Group Title: effects of rate of point reinforcement on human preference behavior
Title: The Effects of rate of point reinforcement on human preference behavior
Full Citation
Permanent Link:
 Material Information
Title: The Effects of rate of point reinforcement on human preference behavior
Physical Description: vi, 61 leaves. : illus. ; 28 cm.
Language: English
Creator: Waddell, Thomas Robert, 1945-
Publication Date: 1972
Copyright Date: 1972
Subject: Human behavior   ( lcsh )
Reinforcement (Psychology)   ( lcsh )
Psychology thesis Ph. D   ( lcsh )
Dissertations, Academic -- Psychology -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis -- University of Florida.
Bibliography: Bibliography: leaves 55-60.
Additional Physical Form: Also available on World Wide Web
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00097646
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000577471
oclc - 13982773
notis - ADA5166


This item has the following downloads:

PDF ( 2 MBs ) ( PDF )

Full Text

I1 E FE-[t C T P'- --F i .ITi_ T :1 iil' *.,,L'LlIL
.-E E Lli : U F.- E T- -,E: IT FI;:ii *. &rEi.i.l.
r i ,tt, -. r -i EE i, BE l- I


Ih..r ,:,: F.:t.' -rr. cri.1j :

A DIS. CFTATII ,' FF CL117iCC T. TnI I r-'.'.! .TE7 I.: rl-.: L TFr.
T.iE I.[JI'.'EF ,I 7 L rI I l F -.rTI.,L FriLFILLi .Fr:T
OF Ti[ ir Li. IFF"iI-iTl F"''F. TiE CEIF-FF :'F
D,-T.,,:T .: .,r Fi-ILL.,-.: Fi i

u'Tii'.'F .,I ,. -Lf- r1.2

This dissertation is dedicated to

my mother and father.


I gratefully thank Dr. Edward F. Malagodi for bring-

rl:, ce',_ _'.- ri ,-i nrial ar /i ;'_ .o f ,,".,5 ,, r r. -, i.irif r:-ir r

of F ; i .,r cr.- ii :ri .:.rl p, r t, a f 3 ri

r:-.: :,r.n :r r. re :e r,h r. r l-.1 s .1 l r. t:: ,r. i ali li ci'_

t..' t1-, -ini Fr--,:-i, ,-:,i, Te ,ac.',inr '.:[ F: l,:,r Inr *.7:.rp.:.r ti i: f.r a

-:iar nt ir, .up-.: c ir r_- arcr.

I alr:. ; .:r. n j r.'y rn: .. t.. Dr. iilli Ls . :.ll",i ,

mI c*:[uiiLr. .- : -cl- l rn, r.: -, lr. HJl-r. C-. [*a'.'a5 r.,

Dr. Henrr/ :. F-rn, : r, 3.: LCr. ; i l a F.r. F- i for rE --

10-r :,- ri/ ::mi Tio' e. .

ar:. iue t'in artrii f r tht ir as.:i -:.a- : ir : rrep:.rici:--

of rh- nar, -r,-..r-.

inalli, I :.,jl1 li': r-.: Fr- ,i/ ,a:pr..i ti In tO

s- '" ril *:tr.lher : p.: i4 i.:.; : rt irJii-ti rn anr.d r..' :.;rt r,'.-t mi-d.

m i' grl du e ed..= J.j.C ,r b:cr, e r ; i plea anr. T-ri s

pc:..p :incl i- Cr. r lic l ". i Ocr. ic r d .r. I 11 ',

[,nrir ? r .i az i 3 J51 .:c : 4., j 3 CDr. M* -i* 3 L.-arnj r.



ACKNOWLEDGEMENTS . . . . . . . ... iii

LIST OF FIGURES. . . . . . . . .. v

ABSTRACT . . . . . . . . .. . vi

INTRODUCTION . . . . . . . . .. . 1

METHOD . . . . . . . . . . . 6

RESULTS . . . . . . . . .. . 13

DISCUSSION . . . . . . . . .. ... 20

APPENDIX . . . . . . . . . . 23

REFERENCES . . . . . . . . ... ... 55

BIOGRAPHICAL SKETCH . . . . . . ... 61


Figure Page

1. Diacram of the mult (chain VI 15 sec VI
'.'i i' ; :LC : l. .- _,r, '.'t I -.-: 1i it =3.:
F iL.Jr-:.. . . . r............. l ...

2. TI' ,. L"- i. i r :':. :n ; ri.- in cr-

r ly .1" r t : f r i n:...n:- nr..:n . ..... 14-

3. Tn, ine d.Ji:n c.:- l5 .: r.a ,.: :.::... ,',t i r, i ti l
C -.C,[-.r ~ r:nrt :,-= :i _r'ju r] ,:,- .: r r r 9.. ar, r l *.'.
re e = r.-= r L>: r = n" . . . . . . 1

41. Ti1- '.. hia rn ni ran-i *:-C rt-. r. i,r1 r [:n-E
rSrte in t n t ,r a -1 r. J AUE.. t L Ir r n l tt
Fr3e i H.i*: d c r :
r. H I l.:.;.. .: i r; .; l | r.: 5 .:. i:r rin
cir . .. . .r . .. .. . . 17

5. A .'.u i.j l it .: r-: : rj : r,.: r ..:n.Jirn : .
Jur i r l ri- r i-i : E- i.:r uFij L :-r: I l',: I "'I
TI 1-1 'i T.1 :

Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Thomas Robert Waddell

December, 1972

Chairman: Edward F. Malagodi
Co-Chairman: William D. Wolking
Major Department: Psychology

Two normal 15-year-old children were exposed to

concurrent-chained schedules with equal and independent

variable-interval schedules in the initial components and

different-valued variable-interval schedules of point

reinforcement in the terminal components. Points were later

exchanged for money. A changeover-key procedure was used

in which responses under either chained schedule were made

with one plunger-type operandum. Responses which produced

a changeover from one initial chained component to the other

were made on a second operandum. The relative overall

response rate and the distribution of relative time spent

in the initial components showed a preference for the ter-

minal component associated with the higher rate of reinforce-

ment. Preference approached a matching relationship to the

relative rate of point delivery. These results extend the

species generality of concurrent-chains research to humans

as well as extending methodological generality through use

of a novel procedure.


The concurrent-chained-schedules procedure was intro-

duced by Autor (1960) as a method for evaluating the

relative strength of conditioned reinforcers. In this pro-

cedure the organism, typically a pigeon, responds on two

concurrently available keys, each of which is illuminated

by the stimulus associated with the initial component of

one of the chains. Responses on each key occasionally pro-

uc.- thi t irjl *:Cr cin -. rE r rr,1 i r.; .rn r t .rri,,

ir. rl .t :. r Fr rr., :r. :.f -r. .r : -1 r, r, i : ,-1 r. 1 r-r .-. -t ,,n

in c. he -r, , :,r l;r ',_'.lq;r : : r.r-: iTJr... J:,ll cirl.i 1 t ,r-; l i:i l-

f .: J. Tr l r ela i. :.r,.j i Cr I ..i ir E :t, : n; r ari i n ,:f trIe

ctii r,,=_i ,rr :,r -, -1r r n: = t r ,-: r- ,-_ rl G r, rn!n r ri t: i en

rir. rr.:d :.,' c: r., :ii c cri -r. s ti rc In r:,.- r. nr
cnr l ir,. .i-: r. i..j rl lr i. j r- ,di r. : y ia' In' .r i, r-. .een

e rc,_ ir-tr ([ji-.r'r,. ],r1 r -chneij--r, 1_ 1 In ,adiitic-rn

un:ri .r' ar; Farnir,:. II 2'l re :rt jed thi r.rrririii l :rjirnle

c ,-'[-, Lr. rr b c-i .',: r.l 2 ij mt, r-c,: c. ,'r r rri s : Lri iFr, ij

segments were less preferred than components of equal dura-

tion which consisted of only one or two chained segments.

Currently, relative response rate in initial components of

concurrent-chained schedules is discussed as a measure of

preference per se, without necessarily implicating condi-

tioned reinforcement as the causal process.

Early research with concurrent-chained schedules

employing variable-ratio (VR) and VI schedules of primary

reinforcement reported a matching relation between prefer-

ence and the relative rate of reinforcement in the terminal

components (Autor, 1960; Herrnstein, 1964a). Herrnstein

(1964b), comparing fixed-interval (FI) versus VI schedules

of primary reinforcement, determined that the matching

relation had limited generality. In that study, FI

schedules were less preferred than VI schedules of equal

average duration. Killeen (1968) reported that preference

between FI versus VI schedules and VI versus VI schedules

matched the relative harmonic rate of reinforcement, also

called the relative immediacy of reinforcement. Prefer-

ence predictions based on the relative immediacy of rein-

forcement give greater weight to short inter-reinforcement

intervals in periodic schedules than predictions based on

the relative arithmetic rate of reinforcement. Later

studies of preference between FI schedules, and between FI

versus mixed-interval (MI) schedules of primary reinforce-

ment, reported that preference matched a transformation

based upon a power function similar to the relative immediacy

of reinforcement but with a slightly smaller exponent

(Davison, 1969, 1972; Killeen, 1970). Davison (1969) specu-

lated that the value of the power function exponent neces-

sary to achieve matching might be related to the number of

intervals in the periodic schedules, but this was later

discounted (Davison, 1972).

In a study of preference between periodic schedules

(FI versus FI and FR versus FR) Duncan and Fantino (1970)

held the absolute difference between the schedules of pri-

7,iarj r i.r. r.r:.i- :rr r.:r ct t n -r in r': :*-- r..-.r. c !.e.iui -

"mai, e in-r .: e] a L: e .:,- j le '.'al.-: i crea ir :i E. .

iiTi l r Efi r,. r, .- re ..:,rr.-.. t., : 1 '" -h. r-

.7mre ---e t,-E aE=.:.iv e :.fTjCi jI t-;T.j njl-. -.F.. r, i .r! nd

apLu-.n t; em-ai=. e-r*L r~ 1 lr : L'J.i -r- : i :r'r.'reir, FI.rJ-
'.'i -=rh- rt.I ,I -- x.ileai~; r, : i cl-, *i r ri TeL.' r i a.ei: r i r-.t r,-

riF t -l '_nj, a; cr-h -rlT-.i 'j ".' l u-; ..r incr-a.m= j, ric, Fra tt-

.r-.5 t trjai r trri.:.:r. the r. lact.' iiuin j : .f reinf :r.e-

mrin r i..'j ,i-i.lo.'a ed.

hr=e '.-alu-e *.r f the inicis: -:'-mFj; nri : r.'I -hedau -l la

f-Dund tj influa.:ie reference (Fant in. 19i. : i- quires and

Fanir. rr I 1 'll i. Ir th frrn.- r i th.= "i :hcdul- of

primary reinforcement were held constant but unequal, and

the values of the initial-component VI schedules were varied.

Matching predicted from the relative rate of primary rein-

forcement was not obtained. In the latter study, the VI

schedules of primary reinforcement were held equal and the

relative size was varied of unequal VI schedules in the

initial components. Preference was predicted in both

studies by a formula which took into consideration the re-

duction in expected time to reinforcement associated with

each initial component schedule, and in the latter study

by a modified formula which also took into consideration the

overall rate of primary reinforcement on each key. Addi-

tional variables which may influence preference include

number of primary reinforcements in each terminal compo-

nent (Fantino and Herrnstein, 1968; Squires and Fantino,

1971), high response rate in a terminal component (Fantino,

1968; Arnett, 1972), and the presence of a changeover delay

(COD) for responding in the initial components (Davison,


Although a number of studies have reported human per-

formance on concurrent schedules of reinforcement (e.g.,

Frazier and Bitetto, 1969; Sanders, 1969; Schroeder and

Holland, 1969; Poppen, 1972), the focus in most of these

studies has been on variables influencing patterns of re-

sponding in separate components, rather than on variables

influencing preference. An exception was the study by

Schroeder and Holland in which subjects matched relative-

eye-movement rates to the relative rates of reinforcement

associated with concurrent VI schedules. Matching was only

obtained, however, when a COD was programmed. No reports

of human concurrent-chained-schedule performance have

appeared in the experimental literature.

The purpose of the present experiment was to determine

the effects on human preference of different combinations of

VI schedules in the terminal components of two concurrent

chains. Responses in both chains were made on one operan-

dum. Changeovers from one initial component to the other

were produced by responses on a second operandum. This type

of concurrent-chains procedure is analogous to the chsnie-

over-key (CO-key) procedure used in many studies wlth

single-component concurrent schedules (e.g., Findl: I'"i:

Brownstein and Pliskoff, 1968; Stubbs and Pliskoff, 19691.



Two 15-year-old boys (F. H. and S. G.), who answered

an advertisement in a newspaper, served. They were both

told that they could earn up to $1.00 per session depending

on their performance. Half of the money earned was paid

immediately following each session and half was paid at the

conclusion of the subject's participation in the experiment.


The intelligence panel consisted of the side of a

wooden box 60 cm wide and 64 cm high which was secured to

the top of a metal desk which was 75 cm in height. Each

subject sat in a chair in front of the panel. In the upper

half of the panel there was a row of three translucent screens

which were 12.5 cm wide and 13 cm high. The screens were

illuminated from the rear by standard-size, 7.5-w Christmas

tree lights, two bulbs for each color employed. The colors

were yellow or red behind the left screen, white behind

the center screen, and blue or green behind the right

screen. An additional screen 32.5 cm wide and 7.5 cm high

was centered above the other three. This screen was illumi-

nated by a pink light. Two Lindsley plunger-type operanda

were located 19 cm below the screens and 42 cm apart. A

counter was centered on the panel at a height intermediate

between that of the operanda and that of the translucent

stimulus screens. A small white light was located immedi-

ately above the counter. Two speakers, 10 cm in diameter,

were located behind grids in the panel on either side of the

counter. A tone and clicking sound were presented through

one speaker and white noise was presented as a masking

noise through the other. Additional masking noise was pro-

vided by an exhaust fan in one wall. The room was dimly

illuminated by a shielded flourescent light located on top

of the box, above the intelligence panel. The subjects were

cr.--*r j, ,ri j ..- rir rni jl ;:r :.n? 5hr .:.u h a p- h,:,le

.. Lrh a i r, l l.:r, i.:.. 'ar .j ir r. 3 :..:.r c,: r:,*r r.: .r

in,: r:n : ,jre

, tcrn ..- 3:;.r, rr,:, '-,i rj e *: -: -:ceJ in rfrcirit ,:. Lh.-

ir r.ellae r .: r.-er,1. L.jririr ru,-_ ir irisi esi;:i.r, tithe rie ht

.:._r a ...T, :,c.." r. Tri ri ]ht *p--rrJ , '11 t- reft' rrcI.

: th- .: .: 'r (. ri -.t tr. e 1nr ni .:

the f, r E i: r, ,i a:l- ur.. j_.t L.Z 1 i' r-rt. i.. t-,sr r. C wouldd

pui tnr left .:.r me .-F j r nriduwr Ind rnr, =':. eti=.5 .her he

p.ullej Ir, t. .:u rld a r 5 ps rn :.r t ~ cC:.urL '. He ..:a

furcl, r in trrh-tr i:r3t ne ..eouiId :.- p31ij Crn -at frcr eaich

.: rr. rr,-... E"-:r r. p,:,rc fr-.ji, j 3 O ,.; -5- -.: fla h c-f

the pink screen. The point-delivery cycle consisted of the

presentation of a 2-sec tone and simultaneous illumination

of the white center screen followed by a 45-msec flash

of the light above the counter and a 1-point counter advance.

Sessions were scheduled daily, five days per week, but

occasional subject absences occurred. Each session was

begun and terminated with a 5-min blackout during which

there was no illumination of the screens on the intelligence

panel and responses had no programmed consequences. Neither

subject showed any responding during these blackouts after

the first session. Each session was terminated after 100

points had been earned or after the passage of one hour,

whichever occurred first.

During the initial sessions, a clicking sound was con-

tinuously present and either the red light behind the left

screen or the blue light behind the right screen was

illuminated. In the presence of either the red or the blue

light, responding on the main operandum produced points

according to independently programmed VI 15-sec schedules of

point delivery. All schedules in the present experiment

consisted of 15 intervals selected according to a progression

(Fleshler and Hoffman, 1962) that generates VI schedules in

which the probability of reinforcement is approximately

constant with respect to time elapsed since the last sched-

uled outcome. The light color and associated VI schedule

were changed after every point delivery.

In the third session, responding on the main operandum

in the presence of the yellow light behind the left screen

produced the red light and clicking sound according to a

VI 15-sec schedule. Responding in the presence of the red

light and clicking sound produced points according to a

VI 15-sec schedule of point delivery. This sequence will be

designated as Chain 1. Responding on the main operandum in

the presence of the green light behind the right screen pro-

duced the blue light and clicking sound according to a VI

15-sec schedule. Responding in the presence of the blue

light and clicking sound produced points according to a VI

15-sec schedule of point delivery. This sequence will be

designated as Chain 2. Point delivery in Chain 1 was fol-

lowed by presentation of the Chain 2 initial-component

stimulus. The entire sequence constitutes a multiple (mult)

,jle. Fki ,jri 1 sIr. a dl air~,m :.t crx: .fr:.-..iJr-.

in -i. :'Jr -i r -u - --':.,r C.Ir^ c _r:.-i jr- .'i r.'n a ju :i"

ct'. r.: i. r,-.- :r .: ,- :r [ r. ,r.i .sur in? ri tIL _r

iriiLL l-c' _a'rI.- d r._ -:, r, r.t rr:.d, j.- J r.Ili _--.LiTuj2l' ; :- :,:iox r.,.J

'.:,r c .t c'r, i r, c i al-::.], [.:r, r., "I i' -"--ec --.sh_ I jil.-. .3 I a- T ,J

Wi tr-. h ch a j i J i t~, in : f tr_ t ,r .- :i- :r .-c r x u ,, r rl-- .r t ,- i u if t _. .:.m -3'

a .:orcjrr-,- ( :r,: .:rin '.r 1i-- : 'I l'-'-se i i,:rn "II
it'--.-:,: -'.I L1--e,, s-:~ ~ l=

Figure 1. Diagram of the mult (chain VI 15 sec VI
15 .sec) (chain VI 15 sec VI 15 sec) procedure.

After stable responding was observed under this sched-

ule for each subject, the schedule of point delivery in the

terminal component of Chain 1 was changed to VI 60 sec.

After stability was observed a second time, the Chain 1

terminal-component schedule was changed back to VI 15 sec and

the Chain 2 terminal-component schedule was changed from VI

15 sec to VI 60 sec. Responding was judged to be stable

when the relative response rat- in rc, ~rn ial anJ i i t irinl

components showed no systemati: Lr i. Ti-, r:.ii"' r'::ra.

rate in the initial components -: j :uI ,1j. r.i t5-.. 1 r,,-k.p r

of responses in the Chain 1 initial -.:p-:-.:nrce 3jfi~-e bt/ r,.

total responses in both initial,nrntr ThJe :.rnae

rate in the terminal component a al-ulaeji .. tr, 1,o-al

response rate (number of respor.ns i'.'ii .,/ r~ r in tha

component) for Chain 1 divided t:r trhe. i.:a) rate

for both Chain 1 ari Ci !i, 2. A zu-air; :f cnEe .r*:j'l. s *:f

point delivery arin nr rn~lm-r .f r s. ins ar .;:r emil :*; i

is presented in Tabcl 1.

(N 0C W


Ea En U)


U; 1D


Figure 2 depicts the median relative response rate

from the last five sessions at each schedule as a function

of the mean relative rate of reinforcement for both subjects.

A matching relation is represented by the diagonal broken

I n._. Ir,, *i f,:.r u. jc i.: r -. ... -h' ca cl': F: *".tc ...i 1 -

tiir, tr.:. a .li r. ich :i1 arIr 1:. [.r. f:r:r r : f.: r tr :

he ,i-ad 'f r 0 ,j i "er' it-. tri, : s-,all-r al-:, thas, 1i

Fpr di.Cr-., t, thE reisa a- r'te r :.f r 1-iL'.r: -rt. l,, ar

i.:.r -: i-::r. F. H. rfr-:.:r i nr, f :-r-:n..e r-t:. e- rn ti-' e -ijc "-

va)u:'i :"l-; i lj ; *:.f ,lint i i r; tI e Er,:.' .are Fr-, :r-

-:r.e f-. cr rnr: E,, '3al4 : i-:; .3 j -l trn -uli t:r pr---

dij'-r :.n l --' ":' cer- r l a" r r.. :.1 rl -- 'rf: r : p,- [-ri nt.

Fi-ur- d: pi.:r.c cI t.r relar t r-.i- I-- nrt in rn

ir. i r.i l :.:.T.p r, r a r j t.-.i n .:. r r: r,.a r-iati-' r-t'

:'f r n fir:.rc:a :nt fr[r br tr suJbj-:t during the last fi-e

5 in.:.r:- e :r arh I rie'1 lI. n* c n r,. aSin. tr.r pref -ererre f

,-s r.]-:. t ,3. ,:l1.-:.e;i- Fpr:.
wich .nir Eli irv i" i.. E preferen.-: rtrian i. pr-diCi d t 5

Fratc:r, in ri) ation f r the t mjil-r-'.'jluej t-clh'~iul c- pint

di '.-r f* r r eie s.-:.iel .: zJ.Im i...n,. C Lbject F. ii. ho1..-ed

gre- ater prf.rr-n-: f:r r-he srrali-er-.alued s':h dul- *:f point

dEliver, r-thn pr-i.rc.1 L;' mr -.:hin .

U2l. 00





0 .40

0 40

g .20-






.20 .40 .60 .80 1.00


Figure 2. The median relative response rate in the initial
components as a function of the mean relative rate of reinforce-
ment. Values are from the last five sessions at each schedule
for subjects F. H. (closed circles) and S. G. (open circles).





20 40 60 ) 10,0

Figure 3. The median rl3ti'-e ti'm -'rent irn inti
components as a function of rr,- m-i: reilac. rcr.- .-. r;ifn-
forcement. Values are from tr.i, lasr. 'e.i :-:- isn aC .-3-r
schedule for subject_ F. H (.:l.:i d ir. c -l i n.j .3. :.p-en
circles) .

Figure 4 shows the median and range of the relative

response rates in the terminal components from the last five

sessions at each schedule. Subject S. G. showed a slight

increase in relative response rate followed by a decrease

under the second and third schedules respectively. The

median relative response rates under the second and third

schedules did not vary outside the range of relative re-

sponse rates obtained under the first schedule. To the

extent that such a trend can be detected, it would reflect

a higher relative response rate in the terminal component

with the lower rate of point delivery. Subject F. H. showed

consistently higher response rates in the terminal component

of Chain 1.

Figure 5 shows a cumulative record of the responding

of subject S. G. during the final session under cone

(chain VI 15-sec VI 15-sec)(chain VI 15-sec VI 15-sec).

Records A and B show responding in the initial components

of Chains 1 and 2 respectively. Slash marks of the response

pen indicate presentation of the terminal component stimuli.

Both cumulative recorders operated during either initial

component. Record C shows responding in both terminal com-

ponents. Responding in the terminal component of Chain 1

is shown with the response pen deflected upwards,and respond-

ing in the terminal component of Chain 2 is shown with the

response pen deflected downward. Slash marks of the event

pen indicate point delivery. The cumulative recorder


- i; 1 "i I -E "i ;,'J "i i5 EE
, '.i I'i 1" "I "1 :- ." "Ii i5 ': "Ii :,i .E C

JCHE i.LE L E F Pr:.iT L.rLI'. r

rFijuc r TI-h i lrzi.; a j rrani e Cf LIt-- reiatr -i
r I:p.,: F n ; -i -..n- r i ir,1l 7 i:.i., .. r : .1 itals ir
L Es l:', at : a'. I.jI:t t.r u l :. lt t.:c F. i 1. 11 .-. :d
nircl-: an. ,. i .n 7ir:i5 1.



5 rA1H



Figure 5. A cumulative record of the responding of
S. G. during the final session under cone (chain VI
15-sec VI 15-sec) (chain VI 15-sec VI 60-sec). Records A
and B show responding in the initial components of Chain 1
and Chain 2. Slash marks of the response pen indicate
presentation of a terminal chain component. Record C shows
responding in the terminal components. Responding in Chain 1
is shown with the response pen deflected upward, and respond-
ing in Chain 2 is shown with the response pen deflected down-
ward. Slash marks of the event pen indicate point delivery.


operated during both terminal components. The higher

response rate in the initial component of Chain 1 is re-

flected in the steeper slope of record A. The pattern of

responding in all three records is one of brief periods of

high rate responding followed by short pauses. Pauses in

records A and B are often the occasions for responding in

the other initial chain component. These same pauses are

seen in record C, however, when there are no programmed

consequences for alternative responses.


Briefly, the results of the present experiment include

some of the following points: 1) The relative response

rates of human subjects display preference for the schedule

in the terminal concurrent-chained component with the higher

rate of reinforcement. 2) The distribution of time between

initial concurrent-chained components constitutes a measure

of preference showing a very similar relation to relative

rate of point delivery as relative response rate. 3) The

changes in relative response rate as the schedules of point

delivery are varied are seen only in the initial chained

components rather than reflecting a generalized change in

response rates throughout each chain. 4) The present find-

ings generally replicate those in the concurrent-chain

literature through use of a CO-key procedure.

The first summary statement must, of necessity, be

conservatively worded. The measure of frequency of rein-

forcement which most closely matched the measures of pre-

ference for both subjects is one which has been found to

have limited generality and questionable reliability.

Although relative rate of reinforcement most closely matched

the preference of both subjects, relative immediacy of

reinforcement provided at least as good an approximation to

matching for subject S. G. The relative immediacy of rein-

forcement would suggest less preference for the schedule

with the lower value than predictions based on the relative

rate of reinforcement. This was the direction in which the

preference of subject S. G. deviated from matching the

relative rate of reinforcement. The preference shown by

subject F. H., on the other hand, deviated from matching the

relative rate of reinforcement in the opposite direction.

One variable which has been shown to generate more extreme

preference such as that seen with subject F. H. is large

absolute magnitude of the schedules in both terminal com-

F:r, r r -; (,.'. --i an : r.. Fr ,r_,_r,':, I ?7 : 'I ..: r, 1 7' i. I it

di If.:.ji t t -.[1 i- r -I : *rj El -;- :r: l -:

. : .' ~ : f f i ri r f- i ': ri- .,- a I a r -- r-r.: -1 .i r

t Lt:r -rs t r f -i i_ _. : t:.r- : ,i ,1' .,-- r i -.r rJ'-.,_ -

i r, t : r i r r r r -r_ a fr t : r r,.

Er- r 4 -,li r, (t ' hr n r:.Lli r :, ] j.'-;- i tl, .

Ti'i.: r : ,- : ,, r, ,:,r, 3 r r l : rp i,5. [-: ".:rij : : t

C..- ,6 ri_--:r, r-: c : :rth tr, :e ': ,--'. ,s : . It ?:h t: l, '.l,_t.:

',i_-,,',_i r [1 I' [l- 1 -; "-. t. : .t": i r'/ ty I t i i to rli, rl. Th'-

Js:r.i.:r, r,. "-,- a Lir i_ [hE: r r:"[-r klca:e. i.:.r a :ri rinli

*7: ~i.:r ;: *r tn-i a t.: [; *. ,Isa-;.'ir.: : i:- t *: t'j14 i'i r. i' . n

of experimental control which characterizes the better re-

search with pigeons makes that species a far better choice

at present.

The fact that the present results have generally

replicated findings from other concurrent-chains research,

suggests that direct comparisons can be made between

concurrent-chained-schedule experiments using the two dif-

ferent procedures. The same advantages which have brought

the CO-key procedure into widespread use in research with

simple concurrent schedules recommend it for use with

concurrent chains as well. Schneider (1972) recently has

urged that caution be used in interpreting the response

rates maintained in initial concurrent-chained components.

His research suggests that it may be difficult to evaluate

the effects of delay of primary reinforcement as a confound-

ing influence on relative response rate as a measure of

preference. In this regard, time allocation in the initial

components of concurrent-chained schedules could prove to

be a useful addition to current measures of preference.



This review is intended to provide representative, if

not exhaustive coverage of the research into human operant

behavior in which intermittent schedules of positive rein-

forcement have been employed. The organization will reflect

Cthis Ei-i, m I : I :.n ti e "a .r .-,u: : 'rI e.Jl, i ,_-f rIf.:.r..---- rr t.

r.'inin :ic- n ra r : .r ut,. rEl;-i.:n, ri :.r er --' c.: r. r.-

t' rer subT jE.:r Ij=. in 3-.e : erli r.n r.-:r a- rC.; a3e, arn

['rr-'T nr:rrr.aii -.- r er. rjar. ar.i g-. :rnir-c s.

fo ilkz i E.' -r. -in -r.e. t-',-11 Ti-,i I ,t, , : --a l,-_r E ,in- :,--

t.pi-all5 / c: rare. ,' FF :he uies in rinf r ur.arn pecieE i

c riactE.r. i d c.' a rn qr, t.Irninal res! :.nn-e rate. A- r Cn-

'.lu- :f r.r F ec..Jdui e i inc're.a.i, pau_ 1--. *:c r ty .zcl 11

afe r r i e c r- nt, pr juci ng a r t-i- -lue3d praern .:-f re-

sprce rat-c if -rtcr and inner, 15".

Holland i1953ii erplo-.lyd FFP chedules in hia in-rti:a-

ti.-n f hiii.arn ,br.:irn rr nm,.nilntnr leha.'ic.r. In tris

study, normal adult males (Navy enlisted men) were required

to press a key to briefly illuminate a dial so that they

could report deflections of a pointer on that dial. In this

study Holland demonstrated that the observance and report-

ing of pointer deflections (which reset the pointer) rein-

forced the behavior of lever pressing to illuminate the

dial. Thus an FR schedule was programmed by following every

nth observing response with pointer deflection. The

behavior generated by FR schedules of 36, 60, 108, 150, and

200 responses was characterized by high, steady observing-

response rates with occasional brief pauses which occurred

after signal detection. Holland (1958b) has also reported

a different and atypical pattern of responding generated by

FR schedules with one subject which was characterized by a

high post-reinforcement rate followed by a lower rate, which

was in turn followed by a high rate immediately prior to

reinforcement. Upon inquiry, it was learned that the sub-

ject was counting the responses as he emitted them. Holland

then instructed several other subjects to count. The pat-

terns of responding displayed by these subjects showed an

accelerated rate prior to reinforcement, but not afterwards.

In another study in which normal adults served as sub-

jects (Weiner, 1964a), button pushes were reinforced by the

addition of 100 points to a counter mounted in front of the

subject. An FR 50 schedule generated high terminal rates of

pushing without any pauses. Pauses were produced, however,

when 5-sec periods of punishment were programmed immediately

after reinforcement. The loss of 1 or 2 points from the

counter (response cost) constituted punishment. When

response cost was programmed continuously, response rates

and patterns were essentially the same as when no cost was

programmed, as long as the cost was less than the reinforce-

ment. Rapid cessation of responding occurred when con-

tinuous response cost was made equal to reinforcement.

The behavior of normal children from the ages of four

to eight years under FR schedules of reinforcement was in-

vestigated by Long, Hammack, May, and Campbell (1958). The

response used was a telegraph key press or a Lindsley manipu-

l rr, 1 ,Fu:. i. I :r. F:, r. : .r .. -I, r : , :, r r .. ,: ,

r1 *: : I :. 6 : r -h, I I [ ; I-I : l I

;-[,', !".-._' 3: c rr:r" -; fr:m i'i 'Fr : :h l -r1 i,' r,-, ; r r-;-ir :-_"

r.-- F :.r. ir- L : F L r :.*fj. i r r i P L'-u 1.

r,:.:.rc ic f: _r tr : r r.i_ T. r-r :.. i-,j.:,-, ( l r.r,,_

i -r, e l i r. 5 .',-,r Il..,., : .':i' l.l . r- jr. r -_ r_ r j. ,:27 :,ir

: ',_-. i.,r.: r [ -...-.j[. r ', .: -r r-i." i..,r.'i-- r,,- L i l'-t ,j :,.

I I: r, r i ..i1.1- i -. T. [ :r, : ri .-. r2 !! al l ,. ,

-;ri-ir r :- i -I i1 i I._ :1" : '. ;--* : i'r, 1 :.L ." ;-',j--r- i.-itr-n,:Iatn :e

'r -: ;, 'r 5 '-*i-a -- ] r, J .J L-i r r, Ir."'-:, Lri tii."

-' i-.lri .5 rr, r L r ,n r -. p :r'-.rrrr' : r C r

subjects who showed good ratio patterns of responding at the

higher schedule values also began to show deterioration after

five or six weekly sessions. It was demonstrated that this

poor schedule control at low FR values, or extended expo-

sure to higher FR values, was due to satiation effects, by

the temporary improvement produced by increasing the period

between sessions and by introducing new trinkets.

Performance under FR schedules in normal school children

has been investigated by Rhinegold, Stanley, and Doyle (1964)

with subjects age two to five years and in infants 11 to 19

months by Simmons (1964) and Weisberg and Fink (1966).

Using short durations of music and movie projection as rein-

forcement, Rhinegold et al. observed high, steady rates

at FR schedules which were gradually increased to FR 18.

Using chimes as reinforcement, Simmons only marginally demon-

strated conditioning on an FR 2 schedule. Moderately good

FR performance was reported by Weisberg and Fink using

cookie bites as reinforcement at schedules of up to FR 15.

Deviations were primarily in the form of variable-length,

post-reinforcement pauses.

In a study in which adult retardates served as subjects,

Ellis, Barrett, and Pryer (1960) reported FR performance in

several subjects which closely resembled that of lower

organisms. The cumulative records of most subjects, however,

showed various atypical features, the most noticeable of

which were broken "runs" and other indications of "strain."

Post-reinforcement pauses under schedules as low as FR 10

were also observed in many subjects. Another unusual

feature was the high resistance to extinction shown by many

subjects. Perhaps the most striking aspect of this study

was the considerable variability between subjects in perfor-

r, 3 r c = .

E.y *::.F ri Son .t t': jitD r : rF. c a.b:.'' cr,t,

g.r,. tii j 3: hr :l n I i nri, :,- rrniii: ., = -,-r, r,,-.r- c_ p a:l1

FF F rf:.rin .r: Hlur.h ir.;:r i a rir,, 1. 1,i. P Ju' Iri a3 1 .,

F F .'ajlu s l. rare, jrn ~-n:r t "ji ur s u. c,: FF ::1, pju 3s

.' c r tri f ;rn I rir u r r r.- ;:-.:jul . u Le rih i l; .- n

a:i i ; inr r: : I.:.. Tr.:.*2 in Cjrnc.e -.u. irnIl 3rn d r.rair

.* :*, -r -.._;r' :3 j ,:re l r eil' 5ctesbt tr. ".,'.'erl: rj[a ij

1ir ri irn 3 ~f tri,; FF : reI.dji r- uij re.mrr.t -be. u-iU_, tr

.at..:r :r .: r". 3 r r:i.J.1 s1 '.rinjti .a t rirn t'.-= FF

__-I;_.j.l: r,.i m,.,r,. grJiaJ ll i-,Creasir i3.- r a r- .7 t n:per:-

*lahl.,': 3 ir.,1 r c: th:. -: 3L*tC r.n -,r i -n Studi,. s ,: i* ,nfra-r1w-Lar,

anrd r..:rn.51 hum~ rs. Th- :'rCe -: Ir. n C.:. t.nis 3 : :n:tial arit' is tlh n pr fl tire r..:zi; ary to pr:.-

duzr nii ri rjt.: :.f resFJ.: irJ n .

Ir i a iu-' .irth inpatic nt re~ .arj-j ildr-n, Barratt

jand Linj' l-'. r 'ii rep'rt:-d p-.:r FF prr n..anr Trne :zuse

f:r tris i.3s nri.:o clear, hb: "e er. First th. pr.r-..* jur,

.Fempl,.,; i t3.. r.anipul nji j-ach *f '3. r:i n i.j3 b.el-.-' a liqht.

i~.CpCiiri-*3 *r r ne nanipui ar-i' .3 cr re.':r reinrforc d,

regardless of whether the light above it was on or not, and

responding on the other manipulandum was reinforced only in

the presence of the light above it: hence, a cone mult FR

EXT mult EXT EXT schedule. Higher rates were reported under

FR 1 and FR 3 than under FR 10, and intra-session varia-

bility was considerable. The analysis of schedule control

in this study was complicated by apparent stimulus-control


Rather typical FR performance was reported by Ferster

and DeMeyer (1961) in a study with two inpatient autistic

children. The schedule values were never very high, however,

being increased only to FR 30. Pauses were quite frequent,

although rarely occurring during a "run." A different sort

of indication of usual FR schedule control was the disap-

pearance of some "superstitious" behaviors when the schedule

was changed from VI to FR.

Variable-Ratio Schedules of Reinforcement

"In a variable-ratio (VR) schedule of reinforcement,

the reinforcement occurs after a given number of responses,

the number varying unpredictably from reinforcement to

reinforcement" (Ferster and Skinner, p. 391, 1957). The

value of the VR schedule denotes the average number of

responses required for reinforcement. Orlando (1961b)

employed a VR 100 schedule of reinforcement in a discrimina-

tion study with retardates serving as subjects. Details of

the experimental findings will be presented under "concurrent

schedules," but in summary, the pattern of behavior observed

was one of steady, high-rate responding, with little paus-

ing. In a study by Scobie and Kaufman (1969) of intermittent

punishment, one baseline was key-pressing behavior main-

tained by a VR 210 schedule of points exchangeable for money.

College females served as subjects. Again, steady, high-

rate responding was obtained.

ri--ed-Tnter---l schedulee of Pcinfr=er-r nt

in a f'*.j- r, -i 't'r"al IFI) :*._eJle ..f r r 1f:.r:a' r t, r.he

f i r, r : ,- err. fc :.i : t e fir- r, p. nse 5eF r 3 Ir, n- r'ar i r Ia
in Ier. 1a ._ ti : ,e ,el [-.i.: tir r-:ter ar,.i k j rinre- 1i '-;ji).

i1, the pre-" i:-u. di :u eC i n' t i lc :r, *:*f uTrn I,:-i't. :r-

in1 t eh="' :,r, 1 h:1.1 d Ii)' l i s: eT. pi:1 ei fc j~ r n .er'.' i

h,-, j, le *: .:.f _reitr_ :r:el-, ,r., i.e. che ftirr Ji l-lll -irinlt-

irqr r c r-l=e ic:t r r 1 rE er.n.: af'er ia onr= ri t e-ri r

OF r.i ff r.- r.h 15. r. .j tr.- r. Ci r. l J -L i 2-.d, -r :,"3r: ,J .3

po:incr r J' le r .' h.'r al a: j,. r r l-, a .=rc e: .p: e- j t.-

ight. 3li-m' _ir, j Eir,., .t e-ja h jlue, ..rni h ',-re :-. 1-.

2-, 3-, arJ J-1i r FI :h le of -h Jul r p 'nter J fi:t i:, I'

patj rnr re -pcr niiin 1 r enri rate.j uri-Jer aii scheduled ,I lu ,-=-

i s i rti.: ji r .hT th 't :r:duced bc ri'-*l r i rtIlarhurl rn spF:e ie;

*frfj,-r FI .:r- i.:dul= :'f pc: .it i r r--; l ci; rt n IF'er cr. r nri

5;inr-er, i ~' l F sing occu'rr,-ed iiurirFji el:,' ja ter rein-

f.-: r:e enr.. Tnr n r;e ons.; rat-s ifT L 'e r i hlh e rminal

rate -hiicn ,'s [:,aintjiri-i jntcil ieflectcon occurred. The

:-rii c frcrm ia cr.:. c:. high r :e j uIjall radJal, tut was

often rather abrupt. As the FI value was increased, the

pause duration increased as well.

Later investigations have indicated, however, that this

"scallop" pattern may frequently not be observed and have

determined a number of variables which may influence the

pattern of responding. Weiner (1964c), in a study with

normal adults in which key pressing was reinforced with the

addition of 100 points on a counter, found that the pattern

of responding generated by an FI 10-sec schedule of rein-

forcement was greatly different for subjects who had pre-

vious exposure to an FR schedule compared to subjects with

previous exposure to a schedule differentially reinforcing

low rates of responding (DRL). The former subjects re-

sponded continuously through the interval with high, steady

rates, while the latter responded at very low rates, typi-

cally pausing until after the 10 sec elapsed before emitting

the first response. In an investigation with similar sub-

jects and the same response and reinforcer, Weiner (1962)

studied the effects of a response-cost contingency on

responding under an FI 1-min schedule. When each response

did not produce point loss, a constant rate of responding

was maintained throughout the interval. Under the response-

cost contingency, however, the rate was greatly reduced,

with pausing after reinforcement followed by a low, slightly

accelerated response rate. Removal of the response-cost

contingency produced a reversal to the previous pattern of

of responding maintained throughout the interval. Imposing

the response-cost contingency in either the first or last

half of the interval produced pausing followed by sustained

responding and sustained responding followed by pausing, in

t-..: '. --1 :li :-.r.diti.:.r. r.:.l ..iri p-'rf -.r5,:]n,: iril!ar t:

l- ,- r_ r.:- a b:"- i n 1i rr r L in t' F ljr.-- 3 [-:.- iL' :" ;

' :'- r. .:r n : --._ t ir.l-, j, t_ i n. r.;-".; r. i,,.,l p i. L'-. ..r.i

r i ._ ', r_ : r- t i' anr i r.: -fric .. ri r. i rCi 'i I : p -' r. l : r

I in L i' r : r i -:r_ :,ri 1 1 Jut.;i : r, :. ::na -

a i r ; ', : r l-;: ._-, -, ., ,r -n.i, r n rj-,.; :r,:.-- '.:* t ,; -nL ,:.. tr_ r:, [,Fr .
, _'r -: :[.. r...i: '., -i r, ': ri;-: .;r *:*r C .' r r inr

P r,-'.r, ',r : r.anr ar..j .t ..:r i : 1l. r- u t .: r i t, i:'

c: .a : r urd._L 5 Fa rC' F , l Ci ., L ff :-r. ,L FI

i:. j'l.: f rr n tar- r Sin r -rz.-:ri.r,. i .: .'j- rU ..r

,-'i r t ,rj .-i t: -,u Lr I- l -h-,i ul.. r:'--:-nii at i. .-:, :nr,

!:::+<511a u _-.En U 7ar r- .f th: r.itri- 'ilUr

p:ri rr, F, :: i- 'r : u .iLb: .:tr a i'.'n r i n trF:r C li.n abcr-

trt,. z r, ''l: r, r.:. :d. ain :r-erl [ r. ir- -i:n -f r;t:F and

r. ,p:.tai n a fun,-rcti- : -. f t e', ir, Le'-r.- n'cr:.irFri

inL- I. rLitj -i r.:- r' :ri.f --- .:i3 Imr[r:. :-r.h i': t r-E-JC

ni uni nsl -ru Cr.- i t ts:: si-- 1.- .'.' 3n .di nfr-r-r:ntit:i ,--

r t-. i. i r, r r (lC. r-tp-rt-l at r,-.t ':inl_- Jo. aut'.r.:C

Cr., .er.:ri_ urj-.:r FF :1-. -i laj : 5,i.d [,Lj_ 's:F 5.-i S resp -ri.

,n t r rr I:'i .-i rt j-: f -ii n unrJ r i I .:'I hed-! l s, bLut

hals.: r..r u:,..i ::t" iLh l Lt r LFL and rF h-..;.-r i'-s LEnd E:.

raJ--: after re f :,r, '..n.:rr. anr, r -'- r1d at r-i ati'.'-]' l-"

rU .

Azrin (1958) reported that increasing the force re-

quired to make a response produced an effect similar to

increasing response cost on responding under FI schedules.

When the key- or button-pressing response required little

force (about 12 g) responding occurred throughout the 3-min

fixed intervals employed in that study. When the force

required was increased to several hundred grams, much lower

rates were found and greater acceleration of rate within

each interval. This is consistent with Weiner's (1962, 1964b,c)

findings under no-cost conditions where the force required

for key depression was reported to be 20 g. The force re-

quired in the Holland (1958a,b) studies is not known.

Laties and Weiss (1960) employed a vigilance procedure

(ref. Holland 1958a) to investigate the occurrence of re-

sponding immediately after reinforcement under an FI

schedule. Identifying the variables controlling this post-

reinforcement responding was of interest because pausing

usually occurs after reinforcement. Laties and Weiss found

that the frequency of the occurrence of this post-

reinforcement responding could be increased by either

occasionally causing the pointer-reset response to be

ineffective or by suggesting to the subject that the reset

button was operating imperfectly. Post-reinforcement

observing responses could be decreased by accompanying the

previously silent reset with a loud auditory stimulus.

The response patterns under FI schedules in another

study by Laties and Weiss (1963) were similar to those

reported by Holland (1958a) and Laties and Weiss (1960),

with pauses followed by abrupt shifts to high terminal rates.

The particularly high terminal rates found here were pro-

duced by the addition of a "limited hold" contingency where

the deflected pointer was reset automatically if not detected

within a limited period of time. These subjects had also

been instructed as to how frequently the pointer was de-

flected. After a few sequences of responding in the above

described manner, subjects were instructed to do successive

subtractions backwards from 1,000. The pattern of observing

responses was changed, in that pausing was shortened slightly

and the increase in response rate was much more gradual.

Lir.l. i ; ll.i ~- .:.:.r .:lu.j.j l fi h ir ;ue n r l .,. -r' *-.

n r. *rrr r ,r F E rir*-..r' Ji l il ri [iiir-jri ; i r -.j f ,' !c ,re ,- r-

.f :.r .'jrirn :._r i r.n-i ,i i:.r. OC'i rup :r. :r r.r. i: t -

F. 0 .-r1. Tr : 31r:: nri"-:- .. ,:i ;.r r,m"e i,. --r.iir,:. 3 ren.- lr.F,:-

1. ),',' T it,:' :,_ . ri-'..j i-,^ r r r r.ri.e iini .al [.Pilrr ,rer' -

:.r :rl ri n .- l" .-~fi .e t fr, j i. .-.i crnc,- te-7.:.rd r -e3F)n.s--

The icrni ulp i r r, :. rin.s .r r ii Li ri.j .'mZ .-r-rher'.i'r .

FI 20 sec. They concluded that neither variable exerted

significant control over button-pressing behavior. Leander,

Lippman, and Meyer (1968), using the same apparatus with the

"light" (20 g) button, varied the FI schedule from FI 20 sec

to FI 40, 60, and 80 sec. Results indicated that the pattern

of responding was highly correlated with the subjects ver-

balization of the reinforcement contingency. The subjects

who were instructed or who afterwards reported that points

were delivered after a certain period of time showed lower

rates and much pausing in their terminal, if not initial,

performance. This latter group was more likely to show in-

dications of scalloping. Factors which may have influenced

the Lippman et al. (1970) and Leander et al. (1968) results

include the rather brief exposure of each subject to the

contingency (approximately 50 reinforcements) and the short-

interval durations used (the longest was 80 sec). In addi-

tion, the "effortful" response required 35% less force

than that used by Laties and Weiss (1960, 1963).

Long et al. (1958) investigated the behavior of normal

children reinforced with candy, trinkets, and color-slide

pictures on FI schedules ranging from 0.5 min to 3 min in

duration. The subjects who were begun on a 1-min variable-

interval (VI) schedule and switched to FI 1 min, or begun on

FI 1 min, showed more evidence of scalloping than subjects

begun on FI 0.5 min or a low FR schedule and switched to

FI 1 min. This is similar to the historical influences on

FI performance reported by Leander et al. (1968). Children's

performance also frequently showed deviations such as knees,

multiple runs, second-order effects, and negative accelera-

tion. The difficulty experienced in the investigation with

"weakening" reinforcers, which was noted earlier, may well

be related to the variability in the patterns of responding

reported here.

Bijou (1958) compared the performance in extinction of

four normal four-year-old children who had been responding

on an FI schedule with trinket reinforcers distributed

according to an increasing ratio, i.e., requiring increas-

ing numbers of FI sequences to be completed before trinket

delivery. Interval durations varied between subjects,

r _it I-r ;:,, :3,i, .1. ,;0 r -. F .-' rc.n- r ar- ai, i r ,- i r r:..rt

rt: -.r,;- _-- n r t'.ln :tiri n .-*r iire.Tz.t :.:.r :- jjrar-i.:.,

C.j.mJl r..j l r c::rrj; r,.:.r-: -. / r: -'--r, .i rhat :.r r* F. .--r.i r.: -
t c.ri trl i-i F I 'r.i l:ri r tir. tli r,.L' 5I- j r.- F p : 1-. iI

*~ ,r. t -irn r._ r Fcr i j, rI5 -. ir. o r iir r .nji r ir,' FL

1-r in 5:l-.F.= -l c*:. *: arJ *:r : m re-tte. r.:;r'f.-.r: .-rn it il1 ---

Er. al., 1 ..r.: =- r.r.rm.L, : ri ,.- L.r... ra .-' jE _13 Tr e

r.tir.:.-ikl hi r, r r.:.' r, b, ,ran, :wIDo-:r.r m DCe r- l

tr.:. ,-, f-:r. .:.r 11 :-f r.r 12 t-.ion- prx.:.r ..:. init13-

tii_-n -.f r- F!i l in ].-Tir 1 ; FF ?.:l. 'ijl-; si b.-n in

erf-=tL. T1 r"- i i [: 1 .- -.m irlir .-_ D L..-F -n :Y-, .ur 1-, l1Ati .'-

records of these subjects under FI 1-min and FI patterns of

responding of humans or any other species reported in the

literature. Orlando concluded that "temporal discrimination

observed under FI schedules of reinforcement are related to

at least two behavioral components: the use of the temporal

interval as a discriminative stimulus, and withholding from

responding in the presence of stimuli associated with non-


Variable-Interval Schedules of Reinforcement

In a variable-interval (VI) schedule of reinforcement,

the reinforcement follows the first response after a vary-

ing interval of time has elapsed, the interval varying un-

predictably from reinforcement to reinforcement (Ferster and

Skinner, 1957). The value of the VI schedule denotes the

average interval of time between the availability of rein-

forcement. The patterns of responding generated by variable-

interval (VI) schedules in lower organisms are typified by

those of the pigeon reported by Ferster and Skinner. This

pattern is characterized by an extremely stable intermediate

rate with almost no irregularities; i.e., the cumulative

record approaches the appearance of a straight diagonal

line, the slope of which depends upon the schedule value,

the subject's deprivation, etc.

The cumulative records of a normal adult responding

under various VI schedules (Holland, 1958a) are qualitatively

identical to pigeon records. Perhaps the only unusual

feature of the human records is the rather high rate, which

is more nearly like what one might expect of VR rather than

VI responding. Holland also reported a slight increase in

rate within sessions at VI 15 sec and VI 30 sec, but a

slight decrease in rate within sessions at VI 1 min and

VI 2 min. Between sessions but within subjects, a decrease

in rate was observed as the VI temporal value was increased.

Weiner (1962) reported the same decline in rates as average

jdjr ti'.s_ i= incre.a;-J r:rn '. I 1 mT t.o "'I 9 mir. in 31 -i-

ci.rn, r.h5 iip:.:r it-i r : f r:., ,.: r.isp:.r,:- ,-c r. c, aus-i I

jd crea.E i.i rir, re l j t:, rat ;i J r c r. : r.-c. t ,_in-

ji ir.i.:.. [. j. crupti .- t, t ie r rrccrn :C.-. jrr- '-, -

- r., t~er. .:.rl/ ria. i in i-: e.

i. :' ;r I' l'-' rei fr.:.r.:i j t i :lr r..r.r rr nl r i r rli i l

j.-1ua .:.r- I 1;-r -:.: -:ir-dj l. : t r 5 -i t.:r; .:r r. ;p.:.r.- r

uni r -:-l:n-:r FF. '1 *-.r ,.iL 1,) -e:. '. r, l i rl- qir titi ti "a

crAn : arj r,.i quil titic- : 5-i.-h .: :.:,.rr.:d r. i ri-- [ a-ti err *:f

r p.-:r.Jin;i :.f r : -, ]..:r ; ..i h ; i t r, .. r c, c :.-. r,'i,,I.:.r E i. ; r

r:-t- d tr r. lil. tLr. :ubr-act? ir.h trc FP hir.:.r rc--. l';e d

r,.= rl rhE i, ir ri p. ~ritl- r . f r -inrf.r: i r. n r.h,- sut:-

*l-,cr.= rE'-,i LJ l-. i, 5t,.:r r- -. ".- J 5 .-:, d ,rat1 I0 r

rac r f r-sinr:r.-..n..rr- r.t n the-' :oul- r&-'l r,; : -1.'e'.

Ui n.i -ia pr,-:-; jr- ii-i r t:. tiit r.p .rt--j by' i tiner

l ':9 ,2, 19645,b. ", 1 .'i ), C '3u ffi[mr., Ezr.:ri, ani i':.-p (Cl'.

rinrfi-.r-;- no.:r-ii ajults' k.- pr .- ir,,i rith pi'--nt on

'.'I 1-min :rnc i .n. Tie 3ru.n-ii -- :'r.n E' -.'i diffzr-ent

sets of instructions prior to exposure to the experimental

task. Group M (minimal) received virtually no instruction

other than to accumulate points. Group R (response) was

instructed as to the relationship between responses on the

key signal lights and points. Group FI was given a descrip-

tion of the point contingency depicting it as an FI schedule.

Groups VI and VR were given a description of the point con-

tingency depicting it as a VI and VR schedule, respectively.

The FI group showed a low-rate, spaced-responding pattern

similar to that displayed by adult subjects who verbalized

a temporal contingency in the study by Lippman and Meyer

(1967). The M group displayed a pattern most like infra-

human responding under a VI schedule, i.e., an intermediate

rate and steady pattern. The subjects in groups R and VI

showed patterns similar to those reported by Holland (1958a)

for subjects under VI 1 min. Subjects in group VR showed

rates and patterns of responding very similar to those of

retardates under VR 100 schedules (Orlando, 1961b). Thus,

control exerted by instructions was well demonstrated.

Long et al. (1958) reinforced the button-press responses

of normal children with trinkets and pennies under VI 0.5

min and VI 1-min schedules. Patterns of responding during

early sessions were similar to those reported by Holland

(1958a) for adults on comparable schedules. After three

sessions, however, overall rates began to fall and pausing

increased and grain became much more rough. This deterioration

tended to occur earlier where VI schedules were used which

w ,re .- or sitr.jt- .;ith f ej s ort int r-r-ir f,:r.ieTra erI inter r-

'.val Ti-ie 3a~ inirifr a f ~ ; t ,eras f trir,.ic aRri, pAF-enn'.

r:infrcer-r in -.arincrainri F.R and F[ rea;nrdirn itn this 3 ruj

3;: described j re.?..:.u.j l .

Th-e perf- rn:nr -: -,f aij.le recarjate: uriner a .'1 1-min

- :h-d l.ji. ass :.n-:rl ai ;ii lar o t.iat reFortc f,:r rnri--Mal

ajkjIt ,.i th ".ar' rou r rain a r 30 s irr:i1.i ar pFa'r?

(Elliz : a ii., i9601. This a-ar.- irr..iular p.u,3in-. wi

r.-F.rted in tre :m ulaic'.' rz-or-. s -f adult inprF tiiit

:h.:.: rc-ispor'ii r, irun-.r a '.'t i-rin -cn iuii Lirri- .

l) l>) Lird e;. a i o r ;.:.r -d ii':h 'r:ac ur tP...- *c t -- -?r- it jt :t.

Jiffrr-en:r;s recn rcr car. : than ceac founr, a monr rinrrmals.

TIr '.'i p.rf.:-r, ar_. .o the r'..:. 5,t. st.: .:hiliren r .P-.rt.ld Lt

F. r cter anr .r (1) 1 11 .a rrnara-c ri.ze try. both lo.

ra^t- a.d irr-egl ar pauCi .

rDifferentci Feinfr:c.reenat cf ates

1rc: this r.-ei..i'--r n:-: urntere no stuills u ih hw ian

-ib]t-c irlp: ir: a di; ra inrorcrrement of high-

raceas ( P) arh;eul;., oni' differe-nrial reinfircerrent of low

rate IDFPL) .ill .:,-o'.-ered here. rr i DP-L sch-eule of re-

infcrc:m-'eit. tnriccnforce.-rmeint follows tne fire respcrnse after

.a paus in rcspondinji rich or ::ecid; 3a fixed period

of The '.'aiu f the- DEh L czhedijl- sp-ecifh-s the.

imirnIT.uj Jjurajtio.n of tj ht pause.- Th. inr-.'stig3at'ona reporting

crt i- uS, of DPRL 3creIujle pres=ntr data indicai.ina excel lent

schedule control; i.e., normal adult subjects rather quickly

increase their pausing between responses until a high per-

centage of responses are reinforced and near maximum rate of

reinforcement is received.

The general procedure used by Holland (1958a) has

already been described. A DRL 30-sec schedule was pro-

grammed by following an observing response with pointer de-

flection only when at least 30 sec had elapsed since the

last reinforcement was received.

In a pharmacological study of Dews and Morse (1958), a

type of DRL schedule was used as a baseline for evaluating

the effects of d-amphetamine. A second-order schedule was

employed which formally would be denoted as a tand FR n

(DRL p), which means reinforcement followed the n response

which satisfied the DRL p contingency. Two schedules were

used: tand FR 100 (DRL 2.5 sec) and tand FR 10 (DRL 25 sec).

During "training" the contingency was described, the sched-

ule values were announced, and a sequence of clicks either

3.5 or 2.5 sec apart were presented to demonstrate the ideal

spacing of responses. Not surprisingly, subjects emitted

few early responses and performed at a nearly optimum level.

In the investigation conducted by Bruner and Revusky

(1961), four response keys were available. Money reinforce-

ment was programmed for responding on one key with inter-

response times (IRTs) between 8.2 and 10.25 sec. Results

showed that subjects mediated this restricted interval by

systematic responding on the non-reinforced keys, and indeed,

after the experiment they reported the reinforcement contin-

gencies as requiring complex patterns of responding using

several keys. Similar "timing behavior" was reported by

i :- :I ir, "!' '. -. = bi::Z :r ir rru,.: : r:-I r.n ,

ri . ? i ..f i- ,'.:. u l .- iA "i. i :.r r .- r ri

* ...* li c : .... i 1 t. t I I- ;r..i: r .i ._ t r r -

*. ,. l .u l .- F,, i.... :. .:. : LF h -, , 3 r. 1 :. t' r

h-.: -r = --'.: L" r r._- :. -- .. :i t rbi: r. r i : ::r.e ,

S :. .: - .. E r .r .S r : r i r ir ".:.r r t r r- a jn, .

i- : .:,lr r : :.' i rr i 1 : : ,- -ri jr. ,:. .: .r: ei e i "I ,_. ..' _-_rEr 1 'a r_

.I'I? '. t ri P t -: 1 i- rn ir,-r.:- ..I r r iT, ri '.,i :-

rt l-, :a r : : f n 1 -,' t lr r n c t h:.: r .
S.j l t ;i . 7r lj crl : L-

ir, -,, l r.;i .i= ; .:1.- i jl. lt j,, _.), I::..:. Lur f.,ore _-:rr ul._ :- .:.f

:r r f.:.r.:,-,=r.t .re _--T r-tl3ai. 1" .'r :*. r:mF,-i -ir, i :i J, s r. r,:C' .'z:-_

"-.r r :,:* r. i ': ;ri.T i 'jl.j i.;- [-' r*:,i '. 'i. -h *:- h -Fc r -" ** w l-5,

in ti:.--: i. . *: : .:.f a r -1 Ll Ti t, cr .:. [ 1 1r, .,s ,.. r: ,f :*r.: ce j

or, c.n F F Ju : ".:..J.jl*, an i. ir, LiF,- fr,- r--'*n:- :.f a j r .n l lht,

r -; r .i ria ,. ,* t: r--ari, _,r:.r .1 .,rn a 1 :-niir, 2.-h.=duli li' arni],

I : 5. 1 I. I n :d t i l -; r a ,u l t ,- l -c -- 7 ,u l e. ,-: l 1 a n a p ': '. ,L .

S i;, i. :"l F[: pattr n:- :f c-: :.r -11-,. ,_-, r.-,h: pr.'.'e .'
l i ,t, -., r, ;:-. t:r...r. 1 r:.r]...r,=.- : a --:- 'it ,h F ij_--ir cI Crn ',

af -t r r- r. f:r:.- rr. In tr- pr .r, .: :f th.- r'.:rn li' ,at,

'it p r r, .: r r: ::.n.-ir, i r.. an ,. .'. ra[L aCC.: l ~ 'r e .Ln

of responding during the interval, but the "grain" showed

the apparent influence of the FR history, i.e., within the

scallop there were bursts and pauses rather than a gradual


Dardano (1965), using normal adults as subjects and a

vigilance task procedure similar to that of Holland, inves-

tigated the performance maintained by three different mul-

tiple schedules. In one two-ply multiple schedule, both

components included a VI 6-min schedule, but the second

component also had the additional requirement of a brief

limited hold (LII) terminated by a bell (mult VI 6 VI 6 LH

bell). The second multiple schedule was identical to the

first except that an electric shock replaced the bell (mult

VI 6 VI 6 LII shock). In the third multiple schedule,

conditions like the second component of the previous sched-

ule were alternated with a time out (TO) during which the sub-

ject was instructed to relax (mult VI 6 LH shock TO). In all

cases the patterns of responding under the various VI sched-

ules with and without limited-hold contingencies were

characterized by rather steady rates with smooth grain and

no pauses. Rates were lowest under the VI without LH and

highest under VI 6 LH shock. Responding ceased under TO.

As previously mentioned, Weiner (1964b) alternated

periods of signaled cost and no-cost conditions superimposed

on FI 25-sec and extinction schedules. Behavior was rather

effectively suppressed by the response-cost contingency

under both the FI and extinction (EXT) schedule.

Scobie and Kaufman (1969) employed a four-component

multiple schedule with VI 30-sec and VR 210 schedules of

money reinforcement for adult's key-press behavior. In two

components an intermittent schedule (VI 10 sec) of electri-

cal shock was also imposed. The patterns of responding were

similar in the non-punished components with higher rates

being observed under VR 210 for subjects instructed as to

the contingencies. Minimally instructed subjects showed

steady, high rates in both components. In the punishment

components all instructed subjects rapidly reduced response

rate in r,e '.'I .; :.,-.I.:-.r Le .r le z _;-.: .ri inrcre- d r.nd

other je:-re -se reApicne,- rate in the '.'F r:Tp e nr. lini-

mally' inacrjucr.ed s.'j e C :r sn.h..ed-j 'upFr-: n tu no. r

difference: in 'su:frssi:n r furnci:-.r :- reirnfcrc--'meint

scr ..eule. Bar:-.r at 1a. (19;'l r n" r i adultr. l y-

pr- : ain behr'. ir under a Cfie- c=-:.,. .:rner nt ultiple si-cdule

of ii.r.ntEar, r inrf:rca-rii-nt i.icn Ji ,fferer FI ''alu in :iacr

coin7.nenr.t. Tre = .itj e.--s ..nD -re i-st ru red abct jt r.he

schnedul.- s- :.ed ar, .:.rderly pr.-..-re: i-rn : reFzon'r race;

and .pa tr-rnirn a a Frincr.i:-r of tr.e irn t r-reinrfcrcemr nt irn-

ter'als. lrini 1mal, inrr.-r,:ceJ suit:)e:t -n.-:.. uniforu.i ,il'

hit-r rate_ Inr ail :.:.ripi n r.n : [nir .: .n raliy [-:rszist d '.'en

after inr cr'jcti:n 5.'- ajj.d.

Usinn nr:.,-i l :rildr-i n a-s i~ .rj:cz.5, Lon (19391 Initially'

Exp.:-d tn im c:. aE .ult F E FF 10 r.ith EJ c : PFI ralue beinr,

eiLher 1.c. -r :.0o min. heT no ind ica1tir fi diffirericial

schedule control was observed, the classes of reinforce-

ment (penny plus trinket or trinket alone) were made dif-

ferent for the two components. The control produced here

most closely resembled that found under "mixed" schedules;

i.e., some change in pattern occurred after the first rein-

forcement following a schedule change. Ratio components

were characterized by responding through the interval or

pauses followed by a rapid transition to a high rate.

Greater evidence of schedule control was achieved when a

Lindsley, plunger-type manipulandum replaced the previously

used button. Long (1962) reported good schedule control

with normal children under mult FR EXT, mult DRO FR and mult

DRL FR where FR values were either 5 or 10, DRO values were

from 2 to 32 sec and DRL values were from 2 to 16 sec.

Although some evidence for differential schedule control was

shown in some subjects under mult FI 1.5 min FR 10, in most

cases remedial procedures were necessary to achieve stimulus

control. These procedures included presenting FRs in

blocks, increasing the value of the FR schedules, adding a

DRL requirement temporarily to the FI component, or tem-

porarily replacing the FI schedule with a DRL contingency.

Adding an external clock to the FI component of the mult

FI FR produced differential patterns of responding. This

difference was particularly evident for subjects whose

exposure to the mult FI FR with clock was preceded by ex-

posure to mult FR EXT or else no previous experimental

history; i.e., the subject was naive. The improvement in

control achieved by these various techniques was accomplished

by reducing the response rate during the FI components.

O'Brien (1968) investigated the behavior of retarded

children under mult VI EXT schedules. Five-minute periods

of exposure to the VI or EXT schedules were presented in a

random order. Besides getting characteristic smooth, steady

rates in the VI components and little responding in extinc-

co.r,, r; c._ r.-: rrt- i-n '.i p ri.'..5 I1 .. in, E..T r-eri.:i:

,r- 'n- r .r s .w"'r r-f r.r ratec in Fp.riod- L.' 1 j, rL I

rtnerr .I peri.:d This *-*:.nrcrac e r. .-r"-d c, 0;

cr- ri- t.r"' n-:...e"Er, and a- nor aFparcrnr. tb: n rntrih

z i -' i:,n. ihr. ran C- :,rne t11 7: *-.c lo', ir. '.'I '.I .and

i'ult "I T s.-rn..duljui .c n r.l ri.:.r l cjij c)ecc, re-

Fp r.ji s cr.irie.j *c:.c r.r t .: -r 'h.n r crt ;criidu in n:.n e

r:rEr.n ,nn t .'ai chanied.J fro .rm E :T, ,an inr:r- = in r-pon se

ra c in ,un:ranric.jd '.i co:.rrp.:.r- ..a ; ~ '.'d -rni h per-

rirccd fr.:,, .o ir si ii .n A '.milar d. re3r. in rc-

sponi-: ratr in tne ,nrcnan. .d '.'I .im-Fponer, r." d *rn

tLn ch e-Ia l ir r. nre other r compoF.jnrrnent cnrng-ed from E%.T .o

. Tni. :ranri; in re pF.rni rac- in i re ,unrncn nai '.'

c.:n.duile ;.:cr.-o.:.nenr d.i n.:c a.fecec eniracceri 5 ic 'I

pacttrnr o*f r ecsPoriin:, tut. ,niy .-r ld be scecn as crang. s in

the s3lFc p r.!-c hc i-rlatices ra:.rj.

Bijou anrd i*'rlanJd (19611 des.rribed 3 prccedure for the

rapid d.:.'IlcrrmCnt Oif til ci. e--crieJue r.-pcrforr'cnce in

retarded children. In this procedure stage one terminated

when subjects responded at a rate of 20 responses per minute

or higher on FI 15 sec. Failure to meet this criterion

resulted in exposure to an increasing ratio schedule de-

signed to increase response rate. The description of stage

two was unclear as to whether subjects were exposed to a

chain DRO CRF or chain DRL CRF, with increasing temporal

requirements of the DRO or DRL schedule. If the latency of

responding upon presentation of the stimulus associated with

CRF remained short, and pauses of 30 sec or more occurred in

the presence of the other stimulus, stage three was begun.

Stage three merely repeated stage one, with the same criterion

and remedial procedure. In stage four, the subject was

exposed to a two-ply multiple schedule with one schedule

being extinction and the other being an FI of 1 min duration

or less or a ratio of 50 or less. The data shown for sub-

jects shaped using this technique depicted rather typical

FR and VR patterns of responding together with pauses in the

presence of S-delta. Under FI schedules, responding was

maintained throughout the interval, but stimulus control was

well demonstrated. The steady pattern generated by a VI 0.5-

min schedule was typical, but the rate was unusually low for

a VI schedule of that value. The stimulus control with the

schedule combination was not well demonstrated as no extended

periods of the extinction component were depicted. Whatever

advantages this procedure may have would relate to the

rather quick development of multiple-schedule stimulus con-

trol as behavior similar to terminal performance was

achieved by the end of the first session. This statement

must be tentative, however, as session length was never

given and appears to have been variable.

Attempts by Ferster and DeMeyer (1961) to bring the

responding of autistic children under the control of mult

FR VI schedules were unsuccessful with responding in both

c-ip.-.r,:nr. bhiln cl- ~f cterizd t; h4 ah r tLe in pn F t rein-

t, r. ei.fnrt car_ j:

C ir, anr Trndieim :h 3jule-: C' Frein..rc.Trt, r.

., :rl-ain ; -,m ilar cr- a mulripi e! *-.c:ct r.that

rsp-,jirn: in ti. Fr.sence of one*:~ r. riLus *r: nrr.

prm ar: reirnt f rc s.ner, n bu .t a --::i i ',,ulu.S, ran resF

inr in the rre -r.- t: h t r ,ir. : t.. n mr. uliu Fr-du:=-e ri-ln-

for:...--:nr.. TarFi,- s.-h-i ,'l _- arc 1 .1 :heairn sche: Jl s uL: ir ,

rno: 'r.c r .:~ -r. s ii'.' u l u pair: .. i h ei ,h c r nF n-rir. iLon

(19'63i1 In"r .- .ti -J .u ,rn pe.rf-.rr'an.:e irjer ciiairin _an t.n-

,ii_ .-rcI e u e i p.t 5.-iti ".- r r inf:trc.- rnt '.i- h n.rmrrl ihll-

dren -'r"irn a E-'-n.jc. ts. Criain OFL FF ran .c:-h n DC,rO FF.

.- : ju, r : i ,l].:.s I C .:-=E pr:.ju. :5 tr.r:rn s n.j.jule an'i z i ,lJs

crnt.rl, tut :hair Fl FF rar.l: id if Jriti .:. l ce:hniq--

werE nr: t us.ij. control wais imlpr:'.'.I rih ch3i n FT FPR -he-d-

ules if LrTu FP ::,r.p.nernt ...a incr.-r ed in n iz-, if rch-eiule

anJ stimulu, s .::,rtr.':l as- first e t. a. .iis e i th h-,.in EDL FR o

-hain C.F'. F FF E.h=-Ji le- L.efor-: s hiftln: the chain FI FF., or

if an external clock was added to the FI component. Tandem

(tand) FI FR schedules never produced regular or repeatable

patterns of responding when additional procedures were not

used. Response patterns resembling those of chain FI FR

were produced by tand FI FR schedules if an external clock

was added to the FI component, or if control had been pre-

viously established by tand DRO FR.

Concurrent Ooerants

Ferster and Skinner (1957, p. 724) defined concurrent

operants as "Two or more responses, of different topography

at least with respect to locus, capable of being executed

with little mutual interference at the same time or in rapid

alternation, under the control of separate programming

devices." Concurrent (conc) schedules will thus refer to

two or more schedules operating simultaneously but inde-

pendently with respect to two or more concurrent operants.

Schroeder and Holland (1969) conditioned macrosaccadic

eye movements of normal adults to two areas of a four-dial

display employing conc VI VI schedules of signals. Rein-

forcers (signals) were delivered to the two right-hand

dials according to one schedule and delivered to the left-

hand dials according to another. Schedule values ranged

from VI 9 sec to VI 60 sec, and sessions were conducted both

with and without changeover delays. In the presence of a

changeover delay, subjects matched relative eye-movement

rates to relative reinforcement rates on each schedule.

Rate of crossover eye movements, with a changeover delay in

effect, was also inversely related to the difference in rein-

forcement rates programmed on the concurrent schedules.

Sanders (1969) reported normal adults' performance on

a button-pushing task under conc FI FR schedules of monetary

reinforcement. Manipulations of the FR requirement from 50

to 1,000 responses while holding the FI schedule at either

a 3- or 9-min duration produced the following effect: When

the fixed ratio was small, more fixed-interval responding

occurred per interval than when the fixed ratio was large.

Bursts of responding under the FI schedule were often ob-

served after reinforcement under the FR schedule, but little

post reinforcement pausing or ratio strain w3s- --r -e3.

Twenty-four normal adult subjects wer.- .Er ui.- f-,r

10 one-hour sessions to determine whether .:,L-.r .-r'Z

i i jl iri.L r tr ::" r. indi id-jl ia";Lt.t= in :r:.riTFl JI.j pla;

c:., lj t- if f-rernc i. a l I' :,r r!l- j 1 _:.r-n jrr n-n -i. -'_diJ l i

C*' 3Im. rI,:. ri.-i r ,; rr-'i i F ra: 3trn it. ,i r .- :, -i t .ct

~rar, Eil:rilL; ;r. Pl;Ci.. I 1Cb.riedul i uri eLji FI 30

_._ mr, i; l-i ".1-.-, "i 3' e: ,r. 'i:0 e. ,fDEL -.S .--: jril 1') _e ,

and ER -:9 ril 3,0,. .', li,,.r.=dj "n:.'L.j ,:., 10 s. .: .; s r, erfecL t.or
arJet :,n F .:, all 5i r, 1-. P|:. -0r, a :, ":. r d,- ol z, ,ipl,. -, .

Fe-lt rrin car.:.. thaL L-. ?crl, r 3pcr.i t c:. Lrbl_.'..iual

air.t- r3 *-.rr rordi..i rto ri-e tie jral paictrrr! if crspon iii.

c-s ;crv-id rider trI;- s: cr-.dules *i L :i lc.wJr *flrclariSriS *:nl; f.rI

the group of subjects exposed to an FR schedule in one com-

ponent. The group who was exposed to a VI schedule in

combination with an FI and DRL, tended to exhibit the same

pattern of observing responses to all three meters during

any given session. Schedule control like that observed with

infrahuman organisms was produced when additional subjects

were given instructions as to the reinforcement contingen-

cies, feedback in training sessions as to missed signals,

and an extended number of sessions under the conc FI VI DRL


Poppen (1972) studied the performance of normal adults

on a lever-pressing task for money under cone FI 1 min DRL

20 sec and under conc FI 1 min FR 100 schedules of reinforce-

ment. All subjects were exposed to both concurrent sched-

ules. The patterns of responding generated in the DRL and

FR components were similar to those reported in animals and

other human studies, i.e., very low-rate, spaced responding

under DRL and high-rate responding under FR. The FI per-

formance when the concurrent schedule was FR 100 was charac-

terized by long post-reinforcement pauses, the latency of

the first response usually exceeding the 1-min FI require-

ment. When the concurrent schedule was DRL 20 sec, the FI

performance was characterized by high rates with no post-

reinforcement pause. These results contrast with those

reported by Weiner (1964c) for performance under FI sched-

ules following a history under either FR or DRL schedules.

Favell and Favell (1972) employed conc VI 15 sec VI

15 sec EXT EXT and cone VR 4 VR 4 EXT EXT with normal chil-

dren in a matching-to-sample procedure. There were four

possible responses, with correct matches possible on either

the color or form dimension. Either correct match produced

token reinforcers (exchangeable for money) according to the

VI or VR schedules. Every match on one dimension also pro-

duced a stimulus complex which preceded every token delivery

(the paired stimulus) and every match on the other dimension

tr.y -, ji l or,' (cri ..n aire -a,,,luj i. Th unn.apair ,-d

Sr_;r1 .j;r F c;_r-r- rL'-, m.=. *:, r. r.: .j '.c .".*,jr f'r r.lh.:,-c, r, r. :l-3;_

~ tilacr-.- ;rr:..Ju:-d bri f t :i .. : njLE .:rnJti:r:n. Ti-,rc ..f th:

fi '.'- _-'jb*-.:cr. r ,on.J_- primri5r l. c:, ii,,.-n-'Ecn r-Ai t ..'as

l l' c r-:e -:'. ll -:rirnj i n "-r.i a r. .:r ii n''-:l -d s':,e

r; p. .:.f d i rirt.-niarrr n r s .. rnr t :. 5', l r.l r e= ,

r.i ih *c n.ric- r 51., L, 2' Fr 7L d c3 r 5r. l 5 sr. n f tr:

crs;F nnea ihe .-i) e:r..C in all .rree :f tler- e->.p ri ,,;nts

,.-re r-r.rij-j c'-il ren. Earr .ct anj L ndil- ; (i'1.,2 ured a

Eraze:. jur in ;n- :-r r :p.:rn n :n r:-n r.,: lIc,'.r: :ere r in-

f.:.r: J an .ran FF s.:rniul ir, --i. e press: cf lli iq :a .,tb e tnh

m ,iFan '.i lanJ,.n,an, i *..cin, y i n, .'az p,:..r" -i -j i hen the .In hr.

was *.ff. Er.tin:tc: n wa.s pro-iramme fr born 1; hc-onr ,ani

li ht-;if -rand : 3 tion :n r the :,t ha rranipul.arinumj TF ii- light

was always on over one manipulandum but never both. The

schedule was thus cone mult FR EXT mult EXT EXT. Results

indicated a great deal of between-session and within-session

variability and extremely slow acquisition of stimulus con-

trol. The higher the FR schedule value (up to 10), the

poorer the stimulus control. Orlando (1961a) used a pro-

cedure similar to that described above with the schedules

being cone mult CRF EXT mult CRF EXT where EXT components

are of 40-sec duration, and the stimulus associated with the

CRF component was terminated after one reinforcement. This

is programmed like an FI 40-sec schedule with a stimulus

indicating when a reinforcement is "set up." Ten of the

twelve subjects in this study responded with a latency of

5 sec or less in the CRF component on 90% or more of the

component presentations, and paused for 30 sec or more on

over 80% of the EXT component presentations. Orlando (1961b)

employed a cone mult VR 100 EXT VR 100 EXT where at any

point in time the schedules programmed for the two available

responses were never identical. Under terminal performance,

99% of responses occurred during the VR components at high

steady rates with virtually no pauses.


Performance was obtained from normal and clinical sub-

ject populations which closely resembled that of lower

organisms on comparable FR, VR, and VI schedules. The

atypical patterns which were obtained were almost exclusively

in clinical populations and were characterized by unusual

pausing, high between-subject variability, poor stimulus

control, and, in some subjects, inability of the experi-

menter to maintain subjects' responding.

Performance under FI schedules, whether singly pro-

7rramned or in Trultiple. chain, -r c-nzurrent Tchzdul-7. -.:--

ra h-r riatL:. Th.e '1l-:..r i a li t f c' : r.:e. .:rf thi

.'jriabilit j n fr:. it- u die: r --.:. _d, '.l :h ffec.-

paF3t~err .if r sr:.njrin u*rider El =chedJuie-

1. Perha.s CTi ri n l.: -.:rr. i.porr. r i-iL '.iri l,! e is cre

:ubje; '. *c eri ,enT a~l r is r.:.r,

2. A "ari t.l.: relia J co e:.*r- : r .T i n i rlis. r' r- is rar.

*ft in r ru : c :.n:.

3. Ccrnc'jrr-nc -c:heluie.- ,:r r.: in r :A=im nL r. :1-r ,n

t': afi-:t r-e paitr. r :. f E i.,' it diff-:re nt. ar er

tran a hir.:r, *.r .f cr,.:. i- sae .heJul- .

1. Inicer'.'i al duratLi .r ..e ll t.: a rCle -a.nt .*arjiable

in Lhar. ho; rt irticr-'l- are FarticularI; likel, to. -.aintcin

*co rTi n-. r -., re :- p .-.Fdii r, .

5. Funir -umnint. .:.f re-p-ooriin fr'r e:-.:imT l b, p.oinr t l =e s,

no, -1h. r, to ifferer cii'i effect. the pa ttr rn '.f re=-s .ind-


6. A .-a ri at l- :nitch r, ., ce relat j t6 psauni-n:nr e *of

rc;p--ndin3 1- ctre C r.: r-equirel t. erit r,.- r.e Ion;e.

7. A '-ar abl- uige-t-l: fcr .iata 'ith, DEL 3 11 ell a-

FI schei dules is c* il-ti ral timing. r.Lt m-.-iedlar in

beha~"ior of the


8. The final variable is an external, time-mediating

stimulus, e.g., a clock.


Autor, S. M. The strength of conditioned reinforcers as
a function of frequency of reinforcement. Doctoral
dissertation. Harvard University, 1960.

Arnett, F. B. Inverse relation between choice and local
response rate with a schedule of response-produced
blackouts. Journal of the Experimental Analysis of
Behavior, 1972, 17, 37-43.

Azrin, N. H. Some effects of noise on human behavior.
Journal of the Experimental Analysis of Behavior, 1958,
1, 183-200.

Baron, A., Kaufman, A., and Stauber, Kathleen A. Effects cf
instructions and reinforcement-feedback on human
operant behavior maintained by fixed-interval rein-
forcement. Journal of th-. F:. 1 i-l : I: f
Behavior, 1969, 12, 701-71-.

Barret, B. H. and Lindsley, O. F.. DCfi'i: i n r, ~7-u:C,
of operant discrimination a-.] jf ~rentrirci :n :'r.:* t2
n1-i-tituti rn3l r-t3rd: '-iilr r :,-
!_-rit a f .7.. r._-.', 11 ,' J -.- '7 .

Eim, 'v ;i. r n.j r: i ia r., H. C. .,:.- : rim. :- 11:.. ticr.

i! *- 5i.

i ,': r,,:r :,-,i r i:.- i ti'E.: L"'. '-. 'i :.r: : :' .ri-
e r i ur1 -C-,r.: rt : i.'.i .

Ei'.:.u, ; i ara ,:,rl ranj.., F.. f.-pil j e,.'el:.-, nr. .:,r l' .i l -
TIi,, F: er: f rm n.::= L'-, r-. rJ ,j *:-I iJLL--r :o r u ,r al
*.f :i l ,= t l r' rl"_; ,.: ,-f E i- '- F r lA'A, 3 J

6r* .,.r, : rin A. J. ,ri L 1 I. r. r. C, ':'. -,: C r .-ct: :f
r l. ,ti3.: r -ir, .r-: ,' ,: ra, r ,j :r r CJr, '':" r ji la,' in
-r,.. J.. irr, .-: rhr, [.' .-rim ri l .p 1, ':- -f E r i.r,

Bruner, A. and Revusky, S. H. Collateral behavior in humans.
Journal of the Experimental Analysis of Behavior, 1961,
4, 349-350.

Dardano, J. F. Modification of observing behavior. Journal
of the Experimental Analysis of Behavior, 1965, 8,

Davison, M. C. Preference for mixed-interval versus fixed-
interval schedules. Journal of the Experimental
Analysis of Behavior, 1969, 12, 247-252.

Davison, M. C. Preference for mixed-interval versus fixed-
interval schedules: Number of component intervals.
Journal of the Experimental Analysis of Behavior, 1972,
17, 169-176.

Dews, P. B. and Morse, W. H. Some observations on an
operant in human subjects and its modification by
dextro amphetamine. Journal of the Experimental
Analysis of Behavior, 1958, 1, 359-364.

Duncan, B. and Fantino, E. Choice for periodic schedules of
reinforcement. Journal of the Experimental Analysis of
Behavior, 1970, 14, 73-86.

Duncan, B. and Fantino, E. The psychological distance to
reward. Journal of the Experimental Analysis of
Behavior, 1972, 18, 23-34.

Ellis, N. R., Barrett, C. D.,and Pryer, M. W. Operant
behavior in mental detectives. Journal of the Experi-
mental Analysis of Behavior, 1960, 3, 63-69.

Fantino, E. Effects of required rates of responding upon
choice. Journal of the Experimental Analysis of
Behavior, 1968, 11, 15-22.

Fantino, E. Choice and rate of reinforcement. Journal of
the Experimental Analysis of Behavior, 1969, 12, 723-
730. (a)

Fantino, E. Conditioned reinforcement, choice, and the
psychological distance to reward. In D. P. Hendry
(Ed.), Conditioned reinforcement. Homewood, Ill.:
The Dorsey Press, 1969. Pp. 163-191. (b)

Fantino, E. and Herrnstein, R. J. Secondary reinforcement
and number of primary reinforcements. Journal of the
Experimental Analysis of Behavior, 1968, 11, 9-14.

Favell, Judith E. and Favell, James E. Control of prefer-
ence in children by conditioned positive reinforcement.
Journal of the Experimental Analysis of Behavior, 1972,
18, 107-112.

Ferster, C. B. and DeMeyer, M. K. The development of per-
formances in autistic children in an automatically con-
trolled environment. Journal of Chronic Diseases,
1961, 13, 312-345.

Ferster, C. B. and Skinner, B. F. Schedules of reinforce-
ment. New York: Appleton-Century-Crofts, 1957.

Findley, J. D. Preference and switching under concurrent
scheduling. Journal of the Experimental Analysis of
Behavior, 1958, 1, 123-144.

Fleshier, M. and Hoffman, H. S. A progression for generat-
ing variable-interval schedules. Journal of the
Experimental Analysis of Behavior, 1962, 5, 529-530.

Frazier, T. W. and Bitetto, V. E. Control of human vigil-
ance by concurrent schedules. Journal of the Experi-
mental Analysis of Behavior, 1969, 12, 591-600.

Henrir (, ,d. (L.J.I I :r3:.. ,re r.-ix r rc,:eT rr. H:, ::i ,
i11.: C._._r .:_,' p r --_: i 4.

H-f r nr-ir,, F. j. :=,:,r. jn r" r_ ril.*-.._ rr,- r. .r-. r_' r .: I

1 -r!r, :. 1 I4 ,,r r ie r r r it g P

H :. ll. r j J -. Hiizn i l'z 1 r c-.- .:ci rr ,--, 1l- :. 1 2. 1 -6 .
H :-ll;r, i, 1. :* r' r, l-, r, ,_ :. -.r "l l .-
ui :. reinr :r. e- ,e r-. J 'rr.,I :f r-r.; :. .r r'-rr.rr i

Hiuir..1-1i : r r. iI j .'.r, r,, H. H. C : r, ji irr 1 rn-,r: r il
h:,z[ il! cl *' _r-.r t.: i..= rjic- c l',e-Jl l f; rein-
r:.r.-.:pre:.rt. JIv -rri -.1 f m r. E rL .rier -cr tl ,,- iS :, E
B.-',i '-i:. r, 1 1 i 4, -,t..

' Fr.:ztIr. E. L. Thle erffe -: :- [,3_ c L l '. l. :n r--,:n
C ,i ri r.- '-- :r .:.r 5 rrir.i:.:. J :rJ n.l :), tr,-
E.. F:-ri.-,-nr, t-3l ".r,s1l 'i- -f 4h3"'i *-r, 1'K)3, b, 2i-19').

Kaufman, A., Baron, A., and Kopp, R. M. Some effects of
instruction on human operant behavior. Psychonomic
Monograph Suppliments, 1966, 1, 243-250.

Killeen, P. R. On the measurement of reinforcement fre-
quency in the study of preference. Journal of the
Experimental Analysis of Behavior, 1968, 11, 263-269.

Killeen, P. R. Preference for fixed-internal schedules
of reinforcement. Journal of the Experimental Analysis
of Behavior, 1970, 14, 127-131.

Laties, V. G. and Weiss, B. Human observing behavior after
signal detection. Journal of the Experimental Analysis
of Behavior, 1960, 3, 27-33.

Laties, V. G. and Weiss B. Effects of a concurrent task on
fixed-interval responding in humans. Journal of the
Experimental Analysis of Behavior, 1963, 6, 431-436.

Leander, J. D., Lippman, L. G., and Meyer, M. D. Fixed
interval performance as related to subjects verbaliza-
tions of the reinforcement contingency. Psychological
Record, 1968, 18, 469-474.

Lindsley, O. R. Characteristics of the behavior of chronic
psychotics as revealed by free-operant conditioning
methods. Diseases of the Nervous System Monograph
Supplement, 1960, 21, 66-78.

Lippman, L. G. and Meyer, M. E. Fixed interval performance
as related to instructions and to subjects verbaliza-
tions of the contingency. Psychonomic Science, 1967,
8, 135-136.

Lippman, L. G., Leander, J. D., and Meyer, M. E. Human
fixed interval performance as related to response
effortfulness and to initial point. Journal of General
Psychology, 1970, 82, 57-61.

Long, E. R. Multiple scheduling in children. Journal of
the Experimental Analysis of Behavior, 1959, 2, 268.

Long, E. R. Additional techniques for producing multiple
schedule control in children. Journal of the Experi-
mental Analysis of Behavior, 1962, 5, 443-455.

Long, E. R. Chained and tandem scheduling with children.
Journal of the Experimental Analysis of Behavior, 1963,
6, 459-472.

Long, E. R., Hammack, J. F., May, F., and Campbell, B. J.
Intermittent reinforcement of operant behavior in chil-
dren. Journal of the Experimental Analysis of Behavior,
1958, 1, 315-339.

MacEwan, D. The effects of terminal-link fixed-interval and
variable interval schedules on responding under con-
current chained schedules. Journal of the Experimen-
Analysis of Behavior, 1972, 18, 253-261.

Neuringer, A. Delayed reinforcement versus reinforcement
after a fixed interval. Journal of the Experimental
Analysis of Behavior, 1969, 12, 373-383.

O'Brien, F. Sequential contrast effects with human subjects.
Journal of the Experimental Analysis of Behavior, 1968,
11, 537- 542.

1 1 -r ,iIT _. r, in r .; ,', .'... _r ", r, J ri l .Tf 1, r, rr r, I1 f t, . E

'L r ., r [,J :r, ;I ,.i .: j .' i, rr L -I: ,,li

: .: _* F ;.x i [. T ,--, : ,-, r_ J ,: r *i_. r i,_-. :- '.id .-n .

_i.''I [ ^: *-, '1_ :[ l' L |_ r ,,r-dr-_ I". :- r .i-.i_ -'J_. :-, Idr .:n .

.' -r :-, r _E : . -. r ,- l : _i l 1 -- l- .,: i i .

r, rr. Nr. r.

1:rH, L r C ,_,[ r 1 .- : r l r. ,i .

',r, _-.i _. 1 l L E--li r, '. ',. 1 ,1.1 .',i '. 1

,- i', LI r 7 I,_ : r r il .r I : ..-_ .r r, -,t .'* 1 ,
r. i i .* A t.

r:riT .:irl:- irn l lt_ l-ILu rl :.L. ?lIS,-r-r J,-u .rr, t1' : 'l-h c
[ i. -i i,,,i .t. ^ l r,.j l, ; ,.*. ' E.:l-, :r i- t';i L, i -_ ,:'t, :1-r : ,4r.

t.-ir, ji-,. l .:_-. J:ijt'r ,-l t-,: iL .[rr Iri ,-_rir. l "n1 ' -.

Schroeder, S. R. and Holland, J. G. Reinforcement of eye
movement with concurrent schedules. Journal of the
Experimental Analysis of Behavior, 1969, 12, 897-903.

Scobie, S. R. and Kaufman, A. Intermittent punishment of
human responding maintained by intermittent reinforce-
ment. Journal of the Experimental Analysis of Behavior,
1969, 12, 137-147.

Simmons, M. W. Operant discrimination learning in human
infants. Child Development, 1964, 35, 737-748.

Squires, N. and Fantino, E. A model for choice in simple
concurrent and concurrent-chains schedules. Journal
of the Experimental Analysis of Behavior, 1971, 15,

Stubbs, D. A. and Pliskoff, S. S. Concurrent responding
with fixed relative rate of reinforcement. Journal
of the Experimental Analysis of Behavior, 1969, 12,

Waite, W. W. and Osborne, J. G. Sustained behavioral con-
trast in children. Journal of the Experimental Analysis
of Behavior, 1972, 18, 113-117.

Weiner, H. Some effects of response cost upon human operant
behavior. Journal of the Experimental Analysis of
Behavior, 1962, 5, 201-208.

Weiner, H. Response cost and fixed-ratio performance.
Journal of the Experimental Analysis of Behavior, 1964,
7, 79-89. (a)

Weiner, H. Response cost effects during extinction follow-
ing fixed-interval reinforcement in humans. Journal
of the Experimental Analysis of Behavior, 1964, 7, 333-
335. (b)

Weiner, H. Conditioning history and human fixed-interval
performance. Journal of the Experimental Analysis of
Behavior, 1964, 7, 383-385. (c)

Weiner, H. Conditioning history and maladaptive human
operant behavior. Psychological Reports, 1965, 17,

Weiner, H. Controlling human fixed-interval performance.
Journal of the Experimental Analysis of Behavior, 1969,
12, 349-373.

Weisberg, P. and Fink, E. Fixed-ratio and extinction per-
formance of infants during second year of life. Journal
of the Experimental Analysis of Behavior, 1966, 9, 105-


Thomas Robert Waddell was born on May 7, 1945, in

Florence, South Carolina. He attended public school in

Melbourne, Florida, and graduated from Melbourne High School

in June, 1963. In September, 1963, he enrolled at the

University of Florida. During the summers of 1963 to '19

he did road construction and maintenance work for Brevard

Cij-L , r I r. r i i rh- : -i -., : i a'- i4 1 ` ,.

rL I.* .J t f .-r 'r h-: L'r -.: -I C-' r1, *: L tLr L-j -,;a r ---r t. r t a -

F rod ir :. r r i. rl .',-ir r, r,.c _: i i -I, ,, r.,

,L. .:i' f r a -.IT, t-.; 'it :f F .ri J r ,rid ir, r-'cre *.:.f

tL-ti r- 5 r ,rI r lJ ^r j.4 r., :'.I..: l in rri. l'.' r.:-rtT 1 r.t f

S C r, l- i ,, n..r.- lrj r i r : t", _r:- l i ,r *.. -. r in ra.J3'- 5 t'

.7:C: l r': i_ r.r!.. -i.... itl Feill.:. r'.

Hi re ci 3 ir.- Ila 7. r :i i .:'-r :ieire iP -uI. i, '19 Fr :

.-- .rt,.i .r., 1.'1, u.ri l .'-u.ujJ 19"' I a a *li-iCal

. : 1 .r i :. : a irlr ni' i =. r i .,'i i r. i ii i j l

l in i,- j. c.r i;:i s ir F.1

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

E. F. Malgod Chrman
Associate Pro ssy of Psychology

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

William D. Wolking
Associate Professor of Psychology

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

Rugh 4 DavisdJr.
Profesor of Psychology

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

erry S. Penypacker
Professor of Psychology

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality
as a dissertation for the degree of Doctor of Philosophy.

Wlliam R. Reid
Professor of Special Education

This dissertation was submitted to the Department of
Psychology in the College of Arts and Sciences and to the
Graduate Council, and was accepted as partial fulfillment

F7. 1 I-r':'l:- rF, 1::'r'-

C,- r-, r" j. 5 ; h 1r t :.:*

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs