Response rate as a factor in choice

Trang 1

ILLiad Transaction Number: 1153669 Call Number: BF1 P88

Location: STORAGE

Journal Title: Psychonomic science Article Author: Killeen, Peter

Article Title: Response rate as a factor in choice Journal Vol: 12 Journal Issue: 1

Trang 2

Drug and Magnesium Pemoline: Effects on Active and Passive Avoidance Chemical Effects Conditioning in mice by Cynthia Wimer, Susan Donner, and

David Martin 23

The Effects of Hypoxia on an Information-Reduction Task in Humans

by F Ledwith 29

Stimulating Effects of Chlordiazepoxide in Rats on a Food Reinforced

FR Schedule by Paul W Wedeking 31

Drug Withdrawal and Fighting in Rats by John Florea and Donald H Thor 33

Reinforcement Response Rate as a Factor in Choice by Peter Killeen 34

Partial Reinforcement Acquisition Effects Within Subject by J Dutch 35 Facilitative Effects of Reward and Decremental Effects of Non-Reward

in the Initial Trials of Partial Reinforcement by James R Ison and

Linda Anderson Pinckney 37

The Effect of Partial Reinforcement on Extinction in Avoidance

Conditioning by George Marsh and Ned Paulson 79

INFORMATION FOR CONTRIBUTORS

This Journal publishes studies that will contribute to the development of experimental psychology and that will fit into a two-page format Its publication lag is 6 to 12 weeks The evaluation of papers is relative, rather than absolute; that is to say, of the articles on hand at the time of making up an issue (or part) of 20 papers, those are selected for publication that would appear to serve best the Journal’s objectives

Manuscripts should adhere to the conventions concerning reference citations, preparation of tables and figures, manuscript format, etc as described in the Publication Manual of the American Psychological Association except that in the abbreviation of physical units the conventions of the American Institute of Physics are followed When in doubt, authors should examine a recent issue or write the Editorial Office for a copy of this Journal’s Style Guide All manuscripts should be submitted in duplicate to the Editor, 1200 West 34th St., Austin, Texas, 78705 Manuscripts must include an abstract of 50 to 100 words

PSYCHONOMIC SCIENCE is published three times monthly in four volumes per year by Psychonomic Journals, a non-profit organization affiliated with the Psychonomic Society, Inc Its Editor is Clifford T Morgan, 1200 West 31st St., Austin, Texas, 78705; its Production Department, of which Anne E Dossett is Manager, is at 1911D De La Vina Street, Santa Barbara, Calif 93101 Printing is done by the J J Connolly Printing and Publishing Company, Santa Barbara, Calif

SUBSCRIPTION RATES are $10 a volume, $40 a year, for institutional subscribers, or $5 a volume, $20 a year, for individuals (when paid in advance by personal check or money order) Sectional subscriptions are available for $20 and $10 respectively Single issues are $1.00 each Subscriptions and changes of address should be sent to 1200

West 34th St., Austin, Texas, 78705

Second Class Postage Paid at Goleta, California COPYRIGHT © by PSYCHONOMIC JOURNALS

Trang 3

age

gg

Response rate as a factor in choice’ PETER KILLEEN, HARVARD UNIVERSITY, Cambridge, Mass

02138

Four pigeons were trained on a concurrent chain schedule, in the terminal links of which food was delivered according to variable interval schedules, For one key in Experiment 2 reinforce- ment was contingent on not responding Three control experi- ments were also run: baseline, return to baseline, and unequal reinforcement frequencies Although rate differences of over 50 to 1 were obtained in the terminal links of Experiment 2, the pigeons remained indifferent between these schedules The only change in preference occurred in Experiment 4, where preference matched reinforcement frequency

In 1960 Autor investigated behavior on a concurrent chain schedule in which reinforcement in the terminal links was delivered according to variable interval (VI) schedules He found that the relative numberof responses on a key in the initial link equalled (matched) the relative rate of reinforcement for that key in the terminal link This matching relation did not change when he made reinforcement contingent on not responding in both terminal links Herrnstein (1964) also found that preference was more highly correlated with reinforcement rate than with rein- forcement probability (reinforcements per response) in concurrent chain schedules employing VI and VR schedules in the terminal links Fantino (1968), however, has recently published data showing that animals prefer FI schedules to the same schedules with an added restriction on the minimum or maximum number of responses His data for the schedule requiring a low rate of responding are questionable, since two pigeons’ rates were not appropriately controlled, yet they showed a greater range in preference than the third pigeon on this schedule

The following study is an attempt to clarify the relation of the response rate evoked by a schedule to preference for that schedule It employs a more effective method of control of terminal link rates than that used by Fantino Since the rate is controlled in only one terminal link, the obtained differential measure of preference is more sensitive than the measure obtained by Autor Finally, reinforcement rate is held constant, so that instead of looking for a deviation from a correlation as did Herrnstein, all obtained changes in preference can be ascribed to the terminal link rate differences

Method Four male White Carneaux pigeons, all with previous experimental histories, were maintained at 80% of their free-feeding weight

The experimental procedure conforms to the basic concurrent chain design which consists of a regular alternation between two states, an initial choice link, and a terminal reinforcement link In the choice link responses of at least 15 g force to either of two Gerbrands pigeon keys occasionally changed the color of that key from blue to red (left key) or from blue to green (right key), with the other key going dark Further responding on the lit key produced reinforcement according to VI 30 sec schedules in all cases except two For the right key in Experiment 2 reinforcement was delivered as soon as a VI 30 sec schedule set up, but no sooner than 1.5 sec since the last response on that key For the right key in Experiment 4, reinforcement was delivered according to a standard VI 60 sec schedule After one reinforcement with grain in the terminal link, the program reverted to the choice link with both keys blue The occasions upon which a response in the choice link would cause a transition to the reinforce- ment link were determined by separate VI 60 sec schedules for each key Each experiment was mun for 17 days

Experiments 1, 3, and 4 were controls for Experiment 2 In Experiments 1 and 3 conditions of reinforcement were identical for the left and right keys, and thus the birds should be indifferent between them In the terminal link of Experiment 4 reinforcement rate for the right key was half of that for the left key; the birds should prefer the left key In Experiment 2 reinforcement frequencies were equal, but the contingencies of reinforcement differed The schedule on the left key in the terminal link encouraged moderate, uniform rates of responding, while the schedule on the right key discouraged any responding at all

Results and Discussion The schedules employed in Experi ment 2 were effective in establishing different rates of responding in 34 IN FIRST STATE RELATIVE NUMBER OF RESPONSES ON LEFT KEY EXPERIMENT: 1234 1234 1234 1234 1234 BIRD: 25 259 344 460 AVG

Fig 1 Preference for left key (responses on left key in initial link/total responses in initial link) Data are averages over last five days in each experiment for each bird

the reinforcement links: averaged across birds for the last five days of Experiment 2, these rates were 52.3 responses per minute on the left key and 0.95 responses per minute on the right key A substantial number of these right key pecks were emitted just after entry into the terminal link, and represent the last response in arun initiated in the choice link

The effect of this large difference in response rate on preference can be seen in Fig 1 Pigeon No 25 showed a consistent bias for the right key, but the only pigeon showing a marked change from the baseline of Experiment 1 and 3 was No 259 When the data are averaged across birds, there is seen to be an increase from baseline preference for the right key of only 1% The pigeons’ relative indifference to radical changes in response rate stands in sharp contrast to their preference for schedules with higher reinforcement rates In Experiment 4, halving reinforcement frequency on the right key resulted in an increase to 63% in preference for the left key Although the present study provides evidence that preference is independent of response rate, we cannot conclude that an organism is insensitive to the amount of work entailed in the procurement of reinforcement When an animal is reinforced independently of its behavior, it will go through a series of stereotyped but arbitrary responses in the intervals between reinforcements This phenome- non follows from the law of effect, which implies that reinforcement strengthens all immediately preceding behavior, not just behavior that is causally effective in providing reinforcement When only the right key light was on in Experiment 2, the pigeons could be seen to peck the wall of their cage, pace back and forth, and jump up and down with higher frequency than in other phases of the expert ments No attempt was made to quantify the total energy expended, which may well have been the same for all schedules in the terminal link Indeed it may be that total energy expenditure is completely determined by reinforcement frequency, while the form of that work—key pecking or superstitious strutting—is determined by the terminal contingencies of reinforcement

REFERENCES

AUTOR, S M The strength of conditioned reinforcers as a function of the frequency and probability of reinforcement Unpublished doctoral disserta- tion, Harvard University, 1960

FANTINO, E Effects of required rates of responding upon choice J exp

Anal Behav., 1968, 11, 15-22, ˆ

HERRNSTEIN, R J Secondary reinforcement and rate of primary reinforce-

ment J exp Anal, Behay., 1964, 7, 27-36 :

NOTE

1 This work was supported in part by a NIMH Predoctoral Fellowship, and in part by NSF research Grants GB 3121 and GB 3723