University of Massachusetts AmherstScholarWorks@UMass Amherst

Masters Theses 1911 - February 2014

1961

Effects of a redundant informing tone in a closed-loop monitoring task.Harvey A. TaubUniversity of Massachusetts Amherst

Follow this and additional works at: https://scholarworks.umass.edu/theses

This thesis is brought to you for free and open access by ScholarWorks@UMass Amherst. It has been accepted for inclusion in Masters Theses 1911 -February 2014 by an authorized administrator of ScholarWorks@UMass Amherst. For more information, please contactscholarworks@library.umass.edu.

Taub, Harvey A., "Effects of a redundant informing tone in a closed-loop monitoring task." (1961). Masters Theses 1911 - February2014. 2025.Retrieved from https://scholarworks.umass.edu/theses/2025

1 1 iiiiiii

1 11 ill

iiiiiiii^;"',v.,''';

'".

' '

1 ( i » >

,,, .,

-.".

, .•

- >%

'' V'," •

|..*»

V. -iX-.^Vv^^ :!•.;;;;.;.;"

Ieffec

: i^r^Sr

TS OF A REDUNDANT INFORM NG^ fONrt^^^^^

TMIOIOOP MON fBMI TASK-, • •

Ik-vl •('• -* - - .

.

TAUB - 1961

LD32341VI268

1961T222

Effects of a Redundant Ini >miing Tone"^*g "one in'

'

a Closed-Loop Honitoi Ing T»sk YT'S^T^VIJIU

Harrey A. ?^ - J-^' 3HMA

Thesis Submitted in Partial Pulfillcient of the

Reqiiirenents for the Dc^ee of

Mnttflo* of Sciencet

University of KassacJ-nusetts , Anhorst

May 1961

11

The author "tAshes to convey his appreciation to the nem!:>ers of

his thetJis coimrdttee, Drs^ H. Teichnor, A. E, Goss, J. L. Myers,

and Kr. Weaver for their cooperation and s»-'ltiance in the

preparation and completion of tliis manuiscript.

A particular expression of thanks Ic extendefi to Dr. w\ H«

Teichnor, it;ho served aa thesis advisor. Through his constant advice

and encouragement the conpletion of tliis study ijns realized,

t\irthGr appreciation is extended to the Air Force for the

nonetary assistance which he?.pefi mke ttiis study possible, (IMs

study is part of a project supported by fimds froii /IFGSR convract

60-2,)

Finalljr, the r.uthor vdsbos to escprofis his appreciatirrL to his

fellow students and friends i7^:ose participation pjid suggentions

ea^jedited the completion of this experiment.

iii

Tnble of Contents

Pas®AcknowledPTientn — —-~ — ±±

Tr-^ble of Contents iii

Introduction 1

Method ~ 5

Apparatus 5

Procedure 5

Subjects 9

Results ^ 10

Discuseri.on — — IS

Svaamxy 17

References 18

Appendix 20

Introduction

The prirpose of the preset study was to investigate tho cue and

avorsive propwties of a redundant informing tone in a closed-locp

monitoidnf^ situation, Tho problcM derives fron the currcsnt critical

role of human jnonitoring in raan-raachine systems, exaiiples of ;^oh

may be found in radar raonitored systotos, air traffic control systems,

and automated manufacturing processos, CSiaracteristicaU^, the

operator acts as a tJaming device, i.e. he is recjaired to siaintain

search of a display and mintain a readiness to respond to some cri-

tical low probability sicnal. Often the (operator must work oontinnous-

ly tar periods of 1 to ^ hours or more, and in some oases he is re-

quired to perform in isolation and/or other unusual environmental

conditions.

In considering the work conditions, a further distinction roay be

made concerning the type of monitoring systai in which the (^eratoac*

iTiuGt perform, ilickey and Blair (1950), using the framework of man-

machine systems, classified tasks as boin^ either opm-loop or closed-

loop depending upon the part that the operator must play in the

systen. In the open-loop situation, the input to the operator is

not dependent upon the output of the system. The operator acts

slmp3y as a detector and only indicates that a deviation fvm. a null

input lias occurred. Althoui3h the operator's response may eventually

lead to a change in system output, this change is not presented as

part of his input. This is retiresentative of the classical monitor-

ing task in which the stimulus is presented for a fixed duration and

then renoved even if no detection response is made, Tlie general

2

finding of these monitorin- studies is that oporator proficiency, as

•asurecl hj probability of atimilus detection and detection time,

dwireases sygtematically v^th increasing length of vigil. The nost

narked decrease in proficicnc/ occurs near the end of the firat half

hoiir and during the second half hour. Those resulta have been found

tiding a variet:^ of tasks including the clock tost (llacizHorth, 19^0),

radar (Deeae & Omand, 19^3) , and flashed li^htEj (llakan, 1955;

MoCoCTiack, 19^9). In addition, it has been denonatrated that perfor-

mance depends directly upon rate of stimulus presentation (Jenkins,

1958), stinulus inbonslty (Adams, 19^6; Bakan, 1955), duration of

dgnal (Adino, 1956) and rofjularity of presentation (leaker, 1958),

Studies -Hiiich utiliiJod the open-locp model have also showi that

kna.xledGe of results can elininate much of the decraraent tiiat is

ordinarily found in the monitoring task, Mackworth (1950), v&io first

Mpirtad thds effect, told Ss they vJGtq "correct," "rnissed" a

iA0MLl, or nade a "false" response 'when no signal was presented,

TtoCormack (1959), using signal lights 0.3 informers, found tha,t al-

though both no^'mowled^e and lcno*7lodge groups showed increasing

res})onse times to the stimulus as longth of watch increased, the

knowledge gro\^ performed significantly faster.

In the closed-loop task, input to the operator is directly

dependent upon s^rston output. This con be represented by an infor-

Jiiation feedback model in which the operator's response has a direct

effect upon system output, -sdiicli in turn changes iiput to the

operator's display. Stitimli or signals in a closed-loop task are

ronoved or reset to their original position only if tiio oporator

3

makog a detection rospxaiae. Thus, by making a response, the operator

receives direct feedback or knowledge of results,

Althoui^h Iiick(?7 and HLair (1959) point out that the closed-loop

task is a good representation of the real monitoring situation, until

recently this type of task has received little attention. The most

CDctensivo twk using this model has been done by Holland (1957, 195B)

iftio sugsest-ed that an operator's observing response nay bo rnaintained

by detection of the signal. If so, frequency of the obcerving response

should depend upon the schedule of stinulus presentation, Holland

tested this hypothesis in an operant situation requiring Ss to detect

deflections of a pointer. The Ss were in darkness and unable to see

the display except vhen they pressed a key—the observing response—

\AiLch provided a brief flash of light, Cuimilative observing responses

WBr« recorded under various fixed-interval (FI)# fixed-ratio (FR),

and v;xriablo-interval (^), schedules. The results for fixed

schedules showed tiiat the rate of observjji;: responses can be con-

trolled and maintained by signal detection. This provided sorie

support for the hypothesis that feedback rnaintains performance. Qi

the otner hand, the fixed schedules that he used were not character-

istic of real monitoring tasks, since signals occur irregularly in

most operational situations. Under the noi^ typical VI schedules,

as ttie testing session progressed, the cl-jaracteriGtic decrement of

response rate occurred. This agrees with Jenkins* (1958) findings

for reaction time to an irreg\ilarly presented light signal.

In suwaary, monitoring studies using the open-loop model

indicate that performance decrements can be partially eli'iinated by

h

presenting kncfwlodge of results. Studios x:±th the closod-loop model

have shown that the information provided by the task itself is not

aulTicient for r.iaintaining perfoxroance ^ihcn the stiroali aro presented

irregularly. This type of situation eOready provides a conplct©

informational feedback, ilowever, the results of studios with the

opon-loop situation sug;;iost that an addod signal, wiiich is rodundant

in that no now informaticai is presented, might counteract porfoiroance

4MMB«nts, T/hethor mich b redundant eifjnal would affect performance

in a cloG9d-loop monitoring task was the major question of this stiitiy,

'Itie sirt?)lest, most practical redundant Bignal in a visual riOTii-

toring task would be a tone presented siriultaneously with or iraiaii-

ately following the prjbiarj feedl^ack signal. A problm in doing this

is the possible aversive affects of the e?Adito3:y signal. Studies by

both Azrin (195^0) with humans and Campbell (19^7, 1958) viith rets

indicate that as intensity increases, sounds devclc^ annoying or

aversive properties sufficient to develop avcldance responses in

Althou.gh neither study is strongly relevant to the pi^esent task,

they do suggest the possibility that at sorie intensity, the aversive-

n«M oi' the redundant signal tone might cancel out or reduce its

infoirdng value. For tliis reason, the present study investi^^ated

the possible effects of signal tones on monitoring pei^foiioance at

intensities both above and below pre-established tonal annoyance

levels.

5

Itethod

M^aratus> A block diagram of the apparatus ic shoi^ in Figure

1» S^s room coribained all the apparatus for generating signals and

for recording response times. A synchrceiisod motor and Western Union

tape transi^dtter wecre used to pi^ram and control tiie H^t signals,

recording tijwer, and infoon.iin^^ toxies. The schedule o£ light presen-

tation was preairranged wiWiin the restrictions that tliore be a ifiinl-

MB Ox 20 aec, between each resporise and the succeedinfj stiraulus and

that theee inteinrals siiould rqprebent an avoxuge q£ one light Tor

evoiy 2 lain, in each 10 riin. block, Heeponse laterKsies were obtained

by rueane ef a TKmter electronic timer wiich started sinulvmeously

with the onset of each signal light and was stopped by 3»s response.

The sound source was a General Radio beat-frequency oscillator in

series vjith a Hewlett-Packard attenuator. Following each response,

3 heard the tone through riatched iinpedance earphones Tor 2 sec,

Tlie signal liijht was a 7 watt, 110 volt, red, jeweled light

mounted on a panel 7" wide and 9 " long. The ligat's luminous in-

tensity wab reduced by opei*ating it on 9 volts, a veiy diia but

suprathreshoLi level. An electric fan laasked outside noise and

liiaintained a constant ai.ibient noioe levels

A pushbutton sviitch was used by 3 duiang estimations of annoy-

ance tliresholds and latar to indicate detection of a light. i>uring

the Moidtoring session, pressing the s\d.tch shut off the Light and

tiraor and started the tone.

Procedure. S sat in isolation for 90 minutes, duilng which

CO<D

Co

u05

C-Pcd

^3

Cho ho Q)

+>CO t—

1

-P

O O-P

c&

iHE•H •HE-i O

CO

O

t

-P

p. HOJ CO

COS

Tr

0)

hronio-Po

V

CO

co

Butt

itch

(0

to

:3-pCO

UCTS

P.

o

SCTJ

Oo

0)

7

U»« annoywic© oliresholdb wero first obtaincxi and thon bho loonitorinG

MflPloa prwanted. Ss p«rrori,iGd iimvidaal3y and wore divided into

fl¥i iO^tsliMntal ciroups tlial ha4 identical uo iodultMi oi' 8Li)nulas

pr«i«ntation3. ?our of tho groi^jo v^ere proaonted with an Oono cpe

tona limiedlat©];/- follcwinj oaoh rooponne, The.'»e j^roupa viifrerod

with ro^jard to tlio intonaity of the tone as rollows: (a) 1|0 ^U, r©

,0002 dyn9/cn2, (b) an intensity dQrint;d as on© difforanoo linwn

(DL) DoloK- tho arinoyanco threshold, (c) an intonoity vlefinod as one

DL abovo tlia annoymica tlireeliold, and (d) 110 db r© ,^3002 dyne/cra^\

A firth group rocoived no tone tiuougiiout the adssion,

(*) ITii^OGlxold detQvrai)iation . Hi© I|0 db, HO db and "no tone"

ai-uups aat quietly in the dark for 15 raiii, bo beconua dark adapted.

The procodure for tiio Ss in tho two annpyanco ^roupn differod in

lliafc auditoiy bliretjholda were obtained While thcay woro dark adopting,

A edified ruetliod oj" lliita was used to obtain thresholds in

both annoyano* grw^s, 3b heard a series of pulsed bonoa that

started at a low iivbansity and i^adually iiior«a»eii, and were req'iired

to indicate v^ian the tones vere iiitenae enough to bo annoyinti. Five

ascondinc and five doccondiiig aorlea ware presented altomatel/ with

a 5- db L-:top for each s-accoedini; p'-iltio, 3s ii.iagined a aituacion whore

thoy stuJiod for an ii'<iporb;iJit exaiviiiiatioti and i-^ado their annoyancu

ju, :ts in terms oi" tliis iiaai;inod aitiiation, Th^y listened to

eaoh tone and responded by pre3oin{i the button once if they did not

think tliat the ton© waa of sufficicait annoyance to disrupt their

tudylns, and ureased tvdco if they believed tliat the tone HM

8

•nnoorlnG onough to distract rrom ntudyinc. The tliroahold wao taken

an the median of tho 10 tonos that ^rore Just vHuTflciGnt to ollclt an

nnnoyance .ludpsflnt,

Imriediately folloviinj: determination of nnnoynnce throoholds, the

intensity one ^ above or below the annoyance throsViold was obtained,

A DL waa defined as the difference in intonnity needed for a judgment

of annpyanco greater or lesa than the threshold 90/3 of the time. The

procedure coneiated of presontlnc paira of tonoo, the firnt being the

MHputed annoyance tliroehold and tho second diffeirinc from tlilo

threohold by 0, 2, h, 6, 0, and 10 dl). Tin eories of theoe palrincs

were glvwi with the order of prooentatlon heUniz randomly prearranged,

Sm in the above armoynnco threnhold [troup were presented conparlson

tone:- equal to or greater than their annoyance thresholds j 3d in the

below annoypjTce threshold groups only roceivod tones equal to or less

then their annoyance thresholds. In maldnt: these Judgmonto, Ss con-

tinued to use the criterion of annoyance in a study situation,

(b) ronitorlng task . Follovdng dark adaptation or clark

adaptation and dotemlnation of tho DL, Ss woro introduced to the

monitoring; tafik by instruction to watch for the ai(^nal li^^ht nnd to

rospond as qiiickly as possible on dotool inc It by pressing the slipul

button. They further instrictod tliat the llpht would remain on

until a detection was made and co off immodiatoly following the

response. Ne instructions ware given conceminc the px^aonce or

absence of tho informing tone. All Ss, including tlioso wlio did not

reoolvo any tonos, wore earphones,

A 9-mln. practice session, with five signals, vxae given prior

9

to the actual testing. Thic session insured that S3 mdorstood tlie

task and were sufficicntT-y dai^ adapted. The inforndnG tonas were

not pronented diiring practice. No rest was given between practice

and tostijng trials j S3 vjore told that the actual ejqjorlniental sessiOTi

b^an vdth the sixth signal light.

Daring the monitoring sosision, the four tono c;3:'oups received

the sigjial immodiatoly folloTwlng each detection, '^or one gro\ip a UO

db tone was Mstd; a second group received HO db; for both the below-

.-mnoyanc© and above-annojance croups, Girjial intensity vias the ajq)ro-

pric.te lndividuo.1 runoyance thresjiolvl ninus or plus the respective

DL.

The mcB-dtorinc; session lasted 50 nin,, during which tine 2$

stimuli were presented. Watches \mve t<?I:eai fron all Ss upon enter-

ing the eoqperincaital room, go they had no reference to the passage

of time nor of the intervals botvreen signals

.

Subjects . Ten 3s were randomly assigned to each croup. These

were 12 female and 38 male volunteers ©irollod in the sumner session

at the University of Massachusetts. Each S received Cl.I^ ^'or

ppxticipating.

10

IlMltt

L .--ifiSM ^S^±' ^ «iiio/aiico GToupa had cocpal'ablo aniioy-

anco throalioldB and DLo. Tho m6«n threshoM-) for tho below annpyanc©

thr«nhoW and abcm cnnoyBnee tlirijahold eroupo w«re 79,2 db with an

SD of !^,^ dh and 79.7 <-n> with on 3D of ^,03 db, renpoctive],y. In

both {pK>upB tho rango of throsliolda for Individual Sa wao fiwi 70 to

Oi; db.

I'on5.toring daiU . Th« raaponaa maasuva wao tho aam of the

rocipro of luLoctlon tixaaa during ouocooaivo 10-*^djfi, poriodu.

Thin convaroion t.ranaforrus tho dotaction timoj to detection opooda

and reducea tlia akounaaa oonmonlj' asaooiatad with latency maaaures.

Table 1 prmmAm mm md 30b of speada of each group for tht^ five

1.1,. lo r)^>rlo'i8. Sinoe an onalyr-.ia of rarianco (Table 2) indicated that

the B X T "ntornct lon was not eignificont, tone conditions apparently

had no difforontlal effects on tranda over tinio. 'Iille tho F for

tono condltlona waa not algnifleant, tlmt for tir.io was ai^ifioant

at ,01, i5ocft\ieo the offoots of tono conditions and tone conditions

interacting vrLth tiiae wer« not si;jnifioarit, maana of ti.o live ,;raup8

vara ooribinad in aaeh of the time periods and plottod in l-'igure 2,

With the aoGoaptlon of a aU^t iacreasa in npeed between the second

tfid thift period, perfolMnce decreaaed ovor timo, Aa notod, thia

daoraraant with tiiaa waa significant; moro ypocifically, the linear

con^)onent was si^jiificant but not the quadratic and culdc coraponcnts.

Thun tho deoronent in performance aver 5o r.iin, was linear.

Infipeotlon of tho last ooluran of Table 1 shot/a that the 110 db

fToup m.0 the only gronp that pwformod moro nlouly Uian the no tone

Table 2

Suraoary of Analysis of Variance

Souroe df MS F

BetwMn Ss

Ton« Groupe (T; .0325

.0372

Blocks of Tim* (3) .0089 3.885**

lAnear (1) .0263 8.i*71**

Quadratlo (1) .0012

Ciiblo (1) .0010 ——

^

B X T 16 .0026 ] .14] 6

Ss X B/T» 180 .0023

Linear^ (45) .0031

wuadratio^ .0032

Cubic** (if5) .0012

a error term for B, B x T

^•°»** error terma for respective compon?tnts of B

p c .01

1?

Fig. 2. Mean response speed for all groups for

each 10 min. period

•

11*

group. These means also indicate that by omitting the no tone gro\jp,

response speed -ws.n inversel^y related to signal intennity. As pre-

viously mentioned, these dlfferoncos among groups -"jore far from

statistical sifTii^icance.

15

Disousaioti

rhe perfom^nca decranent over tiiae ir^ concistent vith all find-

iijgs or iftonitoring beliavior (I'^ic^-^h, 19!?0; DeQse and Cn^iand, 19^3;

AdaiTiG, 195: J and Jonklna, 193'S)* Thia conHiiaation is especially

l»5>oi-taiit bocaixac anXf three previouG estperiiimts have 3budi3d the

effect of time on response speol (Jeiildjis, V?^6} I^iCGorrraok, 19$8,

19^9 )• Of these otucliec, oaX/ thOL.o ox i:cGon.iack provlaod a direct

moasure of the decrease in response Gpe^d to the pri..iar/ :3i{inal» All

other atudiea have used proportion of detection or rate of observing

roGponoes as the dependant rueasure. n:ie preL.^t rojulta ajpree wi"&h

and extend Jenkins' (191?8) and >icGorinack^s (1955 and 19^9) findings.

Tao five groups da,d not differ in oveiali perfoniiaiice or in

pQi'foiTaaiiae over time, Thoi-efore, redun^it inforirdn^ taie-- did not

maintain porfori.iance in this ciosod-loop roonitoi^iii^^ tad-, nor did

annoying; toncu reduce pei'foi^aanoe. In previous studies of noice

avor^iion (Asidn, 1953; CGX^^boll, 1955* 1957) # noi^ie wac a.iplo/ed

either ar. a response cue or as a noxious stiiaulus for developing an

avoidance or esccpo response. Closed-loop tasks w^ere not u^ed. In

tlic present investigation, the function of the tone to confirm

previous inforaation about rosponrje adequacy. i:iscape or avoidance

of the redundant ini^or^ain^ tone uas not possible j nor, was new in-

fonoation involved. In vie;^; of these procedural differences, while

the present results canix)t be considered contrary to iiJiplications of

those studios, ttiej do not support e^ftunsions' to the type of situ-

ation used*

..Itldn the range tested, performance tas not affected by the

16

cue and aversive properties of a redundant tone aa defined herein.

However, failure to produce annoyance decrements might have been due

to the choice of the particular annoyance threshold and the 110 db

tone. The close agreement between the groi?>3 a DL below the annoy-

ance threshold and a jU above the annoyance threshold indicated that

a fTbable estimate of the annoyance thrashold t^s obtained. The data

for DLs provided further supoort that Ss Txere mking judgments in

terms of annoyance or distraction and not just in terms of the actual

intensity or loudness. A DL based solely on difference in intensity

woiad only be approxiinately 1 db (lioscsibllth & Stevens, 1953) and

not the value of 6 to 0 db that was obtained. On the other hand,

the criterion oi' "studjdng for an exainination" raay not have been as

relevant to the njonitorin^ situation as was hoped. Possibly there

is no sin/3:le annoyance threshold, but a difforait threshold for each

particular class of tasks, 'Ihat is ^jmoylnc in a studying; situation

may very well not be annoying in a task that does not require the

sane responses . In the present study, 3s were not required to con-

centrate in the presence of the tone, which in fact, always followed

their responses. Therefore, if Ss did not consider the 110 db tcaie

and the tone above the annoyance threshold to be annoyintj, no

decrment in speed would be esqpected.

17

Snnmary

iuffecta of a I'edundant infomLn:j; tone in a olo^ed-loqp monitor-

ing task waire studied. Four groups receivod inibmint; tones iinpicdi-

atoly follovdn:; each detection of the tank li^^ht, Tlio intenaity of

tho tones for the groups was UO db, one DL below a predetominod

anncy/p.nce threshold, ono DL above a predetermined afino:',rn.nce threshold,

i|id li*^ db, respectively. A fifth group received no tones. While

ovorail renponnc speed docreaaed over time, introduction of a redun-

dant tone of anv intensity had no differential effect upon perf'onnance.

10

Tieferenoes

Adams, J,A, Vigilance in the detection of loi-r intensity visual

stimuli. J. e3ra. Psychol . . 19%, 52, 20U-208.

Azrin, N.H. Same effects of noise on human behavior, J, anal .

Bdiav . , 1953, 1, 183-200,

BsJcan, P. Discrimination decrment as a function of time in a pro-

longed vigil. J. Psychol . . 1955, 50, 387-390,

Baker, C.H, Three minor studies of vigilance. Defence P^search

Board of Canada, GmL Rpt 23U-2, 1958,

Carnpbell, B,A, The fractional reduction in noxious stiimlation re-

quired to produce just noticeable learning, J, cornp . physiol ,

Pgychol, , 1955, U8, lUl-HiS,

Ciwpbell, B,A, Auditory and aversive thresholds of rats for bands of

noise. Sci., 1957, 125, 596-597.

D««se, J., & Omand, E, Studies of detectability during continuous

vinual search, WADC Tech . Rep . 53-8, 1953.

Hiokey, A.E., & Blair, V/,C, l%n as a monitor. Itoaan Factors , 1958,

1, 2-15.

Holland, J.G, Technique for the behavioral analysis of human

obsei-ving. 3^,, 1957, 125, 3l+8-350.

Holland, J.d, Human vigilance, Sc±., 1958, 128, 61-67.

Jenkins, H.Il, The effect of signal rate on perfonoance in visual

monitoring, Amer , J, Psychol . , 1958, 71, 6U7-661,

Kackworth, N.H, Researches in the measurement of human performance.

Medical lenoarch Council So'.icial Reoort Series 'To. 268, 1950,r - -

.

- > II lit 1— I ff" I » "Ill

H.K, Stationary Office,

19

KcCoCTiack, P.D. Perform^mce in a vigilance task as a function of

intor-3timlu3 interval and interpolated rest. Cana. J.

fidiol., 1956, 12, 2U2-2U6.

ItoCormack, P.D. Performance in a vicil?.nce task with and without

knowledge of results. Cana . J, Psychol . . 1959, 13, 60-72.

Rosenblith, W., & Stevens, S.S. Handbook of r.coustic noise control.

Vol. III. Noine and man. WADC Tech . n£t ^2-20ii, 19^2, 27-31,

20

o

o

CO

C0)

0)

CO

HOCO

0)

col

co

o oP> o4-5 ^^•H(0 bCC G<D tH

c CM OpC3 MOc

.-^

%

O<^CO0)

>4o $^ o

c fl

n0)Oc

1

H(D

db

-dHOCO

a)

col

0000 On 00

00ON

OON On 00 CO

CO CO CO vo CO OD vo vo

O U> XTN OCO 00 CO

CM

CO 00 vo CO

oCO

o oCO

oCO 3 00

o00

xrv vo 00 ON

^ ^ ^ ^ ;S^

vo CO vo 00 00 vo

o oCO 00

oCO CO

vr\ vo OD On OH

21

Appendix B

Mean Response Speeds For Each

3jO MLn. Period For All Ss

Ss 1-10 11-20 21-30 31440 i4i-5o

No tone f^roup

1 .lil^9 .393U .3937 .5381* .h221t .3310 .3305 .221-3 .1331 .I8h7

} .3925 .250U .2571 .1351 .2617k .2599 .1751 .1579 .1628 .285U

I.2823 .2950 .3765 .2932 .2955

6 .2062 .22l;3 .1979 .1825 .1363

T .3^85 .3156 .3071 .3206 .32566 .2681 .ThOx .2581 .2771 .22I4O

.3175 .U005 .3128 .3175 .319110 .335ii .2763 .2600 .2008 .2533

liO db [roup

X .2159 .17h9 .270U .2639 .2212

a .2ii28 .2337 .2826 .276U .2li28

.5163 .6038 .5397 .51*10 .5213

.3ij26 .3129 .U330 .3697 .3608

. 3383 .iiOi3 .3992 .11126 .U027

.2270 .251i8 .3251 .2755 .2U99

T .2632 . 29liO .2311 .2631 .1950

0 .2798 .U013 .38140 .3100 .Ii383

.3902 .l'3ii3 .3'393 .I'.5l4i4 .lt092

ID .3612 .3522 .2976 .3181 .2li67

22

S3 1-10 11-20 21-30 3l4i0 Ul-50

DL oelor

X .iiloU .3592 .a6Ji3 ./1637

2 .U203 .3293 .3612 ,309U .3)-!6U

3 .2915 .2G28 .2711 .2900 .17U4k .3056 .2929 .3760 .32li9 .29285 .36611 .3658 .U9h0 .3159 .U806 .2356 .1962 .1019 .ism .1125

7 .ii235 .l'02U .U32U .3li38 .3510A

•UOOfV • - ri> .-^139 .2324

xu ,uooo .iiiuO

uit aoo V© group

.33^9 • 37yo .3910 .3720

2 .2751 .2S22 .2383 .2789 .2961

3 .3023 .2365 .26U .25oU .2821

.U689 .)4369 .'l50J; .»i071 .5175

5 .2703 .3055 .3503 .3038 .1987

6 .2003 .117-3 .2350 .1228 .0U78

T .3230 ,3660 .3538 .3295 .3592a0 •0 f y • ( • d'/LO

y .239u .2007 oilAo

10 .3300 1 onli

J-xU uD (^roup

m J .J i ^

2m 1987 3G08 J1II3 . ^230 .3132

.U283 .U2li9 . 1-396 .i'U2U .l;25l

I .2193 .1311 .1313 .1261 .1701

5 .3120 .1571 .3.977 .17h7 .2197

6 ,2207 .2335 .2579 .2109 .2166

7 .1723 .1650 •]i;l5 .1726 .Oh80

8 .2987 .1356 .2286 .2265 . 2532

9 .3501 .3Ui7 .3783 . 32l!Ji .306U

10 •U332 .366U .3Uia .3597 .3535