate the men after the task
Participants completed a race BIAT.
In pairfocus9b:
There were 24 learning trials.
The men-roles pairing was counterbalanced.
The evaluation of the animals was measured in the learning task using one question for each
animal and a scale of4 points.
Participants evaluated the men after the task.
Participants completed a BIAT that compares two of the men from the learning task.
Here are the details:
Pairfocus9a:
Procedure:
Participants learned about four target men; each performed all of the following actions: (1) gave them
cute animals, (2) gave them nasty animals, (3) took cute animals away from them and (4) took nasty
animals away from them. Participants were instructed to learn what the men's roles and to form
impressions on the four men based on their roles.
In each trial, a man (image + name) appeared next to an image of cute animals or nasty animals for
400ms, at which point a word was added between them: "gives you" or "takes away from you",
displayed for another 1600ms, until all stimuli disappeared. At the end of each trial, we showed for
500ms a countdown of trials that left until participants finish the task.
The game had 32 trials (8 of each role). The task was divided to two blocks. In each block, the different
trials were presented in a random order.
During the task, one cute animal was always given to the participants, one nasty animal was always
given to the participants, one cute animal was always taken away from the participants, and one nasty
animal was always taken away from the participants.
The two cute animals were randomly selected from a pool of eight cute animals:
The two nasty animals were randomly selected from a pool of eight nasty animals:
Instructions of the task:
Before the first block:
Before the second block:
After the learning, participants rated the four animals (in a random order).
Instructions:
There were three types of questions:
1. How positive or negative does this animal seem to you? ("Extremely negative", "Moderately
negative", "Slightly negative", "Neither positive or negative", "Slightly positive", "Moderately
positive", "Extremely positive")
2. How much do you like this animal? ("Dislike extremely", "Dislike moderately", "Dislike slightly",
"Neither like or dislike", "Like slightly", "Like moderately", "Like extremely")
3. How friendly or unfriendly does this animal seem to you? ("Extremely unfriendly", "Moderately
unfriendly", "Slightly unfriendly", "Neither friendly nor unfriendly", "Slightly friendly",
"Moderately friendly", "Extremely friendly")
The order of the different types of questions was random
The men were
Michael
David
James
Chris
Brief IAT
In the end of the study, participants completed a race BIAT. The purpose of using the BIAT was mainly
to justify this study as a project implicit study (i.e., we need to measure the so-called “implicit”
attitudes). Therefore, if we find any interesting results with the BIAT, we will not use them unless
replicated in another run of this study.
The BIAT will have 4 blocks of 20 trials and will compare White peoples and Black peoples.
These were the categories and stimuli:
White people:
Black people:
Pleasant: Joy. Love, Happy, Good
Unpleasant: Horrible, Nasty, Bad, Evil
Instruction before the BIAT:
Sample size
We planned to run 650 participants to have 95% power to detect a between effect of ηp2 =0.02.
Pairfocus9b:
Procedure:
Participants learned about four target men each performed one of four actions: (1) gives them cute
animals, (2) gives them nasty animals, (3) takes cute animals away from them and (4) takes nasty
animals away from them. Participants were instructed to learn what the role of each man is and to form
impressions on the four men based on their roles.
In each trial, a man (image + name) appeared next to an image of cute animals or nasty animals for
400ms, at which point a word was added between them: "gives you" or "takes away from you",
displayed for another 1600ms, until all stimuli disappeared. At the end of each trial, we showed for
500ms a countdown of trials that left until participants finish the task.
The game had 24 trials (6 of each man). The task was divided to two blocks. In each block, the different
trials were presented in a random order.
During the task, one cute animal was always given to the participants, one nasty animal was always
given to the participants, one cute animal was always taken away from the participants, and one nasty
animal was always taken away from the participants.
The two cute animals were randomly selected from a pool of eight cute animals, and the two nasty
animals were randomly selected from a pool of eight nasty animals (same pools as in pairfocus9a).
Instructions:
Before the first block:
Before the second block:
In the end of the game, participants evaluate the four animals.
The question was "please rate how positive or negative this animal is" and the scale was: 1-very
negative, 2-negative, 3-positive, 4-very positive.
To avoid deception, participants next completed measures to test their evaluation of the men. These
measures are not the focus of this study! We do not have any specific predictions about them.
Deliberate evaluation:
Next, participants completed a deliberate evaluation measure.
Participants were asked how much they like each of the four men (David Michael James and Chris).
Instructions:
Example for a question:
Brief IAT
In the end of the experiment, the participants completed a BIAT. The purpose of using the BIAT was
mainly to justify this study as a project implicit study (i.e., we need to measure the so-called “implicit”
attitudes). Therefore, if we find any interesting results with the BIAT, we will not use them unless
replicated in another run of this study. The BIAT had 4 blocks of 20 trials and will compare David and
Michael.
Sample size
We planned to run 650 participants to have 95% power to detect a between effect of ηp2 =0.02.
Results:
Pairfocus 9a
Participants:
*These results were computed for participants who finished all the tasks in the study, including the
measures that are not relevant to the main research question. However, in the manuscript, we
included all the participants who completed the main task. The results for these participants are
reported in the end of this report (the results were the same in the two samples).
651 participants completed this study. We excluded four participants that did not respond to all the
evaluation questions. The final sample included 647 participants (Mage = 42.83, SD = 18.45, 58% women).
Evaluation of the affective stimuli:
Internal consistency of the three evaluation questions as a function of animal valence and action
valence:
Animal valence
Positive
Positive
Negative
Negative
Action valence
Positive (give)
Negative (take)
Positive (take)
Negative (give)
Alpha Cronbach
0.76
0.80
0.78
0.75
Evaluation ratings:
Mean evaluation as a function of stimuli valence and action valence (cuteg = positive animal/positive
action, cutet = positive animal/negative action, nastyt = negative animal/positive action, nastyg =
negative animal/negative action):
Variable
N
Mean Std Dev Std Error
Ecuteg
647
6.30
0.82
0.03
Ecutet
647
6.24
0.94
0.04
Enastyt
647
1.83
1.02
0.04
Enastyg
647
1.75
0.95
0.04
The effect was replicated but it is very small. The evaluation of the positive stimulus was somewhat
more positive when the stimulus was given to the participants than when it was taken away from the
participants (d = 0.06). The evaluation of negative stimulus was somewhat more negative when the
stimuli was given to the participants than when it was taken away from the participants (d = 0.08).
The 2 (stimuli valence) x 2 (action valence) repeated measures ANOVA, found a very strong main effect of
stimuli valence (ηp2 = .93), reflecting more positive evaluation of the positive stimuli than of the negative
stimuli. Most important, replicating our previous finding, the effect of action valence was reliable (ηp2 =
.007), reflecting more positive evaluation of the stimuli when they were the object of positive actions
than when they were the object of negative actions.
Here are the details:
Source
DF
stim_val
Type III SS Mean Square
1 13005.92585
Error(stim_val)
646
Source
action_val
Error(action_val)
Source
2.9696462
2.9696462
646 371.2803538
0.5747374
DF
stim_val*action_val
1.50257
Type III SS Mean Square
1
Type III SS Mean Square
1
Pr > F
13005.92585 8655.81 <.0001
970.65748
DF
F Value
0.1116692
0.1116692
Error(stim_val*action_val) 646 387.2494419
0.5994573
F Value Pr > F
5.17 0.0233
F Value Pr > F
0.19 0.6662
(ηp2
Effect of stimuli valence
= .98)
Mean evaluation as a function of stimuli valence:
Variable
N
Mean Std Dev Std Error
Epos
647
6.27
0.72
0.03
Eneg
647
1.79
0.80
0.03
More positive evaluation of the positive stimuli than of the negative stimuli.
Effect of action valence (ηp2 = .007)
Mean evaluation as a function of action valence:
Variable
N
Mean Std Dev Std Error
EposAction
647
4.06
0.58
0.02
EnegAction
647
4.00
0.59
0.02
More positive evaluation of the stimuli when they were the object of positive action than when they were
the object of a negative action.
Let's look at the effects for each question type separately:
Positivity:
Variable
N
Mean Std Dev Std Error
pcuteg
647
6.37
0.93
0.04
pcutet
647
6.31
1.02
0.04
pnastyt
647
1.79
1.12
0.04
pnastyg
647
1.70
1.05
0.04
The evaluation of the positive stimulus was somewhat more positive when the stimulus was given to the
participants than when it was taken away from the participants (d = 0.06). The evaluation of negative
stimulus was somewhat more negative when the stimuli was given to the participants than when it was
taken away from the participants (d = 0.08).
Liking:
Variable
N
Mean Std Dev Std Error
lcuteg
647
6.24
0.99
0.04
lcutet
647
6.19
1.11
0.04
lnastyt
647
2.01
1.39
0.05
lnastyg
647
1.93
1.32
0.05
The evaluation of the positive stimulus was somewhat more positive when the stimulus was given to the
participants than when it was taken away from the participants (d = 0.04). The evaluation of negative
stimulus was somewhat more negative when the stimuli was given to the participants than when it was
taken away from the participants (d = 0.05).
Friendly:
Variable
N
Mean Std Dev Std Error
fcuteg
647
6.28
1.07
0.04
fcutet
647
6.23
1.17
0.05
fnastyt
647
1.68
1.12
0.04
fnastyg
647
1.61
1.08
0.04
The evaluation of the positive stimulus was somewhat more positive when the stimulus was given to the
participants than when it was taken away from the participants (d = 0.04). The evaluation of negative
stimulus was somewhat more negative when the stimuli was given to the participants than when it was
taken away from the participants (d = 0.06).
The pattern is the same for all the questions (but somewhat stronger for the positivity question).
Next, let's look at the results (mean evaluation ratings) as a function of a specific animal (between
participants):
Positive animals:
Positive Action
Negative Action
T-test
6.28 (0.81)
6.17 (1.10)
t(137.32) = 0.74, p =
(n=89)
(n=77)
.462, d= 0.12
6.37 (0.87)
(n=85)
6.34 (0.97)
(n=86)
t(169) = 0.22, p = .823,
d= 0.03
6.13 (0.68)
(n=77)
5.95 (0.88)
(n=64)
t(117.29) = 1.38, p =
.169, d= 0.25
6.60 (0.66)
(n=74)
6.51 (0.81)
(n=71)
t(143) = 0.75, p = .454,
d= 0.12
5.84 (0.89)
(n=73)
5.89 (0.91)
(n=90)
t(161) = -0.30, p = .761,
d= -0.04
6.47 (0.75)
(n=65)
6.49 (0.73)
(n=80)
t(143) = -0.16, p = .872,
d= -0.02
6.57 (0.68)
(n=88)
6.48 (0.74)
(n=105)
t(191) = 0.87, p = .387,
d= 0.12
6.13 (0.93)
(n=96)
6.04 (1.02)
(n=74)
t(168) = 0.59, p = .558,
d= 0.09
The results are in the predicted direction for 6 out of 8 stimuli, but the difference is not reliable for any
of them.
Negative animals:
Positive Action
1.64 (0.90)
(n=102)
Negative Action
1.44 (0.76)
(n=88)
T-test
t(188) = 1.66, p = .099,
d= 0.24
1.53 (0.83)
(n=95)
1.45 (0.69)
(n=74)
t(167) = 0.73, p = .464,
d= 0.11
1.78 (0.92)
(n=82)
2.00 (0.96)
(n=73)
t(153) = -1.43, p = .156,
d= -0.23
1.37 (0.59)
(n=62)
1.46 (0.74)
(n=63)
t(123) = -0.79, p = .428,
d= -0.14
2.11 (1.19)
(n=80)
1.97 (1.12)
(n=69)
t(147) = 0.77, p = .443,
d= -0.12
2.22 (1.20)
(n=60)
1.83 (0.96)
(n=109)
t(100.61) = 2.18, p =
.031, d= 0.43
1.73 (0.99)
(n=78)
1.68 (1.01)
(n=81)
t(157) = 0.33, p = .745,
d= 0.05
2.29 (1.06)
(n=88)
2.09 (1.03)
(n=90)
t(176) = 1.30, p = .195,
d= 0.19
The results are in the predicted direction for 6 out of 8 stimuli, but the difference is reliable for only one
of them.
The BIAT results:
Analysis Variable : IAT
N Mean Std Dev Std Error t Value Pr > |t|
646
0.08
0.49
0.02
4.20
<.0001
The participants showed weak (but reliable) preference for White people over Black people (d=0.16).
Results for participants who completed the main task:
704 participants completed this study. We excluded four participants that did not respond to all the
evaluation questions. The final sample included 699 participants (Mage = 43.25, SD = 18.70, 58% women).
Evaluation of the affective stimuli:
Internal consistency of the three evaluation questions as a function of animal valence and action
valence:
Animal valence
Positive
Positive
Negative
Negative
Action valence
Positive (give)
Negative (take)
Positive (take)
Negative (give)
Alpha Cronbach
0.78
0.79
0.78
0.75
Evaluation ratings:
Mean evaluation as a function of stimuli valence and action valence (cuteg = positive animal/positive
action, cutet = positive animal/negative action, nastyt = negative animal/positive action, nastyg =
negative animal/negative action):
Variable
N Mean Std Dev Std Error
Ecuteg
699
6.29
0.85
0.03
Ecutet
699
6.24
0.95
0.04
Enastyt
699
1.83
1.02
0.04
Enastyg
699
1.75
0.95
0.04
The effect was replicated but it is very small. The evaluation of the positive stimulus was somewhat
more positive when the stimulus was given to the participants than when it was taken away from the
participants (d = 0.05). The evaluation of negative stimulus was somewhat more negative when the
stimuli was given to the participants than when it was taken away from the participants (d = 0.08).
The 2 (stimuli valence) x 2 (action valence) repeated measures ANOVA:
Source
stim_val
DF
Type III SS Mean Square F Value Pr > F
1 13997.68813
Error(stim_val)
698
1110.75632
Source
DF
Type III SS
13997.68813 8796.16 <.0001
1.59134
Mean Square F Value
Pr > F
Source
DF
1
Type III SS Mean Square F Value Pr > F
3.3883325
3.3883325
Error(action_val)
698 386.5005564
0.5537257
Source
DF
Type III SS
1
0.1527579
0.1527579
Error(stim_val*action_val) 698 404.4027976
0.5793736
action_val
stim_val*action_val
6.12 0.0136
Mean Square F Value
Pr > F
0.26 0.6078
Effect of stimuli valence (ηp2 = .92)
Mean evaluation as a function of stimuli valence:
Variable
N Mean Std Dev Std Error
Epos
699
6.26
0.75
0.03
Eneg
699
1.79
0.81
0.03
More positive evaluation of the positive stimuli than of the negative stimuli.
Effect of action valence (ηp2 = .008)
Mean evaluation as a function of action valence:
Variable
N Mean Std Dev Std Error
EposAction
699
4.06
0.59
0.02
EnegAction
699
3.99
0.60
0.
More positive evaluation of the stimuli when they were the object of positive action than when they were
the object of a negative action.
Summary:
Overall, we found weak support for our prediction. Action's valence affected the evaluation of the
action's object when people evaluate the object with a delay from the learning episode, however the
effect was very weak.
Pairfocus9b:
Participants:
*These results were computed for participants who finished all the tasks in the study, including the
measures that are not relevant to the main research question. However, in the manuscript, we
included all the participants who completed the main task. The results for these participants are
reported in the end of this report (the results were the same in the two samples).
653 participants completed this study. We excluded four participants that completed the learning task
more then once. The final sample included 649 participants (Mage = 41.85, SD = 18.68, 59% women).
Evaluation of the affective stimuli:
Evaluation:
Mean evaluation as a function of stimuli valence and action valence (qGiveGd = positive animal/positive
action, qGiveGd = positive animal/negative action, qTakeBd = negative animal/positive action, qGiveBd =
negative animal/negative action):
Variable
N
Mean Std Dev Std Error
qGiveGd
649
3.79
0.43
0.02
qTakeGd
649
3.58
0.69
0.03
qTakeBd
649
1.47
0.71
0.03
qGiveBd
649
1.37
0.60
0.02
The effect is replicated but it is smaller than the one we found when the evaluation was during the task
(but stronger than the one we found in pairfocus9a). The evaluation of the positive stimulus was more
positive when the stimulus was given to the participants than when it was taken away from the
participants (d = 0.36). The evaluation of negative stimulus was more negative when the stimuli was
given to the participants than when it was taken away from the participants (d = 0.15).
The 2 (stimuli valence) x 2 (action valence) repeated measures ANOVA, found a very strong main effect
of stimuli valence (ηp2 = .89), reflecting more positive evaluation of the positive stimuli than of the
negative stimuli. Most important, replicating our previous finding, the effect of action valence was
reliable (ηp2 = .06), reflecting more positive evaluation of the stimuli when they were the object of
positive actions than when they were the object of negative actions. We also found an interaction
between stimuli valence and action's valence (ηp2 = .012) reflecting stronger effect of action's valence
for positive stimuli than for negative stimuli.
Here are the details:
Source
stim_val
Error(stim_val)
Source
action_val
Error(action_val)
Source
stim_val*action_val
DF
Type III SS Mean Square
1 3336.382512
648
DF
1
410.367488
0.633283
Type III SS Mean Square
16.8262712
648 242.9237288
0.3748823
Type III SS Mean Square
1.8339753
1.8339753
Error(stim_val*action_val) 648 150.9160247
0.2328951
Effect of stimuli valence (ηp2 = .89)
1
Pr > F
3336.382512 5268.39 <.0001
16.8262712
DF
F Value
F Value Pr > F
44.88 <.0001
F Value Pr > F
7.87 0.0052
Mean evaluation as a function of stimuli valence:
Variable
N
Mean Std Dev Std Error
posStim
649
3.69
0.44
0.02
negStim
649
1.42
0.51
0.02
More positive evaluation of the positive stimuli than of the negative stimuli.
Effect of action valence (ηp2 = .06)
Mean evaluation as a function of action valence:
Variable
N
Mean Std Dev Std Error
posAction
649
2.63
0.39
0.02
negAction
649
2.47
0.42
0.02
More positive evaluation of the stimuli when they were the object of positive action than when they were
the object of a negative action.
Interaction between action valence and stimuli valence (ηp2 = .012)
The effect of action valence for positive stimuli (ηp2 = .07):
Source
action_val
DF
1
Type III SS Mean Square F Value
14.8852080
14.8852080
Error(action_val) 648 174.6147920
0.2694673
Pr > F
55.24 <.0001
The effect of action valence for negative stimuli (ηp2 = .016)::
Source
action_val
DF
1
Type III SS Mean Square F Value
3.7750385
3.7750385
Error(action_val) 648 219.2249615
0.3383101
Pr > F
11.16 0.0009
The effect is stronger for positive stimuli.
Next, let's look at the results (mean evaluation ratings) as a function of a specific animal (between
participants):
Positive animals:
Positive Action
3.71 (0.45)
(n=91)
Negative Action
3.63 (0.64)
(n=84)
T-test
t(149.1) = 0.99, p =
.323, d= 0.16
3.85 (0.42)
(n=85)
3.68 (0.64)
(n=75)
t(125.54) = 1.92, p =
.057, d= 0.34
3.75 (0.44)
(n=63)
3.49 (0.69)
(n=100)
t(160.97) = 2.90, p =
.004, d= 0.45
3.90 (0.30)
(n=84)
3.75 (0.60)
(n=73)
t(101.99) = 1.97, p =
.051, d= 0.39
3.78 (0.42)
(n=78)
3.47 (0.63)
(n=85)
t(146.68) = 3.47, p =
.0007, d= 0.57
3.83 (0.37)
(n=84)
3.47 (0.78)
(n=68)
t(91.70) = 3.51, p = .
0007, d= 0.73
3.79 (0.44)
(n=80)
3.58 (0.81)
(n=83)
t(127.41) = 2.05, p =
.042, d= 0.36
3.73 (0.55)
(n=84)
3.57 (0.65)
(n=81)
t(163) =1.70, p = .091,
d= 0.26
The results are in the predicted direction for all the stimuli, and the difference is reliable for 6 of them.
Negative animals:
Positive Action
1.40 (0.72)
(n=95)
Negative Action
1.22 (0.45)
(n=68)
T-test
t(158.54) = 1.95, p =
.053, d= 0.30
1.39 (0.68)
(n=90)
1.16 (0.45)
(n=95)
t(151.93) = 2.71, p =
.007, d= 0.43
1.46 (0.60)
(n=74)
1.55 (0.75)
(n=92)
t(164) = -0.89, p = .376,
d= -0.14
1.30 (0.61)
(n=88)
1.32 (0.67)
(n=71)
t(157) = -0.28, p = .779,
d= -0.04
1.69 (0.75)
(n=64)
1.42 (0.64)
(n=76)
t(138) = 2.27, p = .025,
d= 0.38
1.59 (0.90)
(n=83)
1.34 (0.52)
(n=89)
t(129.69) = 2.24, p =
.026, d= 0.39
1.31 (0.55)
(n=70)
1.31 (0.46)
(n=88)
t(156) = 0.09, p = .926,
d= 0.01
1.69 (0.71)
(n=85)
1.63 (0.66)
(n=70)
t(153) = 0.59, p = .555,
d= 0.09
The results are in the predicted direction for 5 of the 8 stimuli, and the difference is reliable for 4 of
them.
Next, let’s look at the men evaluation measures (these were not the focus of the current study but we use
then to justify the impression formation instructions in the induction task and to justify the study as a
project implicit study).
Six participants did not response to all four-evaluation questions and therefore were removed from this
analysis.
Deliberate evaluation of the men as a function of man's role:
Variable
N
Mean Std Dev Std Error
eGiveGd
643
5.52
1.58
0.06
eTakeBd
643
5.17
1.73
0.07
eTakeGd
643
2.59
1.57
0.06
eGiveBd
643
2.43
1.50
0.06
We can see that the evaluation of the men who took away stimuli were less extreme than the evaluation
of men who gave stimuli (d for the positive men = 0.21, d for the negative men = 0.10; d for preference
between the similarity men = 2.00, d for preference between the opposition men = 1.58; in pairfocus5,
these two ds = 0.94, 0.39). This replicate the effect we usually find.
The 2 (men's role: positive, negative) x 2 (relation type: give, take) repeated measures ANOVA found a
main effect of men's role (ηp2 = .60), reflecting more positive evaluation of the men who helped the
participants than of the men who harmed the participants. In addition, an interaction between men's role
and relation type (ηp2 = .03), reflecting stronger effect of men's role when the men gave stimuli (ηp2 = .25),
than when the men took away stimuli (ηp2 = .05). Here are the details:
Source
DF
rlt
Error(rlt)
1
6.075039
6.075039
642 1039.674961
1.619431
Source
DF
out
Error(out)
Type III SS Mean Square
642 3242.090202
DF
rlt*out
1
Effect of men's role
= .60):
Evaluations as a function of men's role:
Pr > F
5182.659798 1026.27 <.0001
5.049985
Type III SS Mean Square F Value
Pr > F
43.113919
<.0001
Error(rlt*out) 642 1275.636081
(ηp2
Pr > F
3.75 0.0532
Type III SS Mean Square F Value
1 5182.659798
Source
F Value
43.113919 21.70
1.986972
Variable
N
Mean Std Dev Std Error
posMen
643
5.52
1.58
0.06
negMen
643
2.51
1.26
0.05
More positive evaluation of the men who helped the participants than of the men who harmed the
participants.
Interaction between men's role and relation type (ηp2 = .03):
The effect of men's role when the men gave stimuli (ηp2 = .59):
Source
DF
out
Type III SS Mean Square F Value
1 3085.586314
3085.586314
Error(out) 642 2085.413686
3.248308
Pr > F
949.91 <.0001
The effect of men's role when the men took away stimuli (ηp2 = .46):
Source
DF
out
Type III SS Mean Square F Value
1 2140.187403
2140.187403
Error(out) 642 2432.312597
3.788649
Pr > F
564.89 <.0001
Finally, let us look at the Brief IAT results. I compute BIAT D score such that positive score reflects
preference for the positive man over the negative man. Here are the preferences as a function of relation
type:
Analysis Variable : bat
RltCond N Obs
N Mean Std Dev Std Error Pr > |t| t Value
Take
345 343
0.25
0.54
0.03
<.0001
8.76
give
304 303
0.32
0.51
0.03
<.0001
10.89
The preference is stronger when the men gave stimuli than when the men took away stimuli, however
this effect is not reliable, t(644) = 1.59, p = 0.112, d = 0.12.
Results for participants who completed the main task:
779 participants completed this study. We excluded six participants that completed the learning task
more than once. The final sample included 733participants (Mage = 43.05, SD = 19.08, 59% women).
Evaluation of the affective stimuli:
Evaluation:
Mean evaluation as a function of stimuli valence and action valence (qGiveGd = positive animal/positive
action, qGiveGd = positive animal/negative action, qTakeBd = negative animal/positive action, qGiveBd =
negative animal/negative action):
Variable
N Mean Std Dev Std Error
Lower 95% Upper 95%
CL for Mean CL for Mean
qGiveGd1
773
3.78
0.45
0.02
3.75
3.82
qTakeGd1
773
3.59
0.67
0.02
3.54
3.64
qTakeBd1
773
1.45
0.70
0.03
1.40
1.50
qGiveBd1
773
1.36
0.60
0.02
1.32
1.40
The effect is replicated but it is smaller than the one we found when the evaluation was during the task
(but stronger than the one we found in pairfocus9a). The evaluation of the positive stimulus was more
positive when the stimulus was given to the participants than when it was taken away from the
participants (d = 0.33). The evaluation of negative stimulus was more negative when the stimuli was
given to the participants than when it was taken away from the participants (d = 0.13).
The 2 (stimuli valence) x 2 (action valence) repeated measures ANOVA:
Source
DF
stim_val
Type III SS Mean Square
1 4023.198254
Error(stim_val)
Source
772
DF
action_val
1
Error(action_val)
Source
0.615352
Type III SS Mean Square
16.0831177
772 276.1668823
0.3577291
1
Type III SS Mean Square
1.9175291
1.9175291
Error(stim_val*action_val) 772 184.3324709
0.2387726
F Value
F Value
N Mean Std Dev Std Error
773
3.69
0.44
0.02
negStim
773
1.41
0.51
0.02
More positive evaluation of the positive stimuli than of the negative stimuli.
Effect of action valence (ηp2 = .05)
Mean evaluation as a function of action valence:
posAction
N Mean Std Dev Std Error
773
2.62
0.40
Pr > F
8.03 0.0047
posStim
Variable
Pr > F
44.96 <.0001
Effect of stimuli valence (ηp2 = .89)
Mean evaluation as a function of stimuli valence:
Variable
Pr > F
4023.198254 6538.04 <.0001
16.0831177
DF
stim_val*action_val
475.051746
F Value
0.01
Variable
N Mean Std Dev Std Error
negAction
773
2.47
0.42
0.01
More positive evaluation of the stimuli when they were the object of positive action than when they
were the object of a negative action.
Interaction between action valence and stimuli valence (ηp2 = .01)
The effect of action valence for positive stimuli (ηp2 = .06):
Source
DF Type III SS Mean Square F Value Pr > F
14.5536869
14.5536869
Error(action_val) 772 207.4463131
0.2687128
action_val
1
54.16 <.0001
The effect of action valence for negative stimuli (ηp2 = .013):
Source
DF Type III SS Mean Square F Value Pr > F
3.4469599
3.4469599
Error(action_val) 772 253.0530401
0.3277889
action_val
1
10.52 0.0012
The effect is stronger for positive stimuli.
Summary:
Overall, we found a support for our prediction. Action's valence affected the evaluation of the action's
object when people evaluate the object with a delay from the learning episode. The effect in this
experiment was stronger than the effect we found in pairfocus9a.
© Copyright 2026 Paperzz