Running head: DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
Discounting of Primary vs. Generalized Reinforcers as a Function of Self-Control
Randi Hovey
Utah State University
1
2DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
Abstract
The phenomenon of delay discounting involves the decrease in value of a reinforcement
due to the delay of its presentation. Individuals were asked to choose between hypothetical
reinforcers presented at different values and delay intervals using money as a generalized
reinforcer in one session and food as a primary reinforcer in a second session. The present study
examines delay discounting as a function of impulsivity in relation to the type of reinforcer
within the general population. Data were analyzed using area under the curve and the results
support previous findings which show that generalized reinforcers are less-steeply discounted
than primary reinforcers.
3DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
Discounting of Primary vs. Generalized Reinforcers as a Function of Self-Control
Delay discounting, a decision-making phenomenon which has often been associated with
levels of impulsivity. It involves the devaluation of reinforcers due to delayed delivery of the
reward. According to Ainslie (1974), an impulse is defined by the choice of smaller, immediate
rewards at the expense of a large, long-term loss. Discounting is commonly measured by an
indifference point. Essentially this is the point at which a participants’ preference changes from
the one reward to the other. Many maladaptive behaviors such as substance abuse and over-
eating, may perpetuate as a result of discounting (Myerson, 2001; Odum and Rainaud, 2003).
Theoretically, an individual’s lack of self-control results in decreased value of rewarding
outcomes as a function of time (Ainsle, 1974). Enhanced understanding of this phenomenon and
its relationship with various reinforcers may lead to more successful treatment options for some
of society’s most insuperable fallacies. Furthermore, increased evidence of a correlational
relationship between impulsiveness and discounting may prove valuable in the construction of
preventative treatment procedures.
Previous research findings have suggested notable variation in the amount of discounting
between primary reinforcers (i.e. food, water, sex, drugs) and generalized conditioned reinforcers
such as money. Madden et al. (1997) conducted research on the rate of discounting in monetary
rewards compared to that of substance rewards in heroin addicts. Addicts were shown to
discount heroin reinforcers more steeply than money. While these within-subjects results do
suggest primary reinforcers increase impulsive behavior, heroin addicts also showed increased
rate of discounting when compared to a control group on monetary rewards alone. These results
suggest the degree of discounting is a complex combination involving trait characteristics and
type of reinforcer.
4DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
In hope of clarifying the distinct role type of reinforcer plays in the discounting
phenomenon, Odum and Rainaud (2003) examined delay discounting of food, alcohol, and
money in participants who were previously screened for addictive disorders. By controlling for
addiction, the experiment could more accurately discern variation in degree of discounting for
primary reinforcers as opposed to conditioned reinforcers without intervening effects caused by
addictive behaviors. Using hypothetical rewards in dollar amounts and equivalent units of
alcohol and food, they found that overall money was less steeply discounted than the primary
reinforcers. Results also showed no significant difference in degree of discounting between food
and alcohol (Odum and Rainaud, 2003). Although these findings show significance, no
replication has been reported to confirm the external validity of these results.
The present experiment expands on previous findings regarding the discounting rates of
different types of reinforcers. Unscreened participants were asked to make a series of choices in
two separate sets of questions that measured discounting for money and food. The goal was to
determine whether hypothetical primary reinforcers tend to be discounted more steeply than
hypothetical generalized conditioned reinforcers. Data was used to determine any correlational
relationship between type of reinforcer, and amount of self-control.
Method
Participants
Participants were 40 online PSY 3400 students from Utah State University. Subjects were
all 18 years of age or older and originated from various geographical locations. Over seventy
percent of respondents were female (72.5%), 27.5% were male. Due to the nature of subject
selection, no socioeconomic information was included in the process of selection. Subjects were
5DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
awarded 5 points toward their total grade in the class in return for participation in experimental
sessions which took an average of 60 minutes to complete. There was no pre-screening done,
allowing results to be more indicative of general population parameters.
Materials
In order to examine discounting using similar techniques as previous studies, hypothetical
rewards were used. Hypothetical food reinforcers were presented in units which were determined
by the cost per unit of food. Food and money reinforcers began at $100 or equivalent units and
decreased down to $1. Data was collected using Qualtrics Research Suite.
Procedures
Participants were asked to choose between immediate or delayed money or food rewards
in two separate sessions which were counterbalanced for type of reinforcer as well as delay. The
delayed reward for each instance was always $100 or the equivalent units of participant’s
favorite food. Immediate rewards also began at $100 value and were decreased at specific
intervals over the course of the session. Instructions given prior to participation made the
hypothetical nature of rewards clear and were as follows:
You will be asked to make some decisions about which of two (monetary/food) rewards
you prefer. You will not receive the rewards that you choose, but I want you to make
your decisions as though you were really going to receive these rewards you choose. The
possible rewards will be displayed to you in the form of a multiple-choice survey
question. The options on your left display a reward that you can get today. The options on
the right display the reward that you can get after the specified amount of time. So now
you are being asked to choose between an immediate amount delivered today versus a
6DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
delayed amount that you would get after waiting the delay interval. Select the reward that
you would prefer. The choices you make are completely up to you. Please select the
option that you prefer, not what you think I want you to prefer. I do not expect you to
choose one particular reward over the other. Just choose the reward you really want.
Prior to beginning the food discounting task, participants were asked what their favorite
food was and how much they would typically spend on one unit of this food. This information
was then used to compute rewards equivalent to amounts given as reward choices in the
monetary session. Reward intervals for immediate amounts were presented in the following
intervals; $100, 99, 97.50, 95, 92.50, 90, 85, 80, 75, 65, 60, 55, 50, 45, 40, 35, 30, 25, 20, 15, 10,
7.50, 5, 2.50, and 1. Amount of delay also varied between 1 week, 2 weeks, 1 month, 6 months,
1 year, 5 years, and 25 years. Participants were asked simply to select the reward that they would
prefer.
Data Analysis
In order to assess the differences of outcome between discounting of primary vs.
generalized reinforcers, individual indifference points were normalized and calculated using the
area under the curve (Myerson et al., 2001). This method was used due to its theoretically neutral
measurement of discounting. To normalize the data, each value and delay is calculated as a
proportion of maximum value and maximum delay respectively. AUC is then calculated using
the sum of the results of the following equation: (x2 – x1) [(y1 + y2)/2] applied to each successive
delay and the corresponding values where x1, x2 refer to successive delays and y1, y2 refer to the
7DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
corresponding indifference points. AUC value is then interpreted within a range from 1 (no
discounting) to 0 (maximum discounting) (Odum and Rainaud, 2003).
Results
Median indifference points for money and food are presented in Figure 1, showing the
rate of discounting for food (red data path) and money (blue data path). Outcome for both food
and money values were included for all 40 participants with no missing values. Both primary and
generalized reinforcers lost value as delay increased. However, congruent with previous research
findings, money was less steeply discounted than food reinforcers expressed in equivalent
amounts. For example, the discounted rate of $100 after a 30-day delay was $85 while the
equivalent amount of food was discounted to $57.5 after a 30-day delay as shown in table 1.
Figure 2 shows the mean value of AUC for money (blue bar) in comparison to the mean
value of AUC for food (red bar). AUC for money (.27) was higher than AUC for food (.14).
Discussion
Findings in the present study further support previous research on the degree of
discounting using consumable vs. non-consumable reinforcers. Kirby and Guastello (2001)
found that monetary rewards were discounted at a smaller degree when presented to college
students than when an equivalent amount of pizza was offered on the same delay scale.
Reinforcers in this study were not entirely hypothetical; participants did receive one randomly
chosen specified outcome after completing the trial (Odum and Rainaud, 2003). Johnson and
Bickel (2002) conducted a similar study using only monetary rewards as hypothetical reinforcers
in one experiment, and real rewards in another and determined no significant difference in the
8DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
degree of discounting between samples; further justifying the use of hypothetical rewards in
present research.
Results obtained in this experiment further confirm increased impulsivity in relation to
primary reinforcers, and extend these findings to a general population in the absence of control
for certain attributes, socioeconomic status, or behavioral traits. Future studies should involve
further statistical analyses to determine the significance of such findings within a sample from
the general population. If the degree of discounting primary rewards vs. generalized rewards
within general samples proves statistically significant, findings may have important implications
for determining levels of impulsivity which may put individuals at higher risk of developing
addictive behaviors and could contribute to better preventative measures for substance abuse,
over-eating habits, and possibly inclination to criminal behavior in general.
9DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
References
Ainslie, G., 1974. Impulse control in pigeons. Journal of the Experimental Analysis of Behavior.
21, 485-489.
Johnson, Matthew W., Bickel, Warren K., 2002. Within-subject comparison of real and
hypothetical money rewards in delay discounting. Journal of the Experimental Analysis
of Behavior. 77, 129-146.
Kirby, K.N., Guastello, B., 2001. Making choices in anticipation of similar future choices can
increase self-control. J. Exp. Psychol. Appl. 7, 154–164.
Myerson, Joel et al., 2001. Area Under the curve as a measure of discounting. Journal of the
Experimental Analysis of Behavior. 76, 235-243.
Odum, Amy L., Rainaud, Carla P., 2003. Discounting of delayed hypothetical money, alcohol,
and food. Behavioural Processes. 64, 305-313.
10DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
Table 1. Numerical indifference points for money and food observed for the specific delay
intervals (in days).
Days 7 14 30 180 365 1825 9125Money 97.5 90 85 65 55 35 7.5Food 77.5 70 57.5 47.5 27.5 20 1.75
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 100000
102030405060708090
100
Median Indifference Points
MoneyFood
Delay (days)
Subj
ectiv
e Va
lue
Fig. 1. Indifference points for money (blue data path) and food (red data path) as a function of
delay.
11DISCOUNTING OF PRIMARY VS. GENERALIZED REINFORCERS
10
0.05
0.1
0.15
0.2
0.25
0.3
AUC
MoneyFood
Reinforcer Type
AUC
Fig. 2. Mean AUC values for money (blue bar) and food (red data path).