Velocity Toward Goal Attainment in Immediate

Experience as a Determinant of Affect

JOHN W. LAWRENCE CHARLES

S. CARVER'
Baltimore Regional Burn Center University of Miami
Johns Hopkins School of Medicine

F. SCHEIER
MICHAEL
Carnegie Mellon University

An experiment tested a model in which affect reflects rate of movement (velocity) toward
one's behavioral goal. Participants responded to an ambiguous task and received false
feedback regarding their performance. The pattern of feedback either remained fairly con-
stant at 50% correct, went from low incidence of correct to 50% correct across trials, or
went from high incidence of correct to 50% correct across trials. Self-reported mood
change (from before the task to the point at which all groups received 50% correct) took
the following pattern: Participants moving from frequent correct to 50% correct (low
velocity) changed to less positive mood, whereas those moving from infrequent correct to
50% correct (high velocity) tended to change to more positive mood. Discussion centers
on convergence between these findings and those of previous research.

The notion that human behavior is goal directed is commonplace in today's

personality-social psychology (e.g., Austin & Vancouver, 1996; Cantor &
Kihlstrom, 1987; Carver & Scheier, 1981, 1990a, 1990b, 1998, 1999; Elliott &
Dweck, 1988; Emmons, 1986; Gollwitzer & Bargh, 1996; Higgins, 1987, 1997;
Klinger, 1975; Markus & Nurius, 1986; Palys & Little, 1983; Pervin, 1983, 1989;
Vallacher & Wegner, 1985; Wegner & Bargh, 1998). It is often argued as well that
affective experience is bound up with goals. This idea is certainly not new (e.g.,
Arnold, 1970; Ellsworth & Smith, 1988; Frijda, 1986; Izard, 1977; Kahneman &
Tversky, 1984; Lazarus, 1991; Ortony, Clore, & Collins, 1988; Roseman, 1984;
Scherer & Ekman, 1984; Simon, 1967). However, interest in the link between
affect and goals has re-emerged in recent years (cf. Martin & Tesser, 1996).
What is the relation between affect and goals? We consider one specific pos-
sibility here. Carver and Scheier (1 990a) proposed a model of affect derived from
their view on behavior, which in turn rests partly on the principles of control
theory (e.g., MacKay, 1963, 1966; Powers, 1973; Wiener, 1948). In that view,
goal-directed behavior reflects the operation of a discrepancy-reducing feedback
'Correspondence concerning this article should be addressed to Charles S. Carver, Department of
Psychology, University of Miami, Coral Gables, FL 33 124-2070. E-mail: CCarver@[Link]

Journal of Applied Social Psychology, 2002, 32,4, pp. 788-802.

Copyright 0 2002 by V. H. Winston 13Son, Inc. All rights reserved.
 AFFECT AND VELOCITY 789

loop, as follows: People compare their present behavior to their intentions, goals,
or standards. If they see a discrepancy between the perceived and the desired,
people alter their behavior to minimize the discrepancy (Carver & Scheier,
1990b, 1998, 1999).
Carver and Scheier (1990a) suggested that the origins of positive and nega-
tive affect might lie in a feedback process differing from the one just described.
In particular, they hypothesized a process that evaluates how successfully the
behavioral loop is reducing the discrepancies on which it is focusing. Put differ-
ently, this second loop monitors the rate at which the action loop is moving
toward its goal. The input for this second loop would be the sensed rate of dis-
crepancy reduction in the behavioral loop. Its reference value would be some
desired rate. If a discrepancy is detected between perceived and desired rates, the
person experiences affect. If the perceived rate is too low, affect is negative. If
the perceived rate is above the reference point, affect is positive.
Empirically, this model predicts that two people with the same behavioral
goal will differ in the affect that they experience as a function of the rate at which
they perceive themselves to be moving toward it. A person moving quickly
toward the goal should experience more positive affect than a person moving
slowly toward that goal. This should be true even if the first person is presently
farther away from the goal than is the second. A third person moving even more
slowly toward the goal might experience not positive affect or no affect, but neg-
ative affect. A person who is actually losing ground-moving farther away from
the goal-will feel even more intense negative affect.

Supporting Evidence

Studies bearing on this hypothesis have been conducted by Hsee and Abelson
(1991a, 1991b), who arrived at the idea that velocity relates to affect indepen-
dently of Carver and Scheier. They focused on a specific sort of affect: satisfac-
tion with a desired outcome. In one study, Hsee and Abelson (1991b) had
participants read hypothetical descriptions of paired outcome scenarios and indi-
cate which outcome they would find more satisfying. For example, participants
chose whether they would be more satisfied if their class standing had risen from
the 30th percentile to the 70th percentile over the past 6 weeks, or if it had done
so over the past 3 weeks.
Each participant answered seven questions that paired different outcome
scenarios. The questions tested the role of final outcome, distance changed,
velocity, and direction of change as influences on satisfaction. For our present
purposes, the role of velocity is of the greatest interest (e.g., holding the amount
of salary change constant and varying the time over which it changes). As
Hsee and Abelson (199 1b) predicted, participants preferred improving to a high
final outcome rather than have a constant high outcome; they preferred a fast
790 LAWRENCE ET AL.

improvement over a slow one; and they preferred a fast small improvement to a
slower but greater improvement (this demonstrated the robustness of the time
element in predicting satisfaction, compared to the degree of discrepancy reduc-
tion at the behavioral level). When the change was negative (e.g., when salaries
got worse) participants were more satisfied with a constant low salary than with a
salary that started high and fell to that level; they preferred slow falls (downward
velocity) to fast falls; and they preferred large slow falls to small fast falls.
That study asked participants to read about possible outcomes and imagine
which of a pair they would find more satisfying. A second study (Hsee & Abel-
son, 1991b, Study 2 ) was designed to have participants experience an outcome
actually change in time by watching hypothetical outcomes that changed in the
form of a graph on a computer screen. The computer displayed a bar that moved
vertically along a scale portraying changes in outcome (e.g., the price of a stock
in which the participant had invested). Unlike the first study, in which outcome
scenarios were paired and the participant picked which would be more satisfying,
participants had a reference scenario that they were told to assume was a satisfac-
tion level of 5 on a 9-point scale. They were to make ratings of satisfaction in
comparison to the reference scenario. In this study, distance changed was held
constant, while direction, final outcome, and velocity were varied. The second
study replicated the findings of the first study. Most important, there was a very
strong velocity effect. Participants preferred a fast rate of change when the out-
come was improving and a slow rate of change when the outcome was falling.
The Hsee and Abelson (1991a, 1991b) research is interesting and its findings
are valuable. An important limitation of that work, however, is that the outcomes
participants experienced were hypothetical. Although participants were asked to
imagine themselves experiencing the outcomes, the credibility of the findings
depends in part on the assumption that they actually experienced the outcomes as
having personal relevance. In part for this reason, we felt that further research
was desirable.
In the study reported here, we placed people in a situation with a behavioral
goal: performing well at a task that was portrayed as being relevant to a valuable
skill. The perceived rate of movement toward that goal across trials was experi-
mentally varied. It was predicted that participants in slower moving groups
would experience affect that was negative relative to their previous baseline and
that those in faster moving groups would experience affect that was more posi-
tive than their baseline.
Method
Overview
The present study used a situation in which feedback of progress toward a
desired goal could be plausibly manipulated over an extended period of time. The
 AFFECT AND VELOCITY 791

session was disguised as a study of social intuition, in which participants made a

lengthy series of judgments about ambiguous stimuli. Each block of 10 judg-
ments was followed by performance feedback for that block. Mood was assessed
before the task began, and again at the end of the sixth block, which participants
believed was partway through the session, but actually ended the session.
The phenomenon under study here is subject to an important constraint, which
stems from the relations among starting value, ending value, number of trial
blocks, and velocity. Specifically, these variables are not fully independent. If one
varies velocity, one cannot hold constant all three others. One of them must also
vary. In the study reported here, we chose to keep constant the number of trial
blocks and the ending value. We varied velocity by varying the starting value.
Participants received one of five patterns of performance feedback, converg-
ing such that scores on Block 6 were identical for all participants at 50% correct.
Participants in a neutral condition had 50% on the first and last block, and 50%
on average across all blocks (thus a zero slope overall). Others started with low
scores and gradually moved upward to 50% (thus an upward slope; faster
progress despite a worse starting point). Other participants started with high
scores and fell to 50% (thus a downward slope; slower progress despite an
advantageous starting point). We expected the upward-slope (higher velocity)
groups to shift toward more positive affect, the downward-slope (lower velocity)
groups to shift toward more negative affect, and the level-slope group (intermedi-
ate velocity) to have little or no change in affect. Finally, we expected the affect
changes to be more pronounced in groups with more extreme slopes than in
groups with less extreme slopes.

Participants and Procedure

Eighty-seven University of Miami undergraduates (56 women, 3 I men) par-

ticipated in the study as part of a course requirement. Each participant completed
an individual session and was randomly assigned to one of five experimental
treatments. Four additional persons completed the procedures but were omitted
prior to data analysis because their responses indicated suspicion about the true
nature of the experiment (the decision to omit them was made blind to their
responses to the task).
The experimenter greeted the participant, escorted him or her to the labora-
tory, obtained informed consent, and proceeded to describe the study’s ostensible
purpose. The session was portrayed as part of a broader project on social percep-
tion. This part of the project was said to concern how people assimilate informa-
tion from words, even words they do not know. When words evolved in human
speech (according to the cover story), their sounds conveyed intrinsic meaning.
Although different languages use different sounds to mean the same thing, peo-
ple ostensibly can hear a word in a foreign language and often have at least a
792 LAWRENCE ET AL.

slight intuitive sense of what the word means. This intuitive process was said to
be the focus of this research project.
The experimental task was to be administered by a Macintosh computer. The
program that presented the task would begin by giving instructions about the task.
Afterward, the program would present a list of words from foreign languages that
are unfamiliar to most people in the United States. For each word, the participant
would be asked if it conveys the same meaning as a given English word. The par-
ticipant was to answer Yes or No by pressing the appropriate key (Y or N).
The experimenter continued by saying that evidence exists that a number of
factors can influence a person’s ability to pick up these subtle semantic cues. For
example, slight shifts in a person’s mood can influence sensitivity, as can the fre-
quency with which the person interacts with others in the course of a day
(because of differences in the use of language centers in the brain). In order to
control for these factors in analyzing the data (ostensibly), the participant would
be asked several questions pertaining to some of these factors at various times
during the experiment.
The experimenter gave instructions on how to interact with the computer, and
then left the room. The program started by restating the basic elements of the
cover story, amplifying on the fact that the intuitive process appears to happen
largely outside of consciousness (with people usually being unable to identify
very clearly when it is happening) and also on the social importance of the pro-
cess (a good “receiver” of subtle linguistic cues is likely to have advantages in
many domains, including social sensitivity in relationships and enhanced sensi-
tivity to other people’s motives in business transactions).
The words were to be presented one at a time in phonetic spelling. On each
trial, the participant would be asked whether the word had the same meaning as a
particular English word. The items were presented in blocks of 10. After each
block, the participant received feedback indicating how many responses of that
block were correct. The orienting information indicated explicitly that since this
task is based on deviations from a chance level, people average about 50% cor-
rect. Participants were urged to do as well as they could. The orienting informa-
tion concluded by asking the subjects if they had any knowledge of languages
other than English, Spanish, French, German, or Italian; if so, they were not to
continue, but rather to notify the experimenter (none did).
Mood ratings. At this point, the instructions repeated that a number of factors
are believed to influence sensitivity to linguistic cues, including the frequency
with which a person interacts with others and the person’s transient mood state.
Just as these variables influence the ability to pick up on other social cues, these
factors (ostensibly) influence the ability to pick up on subtle linguistic cues.
Participants then were asked several questions about these and other factors. The
instruction screen also noted that more questions of a similar nature would be
asked at other points during the task.
 AFFECT AND VELOCITY 793

The ratings made by participants at this point included several distractor

questions (e.g., “How many siblings do you have?”; “During the course of the
day, on average how many people do you talk to?”; “What is your concentration
level right now?”), plus the item of interest (a mood rating). In each case, the rat-
ing was made by clicking a point on the scale using the mouse. On the item of
interest, participants indicated their current mood on an 1l-point scale ranging
from 11 (verypositive mood) through 6 (average) to 1 (very negative mood).
It should be noted that the theory on which the present study was based deals
with the creation of affect, and that mood is usually viewed as being a step away
from affect. Mood represents a residual affective quality that is less explicitly
linked to any particular event. We were concerned, however, that to ask partici-
pants directly about feelings arising from their task experiences would be reac-
tive with respect to the study’s cover story. Therefore, we decided to assess
ambient affect as “mood,” rather than as affective reactions to the task.
After making the ratings, the participant began the experimental task. At
the end of each block of items, feedback on that block was given in terms of per-
centage correct on that block. After six blocks, participants again made concen-
tration and affect ratings. The program then instructed the participant to summon
the experimenter. The experimenter and participant returned to the cubicle where
the participant completed a brief questionnaire, one item of which asked how
important he or she felt it was to have good social perception skills. They
responded on a scale ranging from 1 (not at all important) to 9 (very important).
An open-ended item, asking what participants were thinking about during the
task, began to introduce the process of debriefing and probing for suspicion,
which then was carried out more extensively through verbal questioning by the
experimenter.
Behavioral goal and velocity manipulation. The meaning of distance and
velocity as applied to this task are dificult to keep clear. For this reason, we give
this issue explicit attention here. The behavioral goal of the task was to perform
well-to make many correct responses overall. Participants received perfor-
mance scores (in terms of number of correct choices) after each block. The dis-
tance moved toward the goal by the time of the second assessment thus would be
reflected in the accumulation of “correct” responses ostensibly made during
Block 1 through Block 6.
Our predictions, however, were based on velocity. Thus, we established the
feedback patterns of experimental conditions so that they would vary in velocity.
Velocity was operationalized in terms of slopes of the success curves over trial
blocks (Figure 1). Two groups experienced outcomes with positive slopes across
trial blocks, starting with low scores (10% and 30%, respectively) and perform-
ing gradually better across blocks (ending at 50%). These were high and
medium-high velocity conditions. Subjects in a zero-slope condition received a
50% score on the first and last block, and 50% on average across all blocks. This
794 LAWRENCE ET AL.

Y
0
-
0 90 -
n
c
0
80 -
m 70 -
a,
C
0 60 -
c
0
p!
50 -
=b 40 -
a,
0)
30 -
m
c 20 -
C
a,
2 10 -
a,
U I
I I I I I
1 2 3 4 5 6
Trial block
Figure I . Pattern of feedback across trial blocks given to subjects in each of five
experimental conditions. Some participants began with good performances and gradually
fell to average; others began with poor performances and gradually improved to average;
and a control condition consistently performed at an average rate.

condition was intermediate in velocity, compared to the other conditions. The

final two groups had negative slopes, starting well (90% and 70%, respectively)
and gradually worsening across blocks (also ending at 50%). These were the low
and medium-low velocity conditions.
As was implied in the Overview, this pattern of outcomes controls for two
things. First, it controls for the number of blocks completed by the second assess-
ment. This removes the possibility that a difference on the dependent measure
would be a function of the amount of task exposure. The outcome on the final
block was also controlled. That is, the five slopes converged such that scores on
Block 6 were identical for all participants (50% correct). This meant that a differ-
ence on the dependent measure could not be attributed to differences in scores on
that final block (i.e., a recency effect).
Given these two controls, it is impossible to control the starting point for the
five conditions. Thus, the groups had to start with different levels of success in
Block 1. This difference in starting point has a useful function, however. In par-
ticular, it serves to create an inverse relation between velocity and distance trav-
eled toward the behavioral goal by the second assessment. Participants who had
high scores on early trial blocks had the lowest velocity (downward slope)
between Block I and Block 6, but they had also moved the greatest distance
toward the behavioral goal by that time (i.e., they had accumulated the largest
total number of “correct” responses by that time). Those who scored low on early
trial blocks had the highest velocity (upward slope) between the starting block
and second measurement, but they had also moved the smallest distance toward
 AFFECT AND VELOCITY 795

the behavioral goal by that time (i.e., they had accumulated the fewest total “cor-
rect” responses).
The result of this is that participants in the conditions with upward slopes
(whom we predicted to have a positive change in mood) were those farthest from
the behavioral goal (i.e., maximizing correct responses) at the end of Block 6. If
behavioral discrepancy is what matters, these participants should be in the worst
mood. Participants in the conditions with downward slopes (whom we expected
to have a negative change in mood) are those with the least distance from the
behavioral goal at the end of Block 6. If behavioral discrepancy is what matters,
these people should be in the best mood.
The design of this experiment did not incorporate any group for which veloc-
ity was negative, that is, where participants were actually losing ground on their
behavioral goal. It certainly would be possible to create such a condition-for
example, by imposing a penalty for incorrect responses that would be subtracted
from correct responses. However, we wished to focus on conditions that incorpo-
rated forward movement for all subjects-reducing behavioral discrepancies for
everyone, but at different rates?

Results

Initial mood ratings (prior to Block I ) were relatively positive overall ( M =

8.07 on a scale ranging from 1 to 1 1, SD = 1.84). Analysis of these ratings
revealed no difference between groups (as no gender effect emerged from pre-
liminary analyses, the data reported here are combined across gender). Nor did
groups differ on post-experimental ratings of how important they felt it was to
have good social-perception skills. Importance ratings overall averaged 7.23
(SD= 1.34), indicating that participants believed that the skill assessed by the
task was important. We infer from this that participants were relatively fully
engaged by the task.
The measure of greatest interest was the change in mood that took place from
the initial measurement to the measurement after trial Block 6 (Figure 2).
ZWe believe that the most straightforward interpretation of the instructions to participants is that
the goal of the task is to make as many correct responses overall as possible (as we intended). It might
be argued, however, that for at least some people, the goal adopted was having high scores on later
trial hlocks-that is. being skilled at the task by the end of the session. If so, upward-slope conditions
would represent movement toward that behavioral goal (thus positive velocity), and downward-slope
conditions would represent movement away from that behavioral goal (thus negative velocity). All
groups would still have identical behavioral discrepancies after Block 6 (i.e., a score of 50% correct).
This view would change the experiment from one in which fast velocity was compared to slow veloc-
ity (as we interpret it) to one in which positive velocity was compared to negative velocity. However,
this perspective would not change OUT prediction greatly. Upward-slope conditions should still lead to
positive-direction affective change, and downward-slope conditions should still lead to negative-
direction affective change. Under this view, the negative-direction change should simply be more
intense, since there is reverse movement instead ofjust slow movement.
796 LAWRENCE ET AL.

Strong Weak Weak Strong

upward uoward Flat downward downward
slope slope slope slope slope

Mood
change
r
-";t
-1.5

Figure 2. Mood change from pre-task to post-trial-block 6 (standard error is indicated by

error bars).

ANOVA on this measure yielded a significant effect for experimental condition,

~ .009. A contrast analysis (using weightings of 1.0, 0.5, 0.0,
F(4, 82) = 3 . 6 2 , =
-0.5, and -1.0) confirmed that there was a significant linear trend across the five
groups, t(82) = 3.61, p = .001, such that those whose performance feedback dis-
played an upward slope over blocks tended to have the most positive mood
change, and those whose performance feedback displayed a downward slope over
trial blocks tended to have the most negative mood change. By Duncan multiple
range test for individual group comparisons, the two downward-slope groups had
adverse changes in mood that differed from the changes reported by the other
three groups, which did not differ significantly from one another.

Discussion

Participants in this study received a series of messages reporting their perfor-

mances on six blocks of trials on a task. The messages indicated one of five pat-
terns of performance. The patterns varied from a high-velocity performance
(progressing toward the behavioral goal with an upward slope) to a low-velocity
 AFFECT AND VELOCITY 797

performance (progressing toward the behavioral goal with a downward slope).

Participants who experienced downward slopes across blocks shifted their moods
in a negative direction. Participants who experienced upward slopes across
blocks tended to shift toward more positive moods. This latter tendency was
probably constrained by a ceiling on mood ratings, however, as mood was posi-
tive overall among participants at the beginning of the task. Despite this con-
straint, a contrast analysis confirmed that there was a significant linear trend in
affect change among the groups as a function of variation in slope.
These findings are consistent with the conceptual model with which we began
(Carver & Scheier, I990a, 1998). They are also consistent with the results of
Hsee and Abelson (199 la, 1991b) in supporting the position that change over
time is a more important influence on affect than is distance from a behavioral
goal. Our findings extend those of Hsee and Abelson, however, in showing that
the effect occurs among people actively engaged in a performance task with
direct personal relevance, in addition to occurring among people judging hypo-
thetical situations.3
It is worth emphasizing that the findings are not easily interpreted within a
framework that takes into account only the degree of behavioral discrepancy
from the goal at the point of final mood assessment. That is, participants who
reported the most positive mood at the end of the sixth block were those who had
accumulated the fewest total correct responses, whereas those reporting the least
positive mood at that time were those who had accumulated the most total correct
responses. Thus, affect change did not depend on mere cumulative success. Nor
did affect change reflect the last piece of information that was received. All par-
ticipants experienced identical success rates on Block 6, yet the groups differed
in the moods they reported.

Limitations

We should be explicit that there are important limitations on what was

found here. We are not asserting that this experiment examined all potential con-
tributors to affective experience. For example, Hsee, Salovey, and Abelson
(1994) argued that acceleration makes a contribution to satisfaction, with posi-
tive acceleration values associated with more satisfaction and negative values
associated with less satisfaction. The study reported here was not intended
to bear on this issue, nor on several other issues pertaining to potential influences

3A reviewer raised the possibility that the findings support a model in which the perceived likeli-
hood of attaining the goal (rather than velocity) produces affect. That is, groups with an upward slope
might be more confident of performing well over the full task period than groups with a downward
slope. Such an interpretation would be consistent, however, with the model under study. That is, in
this model, affect and confidence about outcomes are viewed as parallel readouts of a single process-
ing principle involving velocity (for more detail, see Chapters 8 and 10 of Carver & Scheier, 1998).
798 LAWRENCE ET AL.

on affective experience (see Hsee et al., 1994, for further discussion of such
issues).
Further, in this experiment we ignored the considerable divergence among
affective qualities that exist in human emotional experience (e.g., Ellsworth &
Smith, 1988; Frijda, 1986; Izard, 1977; Ortony et al., 1988; Roseman, 1984).
Affect varies in both dramatic and subtle ways, as a function of the variables that
surround the situation in which it is occurring (Ortony et al., 1988). This experi-
ment was not intended to explore those variations, but rather to minimize their
impact, so that the hypothesis of our focal interest could be examined with mini-
mal intrusion. We do have further thoughts on aspects of the divergence among
feeling qualities, however, which are addressed in detail elsewhere (Carver &
Scheier, 1998).4
Another limitation is that the findings of this study were not as strong as they
might have been. Although contrast analysis confirmed the existence of a linear
trend across experimental groups, the groups did not differ as distinctly from one
another as we had hoped. There are several possible reasons for this. Presumably,
people vary in their reference values for both velocity and behavioral perfor-
mance on tasks such as this; if we had collected information about such individ-
ual differences, we might have reduced error and found stronger effects. Indeed,
it is possible that the experiences of the six trial blocks caused some readjustment
in reference values in the various groups. If we had had an online measure of ref-
erence values at various phases of the task, it would have made the analyses even
more precise. Another issue here concerns the fact that the theory pertains to the
creation of affect, whereas we asked participants to rate their moods. Although
the two are certainly related, they are also differentiable. Perhaps the decision to
collect these ratings as mood ratings increased measurement error.

Velocity,Gain, and Loss

We close by noting an interesting resemblance between our findings and

those of a 1965 study by Aronson and Linder. In that study, a confederate’s eval-
uations of participants were manipulated across trials. Some received consistent
positive evaluations, some consistent negative evaluations, and for others evalua-
tions shifted from positive to negative or from negative to positive. Those whose
evaluations shifted from negative to positive later reported liking the confederate
more than those whose evaluations were consistently positive. (Conceptually,

4A specific issue that might occur to readers is how affect might vary as a function of attributions
for the cause of the outcome. We have taken the position that the core of affective experience arises
with regard to velocity itself. and that perceptions of cause influence more subtle variations among
feelings, such as that between pride and gratitude (Carver & Scheier, 1998; see also Carver et al.,
2000). This general view resembles statements made by Weiner (1986) and by Smith and Ellsworth
( I 987).
 AFFECT AND VELOCITY 799

this comparison resembles that between our control group and our upward-slope
groups.) An opposite tendency occurred in the other two conditions: Those who
experienced a shift from positive to negative evaluation tended (though not sig-
nificantly) to like the evaluator less than did those whose evaluations were con-
sistently negative. (Conceptually, this comparison resembles that between our
control group and our downward-slope groups.)
It is of some interest that the pattern emerged for Aronson and Linder (1965)
only when participants were asked to report their feelings about the evaluator; that
is, their “gut response.” No such effect occurred on participants’ ratings of the
evaluator’s positive and negative personal [Link] and other ancillary
data appeared to indicate that participants’ ratings of the confederate reflected
their feelings (their affective responses) rather than cognitive evaluations.
Aronson and Linder (1965) interpreted the finding in terms of shifts in multi-
ple drive states (deriving from the valence of the other person’s evaluations),
rather than in terms of contrast effects. They did so partly because an additional
post hoc group of participants who had a shift from neutral (which does not
induce a drive) to positive failed to display enhanced liking. It seems reasonable
to ask, however, whether that really settles the matter. Perhaps the failure to find
an enhancement of liking in that group simply reflects the fact that the contrast
was far less sharp for that group.
The Aronson and Linder (1965) pattern of findings seems readily interpreted
in terms of the velocity model from which we proceeded here. Participants who
heard consistently positive evaluations were at the desired goal (being well
regarded) throughout the experience. Those who heard evaluations that shifted
from negative to positive experienced a change. This change implies a higher
velocity. The result was positive feeling, which was expressed as liking for the
evaluator.

Conclusion

In sum, the findings reported here join with previous evidence (Hsee &
Abelson, 1991a, 1991b) in suggesting that the subjective experience of affect
reflects a velocity function. Specifically, affect appears to reflect the rate of the
person’s movement toward (or away from) a salient goal. In this way, affect ties
the goal-related aspect of motivation to the dimension of time (see also Carver &
Scheier, 1998). Affect is an indicator that the subjective experience of the now
derives not solely from the present moment, but partly from relations between the
present and a broader temporal span.

References

Arnold, M. B. (Ed.). (1970). Feelings and emotions: The Loyola Symposium.

New York, NY: Academic Press.
800 LAWRENCE ET AL.

Aronson, E., & Linder, D. (1965). Gain and loss of esteem as determinants of
interpersonal attractiveness. Journal of Experimental Social Psychology, I ,
156-171.
Austin, J. T., & Vancouver, J. B. (1996). Goal constructs in psychology: Struc-
ture, process, and content. Psychological Bulletin, 120, 338-375.
Cantor, N., & Kihlstrom, J. F. (1987). Personality and social intelligence. Engle-
wood Cliffs, NJ: Prentice-Hall.
Carver, C. S., Harris, S. D., Lehman, J. M., Durel, L. A., Antoni, M. H., Spencer,
S. M., & Pozo-Kaderman, C. (2000). How important is the perception of
personal control'? Studies of early stage breast cancer patients. Personality
and Social Psychology Bulletin, 26, 139-150.
Carver, C. S., & Scheier, M. F. (1 98 1). Attention and selfregulation: A control-
theory approach to human behavior. New York, NY Springer-Verlag.
Carver, C. S., & Scheier, M. F. (1990a). Origins and functions of positive and
negative affect: A control-process view. Psychological Review, 97, 19-35.
Carver, C. S., & Scheier, M. F. (1990b). Principles of self-regulation: Action and
emotion. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of motivation
and cognition: Foundations of social behavior (Vol. 2, pp. 3-52). New York,
NY: Guilford.
Carver, C. S., & Scheier, M. F. (1998). On the [Link] of behavior. New
York, NY: Cambridge University Press.
Carver, C. S., & Scheier, M. F. (1999). Themes and issues in the self-regulation
of behavior. in R. S. Wyer, Jr. (Ed.), Advances in social cognition (Vol. 12,
pp. 1-105). Mahwah, NJ: Lawrence Erlbaum.
Elliott, E. S., & Dweck, C. S. (1988). Goals: An approach to motivation and
achievement. Journal of Personality and Social Psychology, 54, 5- 12.
Ellsworth, P. C., & Smith, C. A. (1988). From appraisal to emotion: Differences
among unpleasant feelings. Motivation and Emotion, 12,27 1-302.
Emmons, R. A. (1986). Personal strivings: An approach to personality and subjec-
tive well being. Journal of Personality and Social Psychology, 51, 1058-1068.
Frijda, N. H. (1986). The emotions. Cambridge, UK: Cambridge University Press.
Gollwitzer, P. M., & Bargh, J. A. (Eds.). (1996). The psychology of action: Link-
ing cognition and motivation to behavior. New York, N Y Guilford.
Higgins, E. T. (1987). Self-discrepancy: A theory relating self and affect. Psy-
chological Review, 94, 3 19-340.
Higgins, E. T. (1 997). Beyond pleasure and pain. American Psychologist, 52,
1280-1300.
Hsee, C . K., & Abelson, R. P. (1991a). The relative weighing of position and
velocity in satisfaction. Psychological Science, 2,263-266.
Hsee, C. K., & Abelson, R. P. (1991b). Velocity relation: Satisfaction as a func-
tion of the first derivative of outcome over time. Journal of Personality and
Social Psychology, 60,341-347.
 AFFECT AND VELOCITY 801

Hsee, C. K., Salovey, P., & Abelson, R. P. (1994). The quasi-acceleration rela-
tion: Satisfaction as a function of the change of velocity of outcome over
time. Journal of Experimental Social Psychology, 30,96- 1 1 1.
Izard, C. E. (1977). Human emotions. New York, N Y Plenum.
Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. American
Psychologist, 39,341-350.
Klinger, E. (1975). Consequences of commitment to and disengagement from
incentives. Psychological Review, 82, 1-25.
Lazarus, R. S. (1991). Emotion and adaptation. New York, N Y Oxford Univer-
sity Press.
MacKay, D. M. (1963). Mindlike behavior in artifacts. In K. M. Sayre & F. J.
Crosson (Eds.), The modeling of mind: Computers and intelligence (pp. 225-
241). Notre Dame, IN: University ofNotre Dame Press.
MacKay, D. M. (1966). Cerebral organization and the conscious control of
action. In J. C. Eccles (Ed.), Brain and conscious experience (pp. 422445).
Berlin, Germany: Springer-Verlag.
Markus, H., & Nurius, P. (1986). Possible selves. American Psychologist, 41,
954-969.
Martin, L. L., & Tesser, A. (Eds.). (1996). Striving and feeling: Interactions
among goals, affect, and self-regulation. Hillsdale, NJ: Lawrence Erlbaum.
Ortony, A., Clore, G. L., & Collins, A. (1988). The cognitive structure ofemo-
tions. New York, N Y Cambridge University Press.
Palys, T. S., & Little, B. R. (1983). Perceived life satisfaction and the organiza-
tion of personal project systems. Journal of Personality and Social Psychol-
ogy, 44,122 1- 1230.
Pervin, L. A. (1983). The stasis and flow of behavior: Toward a theory of goals.
In M. M. Page & R. Dienstbier (Eds.), Nebraska Symposium on Motivation:
Vol. 31 (pp. 1-53). Lincoln, NE: University of Nebraska Press.
Pervin, L. (Ed.). ( 1989). The goals concept in personality and social psychology.
Hillsdale, NJ: Lawrence Erlbaum.
Powers, W. T. (1973). Behavior: The control ofperception. Chicago, IL: Aldine.
Roseman, I. J. (1 984). Cognitive determinants of emotions: A structural theory.
In P. Shaver (Ed.), Review of personality and social psychology (Vol. 5,
pp. 11-36). Beverly Hills, CA: Sage.
Scherer, K. R., & Ekman, P. (Eds.). (1984). Approaches to emotion. Hillsdale,
NJ: Lawrence Erlbaum.
Simon, H. A. (1 967). Motivational and emotional controls of cognition. Psychol-
ogy Review, 74,29-39.
Smith, C. A., & Ellsworth, P. C. (1987). Patterns of appraisal and emotion related
to taking an exam. Journal of Personality and Social Psychology, 52,475-488.
Vallacher, R. R., & Wegner, D. M. (1985). A theory of action identification.
Hillsdale, NJ: Lawrence Erlbaum.
802 LAWRENCE ET AL.

Wegner, D. M., & Bargh, J. A. (1 998). Control and automaticity in social life. In
D. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), Handbook of social psychology
(4th ed., pp. 446-496). Boston, MA: McGraw-Hill.
Weiner, B. (1986). Attribution, emotion, and action. In R. M. Sorrentino & E. T.
Higgins (Eds.), Handbook of motivation and cognition (pp. 281-3 12). New
York, N Y Guilford.
Wiener, N. (1948). Cybernetics: Control and communication in the animal and
the machine. Cambridge, MA: MIT Press.