https://doi.org/10.24265/liberabit.2022.v28n2.534
ARTÍCULO DE INVESTIGACIÓN
Equity Judgments: Context Effects in Gains and Losses
Juicios de equidad: efectos de contexto en ganancias y pérdidas
Raúl Reyes-Contrerasa,*
https://orcid.org/0000-0003-1308-4365
Carlos Santoyo Velascoa
https://orcid.org/0000-0002-2817-3793
aFacultad de Psicología, Universidad Nacional Autónoma de México, México
Autor corresponsal:
raulreyescontrerasunam@gmail.com
Para citar este artículo:
Abstract
Background: Previous studies have evaluated mathematical
models of equity
judgments of two hypothetical
employees with distinct merits. They found that the model of proportionality adequately described the data based on an algebraic additive rule
of information integration. Nevertheless, there is a lack of evidence concerning the effect of a context of
monetary losses on the rules of information integration. Objective: To assess the effect of monetary gains and losses on equity judgments, and the effect
of the personal context in which the participants are involved in decision-making. Method: A repeated measures design with within-subjects factors was utilized:
7 levels of A’s merit x 5 levels of B’s merit x 2 contexts (gains and losses) x 2
between-subject factors (third-party allocation and self-allocation). Results: Statistically significant differences appeared between third- party allocation and self-allocation
conditions for gains [F (1,38) = 216.18, p <
.001, =
.85] and losses
[F (1,38) = 110.45, p <
.001, =
.71]. Conclusions: The
additive rules of integration appeared in the gains scenario
while the subtractive rules together with an aversion to inequity
were observed in the losses context.
Keywords: information integration theory; equity judgments; social psychophysics; monetary
gains; monetary losses.
Resumen
Antecedentes: estudios previos evaluaron modelos matemáticos sobre juicios de equidad en dos empleados hipotéticos que diferían en sus niveles de mérito. Estos encontraron que
el modelo de proporcionalidad
describía adecuadamente los datos basándose en una regla algebraica aditiva de integración de información. Sin embargo,
existe una falta de evidencia sobre el efecto del contexto de las pérdidas monetarias sobre las reglas de integración de información. Objetivo: evaluar el efecto del contexto de ganancias y pérdidas monetarias sobre los juicios de equidad, y el efecto del contexto personal en donde el participante se ve involucrado en la toma de decisiones. Método: se empleó un diseño de medidas repetidas con
los factores intrasujeto: 7
niveles de mérito de A x
5 niveles de mérito de B x 2 contextos (ganancias y pérdidas) x 2 factores intersujeto (distribución a terceros y distribución propia). Resultados: se encontraron diferencias estadísticamente significativas entre las condiciones de asignación a terceros y asignación propia en relación con las ganancias [F (1,38) = 216.18, p <
.001,
= .85] y pérdidas [F (1,38) = 110.45,
p < 0.001, = .71].
Conclusiones: las reglas aditivas de integración aparecieron en el contexto de las ganancias, mientras que las reglas sustractivas, así como aversión a la inequidad, se observaron en el contexto de las pérdidas.
Palabras clave:
teoría de integración de información; juicios de equidad; psicofísica social; ganancias monetarias; pérdidas monetarias.
Introduction
One of the conceptual approaches
relevant to the study of equity
arises from behavioral sociology. Homans
(1958) established the notion of universal proportionality between costs and benefits. However, differences emerge among societies due to the diversity that
exists when it comes to conceiving investments, rewards, and the way in which they are ranked. This led Homans to identify three elements of
conceptual proportionality: a) the
cost-benefit ratio, b) natural environmental experiences, and c) exchange experiences.
Later,
Adams’ equity theory stated that the individuals’ sense of fairness and equity
depends on the balance between their
contributions and their rewards.
Thus, the key elements of his theory are inputs and outputs. Inputs are defined as «a person perceives the way in which his or her contributions affect the exchange, consequently
expecting a fair profit» (Adams,
1965, p. 280). In the present study, we
refer to inputs as «merits». In contrast, outputs are conceived as what the individual receives in the exchange,
for example, profits,
salary allocations, or personal satisfaction. Hence, equity is perceived when the input/output ratio of a person (A) in exchange is like the input/output ratio of
the other person (B) with whom the
interaction takes place. The following
equation formalizes that relationship:
Adams (1965) further proposed that
an inequitable relationship implies
emotional and motivational elements that lead participants to make important decisions during the exchange, based primarily on the assumption that every inequitable relationship is aversive,
and that the affected participants
will employ some strategy (or
strategies) to reduce the inequity. One of
Adams’ most important contributions was his detailed discussion of the numerous
factors that determine the
values of inputs and outputs. Adams’ application
setting was restricted to the industrial sphere,
but the model can be applied in different surroundings, including the academic world (Anderson, 1976). So, for example, the inputs –or
merits– of professors include
preparation, teaching, research, community
service, and other aspects that must always be considered, such as age, personality, and the academic
context. Their outputs
can include prestige
in different forms, promotions, and funding for research,
among others. In the industrial context, these outputs appear as salary increases,
opportunities for growth, and labor benefits
like major medical insurance and the support
of one’s supervisor (De Gieter et al., 2012). All these factors
are evaluated within everyone’s frame of reference in a value system that
makes it possible to understand the emotions that are involved
regarding equity or inequity. In a fair distribution of resources between one person (A) and
another (B), each one can exercise a
claim to the inputs. Those claims
will be valued equally since this division is
considered an arrangement between two factors and it is susceptible to empirical testing. Under these constraints
of the equity model, Anderson and
Farkas (1975),
Anderson (1976), Farkas and Anderson (1979), and Singh (1985) have proposed the following model
in which a proportional part of the output equals the proportional contribution of the input.
Equation 2 presents the resulting equity
relation.
Conceptually, equations 1 and 2 are distinct. Adams’ equation implies that the initial
comparison between output and input
is made by each person individually, followed by a second comparison of those two individuals using those two
input/output points. Equation 2, in contrast, implies a reverse
order in the comparisons; that is, first between
the individuals for each
input and output separately, and then between
them in terms of interpersonal proportions.
Mathematically, however, these two
equations are similar, and one can be derived
from the other.
Psychologically, as outlined above, they represent distinct structures of comparison. This equivalence is based
on the ideal of equity. In the field of study of inequity judgments, the extensions of these models lead to contrasting predictions (Anderson, 1976).
Farkas and Anderson’s proposal (1979) is derived from the previous equation, which allows us to make
predictions about the amount that should be given to employee
A according to the merit level of employee B, since on the experimental tasks they
are asked to divide a fixed amount
(M) between them such that M = A + B. This can be written as:
This research is framed in the field of
decision- making processes. The most important models
in this domain assume that an assessment process of the information
exists, which originates in the
environment (Gigerenzer & Goldstein,
1999; Goldstein & Gigerenzer, 2011; Kahneman, 2011; Kahneman & Tversky, 1979). To study the assessment
of environmental stimuli,
we use the information integration theory (IIT).
IIT was proposed by Anderson (1976). This theory is
concerned with how people integrate information from two or more stimuli
to give a numerical response.
This theory focuses
on assessing the unobservable psychological process involved in making
judgments. IIT is developed around four concepts:
stimulus valuation, stimulus integration, cognitive algebra,
and functional measurement (Anderson, 2013). Stimulus
valuation is simply defined as the process of extracting information from a physical
stimulus and transforming it into a psychologically derived value. Stimulus integration
means that, in a natural environment, most responses are based on multiple interacting
factors. It is rare to find one predictor
of behavior. IIT attempts to analyze how
these factors are
integrated psychologically. These
stages remain unobservable, so with the cognitive algebra nested in the
integration phase, it is the process in which observers combine multiple factors
into a numeric response using algebraic rules. And finally, functional
measurement is the combination of the weighting factors in the valuation process
and ends with the rules of information integration. Anderson (1996, 2008, 2012,
2013) found that there are three rules of information integration: additive,
multiplicative, and averaging. Our study centers on the additive rule. The additive function for information
integration operates when the stimulus and its psychological counterpart have a
linear relationship that is maintained throughout the process of information
integration. This implies that the variables do not interact but are merely
added up. Given the algebraic properties of the information integratio
process, when making graphs of the answer patterns of all the factorial combinations,
a parallel line is observed. However, this use of the term «parallel» does not
imply that the lines have the same slope.
According to the concepts of IIT, rather, it means that the Euclidean distances
are similar foreach arranged point.
In recent years, IIT has been found to
be useful for evaluating
complex cognitive processes like sleep cognitive algebra (Mairesse
et al., 2010), marketing and financial
value (Hilgenkamp & Shanteau, 2010), promoting physical exercise (Brengman et al., 2010),
somatic anxiety (Moore et al., 2010), interpersonal relations (Theuns et al., 2010), recognizing emotions in faces
(Pereira et al., 2016), body postures
and emotions (Silva & Oliveira, 2016), aversion to losses (Viegas et
al., 2016), perception of financial risk (Laskov-Peled & Wolf, 2016), moral development in sexuality (Hommers & Görs, 2016),
ethics in politics
(Mullet et al., 2016), and dilemmas related to public goods (Acevedo et al., 2019).
Anderson (1976) used IIT to evaluate the numerical allocations that people make in an experimental
preparation of the following type: the situation
was of a hypothetical university in which the participants
distribute resources between two professors who differ in their merit levels. The productivity
of the two professors, A and B, was described
as follows: both worked on the same task and
the participants received the information on the performance of A and B (as in phase 1); thus, it was a factorial experiment of 5 (A’s merit levels)
x 5 (B’s merit levels).
The participants’ task consisted in assigning a profit to employee B as a function of his/ her own merit and that of A. The results
of this research indicate that the
payment given to B is a direct
function of his/her merits. The profile graphs
of A’s 5 merit levels by B’s 5 merit levels show a clear tendency towards parallelism that was supported by a test [F (16,368) = 2.99, p <
.05]. The main findings in that
research indicate that IIT is useful as a model
and method for the study
of equity and that the additive rule shows that the
participants add up the values
for the merit levels algebraically.
Using a similar method, Mellers (1982) posed a situation
in which the participants had to allocate salaries to professors in a hypothetical university where the merit level of professor A could take 7 distinct levels (.5, 1, 1.5, 2, 2.5, 3, and 3.5) while that of professor B
had only 4 levels (.5, 1.5, 2.5, and 3.5). Three different
budgets were considered: USD 20 000, USD 40 000,
and USD 80 000. The aim of that research was to evaluate distinct equity models. The participants were instructed to assign salaries
to professor B as a function
of his/her own merit level and that of A. Results indicated
a significant interaction between the respective merits of professor
A and those of professor
B [F (18,666) = 16.53, p <
.05]. The profile
graphs of the professors’ respective merit levels showed
a clear tendency toward parallelism
suggestive of an additive
rule of information integration. The equity
model proposed by Anderson (1976) is thus consistent
since it can explain the interaction that occurs
in a psychophysical function of judgment.
Santoyo
& Bouzas’ (1992) study expanded
Mellers’ (1982) works to various contexts in
which university students
were asked to assign salaries to two professors at a different merit level.
In that case, however, the
researchers aimed to evaluate whether differences existed due to the amount to be distributed, so
two different amounts were posited: MXN 1 million and MXN 2 million, in such a way that the experiment resulted
in a factorial preparation of A’s 7
levels x B’s 4 levels x 2
budgets through a three-factor repeated measures ANOVA (2 budgets x 7 merit levels of A x 4 merit levels
of B). They evaluated the main effects
of budget [F (1,1769) = 2318, p <
.001], of A’s merit [F (6,1769) = 182, p < .001], of B’s merit [F (3,1769)
= 438, p <
.001], and the interaction between the respective merits of A and B [F (18,1769) = 6326, p < .001]. Since no differences were found due to the interaction between the amount of the budgets and A’s and B’s merits [F (18,1769)
= 1.74, p < .32], the researchers
concluded that the model of
proportionality (Adams, 1965; Anderson, 1976) predicts that an individual’s allocation of resources will
be a linear function of the relative
merits of the other people who are compared, as can be observed in the results
obtained. Researchers found that the budget has no effect
on the rules of information integration just like Mellers’ (1982) study.
A second experiment by Santoyo et al.
(2000) continued the study of equity. In that work, the main target
consisted in evaluating the effect of a context of inflation on the process of assigning resources to professors at a hypothetical university
where their respective merits varied.
Applying a methodology like one of the earlier studies
in designing the instrument, in
the new experimental situation the
participants were instructed
to assign salaries to professors with distinct merit levels, but on the same scale as in the prior study and considering a factor not included in the previous
exercise; namely, a level of
inflation. Professor A had 7 merit levels (.5, 1, 1.5, 2, 2.5, 3, and 3.5) while
professor B had only 4 (.5, 1.5, 2.5, and 3.5), all under
conditions of inflation of 10%, 50%, and 100%. That study design
resulted in a factorial design
of 7 (professor A’s merit
levels) x 4 (professor B’s merit levels)
x 3 (inflation levels) and x 2 (budget levels:
MXN 1 200 000 and MXN 2 400 000). As in the earlier work, a three-section written document was used. The first
part contained the instructions for
the participants, the second
presented the items, and the third consisted of an answer sheet. Based on the
repeated measures ANOVA, the authors
evaluated the effect of the variables
of the budget level and annual inflation, but neither statistically significant principal nor interaction effects were found. The authors concluded, therefore, that no differences were found among the
diverse levels of inflation or
concerning the context of the budgetary
levels used. As occurred with the repeated
measures ANOVA, there were no differences regarding the budgetary levels stipulated. The study did determine, however, a tendency towards
parallelisms like those
reported by Anderson (1976) and Mellers (1982), which is indicative of an additive
rule of integration. Finally,
they found an effect in which the low merit levels of employee B tended to be assigned
higher amounts than those the model of equity predicted. This also occurred with A’s
merit levels below 2.5 and, in the opposite
case, with higher
merit levels (3 and 3.5)
where lower amounts than those predicted
by the equity model were given. This seems to suggest that higher merit
levels were being «punished,» while lower ones were «compensated».
Hofmans
(2012) made a partial replication of Anderson’s
(1976) study to evaluate the various rules of integration. With this goal in mind, a study with a sample of 58 participants and a factorial design of 5 x 5 stimuli was designed. For each combination of stimuli, the
participants were instructed to assign a fixed
amount of money to employee A. The results showed
an additive integration pattern for employee
A and employee B that was supported
by a repeated- measures ANOVA: [F (4,228) = 118.06, p <
.001]
and [F (4,228)
= 107.39, p < .001], respectively. However, the main result suggested that
the use of cluster analysis
to identify the different rules of information integration found that 53 participants had followed the additive rules of integration while the other 5 had
assigned the same amount of money in all
possible combinations, such that the results for employee A were [F (4,12)
= < 1, p = .368)] and for employee
B [F (4,12) = < 1, p = .853)].
This could be interpreted as indicating that those 5 participants considered the experimental situation
«inequitable», so they allocated
similar amounts of money without considering the merit level of the hypothetical employees. The profile chart of the average allocation of this group of participants had the
appearance of a group of overlapped horizontal lines.
Reyes-Contreras & Santoyo (2017) had
the main aim to evaluate the effect
of a situation of monetary losses on equity exchanges
by generating two contexts:
one of the gains, and the other of losses. To
this end, they posed an experimental task in a
hypothetical industrial automotive setting. In both contexts, the participants were asked to
imagine that they were human resources
directors and they had to increase their salary in the gains
context due to the profits
earned in the preceding year.
In the losses context, they had to reduce the salary due to the low car sales of the preceding year. In the
case of gains, the participants were
asked to distribute resources between
two employees, A and B, as in previous studies, but A’s merit levels were .5, 1, 1.5, 2, and
2.5, and B’s were .5, 1.5, and 2.5. This resulted in a factorial arrangement of A’s 5 merit
levels x B’s 3 merit levels. The same
stimuli were used in the case of
losses. To evaluate the effect of the context, the authors performed a repeated-measures ANOVA of A’s 5 merit levels x B’s 3 merit levels x
2 contexts (gains, losses). They
determined the main effects of the
factor context [F (1,36) = 6.453, p < .05], A’s merit [F (4,33) = 6.213, p < .05], and B’s merit [F (2,35) = 22.887, p < .001], and found that the
losses scenario influenced resource distribution. As in previous studies, they also observed an
additive rule of integration,
together with a tendency on the part
of the participants to grant higher salaries to the employees with lower merit levels
and lower salaries
to those with higher merit levels, suggestive of a subtractive rule of integration. Similarly, the participants applied lower discounts to the lower merit levels
and higher discounts to the higher
merit levels compared to the predictions of the equity
model.
One aspect of contextual interest that those earlier
studies did not address, however, is the situation of emitting
judgments a third party –not the participant her/himself– is involved. For this reason,
the present experiment was designed to analyze if,
when the participant is involved in the situation, the individual perspective
produces additional bias to the
information integration process.
In summary, previous
research has studied
equity judgments
from an impersonal perspective; that is, the participants were not involved
in the psychophysical judgments they were instructed to effectuate. In general, those studies found additive rules of integration
and an effect that «compensated» lower merit
levels and «punished» higher ones. However, evaluating
a personalized perspective was missing from
those reports; that is, when the participants
themselves are involved in the decisions they are asked to make. We hypothesize that this will be an important
element in the equity model by involving the exchange experiences and the consequences of inequity for individual participants.
The
present study: A contextual approach to the
study of equity
exchanges
For this experiment, we adopted a contextual approach
(Bevan, 1968) because
it allows the systematic
study of two types of stimuli: focal and background.
Focal stimuli are those that a person identifies immediately. Background stimuli constitute the specific surrounding conditions that give meaning to the focal stimuli. Kahneman and Tversky’s (1979)
studies, for example,
showed that the value of psychophysical estimates of money as a focal stimulus is asymmetrical and will be assessed depending on the context in which they are presented; that is, gains
or losses (background stimulus). Relevant literature suggests that the way in which money is
valued psychologically differs
in the case of gains versus that of losses (Kahneman & Tversky, 1979; Krueger et al., 2011).
Little evidence exists, however, on how information on equity judgments is
integrated when the participants find themselves in a context
of monetary losses in which
they are involved and that will be
affected by the decisions they make. Thus, the
main objective of the present study consists in evaluating the effect of resource allocation (salary increase
or decrease) and context (background stimulus) of two psychophysical tasks using the conjoint measurement method: the participants were instructed to distribute salaries between two hypothetical
employees –employee A and employee B
(focal stimuli)– and then between two employees where the participant her/himself was involved with a hypothetical employee
B.
We hypothesized that in the gains context the additive
rule of integration would appear, while in the losses context a distinct kind of rule
of information integration would appear.
Method
Sample
The study was conducted with a
convenience sample of 40 college
students at a private university in
western Mexico City. About 50% of the sample
were females. The average age of the study subjects was 19.2 years old (SD = .5).
Instruments
Four written instruments
were prepared to represent the experimental situations. They
included previous exercises
to familiarize the participants with
the task, the gains or losses context,
the items involved, and an identification code. The
previous exercises helped the participants become familiar with the type of answer required in the
items. We created four resource
allocation contexts, two for gains and two for losses. In the gains context,
the situation was
that the participants worked in the automotive industry and that sales in the previous
years had been extraordinary, allowing the automotive plant to distribute additional resources to its employees. In the first
case –that is, third-party gains with hypothetical employees A and B– the study subjects were asked to determine the salary increases
for those two employees.
In the case of personal gains (i.e., where subjects played the role of employee A), the situation was similar, except that they were told their opinion about the salary increase they should receive
as needed. In the case of
losses, the participants were told
that the company had lost market share, so to
avoid laying off employees it had decided to reduce work hours, though this would have
implications on the salaries that
employees would receive. In the third scenario
–third-party losses– subjects
were asked to distribute a discount between two employees, while
in the case of personal
losses they were instructed
to give their opinions on the discounts they felt they deserved after comparing the
merits of a third party to their own. The instructions were identical
in all four instruments. Each merit level was exemplified and an estimated percentage of employees
in that merit level was distributed in a normal curve to replicate
Mellers’ (1982) study. Finally, the
items were based on the same values of merit
–.5, 1, 1.5, 2, 2.5, 3, and 3.5– used in previous studies by Mellers (1982),
Reyes-Contreras & Santoyo (2017), Santoyo et al. (1992, 2000), and Pulido et al. (2007). The amount of MXN 11,500 to be
distributed monthly was obtained from
the average of the tabular salaries of full-time academic technicians, assistants, and associates at the National
Autonomous University of Mexico (UNAM, 2017) in effect as of February 1st, 2017. The instructions in the items indicated that the participants were to distribute that amount. «Rounding up» was
allowed to keep the participants’ answers simpler. This amount was used, as well, because
the average monthly salary of Mexicans in 2016, according to the National Institute of Geography and Statistics (2017), was around MXN 9900 for 70% of the
Mexican population. The goal was to work with «realistic» current monetary amounts that were close to average
family incomes. This measure gave greater ecological validity
to the experimental task: what Anderson has called «personal design» using hypothetical situations but real values (Anderson, 1996, 2008, 2012, 2013).
The following text is a sample item from the third- party gains allocation based on the comparison of two employees: «Employee A has a merit of 3.5,
employee B has a merit of 2.5. With a monthly
budget of MXN 11 500, by what
amount would you increase employee
A’s salary?»
The following sentence
is a sample item from the self-gains allocation based on the comparison to another employee:
«You have a merit of .5, employee
B has a merit of 2.5.
With a monthly budget of MXN 11 500, by what amount would you increase
your own salary?»
Here is a sample item from the third-party
losses allocation based on evaluating
the relative merit of two employees: «Employee A has a merit of .5,
employee B has a merit
of 1.5. Given a wage cut of MXN 11 500, what monthly amount would
you discount from employee
A?»
Finally, here is a sample item from the self-losses allocation
based on evaluating self-merit versus that of another
employee: «You have a merit of 1.5,
employee B has a merit of 3.5. Given a wage cut of MXN
11 500, what monthly amount would you discount from your own income?»
Software
The software to present the instructions, previous
exercises, the gains or losses contexts, and the randomized items of the instrument described
above was created and
designed in HTML5 language. It gathered
the participants’ answers and presented the
factorial combinations randomly (that is, 7 merit levels for A x 4 for B). It did not allow the
participants to exceed the amount that could be distributed.
Hardware
Computers equipped with Windows
8.1 Pro operating system, Intel® Core™
i5-3470 processor (3.20 GHz, 8 GB
RAM), alphanumeric keyboard, mouse, and monitor.
Experimental Design
We used a repeated measures design
composed of a within-subject factor of A’s 7 merit levels x B’s 4 merit levels x 2 contexts (gains,
losses) and two between-group factors
called third-party allocation and self-allocation. Table 1 summarizes the variables analyzed
in the experiment.
Procedure
Participation was
voluntary, and the confidentiality and anonymity of the subjects’ answers were guaranteed in the
sense that none of the information recorded could indicate their identity. The
participants were free to
withdraw from the study at any time if they
deemed it necessary. A reward for contributing
to the research was offered
and it consisted of a 1GB
USB drive. The same gift was given to the professors
who provided access to the sample. Finally, at the end of the study, the general
feedback was given on the main results of the experiment.
The study began with an e-mail message
that was sent
to professors, inviting
them to ask their students
to participate in a two-session experiment in which each
session would last 40 minutes on
average. After agreeing on a schedule
with the professor, the students were taken in groups of 5 to a computer laboratory for the first session.
They were seated in such a way that they could not see the other participants’ answers on the computer
screen. The researcher read the instructions and the first
previous exercise aloud to
clarify any possible doubts about the requirements for performing the task. Upon completion, they were informed that a second session would take place. It proceeded in the
same way by taking groups of 5
participants to the computer lab and seating
them in the same fashion
as just described.
Table 2 summarizes the procedure. The
sample was divided into groups of
equal size. Note that, in the first
session, group 1 students were exposed to the
third-party gains scenario, and, in the second session, to the third-party losses
scenario. Group 2 answered in the
reverse order, beginning with the third-party allocation (TPA) condition. Group 3 of the
self-allocation
(SA) condition, like group one, was presented
the self-gains context in the first session and the personal losses
context in the second. Group
4 responded to the tasks in the reverse order. There
was a rest period between sessions (PBS) for all the groups aimed at reducing reactivity and the learning of the instruments. The PBS was three weeks between observations.
Results
Data were analyzed with Jamovi software version
1.6
(The Jamovi Project, 2021) to calculate
the repeated-measures analysis of variance
(RM- ANOVA), and the partial eta squared was used ( to measure the effect size of each factor.
First, control statistics were collected
to evaluate the effects
due to the order of the experimental phases, using an RM-ANOVA of 7 (A’s merit levels) x 4 (B’s merit levels), and the
between-group factor called group x
2 (1 and 2) [F (1,18) = .48, p > .05]. No statistical differences were found due to the order of the experimental phases, so groups 1
and 2 were treated as one. The same
analysis was conducted for groups 3 and 4. The result
was [F (1,18) = .23, p > .05], so it was decided to treat
them as a single group as well.
Analyses of the gains context were performed
first, followed by the losses
context.
Gains Context
The analysis of the third-party gains allocation condition
was based on an RM-ANOVA of 7 (A’s merit
levels) x 4 (B’s merit levels). It showed the
principal effects for employee A’s merits [F (6,114) = 651.12, p < .001, = .97] and employee
B’s merits [F (3,57) = 878.98, p < .001, = .97]. The
interaction effects
among the factors
were not statistically significant [F (18,342) = 1.37, p > .05]. The same analyses were carried out for the self-allocation condition: 7 (self-merit levels)
x 4 (B’s merit levels).
This demonstrated the principal effects
for both self-
merit [F (6,114) = 473.54, p < .001, = .96] and B’s
merit [F (3,57) = 808.79, p <
.001, = .97]. The
interaction
effects among the factors were not statistically significant [F (18,342) = 1.11, p > .05]. Finally,
an RM-ANOVA of 7 (A’s merit levels) x 4 (B’s merit levels) was conducted with a
between-group factor called allocation condition
to evaluate the differences
between third-party allocation and self- allocation
gains. In this case, statistically significant
differences were found between the conditions [F (1,38) = 216.18,
p
< .001, = .85], as the
participants increased the salary more in the self- allocation condition. Figure 1 shows the profile chart
of the average responses of the participants in the gains context.
The horizontal axis is divided
into two panels. The left panel shows the third-party condition
with A’s merit on the axis; the right panel presents the self-allocation condition. The vertical axis represents the proportional gains, and each line represents B’s merit. The dotted
line represents the prediction of the proportional gains in the absence of
employee B (equation 1).
Losses Context
The same analysis was conducted for the
third- party condition of 7 (A’s merit levels) x 4 (B’s merit levels). We found principal effects for
A’s merit [F (6,114) = 30.31, p < .001, = .61] and B’s merit [F (3,57) = 34.18, p <
.001, =
.64]. In this case, the interaction effects
were statistically significant
[F (18,342) = 3.21, p < .001, =
.01]. The same
procedure was conducted for the personal
losses allocation
of 7 (A’s merit levels) x 4 (B’s merit levels). It revealed the principal effects
for self-merit [F (6,114) = 49.15, p <
.001, = .71] and B’s merit
[F (3,57) = 119.33,
p < .001, = .85]. No interaction
effects were found [F (18,342) = 1.22, p >
.05].
Finally, an RM-ANOVA of 7 (A’s merit levels) x 4 (B’s merit levels) was conducted
with a between-group factor called allocation condition
to evaluate the differences between the third-party and personal losses allocation conditions. It produced statistically significant differences
[F
(1,38) = 110.45, p < .001, = .71], as the participants showed less willingness
to decrease the salary in the
self-allocation condition. Figure 2 shows
the profile chart of the average responses of
the participants in the losses context.
Once again, the horizontal
axis is divided into two panels. The left panel shows the third-party condition with A’s merit on the axis;
the right panel presents the self-allocation condition.
The vertical axis represents the
proportional losses,
and each line represents B’s merit. The dotted line
represents the prediction of the
proportional losses in the absence of employee B (equation 1).
Discussion
IIT is a model that identifies rules
about the way in which people assess and integrate
information from different stimuli in a single observable answer. In this study, an additive rule of integration was found in the case of gains, which was corroborated by the principal effects
of A’s merits or self-merits and B’s merit.
These results suggest
that the variables
analyzed were assessed independently of each other such that the informative variable that exerted greater control over the willingness to increase
the salary in both allocation
conditions was B’s merit. Previous studies
have similarly identified an additive rule of
integration in the case of gains (Anderson, 1976; Hofmans, 2012; Mellers,
1982; Pulido et al., 2007; Reyes-Contreras & Santoyo, 2017; Santoyo & Bouzas, 1992; Santoyo
et al., 2000).
The data collected in this study are consistent with previous findings in the case of both third-party gains and personal gains. Likewise,
the effect of «compensating» lower merit levels
and «punishing» higher
merit levels were replicated, as can be seen in the average
allocations made by the participants on merits levels .5 - 1.5, which are above the predictions of the equity model
(equation 1) represented
with the dotted line in Figure 1. The opposite is true for merit levels 2 - 3.5, where subjects’ allocations are below the equity line. In the case of the gains context, results
show that the equity model
adequately explains the behavioral data.
For the third-party losses and self-losses conditions, we found
a subtractive rule of integration that was corroborated by the RM-ANOVA and the negative values of the slopes for each
curve. In the allocation conditions –third-party and self– the informative variable was A’s merit or
self-merit; however, the effect of the informative variable of A’s merit
or self-merit was not as clearly defined as in
the gains condition. Additionally, the data presented in an orderly manner and the observed
value of the partial eta squared
showed a «large effect size». The study,
therefore, replicated the findings of Reyes-
Contreras & Santoyo (2017), in which a subtractive integration pattern existed in the case of
third-party losses, such that the
projections of the equity model were not met as in the case of gains. We can infer
from the data that the participants were more willing
to apply lower discounts to merit levels 0 - 2.5 and higher discounts to merit levels
3 - 3.5. Furthermore, this same effect of «compensating» and «punishing» was represented with the dotted line
(equation 1) in Figure 2. Hofmans’ (2012) study found that a subgroup of participants integrated the information in such a way that the profile graphs showed the presence of lines that were horizontal and parallel to each
other. Those results allowed the inference that the experimental situation
could be perceived as
aversive; hence, the data pattern was found. In that sense, and in relation to prospect theory
(PT), an effect called aversion to inequity was found as an analog mechanism to aversion to risk in losses, because the rules of integration do not have a clear negative
gradient or a defined integration pattern.
As for the comparison between self-gains
and third-party gains, equation 2
works as a contextual method of
social comparison because there are both focal stimuli
(A’s merit and self-merit) and background stimuli
(B’s merit). The differences found
in third-party gains and
self-gains, corroborated by the respective RM-ANOVAs, indicate
that the model is sensitive to the manipulation of focal stimuli since changes in the focal stimulus modified
the way in which the information that originated in the stimuli was assessed. This led to the finding that in the self-allocation condition the participants were more willing to
increase their own salary
in relation to the judgment
they delivered in the third-party allocation condition.
Likewise, in the case of the comparison
of the personal losses
and third-party losses
conditions, the equity
model proved to be sensitive
to the changes in the focal stimuli, in the sense that the participants were willing to reduce their salary when they were involved in the judgments they were instructed to
make and, they reduced their salary lesser in the personal
losses condition
in comparison with the third-party condition. And finally, we found interaction effects in third-party allocation between factors in the losses conditions, which
could be an indication of a rule of
integration other than the additive one –probably the multiplicative one– due to the differential effect of one factor on the levels
of another.
While it is true that the differences
between the psychophysical tasks of
prospect theory and the one used in this research
are substantial, we consider that the manipulations of the losses context have a defined
effect in both the fields of information integration theory and prospect theory. The «S» value
function in the prospect theory
(Kahneman & Tversky,
1979) is composed
of two power functions in the gains field U(x) = x and in the losses field
U(x) = -(-xa), where
= .88 and = 2.25. This parameter
shows that the
psychological value of losses is «double» than the gains value. In this way, regarding the
slopes of the lines, the fact that
the value of the gradient of gains is
neither reciprocal nor of the opposite sign to the value of losses is interesting, for it suggests that distinct
cognitive processes may occur and,
moreover, that assessments of gains and losses are not complemented
by one another. In terms of classic psychophysics,
this leads to the inference that these two conditions are found in different sensory dimensions or perceived in different forms.
The contributions of the present study
can be enumerated as follows:
1.
The effect of «compensating» lower merit levels and «punishing» higher ones was replicated in the cases of both
third-party and personal gains.
2.
The same effect was
replicated in the case of third-party and personal losses.
3.
An additive rule of integration was found in the case of gains, but a subtractive rule was manifested for the opposite case of losses.
4.
The general applicability of
the equity model is extended for the
gains condition but was found to be inefficient in the case of losses.
5.
Aversion to inequity can be inferred in the case of losses, thus maintaining the differences between
the experimental tasks performed and the assumptions of prospect theory.
6.
Assessments of gains and losses are not complementary processes; rather, they seem to entail distinct
cognitive processes.
7.
The methodological advantages of using a factorial
design makes it possible to handle different
threats to internal validity compared to the simple comparison studies («one-shot») used in prospect
theory.
8.
The data collection procedure using computer software
and the counterbalanced repeated measures
design permitted maintaining greater
experimental control over the factors.
It is essential to highlight the social
implications of the current study
since it allows a better understanding, at the molecular level, of the distribution of resources to individuals
who differ in merits. This is
important because it occurs in an economic
system in which public access to social, economic,
and financial resources is produced by means
of assessing personal merit (Franco, 2015).
Finally, even though the judgments made
by the participants may indicate a willingness to increase or decrease an individual’s salary, this does not mean that they will actually do so (Ortega, 2017). A second limitation is that the conjoint measurement method we used emerges as somewhat artificial for
the study of the phenomena of
equity/inequity (Hofmans, 2012) because the participants have no direct
contact with the consequences of the
experimented conditions of equity or inequity.
Previous studies did not manipulate direct contact with consequences of
choice. This is important
in equity theory since a basic assumption of the
theory is that consequences shape
equity exchanges (Homans,
1958). Therefore, continuing this line of study requires
generating a dyadic experimental situation in which the participants offer salary increases and others either
accept or reject
what is offered.
The experimental preparation
involved could adopt the logic of the ultimatum game or the gift exchange
game in gains and losses contexts
from a perspective of the behavioral sciences
but maintaining symmetry
in the monetary amounts used in the psychophysical task. Doing this will help us to understand if
the same behavior pattern remains
between the experimental tasks. A
third limitation of the research regarding the discrepancy between the principal and interaction effects
is probably due to the averaging of the numerical estimates that mask the rules of information integration provided by each observer.
It is essential to mention that IIT
is a nomothetic model; that is, it seeks to generate general
principles and it is an ideographic
model in the sense that it seeks specific responses from specific situations (Anderson, 2012). Through
data reduction techniques such as cluster
or latent class analysis, subgroups
with maximum Euclidean distances between themselves
that apply different rules to those reported
by the averages could be identified. These data analysis strategies have been applied
in studies conducted
by Hofmans (2012)
and Acevedo et al. (2019).
Conflicts of Interest
The authors declare that there are no
political, legal, economical, or academic conflicts of interest.
Ethical Responsibility
The present research was conducted with voluntary
high-school students. Every student signed
an informed consent form. The study protocol was conducted in accordance with the guidelines set forth by the Ethical Committee of the Faculty of Psychology
at the National Autonomous University of
Mexico with registration code FPSI/422/CEIP/
208/2021. The participants’ age and gender
were the only data appearing in the manuscript,
which were not linked to a name or identity.
Authorship Contribution
RRC: Conception
and design of the study, collection and interpretation of data,
discussion, and final revision
of the manuscript.
CSV: Interpretation of data, discussion, and final revision
of the manuscript.
Acknowledgments
The first author
wishes to thank
the support of the National
Council of Science and
Technology (CONACYT, for its acronym in Spanish) for
the doctoral scholarship number
333985. The second author wishes to
thank the support of the Support Program for Research and Technological Innovation Projects (PAPIIT)/National Autonomous University of Mexico (UNAM) for the project number 301922.
References
Acevedo, D., Santoyo, C., &
Reyes-Contreras, R. (2019). Integración del plazo y contribución vecinal bajo contextos de pérdidas y ganancias. Revista Mexicana
de Psicología, 36(2),
106-118. https://www.redalyc.org/journal/2430/243067142002/movil/
Adams,
J. S. (1965). Inequity in Social Exchange.
Advances in Experimental Social
Psychology, 2, 267- 299. https://doi.org/10.1016/S0065-2601(08)60108-2
Anderson, N. H. (1976). Equity Judgments as Information Integration.
Journal
of Personality and Social Psychology, 33(3), 291-299.
https://doi.org/10.1037/0022-3514.33.3.291
Anderson, N. H. (1996). A Functional Theory of Cognition (1st ed.). Psychology Press.
https://doi.org/10.4324/9781315805924
Anderson, N. H. (2008). Unified Social Cognition (1st ed.). Psychology Press. https://doi.org/10.4324/9780203837634
Anderson, N. H. (2012). Moral Science. Self-Published. https://psychology.ucsd.edu/people/profiles/nanderson.html
Anderson, N. H. (2013). Unified
Psychology Based on Three Laws of Information Integration. Review of General Psychology, 17(2), 125-132. https://doi.org/10.1037/a0032921
Anderson, N. H., & Farkas, A. J. (1975). Integration Theory Applied to Models of Inequity. Personality and Social Psychology Bulletin, 1(4),
588-591. https://doi.org/10.1177/014616727500100408
Bevan, W. (1968). The Contextual Basis
of Behavior. American Psychologist, 23(10),
701-714. https://doi.org/10.1037/h0026697
Brengman, M., Wauters, B., Macharis, C., & Mairesse, O. (2010).
Functional Effectiveness of Threat Appeals in
Exercise Promotion Messages. Psicologica, 31(3), 577-604.
De Gieter, S., De Cooman, R., Hofmans, J., Pepermans, R., & Jegers, M. (2012). Pay-Level
Satisfaction and Psychological Reward
Satisfaction as Mediators of the Organizational
Justice-Turnover Intention Relationship. International Studies of Management & Organization, 42(1), 50-67. https://doi.org/10.2753/IMO0020-8825420103
Farkas, A. J.,
& Anderson, N. H. (1979).
Multidimensional input
in equity theory. Journal of Personality
and Social Psychology, 37(6),
879-896. https://doi.org/ 10.1037/0022-3514.37.6.879
Franco,
A. (2015). Mérito: construyendo el país de nosotros (1st ed.).
Gigerenzer, G., &
Goldstein, D. G. (1999). Simple
Heuristics that Make us Smart. Oxford
University Press.
Goldstein, D. G., & Gigerenzer, G. (2011). Models of Ecological Rationality: The Recognition Heuristic. Heuristics: The
Foundations of Adaptive Behavior, 109(1),
60-81. https://doi.org/10.1093/acprof:oso/9780199744282.003.0003
Hilgenkamp, H., & Shanteau, J. (2010). Functional Measurement Analysis of Brand Equity: Does Brand Name Affect Perceptions of Quality?
Psicologica, 31(3), 561-575.
Hofmans, J. (2012). Individual Differences in Equity Models. Psicologica, 33(1),
473-482.
Homans, G. C. (1958). Social Behavior as Exchange. American Journal
of Sociology, 63(6),
597-606. https://doi.org/10.1086/222355
Hommers, W., & Görs, M. (2016). Information Integration of Kohlbergian Thoughts about Consensual Sex. Universitas Psychologica, 15(3).
https://doi.org/10.11144/Javeriana.upsy15-3.iikt
Kahneman, D. (2011). Thinking , Fast and Slow. Farrar, Straus and Giroux.
Kahneman, D., & Tversky, A. (1979). Prospect Theory: An Analysis of Decision under Risk. Econometrica, 47(2), 263-291. https://doi.org/10.2307/1914185
Krueger, J. I., Evans, A. M., &
Goldin, G. (2011). The Tangled Web of
Rationality [Review of the book Rational Choice in an Uncertain
World: The Psychology of Judgment an Decision
Making by R.
K. Hastie, & R. M. Dawes]. The American Journal of Psychology,
124(2), 247-251. https://doi.org/10.5406/amerjpsyc.124.2.0247
Laskov-Peled, R., &
Wolf, Y. (2016). Information Trading
in Terms of Brokers’ Functional Cognition: An Exploratory Single Participant Experimentation. Universitas Psychologica, 15(3).
https://doi.org/10.11144/Javeriana.upsy15-3.ittb
Mairesse, O., Hofmans, J., Neu, D., De Oliveira, A. L. D. M., Cluydts,
R., & Theuns, P. (2010). The Algebra of Sleepiness: Investigating the Interaction of Homeostatic
(S) and Circadian (C) Processes in Sleepiness Using Linear Metrics». Psicologica, 31(3), 541-559.
Mellers,
B. A. (1982). Equity Judgment: A Revision of
Aristotelian Views. Journal of Experimental Psychology: General, 111(2), 242-270.
https://doi.org/10.1037/0096-3445.111.2.242
Moore, P. J., Chrabaszcz,
J. S., & Peterson, R. A. (2010). The
Cognitive Processing of Somatic Anxiety: Using
Functional Measurement to Understand and Address the Fear of Pain. Psicologica, 31(3),
605-627.
Mullet,
E., López, W., Kpanake, L., Mukashema, I., Armange, R., Kamble, S., Guedez, A. G., Munoz, M. T., Sorum, P. C., Nieto,
F., & Pineda, C. (2016). Medición funcional en el campo de la Ética en Política. Universitas Psychologica, 15(3). https://doi.org/10.11144/javeriana.upsy15-3.fmf
Ortega, E. (2017). Procrastinación en tareas escolares: características de la tarea. [Undergraduate thesis, Universidad
Nacional Autónoma de México]. Repositorio Institucional de la UNAM. https://repositorio.unam.mx/contenidos/245205.
Pereira, T., Oliveira, A., &
Fonseca, I. B. (2016). Brain Activation
Follows Adding-Type Integration Laws: Brain and Rating Responses in an Integration Task with Pairs of Emotional Faces. Universitas Psychologica, 15(3). https://doi.org/10.11144/Javeriana.upsy15-3.bafa
Pulido, M., Andreu, C., Lanzargota, N., & Preciado, T. (2007). Juicios
de equidad:
los efectos de la complejidad de la tarea. Enseñanza e Investigación En Psicología, 12, 223-235.
Reyes-Contreras, R., & Santoyo,
C. (2017). Juicios psicofísicos de equidad en el contexto de las variaciones salariales. In C. Santoyo
(Ed.), Mecanismos basicos de toma de decisiones: Perspectivas desde las ciencias del comportamiento y del desarrollo (pp. 45-75).
CONACYT/Universidad Nacional Autónoma de México.
Santoyo, C., & Bouzas, A. (1992). Juicios psicofísicos de equidad: algunas implicaciones para la asignación
de incrementos salariales. Revista
Mexicana de Psicología, 9, 85-90.
Santoyo, C., Pulido, M. A., & Bouzas, A. (2000). Juicios de equidad: el efecto del contexto inflacionario para la asignación de recursos salariales. Revista Mexicana de Psicología, 17, 143-150.
Silva, A. D.,
& Oliveira, A. M. (2016). Do
Faces and Body Postures
Integrate Similarly for Distinct Emotions, Kinds of Emotion and Judgment Dimensions? Universitas Psychologica, 15(3). https://doi.org/10.11144/Javeriana.upsy15-3.fbis
Singh, R. (1985). A Test of the
Relative-Ratio Model of Reward
Division with Students and Managers in India. Genetic,
Social, and General Psychology Monographs, 111(3), 363-384.
The Jamovi
Project. (2021). Jamovi (1.6). www.https://www.jamovi.org.
Theuns,
P., Verresen, N., Mairesse,
O., Goossens, R., Michiels,
L., Peeters, E., & Wastiau,
M. (2010). An Experimental Approach to the
Joint Effects of Relations with Partner,
Friends and Parents
on Happiness. Psicologica, 31(3),
629-645.
Universidad
Nacional Autónoma de México. (2017, 01
de febrero).
Tabulador de personal académico.
Portal de Transparencia Universitaria.
https://www.plataformatransparencia.unam.mx/archivos/repositorio/UNAM/2017/percepciones/peraca.pdf
Viegas, R. G., Oliveira, A. M., & Gaspar, R. (2016). Comparing Adults and Adolescents Regarding
the Scope Insensitivity of Value Curves:
A Functional Measurement Approach. Universitas Psychologica, 15(3). https://doi.org/10.11144/Javeriana.upsy15-3.caar
Recibido: 20 de diciembre de 2021
Aceptado: 10 de agosto de 2022
Este es un artículo Open Access publicado bajo la licencia Creative
Commons Atribución 4.0 Internacional. (CC-BY 4.0)