Title Page
 Table of Contents
 List of Tables
 List of Figures
 Previous applications of the pupillometric...
 Research goals and hypotheses
 Methodological improvements: The...
 Implications for future resear...
 Summary and conclusions
 Biographical sketch

Group Title: potential of the pupillary response in business research
Title: The potential of the pupillary response in business research
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00097597/00001
 Material Information
Title: The potential of the pupillary response in business research an investigation of methodology and autonomic contamination
Physical Description: xi, 112 leaves. : illus. ; 28 cm.
Language: English
Creator: Bell, Robert Roy, 1945-
Publication Date: 1972
Copyright Date: 1972
Subject: Pupil (Eye)   ( lcsh )
Psychometrics   ( lcsh )
Economics -- Research -- Methodology   ( lcsh )
Management and Business Law thesis Ph. D   ( lcsh )
Dissertations, Academic -- Management and Business Law -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis -- University of Florida.
Bibliography: Bibliography: leaves 93-100.
Additional Physical Form: Also available on World Wide Web
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00097597
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 - 000582552
oclc - 14124093
notis - ADB0929


This item has the following downloads:

potentialofpupil00bellrich ( PDF )

Table of Contents
    Title Page
        Page i
        Page i-a
        Page ii
        Page iii
        Page iv
    Table of Contents
        Page v
        Page vi
    List of Tables
        Page vii
    List of Figures
        Page viii
        Page ix
        Page x
        Page xi
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
    Previous applications of the pupillometric technique
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
    Research goals and hypotheses
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
    Methodological improvements: The feasibility of computer control and data sampling
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
    Implications for future research
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
    Summary and conclusions
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
    Biographical sketch
        Page 112
        Page 113
        Page 114
Full Text







This dissertation i dedicated tro r. paF rints,
Mlr. and Mrs. Al.ah Po, Lel l


This dissertation is a product of the advice and encouragement of

many people. The author 'ishes to thank everyone who participated in

the study, expressing special thanks to the members of his supervisory

committee. Dr. William For, chairman, has long been a trusted advisor,

teacher, and friend. Dr. Fo,. first introduced the author to the prc.tlem

of psychological measurement, and his constant encouragement and enthusiasm

for this project have been inspirational.

Dr. 'larvin Shaw has provided e-pert guidance throughout the

author's graduate studies. He has served as a source of inforrmatic,'n and

ideas, and has sympathetically led him through the pra-e of [syc'iological

testing and measurementt.

Dr. Walter Hill has provided some of the mo:t rewarding learning

e.'.periences in the author's graduate program. Dr. Hill's willingness

to listen to and explore the ideas of a betrinning researcher is greatly


Another member of the faculty, although not a member of the author's

special committee, deser/Js social recognition and thanks. Without the

interest shown by Dr. William \W1lrot, forrner chairman of the Managemrent

Department, the author would not have entered the graduate program. The

author would lie to e...press his deepest gratitude for the mania hour'

Dr. Wilmot spent giing good advice and counsel.

withoutt the ind iasistanr:e of Dr. Isrret laracan, of the U.F.

r'epartrmarnt of Ps/chiatr the project would ha'm. beenn much more dif-

fi.:ult. The tel e vision DuD1il1orneter he loairned .was one cf the most

valuable instruments used in the stud,,. 'Fimilar , Dr entryy 'Ki patrick

and Dr. Tirel lhalil, of the UI.F. Department of Industrial and S5steims

EnrginerEirin, haie the author's gratitudt for maliMrg the S/stenws Ergine-

ering/Hunan Performanrce Larorator, available.

SeV.eral professors orr.'vided help in the djEsi,'r phases Of the

e. D rnent. DrF Erich Thomn necl' rd ..10hn 1.allace of arinagqem nt, Drs.

W. W. Dam'son and oe. Har-ison of Ophthaimoloa y, Dr. George Barnard of

Psychiatry, and Dr. Pote ,t Is.ac:.on of Psycholo,: ha.e his than ;l for

their assistance.

Tuo student group' in the Colley. of Business Atmirniitiration, the

l A.E.. Club and Delta Siqra Pi, ha.e his oratituie for providing suh-

jectc. The author would li] e to than 'il Robert L.,irns, in particular,

for the effort he put intc acquire" iin volunteerss .

Mr. Alan Copr'e, f the E StFe,s Engineering Laboratory provided

support abo e and be.,:rd the ca311 of dutE,. DurinQo the data collection

pail'cd of the study, Alan spEnrt a lar'-! portion Cf his j3,5, nights,

and we.eIends noCrli on the er ent. e a trirt. H e 3'JtheCr's deeJDest

grati tude.

Several fellow graduate stuJnets hElped in different rhases of

the stud>. In particular, the author would like to thard .Jim Mc'Gregor.

Stu Erownr, I anrgle i'uen. and Fred IL'tt for their contributions. Also,

Mr. E\an Eldridge and the staff of the Eusi ness Admiristration Computing

Laborator ,, and the staff of the C om.munications Department in the

College of E.u:iness have the author's or'atitude.



ACl'l 'L E GrEIrE:;TS
Data Assessment Problems
The Potential of Pupillometrics
Traditional Pupil lmrietri,: Techniques
Pupillornetrics in Cusires;. Research
Scope of the Problem Area
Sumnma r4

Historical Develnoments
Attitude Theory
Pupil Size and Mental Activity
Psychiatri .: Pesearch
Emotional Peactions and Pupil Size
Marletini Research
Oissenting Votes

Research Goals

E:.perimental De :ig
Data Collection
Data Transfer
Data Peduction






Analysis of PuLrilldar. F:*.action, 4

FE A I I L IT i rF Ci'Or!PlITEF' COlT li'L b.riD

Atti tude Thecor; '.2
Fese'rch on Stimulus F presentation 3
Field Pesearch 84
Laboratory" F'eseairc-h 25

CIIHA TEP VIil: SUJrtVi.Pr AND CON;CLUl l :rf. 87
Gboal: 87
I;ethoid 7
/rnal.sis 9
Results 89
Cos tr.'E:efi t Analysis 90
Conclus ions 91

E:]VELlC,'P PHr ,9





-, 1 -

2-1 Summar, of Pupill .T.-tric Studies 25
4-1 Affective Patings I Indceprrndent Judges) and
Luminosity Scores of Slides 37
4-2 Order of Stimulus Presentation .1
5-1 Prediction Variables Used 51
5-2 Anal.sis of Variance of Ouestionnaire Results
for 4 Groups of Slides 54
5-3 Last Step of Multiplc Steop'.ite Pegression 541
5-4 Stepuise Linear Pc-gression 3-Second Percent
Change in Pupil Size 6
5-5 Last 'Stp of Stepwise Multiple Regression 5S
5-6 Sterpwise Linear Pegression, g-Second Percent Chanr]t 59
Pupil Size
5-7 Correlati nrs between Puril Chringe and S.tirnull 60
5-5 Summar: b,, Stimulus Grrouprs of i3-5.ond Hear trate
Fegr erss crs 61
5-9 Summary,' b/ Indi'.idual Stimuli: Steoise Linpar
Pegressior.s with 9-Second Heartrate a:s Dependent
Vairiables 62
5-10 '.um7mar, by Stiiiul us Groups: '9- second PFrcent Charge
in Heartrate 63
5-11 Stepi-rise Linear Pegression of .-Sr cond Percent Chang..s
in Hood Pressure 61
5-12 Stepwise Linear Pe qre.s ior, of 9-Ceccond Percent Charng:es
in blord Pressure 65
5-13 StepiJise Linear Regress ions of 9-Second Percent Changes
in S in Potential 6
5-14 Pvsults by Hypothesis 69
6-1 PDP 1 /20 ,,stem with Detanhone Linkagr to IE;I 360/65:
Ccst I'[.L-nefit rnal1,'sic : Benefits 78,
6-2 PDP 11,'?ij S.,steom r i i h Da.taphone Linacge to IE.r 360/65:
CostEenefi t Anral sis: Costs 79

-\ i i-



4-1 Procedure 30

4-2 The E-periniental Environrment 31

4-3 Equipment 33

4-4 Pecordi ng Equipm'ent 34

4-5 The Eperinent ir. Froqress 35

IHt L l.il ; l IPL UI IHE PUFILL.'Ht HLi-l.[J! h. 1Hi HU.iJL.: HLP.L i.HM:


Fot.ert Poy Bell

August, 1972

Chairman: William 11. FoA
Major Department: Mianagement and FBusiness Law

The purposes of this dissertation were to replicate findings pre'.iously

reported in the field of pupilloinetrics, develop a linear model e:-pla'ni n

the parameters contributing to charges in pucil size, examine the feasibility

of utilizing computer sampling techniques in the collection of ruoillometric

data, and to catalo, physiological responses as coariates of the pupil. A

problem limiting advanced applications of pupillonmtric technique, ha; been

a condition know,,n as autonomic contamination, a phrase describitln the dilation

of the pupil when e.rposed to ansi.ty-arousing stimuli. A goal of the disser-

tation was the utilization of several physiological indicators to define dif-

ferences tbtween pupil responses under autonomic contamination and certain

other conditions.

The study inrolv'Ed presentation of a set of visual stimuli (slide:)

to a group of 29 subjects, and incorporated a randomized block design.

Stimuli were divided into four types: pleasant, unpleasant, neutral, and

an.iety-arousing. Light intensities rere equated for the 16 slides used in

the study. TIo startle stimull there used as hart of the aniety-arousing

set. Environmental factors such as ro:Im light intensity and scund levels

rcre controlled.

A teleiisiion pupi 1 lmeteC'r was used to monitor changes in pupil size.

Other physiological .ariables measured were heart rate, systolic blood

pressure, and skin potential. A dynograph direct-wiring clhart rUecorder

recorded continuous samples on ea:h of the physiological .ariables. These

data were coni erted from analog (c oritinuous)j forrr to digital (discrete)

form, by 3 Eun.err-PF'yrio U'f-', PO r.rocess control computer, which tool. s-amples

of the dita at one-second inter'vais throughout the study. The computer

controlled the timing and presentation of slides during the experiment.

A linear model of the factors contributing to pupil size changes

Was developed for analytical purposes. tep-wise lineer regressions, the

chi-square t,-st, and analyses of verianc uSare use to test the data.

Other data jnalyes involved time-lags and descripticrns of response imai-

ritudes of the phsiologicai :.ariables under differing types of stimuli.

Results r:iised questions about the usefulness of current punillomet-

ric techniques in business research. Variables in the regression riodels

produced multiple correlations of 0.62 and 0.57, respectively, for 9-second

and ?-second percent changes in pupil size. The corresocndling e. plained

variances of less than Jir1 percent indicate that t'lere is a great deal of

reactivity in the pupil '..hich was not e-plaired by tthe variables considered.

Furthermore. the direction of pupil response to the four types of stimuli

did not follow the patterns which have been previously reported by proncnents

of pupillometric techniques.

It does appear that other fi ologcical variablecs e-hi bit s.me

amount of c:nsisLency in their reaction, to an iet.,-arousing stimuli.

covariate with pupil change to irdicate situations in which autonomic

contamination exists.

Although the results tend to question the use of current purillcmetric

techniques in business applications, therE are certain areas of laho'rtory

research where they may still be useful. A cost/benefit study of the feasi-

bility of using a mini-computer in the gathering of physiological data in

the laboratory was performed. Results indicate that the use of computers

as a laboratory tool can be justified in many circumstances.



In any organize tion, the processes necessary for oncoina acti. it

and success of the enterprise are carried out by human beings. The

study and practice of mrna .eernt. therefore, has implicit in it 6 n:-d

for understanding people--hou they or. how they thinkl, ind how- the;,

can be most effect .ely utilized for progress to Jrrd organization goals.

As Lier't has said, "of all the tasis of mranagem'.nt, mangingr the human

component is the central and rost important task because all else de-

penJs upon ho.u well it is done" (19i 1, p. 197).

Information on r many different aspects of the hurian variable has been

compiled. One of the most elusive, most hard-to-get-at arens of human

behavior lies in the area of ittitudi nal .and eriotional feelings held tb.

people. iet these are unquestioniFly morning thr most significant determin-

ints of human beh.avn ir. The rese' rch described in this p.per was an .attetiiF.

to further refine a ne,. and promising technique for assessing these types

of variatles.

Data Assess:tint Problems

The internal states of individuals art prob tbly amoninq the most dif-

ficult of phenomena to measure. Man.' attempts haje beer man t d. .'eic,

unbibase d :,ar'sticks cf thI emotional and attitudinal resr.ronses of pecole.

Coin.ventional m asures of these' responses such as r.naEer-and-cenc:il question-


to differentiate "true" reactions from those containing elements of arti-

fact. Ir, many situations, conventional measures are still relatively

accurate and useful. In some situations, how:e.'er, particularly in industry

and in market research, respondents may have scme reason to conceal their

true attitudes and,in these situations, the traditional approach to measure-

rent may be limited in its usefulness.

Lapiere (1924) was one of the first to report a discrepancy between

hot: indi,' duals, in response to a questionnaire, said they would act, and

how they actually behaved. Cook ard Selltiz (1964) prseeited a riore com-

prehensive analysis of the problem of obtaining honest and 'alid question-

naire reactions from respondents. They noted:

Susceptibility of ov.ert response to distortion--that is,
the possibility of discrepancy between private and overt
response--uould seem to be a functions of three character-
istics of the measurementt] instrument: the e tent to
which its purpose is apparent, the etent to which the
irmrlications of specific responses are clear, and the
e'tont to which responses are subject to conici ous control
(p. 222).

The problems of social desirability respondingg with the socially

"correct" answer: Posenthal, 19E.6), demand characteristics (cues which

convey the goals of the study: Orne. 19'9), and evaluation apprehension

(fear of being negatively e.'aluated: Rosenberg, 1969), have also been sug-

gested as contributors to bias in wost tpes of que. tionnaire measures.

Y'ugman (1964). in a marnVeting research study, surveyed results which show,

how social desirability and demand d characteristics interact to distort

several t)pes of questionnaire data.

Other types of methoidologies used in attempts to assess emotional

and attitudinal responses, such as the projective technique; used by Ic-

Clilland. have pro,.ed to be someI:wh.t useful. Their validity coefficients

are usually: lo,-, houe'ser, perhaps due to the fact that scoring of a sub-

ject's responses is so difficult. The tra'inng required to be ahle to

use pi.jective assessment techniques is long and rigorous, and the scoring

methodology at best preserits mTiany areas where errors of riterpr'tatiorn or

other e.*perimenter effects could bias data. The score ron some t pes of

projecti'.e techniques in addition, is to a lairg degree a function of the

vocabularyy of the respondent Itee lMcClellanrd, 19F9. E.Dth question nn.aire

methods and projc,:ti .'e techniques, therefor- are subject to several sources

of assrssmen-t error.

The Potential of Pupillcmetrics

The dc-.elcopiTment of pupillomietric measures of attitudinral/enmotional

response iHess and Po lt, 19V-1: Hes: 1965) presents the hope of circum',ent-

ing some of the errors associated i tith paper-and-pencil questionnaire types

of measures arid pr'cjecti.e techriquc s. The theory of pupillorretrics is

chased on the hyoC'thesis that the pupils dilate in rFrsp.nse to ple-3surable

or favorable stimuli, and coritract in respr.nte to ne'gati.e or distasteful

ones. Sirice the eye is part of the. auttonomic nrer'ous system, inrdi. iduals

presumrlbly carinot consciously control changres in the pupil If it is
The rcsear.-h on conditi iiin r of the pi ;illairy rerporse his produced
equi'.ocal results S.c'veral investigations ((ason, 1912; ct:rier id Eaber,
193. i. Girdern 1941: Crasilneck and -IcCrarn' e, 1956.) reported cositi.e results
in attempLs to condition the pupil It seemT. that just as man. studies, ho':-
e 'er. 'reportedcrjntrar-y findings (Stickle and Cerishme, 1934; I'edell et al.,
1940: Hilgard. et al .. 1949; i'oung, 1954). Even ,.hen conditioning is achieved.
hlo'.e'.er, the process is quite lro.i, aind appears to be limited to a soecific
type of i rdi. idual.

possible to reesure changes in the pupil, and if the logic rmentionEd above

is valid, then it should be possible to obtain a true or objective measure

of attitudinal responses to a stimulus.

Pupillometric assessment should offer several advant.a,.es over other

t.pes of reaction indicators. The response tc. the stimulus is very fast

--usually beginning in less than a second. As mentioned previously, it is

difficult to fal.e a ouoillorietric response, since pupil changes are autonoric

(i.e., not subject to conscious control). A third postsit le ad\antain may

stem from the bi-directionil changes of the pupil--while most physiolocgi-

cal indicators give only an indication of the size of the response, pupil

changes show both size and direction. Previous 'or'k i'ith phl.y:iological

variables, such as that done by Cooper and Pollac. (19.?) on prevuicial

attitudes and the galvanic s in response (GSP) mi.-ht h\ve deri .ed more bene-

fit from the bi-directional pupillar, response approach.

Traditional Pupillor;etric Techniques

The methodology .employed in pupil llometric e,:perinrents ha- been out-

lined by Hess and Polt (196)') and Hess (1965). Eriefl,, a pupill ometric

e-periment requires some type of stimulus. input device, and some type of

device for recording, pupillary ch.anq. Host pupillcmetric e.,perirments have

involved visual stimulus presentation, and have th.refcre utilized slide

projectors as stimulus input devices. Other types cf stimuli which have

been used include liquids (for taste research), still- and motion-oictures

(for market research), sounds (auditor:, research), and several t,pes usin.

cognitive information processing, physical uiorl or startling csunds as

independent variables. I'ost e:ptc 'imenters use rapid frame cameras to tale

pictures Cf the eve, *and leasu ire changes in pupil size frorr t', pictures.

In the most common type cf pup lClmetric stujdy, the subject is seated

before a rectangiul ar boyv i th a vi ev screen in onf end. In ord er to mrini-

F ize rneasuremenr. erroors resulting r from head mc'. cment the siub.ict't head

is usually nimtrobiilized (placed in a chin irest, sometimes with an elastic

str'.;p holding the forehead aqalinst a bar). EquipFment for prc. jecting slides

into the .I e. screen is situated o utsiide the bt-:. A mirror system a,hereby

a motion picture or rapid frrime camera ca r photograph c anges in the eye

is used. The mirror r is placed below the subject line of Sight, at a

4:." anrle, so that a camerer mounted in the side of the boe can photograph

the ey,,. Since the eye is highly reacti.e to light changes, slides are

usually contr'tolled for ll ght intensity, and photographs is performed using

infra red iqu pment. Ihen developed, the photo,?raphs are Tmeasured by hand,

us-ing calipers or millimeter scales or grids.

Pupillom retiric in Cusiness Peseaich

As will be seen in Chapter II, pupillometrics was first emolc.ed as

a business research r method in the field of .jirl'etina research in the i IO 's.

The potential of pupillometrics in this and othcr bhusirne. fields will be

di-cu:sed belov'.

The Perscnnel Function

She personnel function is in area of niaranerial decision makl:in.j which

may benefit from moore accurate information about attitudes end values held

by potential and present employees. This statement must be qualified

b' the possibility that specific icbs nay riot require that prospective

en.ploy'ees' attitudes/,'alues be evaluated. simply because measures of

thEse variables have beer, shcrn to have little predictive validity--i.e..

have litt'l- or no relation to success or the job (Porter and Lailer, 19C,9).

A question that has riot been ,riswered, hoe'ver, it whether the low validity

coefficients are caused by no relationship between the causal and dependent

variables, or because poor measureirent techniques have produced inaccurate

reading of the variables.

There are sone types of managerial positions where specific attitudes

and values have been shoc.'n to be critical ccmponenrts of success (Le.'inrsor,

1964). Certain positions for e'-ample, require that incunitients be able to

perform effectively ir situations involving suhttantial uncertainty and

high risk. many individual: have to assume responsibility for handling

and committing large sums of corporate furins. Other attitudes and values

which may be important include thc.se concerririn minority groups, corpc.rate

ethics, social respcnasi-bility, or tho-e necessary for ,oriinrg in "c.r'ganic"

forms of org.arn iations .

Although pupillometrics is not today at a stage where attitudes and

values such as those mentioned above could be readily uveasured, it is

this type of difficult ime surement which is conrteriplated for advanced

applications of the technique. It riay be that physiological measures,

%hen coupled with other t.,-pes of .assessment, will provide practitioners

with the type of data needed to evaluate candidates along these dimernsions

of "personality".

Certain ethical questions and-questions of acceptability are raised

when industrial applications of phy: ioloqical measuremTent sys tenis are con-

templated. In organizations where high level job candidates are nornially

conducted through soein tpe of psychiatric or psychological evaluation

procedure, pupillomietrics nay be accepted as a matter of cou-re. In

organizations where the technique m ,'y be deemed undesirable, pupillor.mtr:ic

nay still contribute mTaningfully. One of the hopes of de.elopi n. an

easily useable and accurate method for evaluating attitudes and values is

to use it as a tool for validating other types cf de.'ices which rni he

useful or more acceptable for s.aT t ypes of assessment probleFs.

Another application related to the personnel function lri.-s in the

area of human factors ernginEerring.'orl: design. Pupillometrics has been

shown to be a useful indicator of certain types ot stress and arn.iet,

levels (Hess. 196.b), and of noise levels (Nlunnall., et ai., 19E' ). Pupil

response systeT:s. might pro.'e to be valuablee aids in designinq jobs rnd

work' en.,'ror~nments through tle study of their effects on physic.logical

systems .

Mar.tetina Pesearch

Perhaps pupillometrics is one of the oldest kno'.w.n riarketing tools.

Fess (196L,) noted that Chin ee ijade dealers had emp loyed the technique for

centuries, watchingg their client's, eyes to tell when his interest in the

product ':aS highest, then mailing the s..les pitch. More scientific studies

of the pupil in marketing research r-re conducted t, BSrandt (1i-15), who

used ":cular photogra.hy to measure responses to 3dertising messages.

flore recently, pupil dilation and eye lotion for ouoil track)

have been suggested to be useful in measuring package design effective-

ness ('est, 1962), television commercial interest peaks (Lrugman. 19D a),

and product response (Eusiness T'ec, 1967). Fruaman (l'64b)re,'ie'.ed a

series of studies where pupil response was shoun to be indicative of

interest in products such as sterling silvnrware and greeting cards, and

also presented data supporting inter-subject consistency in pupil response

rankings and sales rank data.

Hess and Polt (1966) performed research on taste stimuli which could

be construed as product preference re.erch, jr.d were able to demonstrate

correlations between pupil size and e'.pressed preferences for certain

drinks. The same study also found that both strong positive and strong

iegati..e aversive taste stimuli dilated tie pupil. This finding lends

credence to the hypothesis presented by some researchers that magnitude of

response, not direction, may be the most important indicdctor of feelings.

Halpern (13S7) presented data on pupil responrss to TV commercial and ,ac -

aginr, and noted that contractions were usually found to be associated with

stimuli which "lac the power to interest or arouse the viewer" (p. 7). Hess

(1963a) reviewed additional successful applications of pupillornetrics to

advertising and packaging research.

Scope of the Prcblem Area

Nethodological Improvements

Methodology is a factor limiting the overall usefulness of iost new

techniques. In pupillomentrics, areas needing improvement are methods of

data collection and analysis. Typically., nhotographa of the eye are taken

continuously throulihout a pupi llIme tr'i- e.peririent. These ,hojtograr.hs

rust be developed. and are then mea ured b, h3nd to determine pupil

diameter onr each photoigrrph. This writer has previously demonstrated.

on the ba.is of a simple study, that a porsibilit., of nmalirn major errors

in data interprettio n e.i.s hen .'perim.enters measure Fupil diameter

b, hand (Eell, 1971). This pro.bl-em of e.,.pectancy error i co;ipoiurided

by the sheer nunmter of lme asur.r.ents to be made. Fori e, nmple a study

involving i subjects, with 1. Stimuli (and 15 control stimuli) sho:.-i

for 10 seconds each, Ijith photographs tia.en it 1-second intervals during,

the e: -erinment, would 'require the devel iopment and riessurmigenlt of t6.0O00

plhotographs. Furtherr, a lag time F. ist: between the time the e..per irf'nt

is conducted and thei tire the e..perli:ernter deter, inEs whether his data

are usable--i .e., whether the cariera and film were operating properly.

One of the mnijor goals of the res erch reported in th;i dissertation

i,.6 the jdeveloprernt of computer sa.ni lirn techniques to circumvent the tedious

and perhaps unre li able mieth.od; of data collection mentionrJd above. The

nmethod used in thi. research involveaja TV Punillorr.ter (a video ciaerai.

c:apble: of continuously recording j picture of the eye and rjmeasurirng pupil

di;aneter) tied into a D.inograph recorder (a de',1ice pro'.idinr charts of pupil

and other iphysiological changes) and a Bunter-Pamo FP-330 process control

hybrid cor'iputer. The comORuter sampled dat5 once peir second, and con-erited

the data from arnlog ( conti nuous forri to di ital (disc-re te form through

the use of an analog-to-diitai con.:ie r. The feasibility, and general

appli cabi 1lt., of this technique for laboratory and ron-l ibor ator, settings

is discussed in the paper.

A major problem associated with pupillor.ietrics at thf present state

of the art is caused by reactions of the autonomic nervous system to cer'-

tain types of stimuli. Instead of being a "clean" indicat:.r of a favorabi le

response to stirruli, pupillary dilation is also causeJ by s:,me ve-ry non-

pleasurable stimuli, such as fear and pain (Hess, 1'?62a) and stroinly dis-

tasteful liquids (Hess and Polt, 1966). Since pupil size changes associ-

ated with most stimuli are relatively smr,all, it is highly possible that

autoriomcrc reactions could oive investigators cornmoletel r-eversif-d data from

what the:, expect. A high dearEe of fear associated with an attitude

object, for e,.ample, coulJ, by causing pupil dilation, indicate a .:ry

favorable- ie.spi ne to t.h stiniulus. To Lchieve a h&L.ite r urnd-rstandinu

of why these responses occur, and how they might be "controlled" (partialled

out of the data analysis), a discussion of the autonomic nervous system

is necessary.

The -utornomic nervous system

The word autcnricmic has been defined as "acting independently) of

volition.' Tlhis definition describes the wor[iings of the autcinomic ner.cous

system ( rnS)--"acti ities are largely involuntary, and we- are usually

unaware of them" (Sternbsch, 1966, p. 14). The AIlS is generally concerned

with the regulation of the visceral system of the body, and attenmits to

maintain "homeostatic equillibrium" in the face of '.arying e.tern-l factor;

aff :tinq the body (Cell horn, 1913, c,. 195). The AIIS can be broken down

into two antagonistic :ut-sySten-, the synr.athti: neroius s ,tem (Sr 4 )

and the parasympjtletic nervous system (P'"lC). The '!JS generally' pro'.'ide

emergency resnrrses., 'while the FPJS attemrrts T. mediate or slo'.' dio'.n autc.-

n.mi.: act.i itie and restore normal metabolism ('ternhb.:h, p. 2").

Autornomic innervation of the pupil

The eye and the pu 1il are "doubly boundd, biing innervated by both

the FP and the .1_.. The dilato r iluscles in the pupil are inrvi': ate.jd b.,

the 514,., while the sphincter muscles, which constrict the pupil. a e con-

trolled b by the FPN. (!ilner, 1970. p. 17 .i. The size of the pupil iC con-

trolled by the antago? risticr intr-ractions of the ta.iO sub--.'Sterrs In

qrner.l, e.-itati.in of the is rrpath-tic system cauI.te contr.icti-n ..f the

dil.itor muscle, ..rhich in turn di lat.i the rin il. whi le parjosynrr pathetic

ex., station causes the sphincter mTi tcle to cons' trict the upil ir;-.ossman,

196? p. 170). Se'.eral investg.atorS telio- e there is an inter.iction

between the P1IS and the Sr:, which c uses dilation. Adle-r (1'50, p. 1 .),

for e'.ample. bel ie'. s that inhi .iti on of the splhiri:ter and crontraction

of the dilatcor muscle Irt toi tthir r to cause puoDil dilatirc.: in mrin, uhile

thr- ,phincter alorine contr cts thc pupi l .

It appearss that .iliticnr of thi- ey? resulting ? from f-ar. howi'..er,

occur not from a joint rworlinn of the SlI; and PF'il. -ut from! a phySi cloir icl

dominance of the S!'. o'ver the r',;i. That is, in timcs rf high statess cf

fePr or pain, the simp.nthetic S) 'tem "tales o.,.er" control of the eIe i rsaac-

scon, et ii 1971, p. 2;E., I'dler, p. 190). This findinri r ill te th tii- ri:

of the e,.pevrimental desi'ln to be dr-scril.:.ld later in thi. pirr.re., it should

be possit-le, if the finding is v'.lid, t, monitor the 'I.'.. anid PFl' and r'b-

ser:.e '.aribles other thar the rpupil ihe- rtrate, tb.looj DIre.C ur e, slin po-

tential. for fe, mple) to define l ,ei'1: of ': iand PF ". a, ti'.'atin'i, arnd

thereby cbtaji n ins i ghts into the puci ll ,-,y resrpon: e. ;Whe.n tot'h .y-,stem

are noriirally aict-i.e, the firndingii h:.poth.;i :i tze: :' Hzes S. hou11 occur. Whh-en

the 515S domain ter- ho;cv'er, w-e sh'IOJld e.'pect findiinss contrary, t.o those

associated with normala" affectie r.-action: of the pupil. If levr el: (of

.activity of other lutonomic .ari .rt -lees such .s the ones mentiron'd .F.c..,*: can

1e .hoi:n t c.:*orre- po',nr d lith the s.,n.*'athetic dori.inanre, then the fear i rnd

pain re-ponwe, can he partialled O'ut of the pupil ret polie t-b :orcurrn-nt 1

moni tor i n, other 3ri.3i]let .



It appears that Fupillonm trics is a fruitful tool for conducting

research on the erntional and attitudinal filings hild ) ieorple. The

technique nma, have apolications in riraogenant .rd narl eating resear h.

IMethodological shortcomings anj autonomic cornt !mina-'tion a re two confounil-

ing variables whichh pr'esent1l limit its usefulness. The gol s of the study

reported in this paper are to impro':e the r,.,thodolooy, of pupi ll om'-tr is

anr reduce tle problem of autonomic ccnt eination.



Onre reason some, researchers tend to disreg.ird the rupil response

as a valuable tool is that it has been shown to -e a general indicator

of many different typesof processes t'unnally, et al., 1967). As this

literature review, will show'., pupil change has yeepn used as a dependent

\ariable in attitude research, psychiatric resedarich, m.ari.etinr research,

and gustatory. (taste) and auditory research. Pupil size has beer, shown

to be a general indicator of ph:yical and merntal acti..ity, Fnotiona I

arousal, and interest value of visual stimuli.

Nlotwithstnrdirig the "generality" of the Teas ure, Curj-ilometrics

still appears to be a potentially valuable tool ior tisirinss research,

gi,.en proper controls. This chapter wil re.ier- in greater detail the

investigations cn which have contributed toc our present 'rncv1ledge of pupil-


Historical f'evelooments

The contention that the pupil of the ePe is a reflector of emotion

has be"n held for somiT.e time. As early, as 1765, Fontana descrii-,ed studies

showirnq that pupil dilation folloe.-d the introduction of fearful or Dain-

ful stimuli, e'ern in the pres Cnce of britaht licht. Bender (1933) photo-

graphed sulbjectS' purils in the presence0 of sirul taneous trig''It 1iqht

,and emrcitonal stimuli and also found a "p-,chic" pupil reac-tion.

Attitude Theory

In 1960. an inrr.estigation t., Hess and Folt uncovered a relation-

ship between pupil changee and the interest .alue cof :'isual stimuli.

In this study, the eperim-rters displayed ,hiotoraonhs of male .rnd

female rpinups a taty, a rn.ther and e baty, and a clands:ape scene to

male a;nd female subjects. Fupil diltrion' was laIrest for mer, when ie' -

ing the female pinup, while e the women sho-wed great'- r ouoil response to

the pictures of the taty, the n'lther 3nd tab ., and the male pinun. This

stud rmaled thie Lte nnini cof a concerted effort on the part r.f less

and his associates to study the eye as an indicator cf er,-ticrnl reaction.

In 19?'., Hess surim ariz.d the results of ni.osit rrior studies of

attitude a'rd pupil size. Some of these in"estigatic.rns, using aversive

stimruli such as Dictures of shares, cross-eyed and crippled children, or

strongly unfa .orat.l e political statements, cause'i [upir i constriction of

over 3 percent change in a.'erag'e diameter. Other negjti.e stimuli, suchi

as sips of unpleasant tasting liquids, also caused constriction.

Hess also suggested, on the basii of prenliriinary data collected

during the 1964 political carmpaigins. that the uCunil response might tbe

useful as an indicatori of attitude change (paoe 5).

Puc1il response w.s; sho-.'n to be caret.le of differ'enntiatina betWeen

heterose-ua s and self-adn itite~ l honrce.uals in another studr I'H,:ss

eI tzer and Shlein, 1935). In this r.ilot e'periiment, subjects were shown

slides picturing males arid females in varying states of undress. Four

out of fii'e homose uIls had larger pupil responses to pictures of males

(some up to :35: larger) while five out of fl'.e heterosexuals had larger

pupil responses to pictures of females. A Droblem exists in this tudy,

howe',er, in that the larger pupil response to male pictures on the Dart

of admitted horiosexuals' mayi be the result o.f a type of erpectrncy effect.

Once subjects hai admi tted homosexual interest prior to pr'esr-ntatioun of

the stimunli. the dilatinns could d be phys biological acti.vation1 caused bK

something other than hiiiher sepul interest in the mTa le pictures--some-

thing sir,ilar to the well-I noun skin flush when one becomes a.-'ar he is

being obse v-l.

In : d':rtstin, ld.r',? (17,?5) CLaluated the puril responrie

as a means of reasur in attitudes toward Negroes and found that affect-

relateA1 responses, while significant during e r ly presenitati cns, diminished

during; repeated trials. He did find a significant difference between reari

responses of anti-'legrc. and egalitarian subjects thus provicdirni Euprp.rt

of Hes:s hpothe'.es The repeated trials inv.esticiation als;c prodi.:ed dis-

heiartening resultS concerning the stability of the pupil response, whichh

was freasured to he around 0.50.

Hess '(19I.b) alio described the peculiar results cf the pupil's

fear response, and discusse,1 invpesitigaticns shuing thai t high g alvanic

skin response (GSP); iee Is ere a concomi .tarnt of urn il di lati on from,

shock ocr fearful stimuli. I When GEP iriFoped out", punil dilation turned

to cons tricticn Ip. ,581.. In another ;rti cle (19F.'h a). he su testedd that

fruitful futuLre research should concern itself with ine, stina tions c.f


the relationship b-et .i en the pupil and ri.n-visual srnrses and t.et -:een

the pupil response ard c.tlher ph,.siol ,? ical rr.e sures.

Pupil Size ,nd 'lental Actijitry,

Data showing that pupil size 'as directly, related to rentl- acti-

vi t, were presented. t.l Htess arnd Pc lt (19 6.1). Thi s e.- oer i ent, r'here

sutiects ,were pre e -nted i i th in.-rtal matheiT tic.al probc lems of .ar,'i ng

degr.Fs of .di ficult, while their pupils itere phc.togr.rphedi shr d,'.-- that

mean pupil dilation wji3 high] co rrel ated ,.;th r .rrblem difficult;,.

1-Jord at. tracltrlne s 'as Eisa o shc',.-n to tb r'Elated toc pupil size when

:i.ibiectt :s er asled to foin r im nrtal iima.-qes rf ab.st i'ct r an1 conCre e it words

(Pa'.io and i:is'.r o n. 19'66). The FP vio anrl S.i.I'pisoi tudy found no relatii'.n-

hip, ho ~"eter, betr4een the rpl'as ntni.s- -un ple.s jtnesi of a ucr'd ind

punil si e (p. 5SE An. he r s tud,' corrob.-.rated the first FPa io and

ips'orn fi'ndir, g in a3 tud on', pupil si-e nrd rimrioiry 1 rIa finding that

"pupil di airiett-:r s r measure of the amrr.unt of material thich is under

acti',e processing at .an tjie" (.ah erirean andE' r.tt 196r,, r. 15: ).

c'ome studies reported in '167 included one t'b I .,rn arnan ard E.ai le,'

repli c tir.g tihe inf.:r.n l tion load end ior.i l I 5 resiul t; repCi'rte1C in

19 6i; ; rep':rt t'y h'.id -d A nderson that ruDil dia.,ete"r mri., t.e a me,.h.n nis

of cerc;ptual rcc.g nition trrethol d. and a stuj:d ty Fes,.ler and !icLauhlin

on the r.elatic.nship between n io.el tiruli and r.upi si ze, all c.f w~ icli

gesnerall, supported the r.su tr.s dlescrlt cd in the aboVe p 'r 'ar, hs Pr. lt

1i 70) descritd another" .-::perient shoniig that subjects e-rtinrg Ir-ater

mental effort had cocresrpond in.-i / greater pur;il dil-ttion.

Two dissertatioiis written at the Urniversi t of O1 lahema (Clar-l ,

1971. and Johnson, 1971) investigated urunillary resrlo'lses duririg short-

term riemiory tasks. Both replicated the finding th t amount of cooiniti.ve

processing ard rpuill.1ry diameter are highly related.

Psychiatric Pe-earch

A stud, in the field of pr'schiatry (Pubin, et al., 196'.) suoG.ested

thleT pupil response ma, tbe a usrcful ps',chiatric diegnortic tool. Pesults

shlo':.ed that characteristic pupil changes in children with cystic fit.rcsis

v'ere significantly different than those of normal chi ldren. The studa,'

also discussed the effect of age or, oupillary reactivit;. noting that

response characteristics tend to change as people- hecorre older.

PF'search rer.orted K,' Sheflin I1969) inr.estigated uriil reactrions

of hnspitalizEd schiz:rphrenics and found that, if pupil dilation is in-

terpreted as a r.ei ire of interest, male paranoid. are not iIIore -e'uall

interested itr m-n than in worir'en--a findinri cntrar, tr. that held bJ y most

psychiatric researchers.

Eoersna, et al.. found that pupillar., response was useful for

studying differences in cognitive information processing irn -educatable

mental retardates aid normal children i1970, D. 143). Purin] the S3am.n

time, Tanci aid Fot.tins invm tigated the [redictab- lity of the r.ur il

response Iusino the Dream Incident Techninue (DIT), the Edijard Per'sonlr

Preference Schedule (EPPS) and the Pepres.-ic.n-.rensi ti action Scale. They

found the [PIT to tb, fa',irly rni"edicti' e of nur.il rsnronse, while scores

on the othie two sc,;l.es sho'.'ea no rel ati 'r.nshir to runil changes .

In another study, Good and Le'in (197r0) fCiled inr an attencrit

to hi:,' that "sen sit, Iers" would e .ert rore rerceorittual '.oilance. orer-

ational ied as cupil dilation than would "r'nre eser,"'. although there'

Jdd co:.rrol'tora te pre ,ious firidin. n on affect ve arousal nr,d puil 1 dila-

tion (c. '21).

More research shoirin' u puipil chrianes iliay t'e useful in clinical

programs was reported b ennedy (1971 ). This tud/ indicated that

pupii dilation coul1 t.":: used as an indicator of the therapeutic irrpact

of treltrcent on.r chronic alcoholics and ma, tbe useful for preJictinq,

r-c idl'is' (rccuirrence of need for treatrmnert).

Einrtional reactions arnd Pupil ize

In the 19l'6t stud discussed earlier. Hess also reported th.at

dilated pupil. in a phCtcgraph of a youn- 00orian cau'-ed grieter crefereice

for that picture than for orre where the woirin's puoilis uere of normal

size. Hicl.k er al., I'1'6 ) in an attempt to rerilicate Hess' findings,

reported that pupil s -Iz .howed- no rel rtion.hip to e'pre.sed attraction,

.lh-ilt facial angle did have ; ianificant effect (D. 3: ). Sim,.s houe,.er

foundL that nuri l 'i ze in pictures had a ;i nifi cant eff ,' t hern 5o... di f-

ferences w 'ere cont rolled. I' arrived suruiects shoiu d .ret.test ilii t ion to

cir.o:i te- -. oicturel tvi th larigrj pur i il and dilatean least to lit e-se,

r.hotograr.is ih 1 rgq urp ils (19 ", D. *~J. Data r, presented bty Etarr

and Jill is (19la.) also supported the dilati crn-.ittractiorn I,-D t.esii .

Ot'.er studies of thie r-elati onshir, bett e n oip'il siz- arnd trhe

ern'Otions w eIr- rer orted inr 19!'6. in a dirss rotation. Guin.-n demonstrated

that emc.tionrlit, (operationali-zed through slide presentations of

emotional wordsds ) did have significant ir.pa:t on puoil size, and that

a significant interaction between time arnd emotionality occurred.

Earlow (1969 and 1970) reported a positive correlaticrn brtteen

expressed preference for photos and punil dilation, arid found that

aversive stimuli did cause constriction of the pupil. In another study,

he found the puoil response to be in perfect agreement ,with ver:, al

preferences in a stud', of political candidate preferences (.arl.w.


The Hess (1965) article also reviewed some studies of the rupil

response as an inde of "motivation". In this study Hess and his as-

sociates d pri ved scme subjects of fond for four or f i e hours. The

mean pupil re.-ponse of this gr-c.u to pictures of food was two. and a half

tiTmei larger than the riean response of a control ,rouc, which had e ten

within an hour prior to t.he epcr;iments. whether or not th- terIn "moti-

vation" is appropriate in this crnte:'t, it does arcear that these data

support the He-s and Polt (1960) contention that r-uoil response is an

ind .e:' of interest .alue of visual stif.uli.

Hess and Polt (1966) presented data showgi n that the puoil ma, be

an indicator of taste sensitivity .nd taste orefF.rfence. Five different.

orange beverages, together w- th a control beverage watere) were Qi .en

to subjects. Pupil response was positively correlated %;ith other rank-

ings of beverage preference. Drjole,, and Lehr (196.7), however. criticized

the methodoloa i emolo;.,ed tv Hcss and Folt, noting that controls iornall,

errlo.,ed in gustatory research were not used. and trhat the results may

have therefore bnern biased. The, also noted that order effects coul:1

have t.i.-sed dat In a rcrily, Hcss arin Pelt (19(7i stated that they

hid irn.esti'ated order of ,resentatior as a source of artifact, but

had found no effects in isual. cifactor.i, auditory. or tactile stiTiu-

lation e'-perir enit .

liar -tinr? Pesearcn

]ri the early 19C0's larp.lar a marketingg research orgrniibTiori.

tbe:a;ai irnterest:-d in :omrii.ercial appl icitions of the HE :s pupil resr,.onsej

imeas.urirn ssteri.. The s.tujdie they conrikicted. usually i n conjunction,

with Hess, were rep rted by lIest (1'.2), 1rugrnr I'(194,) ard by Cnonsor

mai az2ine I196<' F briefly, findings \ere that puP il rrsrponre was r elated

to sales rpatterrns of the products t. sted (watches. r.eeting cards. sil 1.er-

wai r ) rid was also r-?lated to couponr returns or n ad..'ertisemrrtstr testes d.

These dat tended to supr'ort the tlhetsis th'at puil rrespunses anrd atti-

tudinal emotiornl resporise were related, ard led to the hope that fur-

thlr research could develoD a new- and useful marletinr tool.

The Marplan studies also sh hoed that the puril res.orins could be

usrd with mrovingq pictures as uell as still frame photo' A continuous

nmea-urcC~.-nt .of the punil, called the pucil rnterest track was recorded

while subjects watched tele, is ion ori .ercif al E/ n.ronito ,r'r ri,, il

dilatirn and ccstr iction. in ,estijq.:'tors w re able to dcet-rflir, ie where

interest pear s p er'e Yhi hest and i:rere intiere' t had waned. trateic

rpl.acement of the main sell in poirt, it war' hoped. could be achie'.ved

by locating the messae!p dirt-ctly alter the high point on the putiil

interest track:. Puril tra'cl a;lso allo.e-d the- makers of the commercial

to ev.'aluate their oroiduct's capaci t to ILeer, vi ere s 'wItchinq that

channel and that co.mmer! ial 'Sn'orsr.' 1'~ ).

As mentioned in Chapter 1, a further use of rpioililorretrics to

marketing researc.he-rs lies in its ability to ce.'ercc.ome social desire-

ability problems associated with t pical written n or c erlbal question-

ing procedure-s. Pupi change ..'a sho in tb t tir:'liarlan studies to h .--

a mathematical l higher relationsir, to sales than dc ver .l ind 1-ritten

sur.eys althc.uuh the differenc-s tire riot alw.Ays stLutistically iq-

nificant (.ruginan, p. 17)

Halpern described the p.otnti;l usefulr1ness of puni 11o.etric tech-

niqu:e in "bt foi'r:-and-,a:ter" iamar -etrin e:.rE rime-nts. He noted (o. 3201

that, often, oh!eri suti.lc:ts ar- calle d urorj, to iTimaLe "after" rIl- rornnt ,

their re plie are bt iased t.y rino.jlrI-d rni their "befoe" resoorises. The

interv iew and questic.nnai rr techniques conrriFonly usdii hy r,.ari etin j re-

searchers also corne under file as discussed in Cha.nte. r i. Halpe-rn

conducted studies shc,'hing that ptrpi 1 dilation was as "ncod" a resooner,

inid--. as w~ere ques tiionnaiirs for these. tFpes of e'neriimenrts, r .'iJ n rot

possessir, the b.i s and social desirahi li ty dra,,ba.: s inhe-rent in ou-s-

ti nna ire and inte,-rview tr-chnique;. Pull d l ti .r, sas a ain slo..wrn to

be- an indicator of interest in p.c': ar esiqn (n. 321).

A "questicorinir article b-, E1laci-ell, f-: al (190 sur.'eyed. the

state of Dupi ll et ric a.rirraiisal in r, irketir r s research, and noted that

the technique was t.b in, used tefor'e being ,:i 11 uirdei stoi-od. They stated

that there eist both "unfounded optimism" and "ill-timed slepti.:ism"

concerrning the cupilloreter_ ana its acnpilicationr and echoed the c.ill

of m.anr ir .est iator for- more research or, the technique anrd its use-


Dissenting Votes

V.hile these studies were proceedings publications in the field

*.if ophthalmniolo :y appe.ire. d whirh ha .e s rii ficarce for Pur ill o-etric

r'eAe rch. in par ticulr the ocr' s cf Lc.i er; te"nr and Low:enieldd i19%61

and 196.:) made considerahile cortr'ibution to 'no:,1eji e about Duoil func-

tioninil. sr.d sparc ed contro.r .ers, 'ovJer sone of tieiir fin ings. Their

discussion -of the "r.ipll11 5ry refle. dilationn. defined as an increase

irn pupil :1 r e caused t,' sen-.ry :ir er.-jtiran.l stimuli. or bY spoint.irneous

enritloC S :r thoughts. suppor' ts tre affect-,i il ti rr results oF.tainred

b!, Hess. t 'htlialrioologic and phisiological treatises on Fpu[il con-tric-

tion, ho'.e. er, a lm,-.:t unifornly, di sjairee with tie hess assertion that

conr tiiction is affect-rel.itr-d (Lr.,;enfeld, 19C.i). According to Uood-

mansee fi 965), pupillary. constriction can occur for an,' of three rtasr.ns

1. The liqht refle. , hereb' "'an increase of liaht
flu to tlth: retina results in flrow of efferent
impulses from the cciloiTOtotr nucleus to the sri'in. .-
ter muscle, thus acti.'ely co tr ictir. the rpu il."

2. The near-'.'is ion refle. wrhi-h occurs ,hern tle
e.e's shift fr'.,n fc.cusin, oin distant ier.t to
focusing on an otiect less far aw.;y

:.. The relation of the dilator muscle, cause-' ',' y
de.:rease in roual fllraina close on the. heel- of
a high-arou;, l rcfle. di ation condition (cag. 1.

It is this third phenonrenon which Lownifel d lnd oodr.enree hypnthesire

as a possible cause of the cortrictic.n measured by Hess.

Another type of criticism was contributed by lIunnrllv, et al. 1967,!.

They described a series of e>Deriments in which Dupil size 'as shorn to

be related to muscle tension, sound levels. affect, novelty of pictures,

and fear Ithreat of gunshot). They noted that punil si2e seems to be

directly relateW to fear through autononmic contamination (p 154), as

discussed in Chapter 1. The authors conclude. that puril response was

such a genEr-al ied reaction as tc te i alr..est without value for stud,inri

ccrrpl:. ercirtti .ro al stimuli arnd responses.

Summa r

The reccid presented awo: tends to obscure iA, patterns which

ray hrve ererge.d from the past decade of study. A summer, content analysi:

of these articles reviewed in this chapter is pr .enrted in Table 2-1.

Perhaps the mijcr conclusion that car be jdrein at this stage is that the

pupil does respond to stimuli. whether or not it is possible to accurate,

a:seN~ such things a2 attitude and eiroctional reactions is not yet :lear.

nldy studies have presented evidence cf ability to measure attitude_;

others, just as rigorously controlled, have found no consistent relay tiorn-

ship teteern attitudes/emotionw and the pupil re:r.,orse. Pefore the

question can be resolved, a much more precise rnethodolog,'v for measuring

the liariables cf inrteire .t must be developed, and many of the artifacts

nnir associated wiith thr punil response r.ust be catalog e.d and controlled.

The rest of this study is an attenpr to contribute tO these goals.

T/ALE 2-1



Eell, 1971
ElU1 cLwx ll, -t al ,1')70
Le.ersmra, et al., 1970
Eradshau. 1267
Bradshaw, 1'S'
Doole/ anrd Lehr, 196'
Fr;.ncic and Leile.,, 1963
Guin ia 19i' .,
Halpern, 1967
Hess, 1965
Hess. 1964a
H..ss, 6 1
H-ss arn- FPlt. 19640

Hes and l'alt, 1506.7
H -;s, er.i F'olt, 1967
Hess. et al.. 19'5S
Hir.s, t al 1967
Holmes, 19 7
i.ahii-rn ari [ tty, 1' C
S1 unniarn, 1964a
'ru gnar, 19A. 4b!
F irmel 19E
Landauer Feal rs. 19C'
Lo. rinfr l 19, 1;.
Wlurrnn ll at Il.. 1967
P3. ir, an; Eirri on_,. n i9.6
F:uL.i nr, 1961
Putin. et al ., 1?6.3
Si nrns, 196-
T rnci: and r,bt.i ns, 1970
A.m'wc,, ansr,.- 1965
Television 'lagazine, 1962
['. iness 0e9, 1t'6



Att t.ude Studies
Mental Activ.ity
A rous~ 1
Wa'ri t PeseTarch
Enotion .tudirs
Interest Va 31i Pe:Fcarch
la-te PFe-3erci
Conrli t i -n' i n
Po itr"/ F'ese-arch

1 nrpi C

10. 11


1, 1
2, 7 11

I, 2, _

1. 10
4, 10
2, 10, 1 15


7, 11

10, 11
2', 3, 6, 10, 15, 1r
2, 3, 5
I3 11

5., 1?
1. 10

Peasuriment TechniqiuL;
'!ethr.l l0 ,'
Age Differ-tnce Pec arch
utcnTrc 'i : E a 1 E n.:
FPS,'.hi .try P'e r h; -h
Fear PQsc-rch
Fhr sical '.':.rl Stu.li s
Personal i t tLudi;e



This chapter will define the obLectives of the dissertation

research, and will present the specific hypotheses which were tested.

Although they were tested in the null form, the hypctheses listed here

are in the form of expected results. Chapter IV will describe the

methodology of the e-perimEnt, and the collection and analysis of data.

Research Goals

1. The validation and reolication of earlier worl's sho..ing

th-t the pupil res;por, s differentiall, to various stiriul .

2. The developnrnt of a linear model describing the para-

.leters contribution to pupil chan.F-.

3. E.ramrination of several autonomic responses to favorable,

unfavorable, neutir-l, and aro lety-arousing 'tiniuli with

the objective of eliminating autonomic contamination as

a source of error in pupil response data.

4. Examination of an improved n.ethodc.log. for conducting

puFpillo mtric e>oeriments, including conpruter control

of stimulus presentation. analog to digital ccn'version

of ta. and real-tilme sariDling and measurement of the

puri 1 ary response.


5. Collection anri anari is of daai for testing the fol low-

i nj spec i i c I .I.oro.thest

H:,,'othes es

Purpilarv Feaj:rtion

Pleasa;r. stimuli i 1i1 be accor.irpanied b,' drilation, ruhile neutral

stimuli will cause no, chanqng in pupil i.'.

rNesati.e stimuli wi 1 craus a constricti c.n of the p.ji .

Arn ie-ty-' ou irg stirrull w;il di late the oupil.

Hi hlvy rpleasurr.e .: stiruli will orrduce gr ate r r Juni l di l.t i.on

than ill -lss r.ile sa-lit stinrjli.

Autiiorro:ic Conrt.erinat i on

Majni tude ':xv: r i c r-

Eiod pressure charngie will t- simi lar iri re.-ction c hen e.-nosed

to plei.ia a t stimul ii and whern -e posed1 t-, .a n..ie-ty-ar,outsing stir uli.

Heartr a e changes wnll t.e limi lar i when -e, p.ose to. o n..i ety,-a r olsirng

stimuli and pleasant stiir.uli.

St in poteri tit l changes ll t.- sil ilar in react *n r, %lhen e p.:ose

to .n. iet., -.irousiniq and t.: pleasant stilu1i.

Pescr'nse duration

rupillry. dilationr, to fear-arousing sttimuli inll tale loner to

return to initial base 1 .el (at time of stimulus DO-esent.tion) than

will pupil dilation to pleasant. stlnli, for a give.r lcvel of stimulus

intense ity.

Heartrat e will take longer to return to b.ae line in response to

an..iety-arousirng stirr.uli then to ple..ent stimuli. for the same stimulus


Skin rnotential ill take longer to retur to 'asp line in response

to an.iret.,'-arousir ri stirmuli tran to pleasant t i.uli. for the srame

Stimulus level.

blood pressure will take lrnuger to return to base line in response

to an., iety-arousin stir.iuli than to Dleasant :tirruli, for the samie P ieS el

of stimulus intensity .



E>oerimental Design

The stud, ..'as conducted under a randomnized 1c.ck design. Each

ut1bect sE.red as a blocl and received each of the 1I. treac~i rits (sti-

muli), 1i th the order, of F, reseritatiorn cUf tr'EatTl'en.ts random' i ed for E Lh

t.1 cl. The desi,-,n '..111 tbe discussed i n rather detail belw.


T..'enrt,-r inr persons sr-r.vc, as subj.ercts in the ': tudl Ali w re

volurteer- mal-e t-eteien ter he a ie t cf 20 arnd 40l, drawn ar laelly from ti'c.

studert .eroiA. i r the Colle e of Fuiinr,, A.dmi nir ti tion (the '1.. .

Club and Ni elt Sirra i1 ). They i rp ie pre -te ted throw. : h que: t i.rnnriire

to det'ermirne an,,- tr'i.c n r prir" ri-edical rprob'C lem1 heart trouble, ec"

deficienriLies etc. which mriht ha'.'e ffectrcd their resro.nses. Tho i

repoerti no. medical probi-rlrr1 werr .:Led to. ser'.'e a: ut iec t The

accc.impi a, yin fig ur-e Shr.ui the flow oCf the pr 'c. '.durI T the su'b ih- tS 'r:lt

thro nh.

At the t e ii of the s tud,, e..ch :.ublect': '.iual acuity r1as tteted

tc. deterniine his ability to di-i-criiriate dert il on li;de stirruli.

3irice one .o eie:t Cf the re'l earch was to de termirne the ft.Aia lillt., of



.-- L
C r

U U.

- 4 C'
.c r. sI-

u U
44' C- 4-*

L' 4--
L. C

Lu cl-p C
- O '.

LI rC'
al .:- -~
n >,

O r 4444
0T 4-' +-

4- di 1' IC

ci. I-.a

t5 C 3
0 C -- LI
U C (J

0%' 4-4 L',-" ^



C a' -

r-: U r--

In C-
C -C

4- 0' In
ra *-

C- Q.I i

a 3 c

a a-
Li CL *

*"1 C^ I


I'1. C
6- I 4-'
4a' C:

u" *' r.

i4 '2 In

IC n.
LiL- In

L. .
SL >,
LI r, Cr '
C. L1
L *u -

.L C,

y aj GI C
no 0 0

LJ 4 r
n .10 o

C.T 0 -

L L .


Ui aOC

> aI .
C' 4-' o
r- >, LJ 4-'

C i3 3C



s CLi
*r- 4 C

3u U 4'

0 4- *a
C. -1 rL

I n i

*-r C

C *r-
E "

-- 4-



0' I

LI I '0 4-' d'n
O rf re r I
w L E ,L

Cn I- c 3 i'

Q' C- E 3




4-' C
I 0O
4L L-

o .C C

4- a] C'

CU a-
n -n

w- :I r
-- C C0.

'r. r C

L. (. ar5 fc
.- ci -

0 a. c

,0 ('

E CL 4-- a

Ca 1 c c

Er' Cl 0 CT

I 2I

t Q 1 r-
r- 12

C 0 L.

'-diE a
(Jr ;F
-4- r- 0
if. LU II .. J




u:ingr p'Jupill oi etric. tecrhnirl-J s in f ,-Ield r, ,.arcli, tlIe '.._r i errl

:.controls nrern itirind abc. v .' re tlhe onl:; one.s u:d trc, iel- ct .uhject .

it w.as Ecsuined that close controls i.o uld hc. l.ac inn in fill r c' -.' arch,

ucina su ject s ;e lct:le from thi'. Inra: l .. t aul t Dc.aiultionr. .nrd thr r' frre

an, labor atory s -Lu usi ig ri,,rwir pec ific control wc uld not be o.nerjiali-

z2ble to thi' t.,pe of pcop.ul tion.

En. i roirrne nr

The st u idies re :onriduct ,:d i r thl', ./ t ,ri: E[l i ne-eri rl,'Hu.,lan

Perf'coriar, nc- L ;a or'ator-,- at the Un i *.er:ir .' c f F ir.' id Li oht 1-v',' 1

periF pher orund, and teirpera t ure tu 'er- all co tr.ii' ll d. The: sne to.,

persons (Lth- invc- nigt:r .arnd a labor' tor.r, technician) Conduct.'e all

r -sionr of the, e-perir ert. 'ccast ionnalll 1, other persons r .t in orn the

study .s; ob, r.'er:.


Figure -2. T ': E *ei 'T.nital nr ir nr.n trt


Equipment layout follow 1d a modification of the design used by

Hcs. (1)65. p. .)jI. Each subject wa: fitted with electrodes for measur-

ing hrartrate and sk in potential. H:c has cho,'n the measuring eouip-

ment, and the basic scope of the stud,' was described. The subject was

then seated in a dentist's chair, which allowed a certain degree of

frFedom, ihile restricting beod, mo.-ner'ent.

The subject looked into a rectangular viewing bo.' measuring

27 cm. by 32 cm. bt, 60 cm. The bc. was erquipor;i 1.1 th a rear-nroji ction

screen in the end opposite the subject. The walls, top, nnd bottom

of the bot were painted flat black to effectli.el, eliminate all out-

side light. A black cloth covered the boe and the sutiect during

each run of the study. The end of the be. at which there subjW-ct was

seated had a 10 t;, 13 cm. opening which circumscribed the subject's

hePd. A chin rest, used to mninimize head nio.ement during the e.'peri-

ment, via recessed in the open rnd of thr' b...

Ti'o lod,l. E tagraphic Model C50 H Carrosel slide piojectors,

located 25 cm. to the rear of the vie-, screen. projected slides on to

the screen. The size c'f the projected injae was reduced to 4 by 6

cm. to reduce eye search of the stimuli. The projectors i'ere connected

to a BR-330 hbryid computer which controlled the timing and presen-

tation of thf slides.


Fi gue ,4- I.. Equipr, irnt

DuI'iirIq th e .p rirner t, the l; iject :,c.re a :;t of dictattion head-

phc.nie:. The he.dr.hones, c.nrir, ect t r to t( r'e. ,'e- 'oil erIi t pe rccor'der',

prel- 'nted uni f rc m ihite (randei,) IoisE du.-ir,. the -: tud,'.

Just to the left of the sub:j-. t's l;in of :;.jht, a I'hittaler

rlorie l '0, u TV Fupi 1l Tmeter ITVP) iwas situated to tal e coritrtiniOLu I e.d inr s

of the si:e of the subject's left puidil The T;VP is a lc.;l d-: i rcui t

tele.'i Gr i s y s tern, ic i'cor3ip i .rp ti 'ig a hc, o ri, ta'l Es r.nrin r? techniq'.u; .rnd a

signal 'processor uliCrh measurr- arnI dir.pl:ays dupil d.liarmet-r. The TVP

u.s equipped ,ith a t .le '.'i i cn menit cr a-nd remote control s whi ci "l "l ed

the r periiTrert.:r to adjust focuii Ii so that r,,'nCr-r. setting; could he

found for each subject. The eui FIiieit operi'ted in the near inf ra--ed.

with a low' intensity r.ear infrared source illuminating the e.e. With

the exception of one subject, the light source produced no reported

discomfort in this study.

Figure 2--. Pecordin Equiprment

A Eec.kman Type Rl1 Pyn:graph direct rilng recc rdetr. capable rf

rmeasuring up to sit physiological variables simultaneously, was situated

near the maini equipment stand. Sensors on the subject were connected to

the recorder, r hich disoleaed readings on a continuously nmovirin chart and

simrultaneouily transmitted data to the c:,om'uter system described beloi.

Tne recorder contained a timer unit and e.ernt r.arket which were used to

Source: Whitta ter Corporation, Srace Sciences Di ',ision,
Instruction Manual for Series 300 TV Pupi 1 loeters.

4.I i~

note event chariejs such A: the in r' odur tion of rei stiuIli.

-h~ -C

Figure I-.. Th- E. periment rn Proi're' s

The co'n'puter sysl -ten'i or toreli th.- .peri"'nt n t cn.rtinuiOus Y arnd

tranr forrnid the data from dir-ect asr,.lr re-adout of the rph.ysilo, gical

'va riables to d ioit l form. u ing punch.ied parper tal :- a: 5ra output ri d uT:.

The Jdat c.n paper tape %ere th rn Ir arsfrred to r'-r i.et i tape r, l

analyzed us-,n the 1 'I 'CO*6'.5 cc.puter. hi th the .,*-:pti: on cf dat,

tran.rifer prograffmini n). there .wasi no. h'umran, Intr actc iorn w th the data until

after anal /.is lid b:-n completed. The rn a:urrement techninu~i- u-ed in

the ystt-em i.ere .vtaiidati. usii,.j "l licon" r. p:pornr':r Isuch a jartific ia

c.upils o.f 1rno,.n diameter) before the per irrent ,a: begun. Ch r-ses


were measured in tenths of milli eters on a scale of 1.0-10.0, so

the neasureenet system %as sensitive to changes as small as 1 percent

per second.


Stimuli for the studies reported in this dissertation ,ere a set

of blacl and white 35 mr- slides. A series of 58 slides ;?ps ldveloned

for pretestinr. Of these 5E., 16 were chosen for use in the study. In

the slide selection pretest, 1. raters iud'jed the slides using semantic

differentials d -elonred from the mrnifest anriet scai.le and from

evaluative scales used by Osgood, et al. (19'67). A cop cof the rating

questionnaire is included in ADDendi. I of this oaoer. The slides chosen

for the study were those e'.hibitirn high scores al3on .a specific d1rien-

sion and having lo. v'.riab ll it,. The direns:ions of interest .'er.e plea-

sant-unplesarnt, fear/,'n)iety, nd interest value. The slides chosen

are described in Table 4-1.

Scoring the nues.tionnaires

The set of semanrtic differentials shocrn in Appendi I was u.ed to

e-vluate each slide. The set was mriade rup of 2 sutbsets of scales, or


The pleasart-unp pleasant fac-tor was corpos'ed of the f,: lloWin-g

semantic differentials: Disiusted-elated
Gloom- h-he erful
Ho-tt ile-friendr l
Unp l eas3snt-r Fleas. nt
Uinfa .ora.l e- fa.orble
rNegati've-posi ti .'.


Table 4-1

Affect ve Faitinriqs ('linden.rrr..'ent Judgres) d i. Lurri nos it, cores of Slides

s'tiriulus Criiitent

i. Picture, rnarl in g [DC
?. Picture, Defori ed Eat..'
. Pi cture, Ferret (aniiri.i I
4. Picture, Diefor. :-d EaL,;
5. Picture, Err aci.t-d Child
6. Picture [leforC ed [. I .
7 Li rin D rawri';n T r .ia r l -
S. Chiarc. al C letch. Girl
9. Ficture, C hair
0. Charcoal Eletch, [;udP
1. F i.tur Srnilin i Girl
Line Drawinq. Ebl rcl
L. Linri-e bra wi r P-ct nrial
14. Picture, Child n Pain
5. Ficture [I,%iig fri Han
L.. Line Drawi ri 'oir, ii

Controi Slides

C' bla rl
(2 Ela r
C: Llanl
C4 Elanrl.
Cs Ll anl.

Judoed Affect"
Fear Fle3as ntness


17.. n


2-0. 1
:''i,. rI


51. 1


21 .1


24. d

Pre.dict J
Ps r, nose

I.pl a Fle sant
Unpl 1 a t
Ir npleasant
I.'eu .ra ,

Pl eas.ant
PFleasa, t

r;eutra 1
I.i nr. 1 e a ,3 Iit
hipl easa I.it

Luminosi ,i .
rr.'f r ai l
'irih tness'






[,L-utrl 1

range PF.;s ibil i 1ie: f 'r Pe.cnrse Cat.- r.riL : Fear: 5-35 u .tr -I 1 0.0 ;
P leas.ntrn : 12, fleutra 1 =." ;: in .th :a e:or.rc ies, e hi h sc'C-re im-r'i I,
relat'.i PFeac ti r.s i .e., Hi h F'-r ard Ulrr]e.s. rtness, r- rect i'. ..'.
' a, ured irith '.-strn Li hit fl-T.er rnd trari f r-d tc. d oilta.l forr-. The
r aci e '". t,;, -72'- cor- c.ric.dr. ar, r.ro. riat.l tc t e ,.rig,:. 600u tc.
*'5 fuoC't.n.ld'le: re'pectt ivc- ...

Since each scale had a rarrng of from 1 to 7, the possible range of

scores was 6-42. If a person had chec-.ed all scales at the e.,treme "plea-

sant" side of the scale, the score vjlue for the factor would be 6 Con-

versel; checking all the scales at the negati.e end w..-ould have produced a

factor score of 42. Checking all scales at the "neutral" midpoint 'alue

would have produced a score of 24. The middle roint appears to be a true

"neutral" position on the pleasant-urncleasarit dimension.

The arii et\-arousin factor was corCposed of the semantic differentials:
Nervous-ca Irn
Tense- r.elyed
The lowest possible an,ipty factor score for a slide was ind tie

hi hest an. iety factC,'o value as :... Unlike the r easant-unpleasant fact',r,

it appears that the an-iet, scale does not yield a tru'- neutral Fpo, iti on, if

by neutral ire mean lack of anxiety. Father, this scale appears to produce a

step-score, ii th degree of an, i .ty increase iig as the score incrir.ases.

A single ser.antic di fferential interestt 1i -uniritere tingI) as used

to ePaluate slides along the "inteEst value" dimension.

To ensure that the groupings of slides used as independent var'iabll

were indeed different from each other, the medians test i:as cc.omutied for each

d:,ad of grouping, using subjects' que tionnairr- data. The groups in each

dyad were si nificantl.,' different from each other at the .01 level ifor the

pleasant-unplea; ant dyad, "' 131 .a9; for the pleasant-neutral d.ad, '*

13i .49; for the plea ant-neutral d), d, *'=1 .31; for the uri e lcasant-rieutr,'l

d/.d, Y. 1F.-1.9; each d.ad tested m: th 1 degree .f frc.edcni:m

After the 1 stimulus slides ware settled o., Kodak 2'91 G,.elatin

neutral Density Filters were 5ddJd to each slid as neeede,, to de.'elop

closely balanCnd l i' It intensity tes. All sl des and control slide cs used

in the study had the sarm intensity, wi thin limits so that any pupill r:.

resrons; occurring could not tbe construct. as a light refle:,. Table 4-1

alsc presented the range of scrarbjltea liht intensities for the set of


St imr lus Presentation

A hlani control slide waeJ projected or the screen while the suh-

ject was being seated in the chair and adjusted to the equipment. .As

soon as all sensors were reading oroperli, a two-mirnute stud, -imilar to

the main stud, iuas conducted tD fariliarize the subject wi.itl the equiJ'-

ment, rirling, and renerral en.:irrornent he would be in during the study.

At the end of toe riinut-s. he was.allo,.,ed to rpet and was asked ahout his

ccmfort. This time .also allowed the erpei-rimenters to male any necessary

're.dju..tments of the eqluipm ent. The subject was then cautioned about

movinr, his head t..c from the headrest. HP uas: 1: asked.. to contair

suallow.irrn and ,most -yeblinkirn to the tim, when control li des were on

the screen.,' since these types of mrveicnts introduced errors into the

lreasurrenrt Syster. He was inforrrmid that, at sonre tine during the eperi-

ment, he night lhe r a buzzer, and that it vas nart of the e..peiriment.

He Lwas cautioned not to io=e ri'aw from the equipment at the :soundr f the


Tle subject was then aslied to readjust his head into the chin rest,

the camera focus nas checled, a black cloth was pl1ac:d c' er the subject

and over the equipment, and the main e-.pr.-iner mnt was begun. A bl nr.

slide wa: p'rojectc;d on the :.:reen during, tle period described abor'.

The bianl. lide remained on the scr'cn for a period of eighteen seconds

after thie cor puter had tc-n gi.en the. instruction to begin the e. Ieri-

mient. Once gien that instruction, the comr.uter controlle-d both

slide projectors,. and operated them so that no light flash occurred during

chanrge-ov'er from one slide to the ne t. To accomplish this, the ronoutrer

was prograrmmd to start chanqinm the slide to be projected one half sec-

ond before it instructed the other slide projector to remove the s'lij

currently bteinog projected. E.'ery other lide in each project cor w .as a

black slide (one allowing ri. light to Das through) so that vhile one

projector was shioiing a ttiniili;: r- control slide, the other wa sh iowinil

a bltcl lide. This "flip-filop" process was de.-elcoped to allow the re-

duction of thr. light flash nenticried above to a period of les: than 1/20


Each stimulus slide 'wa represented for 10 seconds and was fr'ilolced

ty a blanrl crnt1rol sli de priesentcd for 20 seconds It \is hoped that the

20-'econd hlani side would all c't. response variables to return, to riormal

levels t.bfore the ne-:t stirmuiluS uss presr-nted. The crder of stimulus

pre-sentatior w'as randomly '.aried to control for order effect Table 4-2

presents the order of each stimulus presentation

The slide runs too'. 4IO secondss i,2 slides, half for 10 s"-cornids ecch

and half for 20 seconds). At the end of this time, the subjectt ,6as given

a "startle" stimulus conisting of an alarm buzzer soundino. The brz:er

was- attached to the wood bottom of the arm rest used bL the subject, arnd

produced both a startling sound and a simultaneous. startlirig .ibr-tion in

the arm rest. Appr-. irr.ately 20 seconds later, t ihe investigator' pushed

the "reledC'e" Switch on th', dentist chair. causing it to mr.'ve very, rapidly.

doinuard, .r'oducing a second startle resLCon:-e. Because of the raIid down-


TV>le 4-2
Order of Stimulus Presrent.atic'r,

Sub- Stimulu.: u uniber
1iS 07 01 06 0~ : 11 14 12 16 02 n3 1 15 10 05 04 N:; 17 1I
JR 02 04 15 12 01 0- 16 13 14 03 08 06 05 11 10 01 17 15
E, 04 i0 02 09 11 15 07 06 16 10 13 03 15 12 14 01 17 18
.H 11 03 05 16 01 07 14 o0 09 0S 10 04 15 1I 023 13 17 1-
TH 16 09 02' 05 13 10 15 03 08 12 01 04 O' i0 14 11 17 1:
Ofr 16 05 10 15 03 04 12 11 02 14 01 12 07 0'? 08 06 17 15
1S 0i 02 05 10 04 0I 16 06 j03 14 12 13 01 15 0' 11 17 13
E.H 1 05 02 16 12 11 01 0O 10 11 09 15 '3 04 06 07 17 18
JP 01 13 04 03 11 02 iE6 10 12 14 O0 05 09 16 15 07 17 18
EG 06 05 0: 04 13 02 07 12 14 09 11 16 10 15 01 03 17 18
lib 16 13 02 15 11 03 12 04 05 06 01 0' 02 14 10 1il 17 13
FN 14 11 13. 0. 05 03 0' 04 10 15 06 01 09 11 16. 02 1? 1'
E. 10 07 02 13 15 16 09 01 1 OC' 01 14 03 06 05 11 17 10
TD 01 0. 15 13 Oi 16 0; 0 i 0. 11 02 10 14 12 03 01 17 1S:
DP 1i 04 07 l0 n9 16 03 05 13 O. 12 06 10 11 14 02 17 1S
V 10 1 04 15 11 02 02 01 06 05 03 12 13 16 17 18
l)4 07 I6 11 04 13 10 01 03 14 0r: 05 02 09 16 15 12 17 1 i
oD 16 03 06 02 05 12 07 09 14 02 15 01 11 04 10 13 17 1
PS 01 03 04 0' 14 13 10 05 15 02 0.1 16 06 11 11 01 17 18
ET 15 07 01 05 03 12 10 16 01 04 13 019 03 14 11 06 1' 1;
JU 0" 13 09 12 14 0. 06 02 15 11 16 O; I 03 01 04 17 18
P.P 13 16 15 03 07 14 09 06 05 0 12 01 04 10 02 11 17 12
\.. 13 16 15 03 07 14 09 06 05 0. ; 12 01 04 10 C02 11 17 1
Cil 01 05 11 04 09 10 15 C0 03 02 16 06 12 13 1J 07 1: 1I:
IU:' 09' 10 16 0S 04 02 12 15 06 05 03 01 07 04 11 13 17 13
F:L 01 11 05 r0 15 13 04 0? 16 02 07 03 14 10 (6 12 17 12
L I2 05 13 16 02 10 0D 03 01 15 06 04 07 14 11 09 17 12
PE 10 16 13 OE 14 02 12 15 06 05 03 01 07 04 11 09 1' 1;
SE: 11 09 15 04 A0 12 14 16 01 03 06 10 i. 12 05 07 17 15

ward movement, the subject could not maintain his head in the chin rest,

so pun;l data w'as not collected for this second startle response. ThPsn

startle stimuli always occurred at the end of the set of slides, and weir-

always poesentej in the same order.

Data Collection

Method and Pesoonse Variables

One hour was assigned for processing each subject. Upon entering

the Systems Lab, the silubect was introduced to the e'per'imenters and w'as

shon the equipEn.nt. The type of data to be- collected w',.s discussed in

general terms. He was told that there uould be time after the e:,peri-

mernt to .*ie\ the equipui nt in more detail and to see the charted results

of his study Aft.:-r filling out the pretest questionnaire, electrodes

,.ere attached to him and he w'.as seated in the dentist chair. The elec-

trodes were in turn attached to the Dynogra.pn recr.rder. A light-sensisn

pressure clip for measure g iq stol1ic blood pressure was attached to the

subject': right car. The subject as as'ed to place his head in the chin

rest and acclimiate himself to the efpe irr.mntal apparatus.

While the subject w'as aciclinating himself, the e-'perimenters were

adjusting the in;trumentc and focusing the camera on the subject's left

pupil.2 As son as all instruments seemed to be properly functioning,

'An interesting problem occurred when cne subject noted that h-e as
blind in the left e,E. The ca-eira was focused rn the ioght 1 -'e and no
problems occuLrrd as a result of this mir dification.

-4 -

the subj.ie-t twas 11 '..:ed to ,rel a for a minute. He ia then as ed to re-

adjust himself, i.r the 2-minute famirili ariration run was t.epun. After

thi: rur, tne stutiec:t %,'a al 1:wed to rele.' for' about 1 minute. During

this minute, the c i'arusLrlt wear: rotated so that the stir'uluc and conritr'cl

slides sere in pr .sition for the be ilrrinini rof the main E tid.,'. The tape

recorder ia. recy'ced, and all equip',rent '.'as checl.ed for proper func-

tioning The main stud,- fol lc.ed this rest period.

Le.rel of le1 :,ur'enri t

while e slides rere being piresentel, continuous line chart recordings

e,.ee being made on the i,'ynograpli reco,'rder. The Dy:.rogr-aprh had a recordings

cpa;c ity of 6 criannel.. In crder from top to: btttorn, f the reLcrdir n p-per,

the si> .ariables rec:or.'ed ,ere: beat-b,-tbeat heartrte; 60-second he.rt-

r.5 te (e~trap.ljsated each tecrrn ); r.upi l chance: sl.in potential ; blood

pressure; and a light i rten-s i ty r: adCiin for' each slide, t ien by a Wes ton

r'M del 7-1 li iht meter ins tal led insi je the e le irTig t.i:.

The EF-330 computer c :amrpled the corntininuru dat a tt a rate of 1 sail-

ple per second for e;cin of the .ariab le: being recorded, thus pro'.idi n

a digital sample. E'ea t- h ..-beat heartrate w3s nrot saiipled by the coTimpute r'.

A total .,f El ,5.0 .iniples was c.ll ecte, fc.r each subj.. t (512 samples fcor

each of 5 response ,riables ). These date er,'e stored intert ral l, in the

computer until tin- study had beeri completed, nr- were then transferred to

punrchej tape. Later, the dati iere again trarnsferrd this' t ie to ma -

netic: tarn i. al idationi chec:l s ..,'wre made for each data reducti in con-

.'ersion, ensurinri. that error: in tr.ansfer i-ere not maide. Priimar,

data analyvsi w3a done on the discrete data contained on

the nm-inetic tane. The punched tape data and chart data tiere used as

bacd-up data sources. anJ were not subjected to much systematic analysis.

Post-Test Cue:tiornai re

In an attempt to develop convergent ..alidation rf the autonomic

responses (i.e.. evaluate the character of each individual's autonomic

responses from another tyrie of measure) each subject uas as.ed to com-

plete 3 qcuestiornai re ratirin the slides he :iered during the first part

of the e'pe: iiment. COopy of the questionnaire is included in i ,.rlendi..

of the dissertation. This quertionrnire was used a: a major cc'riparison

inde., of affective responses to the slides. The coi relation between this

iide. and the pretest rat iing riade by irndeperndert .judges a.' al so computed.

Debriefino and Eqjuinrmerit F:ecyclinq

Since subjectss iwere members of two student orgariza tions, the pos-

sibility of discussion of the e-periment prior to another subject's parti-

cipation existed. Subject: .'ere aseiJ rnot to discuss the e..per'iment with

anyone, and were told that the irnestigator planned to give a debriefinrg

and suimmiary of the results to the group later in the quarter. Each ut..-

ject was shon the chart output recorded dur-in his r articiration, and

the responses w'ere explained to him.

While the subject w-is l filling out tie post-test questionnaire, the

.e Prinenters were trarsferring the data to punched tape, rear'rangirn

stimuli according to new ranj. nrumbter sequences' resettirng the ccmiruter

and the rest of the equin.nent, and ,reparing for the next subject.

Clata Tra sfer

lihC o'r; utei'r-G I.hr-red de7ita jSir pe .- tra-n; firredi frcrn punched

a i',per t.-,e to ,ir netic tape and th- r to cunched cari ;. Since thi- e.-

[,eri,] 'i nt 'a: a pi lot s tu.-l, n-r, computer :r r'- in r a gr- t arn.:unr, of t in-

W.-5: cori'currd i ri the do.'e .:r-r'r. t r Cf Fil-'T[.r 'IJ and rrnchire ar. iruag q [.r- o.pra srr,,

ftr con','..rrtirn and .1 clidatinri data gathered in c tal (t.r e ') fori onr

the pap:r tar.e to hr-.-,deciiral (base i6) form,- c, tih rniagretic ..:are .:nr

their to. di:ciril (base 1I0) for'i' for arnac ly i a r,i ,re:. ent ti.rn. The;e

conr,'. rsi cr; in'vol d the ilch n I tap e[ t tc-t e or.-rte r t-he IET 1 F 20

ar,.. t:m B.,'E cc u tr rs. Th=j con erir rion froon djiff..rert t:.b : I.

rnadi neceas ar,. bec:u: of t'- uti i ati crn of th'e F.P'-'.0 .c'r.Orirute i and r i .

date- r c:r.r.liir, -,.t- If I-,al tim .a s m;-lii l.i t I"1 L 'i r-m.-rt and

direct a.a'r,-tic tape or-" punch.-d card oultput had h--en fra:ibil for th;s

"tl.o,. their c c.r ers ion pr-,ce:. ,.,u'ld h3a' ie b en 1es s bOthe,-t:r '.

Da tr.anrsfe :.-rd e,.ilucticr. pr'-in ramTil ir ng tool' limic,:t 2 ri.crrtris to

conp '. :t. llo, th.at tht cor,'.ersion padcla3rt have been "det-uajr.d" arnd

.a1 idated, a cor.i et data tranr'fr-r fro 'canir "on- ii ne sann i of 30

.u.bji:ct duij inn1 an e.ri.?r-irirt to decimal da ta ready: fcr sta st itistica

arn ly 1 i can he accomnpl i -hed -in % thian I hc.ur:. I f the '1 stteiT c "Fl) 'C'

hd t*c i-l u'; -d, Cor if a direct ti-i tie-n to thie 3 0'r.E tnro:ti h s.ome t',:

Cf ter.miinr;i l inl a d'. icr- ,r'e u:ed. data co.l d have t.,.r, tranisf, ,red

to a ft'.r r.Iead, for ana'l, i rr..atter of :ieconls fT0roir, the ti, .- the

dite c: llecticn l t :. c:l.c.l ete The id.arnta,:-s ,and dise dv tanta. oe: of such,

t real-tiri *..:ttm, r.ra-ther wl th SoLne co:t -benefit data, Hili be di -

CU -I d i rn a later r- tiC:.n.


At each stage of the data con.'ersion process, output data wrre'

compared wiith chart data recorded h:, the Dynogreph recorder to ensur'-

that errors had not b-en made. Several programming errors stemm&ing frori

reversed polarity on the D:nco.gr"phr recorderE*'FP-3 0 irterfacing ineire

detected u.ing this te-chnique. In mri:t caies. these errors r w: re corr-

rectablr- with minor modifications in the con. r' ion prorams.

Data Peduction

Sei.eral forms of data ;rF. ?naijl;'zd. In almost all cases, the

10-second tin:me srpa during whicl each stinrulu w..'~S presented formed the

basic period for analysis. Sincr the EP-3j30 served as both a director

of the ei per-inrnt (chang.in slides at iven tine inter-'al :1 and as a

data recorder, it could not tale sanples during thle second it was chinrg-

ing slides. Co.rsequentl 1, no data sample was talern durin;I the first

second aft r'ia slide change had occurred. This left a 9-s cond aminie

of each .'ari.ble when a stimululidE. sid was r'esrnted, and a 19-second

samplee whien control .lides were presented.

T\:o ba5iC mrealluremenr'j. uere used in drfirni n tries of r-5so.nsc..

To t-st hy.,'potheFes cC n reaionitu-dea the first measure rcc'paed absolute

mrna.imumr during the relevant response period with absolute rn.a..imlums dur-

irng the prior control period. Second, percent change measure ue're in-

corpLorated. Using pupil size as an e;arimol,

percent chi; ge in puil si;e pip =_ -a c.a
for stimulus on pr-esen'ttion a FD

wlertf PLI = rean Fupil diamr.eter during pre'sontationr
.a a of c ti ulus ind

PD : = mean pupil diameter during control pr"ic.d
cs.a preceedi n i sti u Ilus r, o rn ir'F entatio rn .

Peylre: .i n c:.r ffi .ients I c-re used to test di fferenice:, i n di rc tion c.f

responses, js Cdicussed in later sections.

Two tir.le periods were ral nezed for each stinulu. A 3-sEccnd

tiTe. span ias!. an alyzed because it was felt that the l.ugjerl time periods

ccn'.er tnion 1i1 / used Iight be losing a subs tantial r'r.rt rion of the rele-

v rait Ipup il r'es Cpose du! to a "da iping" of the r'espn;:.rnrs o.r tihe 1c.rn

per iod. Inr this; cat data collected Iii the first 3 seconds of each

stimulus data simple ,were corinpai'ed with dIta :olle:ted in the last 3

seconds pFccCedi rig the s e.rple. The other tin., rieiod in.c1'.ed a 9-ecc nd

analn si., he re data CC ll iected ir the 9-se:cond rie-iod fo'il l i ng a stim;u-

lus jr r'e corparr.-d ith tl data c.: elected in the r? -'.cornd prior tor the

int reduction iof the stiniulus.

Peduc t 1 n iri .,io-r. I e Size

In s'*.er:.1 c ses during the study. one or mnorc, probilesr,, (.CCurrod

in the d ta scni-r.li rg te:hnioues. Several subjects for e.* mple, ad

er.teliri rates which c ..aused murcl oct the data simple to be errorne'.us. In

a fe' other cases, c blod pressure nrd skin potential recording Y,- ste-m

mralfuri..n:ti c'ned durj ing the e-.perimrr.-nt. Causi rl dita cn the:os.e vir'i, bles to

be lost. Consequentl., for ro,. t ,of the anal.":-ss conrduc'teI. the salr le

size i.iw s reduced to .2?. and the nur.iter, of ot-ser,.ati onrs i .iged from'TI 270 to

-17 per ,ariabl:. These will e t noted. in the apiropri ate pla.ces in results

discussed t el C.



As discussed in Chapter IV, the e.peririrtnt incorporated a randomized

block design, with 2'i -ubjects serviri as blocks and slide stiruli

seririg as randomized treatments. Most Ocf the analyses discussed in

this chapter utilized all 18 stimuli (16 slides and 2 "startle" stimuli)

as treatments. Case: where the 18 were collapsed into C treatment groups

(pleasant. unpleasant. neutral, and fear-arousing) iill be discussed


brnal/sis of 'Puoillar; Feactions

A Linear tlodel of Pu ril Pcsronse

In describing the pupil response to i g~.en stimulus cf the typr

used in this disser tation, the folloininrg 1 near rimodel was developed.

PP' = E + 6 .. t + te 't + i '' + + + E + .

where PP = change in cupi 1 size
= regression coefficients
= the initial eciue of the pupil (base level;
= rrrnvircnmr.ental sound cha.ries
: environmental light changes

v, = the orienting refle' natterns of the pupil
*. = color netterns in the visual stirulus
'; = cormple ity of the stimulus
= interest value of the stimulus
pl. e asantnesEunplessantress of the stimulus
s o t liig t intensities in the .'isual s.tirulus
S= C: ramrrblted l ight intensity of the stir'ulu:u
:I 1 endogenr-us subject 'varla.' les
c = error tern

s ca.n be seen. the parameters of this ter,' simplistic model

ermrihasi_: that n rny things contribute to gi'.ven chsinrg in Pupil size.

The model served as a guile for cortitrol' used in the e perirent and for

the de'eloprmernt of tne set of linear predictors used in the statistical

anal., st. Leti. cConitruls anid the utili action of riore accurate predictors

should have caused an increase in the anount of ex,:pained v'ariance in

thr dependent .ariable.

Tre rn::del alsoit enphasi es some firidinigs which are not a51i5ys ,con-

sidered in e.r rerirent.al deMigns and analyses. lost researchers for

e-ample, have not ta.ern ideo3uate account of Wilder's Law of Initial

Val'uies, ,rii ch states that a variable's "...respr'-nse to stirriul tion is a

function :Af tried pre-stimulus ie'v.l..." at which the .aria.ble was oper-

atinri (SterAbach, p. 44). itien usirn percent charge as 5 dependent

variable, an error is introduced rhe-n this la, is nor t colnsid' red. For

e:amp-le, pupil '*ith a diarneter of 7.5 nillimetern has little cnotential

for large dilationr (the familiar .eiirin effect I iile it has a much

higher propensity to constrict in si:e .and return to r.lore nornral

level of operation. A 2 percent dilation, therefore, might be miu.:h

nmcre significant at this level than 10 percent dilatic.n from .3 pupil

which started out at a diaretor of 3 millimeters. Ey using initial

value as a predictor, and incorporating analysis of cvrarirnce, the

m.del can factorr out" of treatment effects that oart caused by initial

values. Sini larily. anx.irety score'. can hbe facto'red outi by cCv'ari -nce

whenr the effects of plea.santness-unpleasar.tress as a a redictor are to

be inresti 3ate. d.

The specific model used tr. test the ;..-er-iriaental date wa s:

PP = + ,' + +' + / + / + ). ++ + + t
S I 2 3 L 5 5 ? 3 .

+ + + + + + ; +
I 1l II I1 2 1 : i 1 ]L I -

+ i i. + 5 .'.
Jo ]

E.4 ;

+ I ...
I' 1

+ '
d/ dl

. + 1 +
1+ 1B IS 1 9 .

+ Y Y + 6 / + f /
2' 2'1. 7: 2: 24; 6

+I c. + + X + ,V +
267 .. 2' t: ; 32 i 3 1 3 1i

4 a 3 i' -+ +

1,here PP. = 3 and 9 second r'ri'cnt change in pupil dirimeter

,-, = 0

r. =: regression coefficients

. ....X were parameters described in Table 5-1

F = an error term.

Table 5-1
Predicti n Varw blcs UsL-d

Varia ble
Number De criptior, Tyoe
1 3-second pupil percent cnr.-ce dependent
2 9-seconr pupil percent crsn.l dependent
3 initial value moderatc.r
4 slide 1 prCd' ctc.r
5 slide 2 pr'edict.or
6 slide 3 predictor
7 slide 4 predictor
o sli'.: 5 predi tor
9 slide 6 predictcr
10 s idee 7 rp redictor
11 slide predictor
12 slide 9 pr-dictor
13 l ide lu predictor
14 slide 11 predictor
15 slide 11 predic-tor
16 slide 13 predictor
17 slide 14 pre..Iicito
10 slide 15 predi- t.r
19 s ide 1t pr-edictor
20 3-s.,o:,d av.e age-- -lig t i rtensi ty moier- jtor
21 '-secorid average---light i rnteniit ; niodera t'r
i'2 questionnaire score---c.l esratr,-ss predictor
quest on i rre sc,:.re- --interest ,a lu predi ctor
24 interact tion---light initial '.alue pr. dict:,'r
t25 inter.action---irter e.t V'.l. + f ''. pre i.:to:r
26. interaction--- ir, te est '.al. + i'.' e i t
27 interact 'or.---1 light iriten sity ['redic: tor
2 irterectcin----i f.~ predict:or

TablE E-I .orI t.)

Lumber ['escri pt ion Tyr.e

29 interection---1-v Interest ,.alu' predictor
30 intera.:ticn---interest value lint predictor
31 interaction---iv fav oredictcr
32 interaction--- '.' interest value predictor
33 i nteracti on--- 1 iiht i nten.i'v predictor
34 inateraction--- iv. 'f; predictor
35 i nter.action--- i v intere- t value predictor
36 questionnaire score---irni et., mFod-erator

Several r'cstr'ictionrs are inherent ir the li-nealr niodel: -ipproach

cdcptcd above. First, th, assu;,, i 'p ion of li ne r'ri t,' and linear effei ct.,

ma:, not be a good appro.li1, n nation of re l i ty fCor ph,sioloc ic, al r'eI .porn.e.

Second. as can be seen in Table 5-1, ana lysis ,ras limit ted tOc the fir'st-

order interactions of th'r mrdel parameters. It doer. appear, ho'-,ever,

that the d-ta fit the par.amnetric requiiem.-nts nrecessar5 fcor the employment

of the statti: c.c l techniiue1 de.criihed below. t' noted earlier, of

rarticiilir concern ,'as the time period btet.,een r ides, since the e' e'l-

mnenter vwi.hej to ne?.ate the impact cf the preiviIout stimulus r' lid: on

the response to the ne t-o.curir'in slide. To test for serial effects,

the Durbin-.'atson te-t for auto-co relation tiree .S.erics in the dati'

ias utilized. Tre Durbin-iat'ion statistitc r ringed from 1.92 to 2 .'0' for

ill of the data used in the analysis' irinlifatring that the 20-second time

period betjee n il de succeeded in el iin itirg the effects of the previous.

re ponse.

Find ingj

A :Spearman r.nl:-c-rder corril tion ('fie:jgel 19-6) ) between the oues-

ti:onirr ire scC.res of i nde-pendent judges and subt.ect's cues tionria ire

s:ore: ,:' '. .orrinute,.l in order to dete,7rin e i the aIreemert t :'. to lide rank:

te tet. r een t u le arn the uu ee ts in the .tIrujy. Tie pleajsantners;-

unple asantn ii;e score i ts the va riable in r.e tii .atei d.

The ub.iects .and piretestt judges tended to aqree -l.,out the emotional

corntert of the sl id-es. ieldinr a correiationi coeif icient ,.f c0..1 (sig-

nificant it th- .[S i', l ). The ranr order of the ;lides in the r,-e-

te-t (indererndt jud-rs grrur, nr.er di offered ty nore than 1 po.iticn

fiorr thi ra.rJ or.:jer of the slide: ir, the post-e.. r rimert.31 (1 ut jie: ts)

nrouL p. Iith o-ire e-cepti or, in the rieu tr Cl :.cat 'Lor the r lides. alo n:s ir,-

tained the same ranl: when cnimpaired by rmdi a3n (nor~noriamietric.) or by the

:era. e: reported in Tatble 5-1. It c.ri t.e inferril frmc-n the hiih degrPe

of inter-rater relitability implied thu t a t,'ce of cr',ss-..alij ticn of

:.timilur tt;, re h. b een .cciomrplishEd.

T. ble ,-2 pre--rit: 3 summary of n41ls. i of i aiarice for the A

*lren r'al qroup.:, of :1 ide: u.ed in the study. The differences here are

alo si'jnifican at the .0IJ le.-el of cornfid-rnce. As .ill t. sr'een later.

how-ever, irid;.. idual slide differences, when jud:e, d ti the deperndenrt ari-

tlble;. do not corfrrm to tho result, hypothes.i:;d or to the re-ult: re-

ported by proponent: of pupillomrt'ic techniqlue-.

Table 5-2

Analysis of Variance of Ouestiornnaire

Pesults for 4 Groups of Slides

Sum of Squares D F llean Square F F:atio

Between Groups 1130.27J2 3 3'6. 7"E.:1 3.3556

Within Grcoups 1lS6:33.04F 1F6 112.2774

Total 197 8.3203 1E9

Tables 5-3 arnd 5-4 present tsumrner, statistics for stepu1ise re res-

sion lth "-second pupil chneri as the de-pendert variable. As car be

seen, the rmoJel has a multiple correltionr coefficient of 0.61, eoplain-

ing 37 percent of t'he variatbility in the dependent :arible Ar analy:1i

of the correlation coefficients associated with the independent and rioder-

ator variables yields rcn consistent relationship betu:rn the 4 stiraulus

types. 1'hEn anal/;ed separately, the startle stimulus was the only one

causing a consistent dilation in the pucil. It doe' aopear that jnalety-

arousing stim-ul do cause significantly greater dilitioris than do Other

tpPs of stimuli. :;o consistent findiniqg were associated with pleasant.

neutral, or unpleasant -tinuli, as the titles shjo .

The 3-seconj time period, used inr the present analysis, may cause

part of the confusinri results obtained. As toodmarnsee (1965) has noted,

the oriierin i and 'lear-'.'ision refle:es of the pupil cause a constriction

in size. It ma; be that subiects allo ed their eyes to lose focus during

Ta.le 5-3

Last Step of multiple Step-'lise F:egression

Multiple F 0.6002

Std. Error of Est. 10.1773

Analysis of Varicnrce

F'eoressi c'n
PF sidual

Sum of 'qu.res

16665. .'3

[IHean Squarc F Patic

793.612 6.9 1

Va\ iatbles in Equ.ation

Std. Error F to Pemo.e

1 Fr. 79

?7 .2-1
71 .92
7 .277

r0.6 7

0 02


O. 9i0

0. 41



C0. 4


(lons t-ntt


1 C.




-4.1 ..0
-O.S. 76


54. _6



U. -..


Table 5-4

Stcpwiise Linear F'eqression
3-Second Percent Change in Pur il Size

Step Varlable Multiple
Lumber Entered F' F'S
1 2 C0.4A6 0.149
2 12 0.490 0.24.1
16 0.523 0.274
4 24 0. r. 3r'i
5. 10 0.55: 0.312
S.4 0.570 0.3?5
7 1 0.5. 52 0.339
9 0.57j 0.351
10 11 0. J7 II..57
11 19 0.599 0.359
12 34 0 600 0. F.1
13 3u 0.6;2 0.363
14 17 0.603 0.36.4
15 36 C0.03 0.3
16 11 0.6'. 0. 3:
17 reMc.'e 0.601 3.4
18 22 0.604 0.365
19 15 0.f04 0.365
20 rrm e 0.C .04 o0. 6'.
21 7 0.605 '. "
22 32 0.605 0. : .
2. 35 r0.605 0. '6
241 re mo .e 0.605 O. 'j.6
25 9 0.605 0.67
26 0. 606 0. '67
27 remove I.606 0.36
2.1 13 0. 6 O.Jr
29 rem-ove i .606 O. 0
30 24 0.606 0.:'6
.1 rem,.e 0.606 0.366
32 14 0.60 7 0. 6.
33 remo..e 0.607 0.;6
34 19 0.607 '. -.,5
35 reci.,-.,e 0.607 r'. 36.8
36 23 0.607 0.369
." 11 0. 61I 0. 6

the 20-seconj control period bet,:en s lides and then had to refocus

rihen ,; n-w slide ,rpeared. The nari r-'.isiori refle:-' associated with re-

focusing would e.,plain at least cart of the rFsults found.

Tables 5-5 and 5-, present the results obtaiind from the regres-

.ion with the 9-second percent change in pupil size ls de, perldent variable.

The rnultipile corr'rlti,:,n coefff icient, and corrieqiiently, the eF planned

.variance, is lower" for this model. Again, the results of' indi idual

slide: rei'es..ic ns are nequivol cal. As tbefc.re, niar ly is of co-, ariarce

should ha..e el-iiriated the effects of light intensri ty and initial alue

on the final results. ThPse firdirngs Iaise questions .-.',,.ut our j tbilit.,

tc. deduce an,' r.,-arniingful conclu-si ons from resu. lts of rjupiliometric

as-essr, ent :tudie The low eil1.ained ..aria iiit., toae other iith the

inicons istFncy associated wi th rr-'pon, ses to thF 4 t r,'es of stimuli, raises

lman que: t i ons abt.out th: usdfulness of pu ii lonetri: tecr-rniques.

The fc.ll inri tables present i-e ult s of step, se li rinear egreres-

sic.ns on rhertrate, blood prressure, ra d slin potential. As would be

e-pecteJ in monel which do r'.t l1o these ; ari. a.les, the rultipie cor-

relation coeffficients iad e*plained .ariance are lower th-n th,'.se for

the pupil rsporns' equE tiions.


Table 5-1

Last Step of Stepisc Multiole Pegression

Dependent V 'riablc: '-Second Pucil changee

Multiple P 0.r575'
Std. Error of Est. 11.0520

Analysis of Variarnce

(Constant 0.0.




D F Sumn of SCqu.'re
22 15019.4'26
2.4 3029?2. 547

Variables in Equation



291 .01
241 .07


Std. Error

221 .2.5


7.1 2. .

r, .00
1 .
S1 C,


F Patio
5. 50,

mle.rn Squsr"

F to Pemove

1 .2


1 .96
1 .91

2 .6.
2. 0I

1 :
1 .29


Table 5-6

Step: ;ise Lirneor" F'.For Sion
9-Ee-.ond Percent Chrnge, Pupil Size

.tep vari ble Multic.1
Number EnterEd P FP.

1 16 0. 18 0.101
2 10 0.416 0.173
S12 0.4l 4.1 0 .198
4 E28 0 .4- 0.227
2.6 u0..501 0.251
13 n..09 0.259
7 19 0.51. 0.266
0 16 0.. 24 0.274
9 9 0..3 0.'281
10 32 r0.i -l 0.2.
ii 21 0.5 9 0.291
12 31 r0..5 7 0.311
i3 33 0.5t.4 0.31?
14 4 0 .567 0.322
15 223 0 .6E, 0.323
16. 17 0..69 0.324
17 e 0.570 0. 24
IS 35. 0.570 0.325
19 14 0.571 0.326
00 0.571 0. 26
21 1'' 0. 71 0.3't
22 E 0.,571 0.32

Table 5-7

Correlatiors Eeteen Puril Chang:e and
Percent Change

3 sec -
9 sec t.,
Initial Val

Anima 1

Hargi ng
IOml'ia r


Percent Ch anrq

0. 1




-0.0 s
-0.1 E6
- 0. 03
.-o K ,


Step 'lumt er Entered. Peno,.ed
1 &.

2 3

3; 7

4 ,

5 4

Table 5-;

by Stimulus Groups of
Heart ra te Peq r-es S i ns


0. .155


O. 190

i, #j7






V r i..t.1.: s are- 3 Initial V-lue

4 Ple s;nt Stima.li ir, n rroups)

5 eutir l Stimuli r, i r r up: )

n 1Unpleasanr. timuli in groups)

7 rn, i et,' r .u in Qr St iol i i u n i ( roui )

Tahle 5-5

Sumnart' Ly Individual Stinuili:
Step,.-ise Li near PF'e ressior ,i th
9-Second HEartrate as Deprenjet Vjriable

Number r




2 C'

lariablr ~ i
Variables 4-:


0. 11
0.51 ;


ilul tlole
ri. 2; 3

0.2 34

ri.: 5



is Initial Value
1 are irdi .idual sl;ide
and startle stirjull

Tblie E-10

9-Sein ry y Ser tiu Charm n: Gr-ro:

9-S5-cond FPerccent Change in Hir-j'trate

Step NrilUer






Va ri ,le





In iti l Value

Pieisant Stimulus Gro'up

re-utr 1 St iimuli groupp

Unpleassant Stimulus Group

n.r! iety-ProuIi ng Stim ulu" i"l' ,rcuD


0 .4-'7

0. 439



0.J4 -,

F Sr'

0. 19

0. 197


2 l l




6 -

? -


Table 5-11

Suriniary by Stimuli

Stepwise Linrer Pegression of
3-Second Percent ChJqnes i F;Blood Pressure

Ente rd.'PEF'r cri d

1 1




'*lul t i.le

O.'397 0.009
0.12?7 0.010

0.148 0.021
0.1 5 C0.027
0.174 0.030
0.178 0.031
0.179 0.032
0I. 180 0.0'1
0.180 1 0.032
0.1E1 0.032
0. 182 0.033
I. 12 S '.023
0.123 0.033

0.133 0.033

0.1 3 0.033
0.1 S 0.033

0.14 0.033

Variable 3 is initial value
Variables 4-21 are stimuli

Step Irimbc, r







Tale 5-12

Stepuise Linear Fegressiron of
9-Second Percent Changes in 1Elc:d Pressure

Variable Ilultiple
TlbPr enter td.' ro te:i: .ed P PS(

21 0.115 0.0
15 0. 162 i .0O
8 0.191 0.0
19 0.200 0.0
4 0 .20Z 0.0-
14 0.212 0.0
3 0.214 0.0.
18, 0.215 0.0"
6 0.215 0.0-
10 n.216 n.:o
1' O.21 0.04
9 0.216 0.04

11 0.21 0.0
16 0.217 0. 1
13 0.217 0.04

Step Nur


1 1



Variable 3 is initial v*alu
Variables 4-11 arre stimuli







Table 5-13

Stepwise Lineer Pegressions of
9-Second Percent Chanaes in Skin Potential

Step Nunbe,-

Entered/Pe o'. ed




Vari. ble 3 it initial value
Variable: 4-21 3,'e S.ti r-uli
Niote: [ecai e of the res:ponr se l:tern in s[in Dotential,
S-second changes were not anrl. 2ed.

ilultic le
0. 205 0.042
0.275 0.075
0 298 0.0S9
. 299 0 .0_:9

0.330 0. 090
rj0. nO 0.09C,
0. 300 r0.090

0.301 0 r.r.,

O. 301 O. 09
0. 301 0.090

0. 3:01 0.000''
0 .301 0.0 l

0. 301 0.091

0.301 0.091

0.30'1 0.091
0.301 0.091
0. 30) 0.091

The re;u1lt e aeicri b.--d atxove uwre applied to the h Dotihei'C-

li;t-d in Chaptc, I H.,pothesit, 1, "Ile-asarnt stimruli will be accoT-

pani-d by dilation, while neutral stiriuli will caaj rno .:hanr. e in Durtil

cize", .a. ri .t surprrted. : can be .seer, ir the correl aLion nratri nre-

senrt.d in Table 5-;, all p l asarit :tirrn jli have a slight neatii.'e co: r e-

l ti crn uith 3-seconi rupil char.n lt.thoua h the cor'relat ior.: were Eir:

negqti' : for the ?- record cupil chr.-nrge dat the, .re I- tili neqati e.

Sni ilarl., r, eu ral ;tinrul' i general ly h.ad rnenati ve corre lat or, s. These

re:ut1 s i ir lu e tl u i fi Ta t ir.in o ut cf .,rn.iet., scores tLv analysis of co:-

v arij a rc .

Hr.pothe:i : :, "n gati'. stiiul i will c ause a cor.strlictior, of the

pu.il" as n.t supportr.d, :in:. ti: puril r acti .r, to th s.e stimuli

waj liarer 1A a po ti'e :ch.iange in ize. Hao, t" si; 3, "a. ' iety- air'ui 1 ng

itiriuli .i 11 di late the curil", \ ':s s upp:,rted.

Hypotlhesis; could not be tested due to the inrConsisT.encies inr

the resrorise to p. easart ,t r I irca ri consis te -nt r :r,n to r. le a-

sCarnt ;ti iul i \ s f'urnd, th, dearce-e .,f dilatio.rn could rnot bt cor;Npareo

with dilatiorci;s a : ;.:iatec i th ari ety' stimul i. Hi,'pcthesi 5, "hiihlv

pleF isurat le :stimc uli *,ill podJuce greater pu i l di l3tion than will less

ple tas:it ;t lrj li" \was ro, t s uoportied, 5a tl h cori.,elatjion matri.i. shis .

Scr.r-e poi ti *'. results; wevi oLt.i-.in- ed in th r ari aly, i; of flarnrniltude

co. ari n ace. Hyp-th-sis 7, "t loc.d pressure change, .; 1 be sin il r in

reaction when ri .poFIed rt pl a-.:s r ant : tin,l) i r1 1 r .l;en Ec po.j:ed to ri ety-

ar.:.u:ing ;tiruli", w.as not suJported. !!ii e blood pressure sho..r:ed no

consistent relationship to positive stimuli, it did sho i z positive

relationship to an:n--ct;,--arousing stimuli, with a lag of sw\eral seconds.

Heartrate decreased as a res;:onse to startle stiAuli whilee shoWrin no

consistent response to pleasant sttiuli, so h,'nothesis 8, "heartrate

changes will te similar whern exposed too ain:iet, -arousin.] stimuli and

pleasant stimuli" was not Eurprported. '.kiin potential had the largest

time laq (t.etwreen C and 10 seconds) to startle :stimuli, and had fairlJ

large mnagnitudi c.h nges (un to 70 percent. It shoir-d nc consistent

relation to pleasant stirrull, co Hnothesis 9 was not supported.

l.pothesis 10 could not be tested, again due to the iny..,risi tencyi

in responses to pleasant stimuli. H,pothesis 11 a's su reported, since

no consistent r-lation hFtaeen heartrate .anj nositi':e stimuli were found,

while the average time frcm beginning (of a startle resporine to its end

h.'as 6.2 seconds. H.,pothesis 12 ;a: not TuppoIted, sinCL no consistent

relationship was found for resp.inse latency or return of s;'in potential

to either type of stimuli. Hypothesis 13 was supported, pr'o'ah l; because

of the relationship t 'isting between heartr-ate and L lo.od Cressure.

Table 5-14 presents a summTar, of the hpiCithEscs and the results


Table -14

Fa ults t: I','potheP is

:upper td!PF.j ec ted

Puli ilary Peact':n

.1 Pleasant tlirr, uli .ill bte acc=nl'ianied tb,' dilation,
while neutral stimuli i;ll Ci.. u nro change in
pupil size.
2. Il:.ative stimuli will cause a constriction of the
3. ,An..ircty-aro;usiri. stirull will dilate the pucil.
4. Anri ty-arcouting. :.ti.uli will c:u;e IrEater di-
Itti nr than will pleasant stinuli cf the rame
inten i ty.
5. Hiqhly ipeasurab, t tiruii will ro'r d.iucoe gro.Ater
pupil dilation than l 1 less1 pleasant stimi uli.
6. Highli negati. st ilul will p'roducE or-e.Ater
pupil conr tricticri thar will iess ne.lati'.e
s t iriul .

'u Itonoric Co rn ita ri3 t i r.,
A rllonitude Cj. riance
'. lood. pressure chari-'no.. ill he ~ si~ ilbar in reac-
tirn whenr e posr ed t. plc;arant :. irruli anrd irher
C-pos ed to an- i ty-arciusing str.iruli.
C. Heart'.te crnii e' will be simi lar .rwhri r po:p.d
to anr iety-arou'.l in stirrul; r.jd pleasant stirnuli.
9. 1 in potential cha ,..i11 :,be similar in reac-
tion ronen r po:.Fd to ajri et, -ar-ou: ing and to
pleasant stiniuli.
B:. Fesp r:r. Dur ati. ri
10. Pupillary dilation ro n-iet*,'-arousino stimul
will t 'a e lone, to r-tiu r to ini tial base 1e .i
(at tine of stimulus rres nitatio:.n) than till ruril
dilation to pleasant c stimuli, for a nien le .cl
of t in u lus irnte-ri ty.

rot suppcr'ted.

not sup.poted

not suoport-dl

not SupOCrcted

not supported

not .urCOr ted

not sur.rrtrd

Until: able with ores4rit data


Tablt 5-14 (Cont.)

Hynotheses Supported/Pe iec ted

11. Heartrate will tale longer to return to base
line in response to an-ict.-arousing stimuli
than to pleasant stimuli, for the sanre stimulus
level. supported

12. SK.in potential will take longer to return to base
line in response to an. lety-erousinq stirruli than
to pleasant stinrli for the same stimulus level. not supported

13. Blood pressure will tate longer to return te base
line in r'sp nr.se to' an.iet.,-arou'sing stiniuli
than to pleasant stimuli, for the :a',e level Cof
stimulus intensity. supported

Shortcomi nis

It would be useful to attempt to e-,lain the disheartening results

described abo'.e 6~ the product of poor e-periniental dt-siqn ,or analysis.

The controls used in the study, hoev,'er, were at lFast as rigorous as most

reported in the literature. Also, as mentionedd previously, the ddta w,re

collected in 2 forns (chart and digital) with chart data being used to

%alidate digital data at all points of con'.ersion, so data reduction

errors seem to be negligible.

Although the step: taken sDekP well for the data collecCr.d, the e

are several areas where the eperinrent could h '.e been imo.,rcO.ed. Better

care could have been taken to de'elor, a set of slides with more closely

balanced light intensities. Spot intensity on eecli slide should hiv--

been controlled. At the data recording le.fp better rteasures of

11-1ht intensity could ha've been made.

It appears that the scc.r-es on the pre-test questionnaire ram

ha'.3e lid to e rIrorne:uI conrl]us iois c, about c.ne c r" more slides. For e>am-

pie, the s11.ie show ring an attac:l ing pq lice dog sjcoredJ hiuheSt on the

.nrr ir-ty-arous n u Jiirien.-iocri arid wa.si therefore included iri the -rii ety

5st. Sub jects wih. lere debriefedj inforrn'll, ho,ewe-,r tended to ag'jree

thjt tnriIre i '% s little :an.iety,-.3rjuusinl ru llit.,, inr a Islide c.f a.n attacl'-

inrig d j o the questii-irin.aire r scaiF ri, v ha3 e been iiiilei in .

The nme-thoad b. ir ich rc mbl in r s r.;,re rdi ted in thir computer data

c.-ulld also be imripr'.'. id. ThE .it.a r edu:tion priorn's inrc rr.rorted a

-:.*r jutiri.? th.r *. i;ch r .? di ng ur.der .. millir, e er: '.' e1 li in ted

.nad the i:te coint imniredJiatflv preceedinri the eyeb.linr was .utstituterd.

eriih ps the r,:c..t ;apprlFpri ate procedure iU'uld haj br--n the total elijiri-

.1i Cf that dat, point. Alternaiti ely, :e'...- SIT.I thir g t echni qures

are .a. iiab ic- which iiould hev'.e b: rin mo.re ar, rroic.iate th i'il the mI thrlo used.



As discussed in previous chapters, the methods of data collection

ir psycholoc gical and n.cdical research on ph)isioloical variabls .re

often tedious and su-ject to :t-'eral types- of error. Onre of thr most

productive results of this study \:a the demcnstrition that co-rput1r

sampling i a viable and ad,,ante.eous niethio for collectiri dc:ta ,on

physiological *.'ariables. Computers today are capable Cf performing mrn:,

diffiertit functions nd .re availa. le ii rnan forms and si ze:. Their

costs are now at a 1,he1l where rniost rets archers should take tine to

examine their" usefulness.

Adva i ta qes

Several adiantagies accrue to the Iresearchers oho utili-c the

computer. If he can de'.elor a miethocd hererb the computer cari be rro-

,aran .red to contr'Cl the presentation cf stimuli, as in this study, the

experimenter can significantly reduce the uncertainty and inconsistency

with -hich hu.ansr perforni such activities. The comrriuter c:an overcome

limitations cf hurlnn perfor-,mance cap.3'bilities, such as the split-second

changing of slidfc in the, present study. Because of their hi h r.o-

messing speEds ciomputors ari-, cLrapbl' (f perform rin several ope'Fjtionr

simultaneously,I such js the chanirin, of li es anr ri col action of several

data samples it the span of 3a second.

As a data co llector, the c':.r.putter offers the a'l',antages associ-

at.ed with unritiased srpling and recording of results. As the epei-

mrnter has shown earlier (Eeell, 1971), controls oft this tpe must be

incor-pV.rated in ph,'-iological diata-g lthe-ring. methods. Cori-ut.--r also

pro' vi'e' a -ijgnrif-icunt reduce tl[ n in thm_ tinmer it tl e: to pioduc.e cat

in 3 form read. for anale 1is.

Perhaps on o of i he major ad '.aticeI in i:'cc uter h.:idware in reircent

,ears ha tiLeri the mini-crom. pu -e~ These- machires irnco.rprrasti ng solid-

:tace iiir. ocirii ui tr can b' obtjir;n F1 in :i:r th.a. .ill fit 3lnost

ar..sh re and that are hihl, .-.ptl .'. to the r_,ecific re.I qui rem-nts of

tnr purcha.cr. Vendors will, in srone ca-se 3asist in the acquisition

of rqui nrerit confi'uritions t hich :, :t mr'-et th-f needs rf thie u .ser. e.'e' a1al

mini-cc.riputers a e- "con..ers iti rnal" (he,- the ri 1 ity to corinunicat e

v ith larr'.er- scale niachirnet phyvsio .?lic rial resEarcher could obt in

a PDP-11 ri ri i-c'omput(r D (Di it l Equipmr.n-nrt Cor''orationr), for e .nmpl e

to control his eIpFri merit and collect. hi.s data. The machine ciuld be

cor. ne tet d vi a phone linkr to a lcar?. -scale E,: tern such as the IErIl CO/'..

w'h--re direct data i'ea in, anrd araly' is coj ld t.:le r.lace. Thus the

riachinre at the local 1;e--l could be %dapted t, 3 sprecific re-search

need, r.'ilt raiiltatirilng the abil ity to use ler e-cale data ara al is,

packages, and all of the peripheral equipment associated with the

larger scale systems.

Finally, ri.;ny persons associate the word computer ..ith high

costs of operation. while not fully realizing that computer until ization

may reduce ior eliminate other costs related to the pro.iect. This will

be discussed in more detail later in this chapter.

Di sadvantatqes of ComrPuter Lisag

One of the niaor disad'antagnsF of computer usage to behavioral

res'rarchors is the dependence on the "systems man" such us.ge creates.

Lefcore system can bte utilized. hardrar' must tbe acqiui'rcd and made

operaoticnal anJ software (progrnrms) mus t be developed and dehuuged.

Once the systemrn is cperaticnal .nr. changes to be irmJe or any, errors to

be corrected require that someone familiar with the computer be called


There may also be a tendency to accept the face validity of

cornputer-generated data without questioning the data collection, trans-

fer, and analysis. When using the computer, an e:.perimenter can be far

removed from the data he is collecting, and may miss important happeningrjs

during the e.eoFrirment that could influence the data. ThF use of a com-

puter does not reliiee the e.periliernter of the responsibility for accurate

The use of an earlier model Digital Corporation Hdachine, the
PDP-S. in clinical and diagnnostic nodic l research has been described
earlier by Cov.vey (190,.C .

d-et collection., ari.lysi;, and reporting. An important cornsidPration

is the frequency nd rcomple -;ty of the projects to be underta'i'.enr by the

computer. Coriputers male mostly nieces of furniture- in r research 1 aor-

atcry if the e tent of their usa,3 e does not justify thi-ir" ac. ui;sitio'n.

In uri i'vrsit. sEr.tinrgs, rese rchersl many times share a computer wi th

others w.ho need one in orper to reduce the per-projec:t utilii3ation

:co ts.

Scome Cost-Eenefit C cns idrat i ons

It is 5 rela ti vely simple mai tter to d-iscuss thl' h dvanta.je. 3and

1 mir t .:tir o s of co'rrputer s,':tens in giereir.l1 terms. L.efo're the reco ni-

menridation to puichl ie a 5.,-stem can bte .:cepted orr rejected, ho':.ever,

the \ery difficult tisl: of i ;ss',,ining doll r cists and b-rnefi ts cif th

s.Steni rnust be undert.a' n. !lin;,' cost f'i-ures c ri, be readily obt.airled

fro im pr-ice lists of mariufda turers. Other. ;uch as- the cost of install in

the .,':te and mal irng it operational are less ea.:, to pinpoint. and

ma, require a qreat deal more re r ;earchi rid "uC-sIt im a tl or,". Even thi.se,

ho:everi, may be estiriateid fiorm the files cf ui ers. of s ii ilar equi ment,

or from th.- '" i iUi'.:tuel .

Thi most difficult problem lies in the generation of accuratee

retiriites of the benef ,it to t.: der"i.c- from tlh coTIpiter r .,'tiTe Before

the system m isr installed, it is imipossitble to: foresee al1l of the rp:; ible

.-pplications it n,',:. have. It inaj, also hr ppern thit so:,e .anticiFp ted uses

of the system are not realized after installation. bout th: best

that can tb done is an attempt to foresee as ran,: Dotential advantages

(and limitations) as possible, and to treat these data in a ray that

recognizes the uncertainty under hhich they were generated. For this

reason, this section will treat data it, a r-.rncr proven useful in other

situations iriv.ol i-in uncertainty. In particular, vhenr uncertain nt,

exists, data will be derived from the familiar weighted-average formula

used in PERT and other project planning and control techniques.-

P co:rruter system for collecting and analyzing data of the

type descritbed in this di.sseration, would at a minirrun, coritain elerrents

for acquiring, storinq, and processing data. The hypothetical s.,stemr

described beloi attermpts to re l isticall, neet those requirements il thin

the frame-iorkl of decisions made by a uni er'i ty researcher. it is

assumed that the researcher it norv collecting data manually (as iTos-t

pupillometric researchers do)', transferring it to punched cards and

anal.,iing i by computer.

It is further assumed that meas, urino instru.Trents such as the

television pupil i oneter and devices for assessing heartrate, blood

pressure, ar d 5sin potential are readily available. A. central orocessng

unit .C.P.U. ) dith at least 4-thousan.i (4- ) bit men'.ory vould be required

to store the data. Ana'log-to-digi-.il conversion n requirements call for

llallace. J.E... "lr oroving. communications between s,stenis
analyst and u e-~r Data larniaerene t June 1972, cr,. 21-25.

the purchase cf 6 channels of con,.erters, and the input .output oDera-

tions of the s.sste- require a 4-k 16-bit read/wri te computer core.

The harVdiars de:-cribeJ abo.e form. the nucleus for a system capable

of acqui i nr anJ storing an amount cf information cr. omE.ir- t le to that

gathere.' for each .ubject in this dissertation. If we asi.ure that the

system;. is being desiqnedr for a university t research laboratory, then

the mo,;t efficient use of additional funds for analyiS dvs3i5 t.a eq,'liFp nt

would be the purchase :,f a dataphroe. This rould allow. the .Oriputcr

tC, .:omnT' ic'te with a larqer- .a lae ctrr,. The d!ataphone would tie

the small cc.'puter in with the C P.U. cf th,- larger tma,:hine, thu: greatly

enlar- iinri the stor'a.e capacity :.f the s s,'tern, incr".asing its analytical

rz,pabi i tic:. -ind ccnn tectin? it to all of the r.:-i'p er.l devices on the

larger s.; .stem.

Tabl E -1 de;cri.es the .d llar iestiniates of annual benefits to

be gairoid fr'~ci the new computer system. Table 6-2 d_,scr'ibes estimated

c* ts. Esti mates icre nm.ade by this wri terI ain by prc'fe s.o.r cL'r entlcy

en.aed in psychiatric res-earch Lsinr pupillometrics arn heartr'ate anal.si .

3'mon of the sources cf dollar cr t reductions or- added -enr efi t are men-

ti,:ined iefl, in the ta.bl-s.

Fs can be seen, the simple ,atio of annual rbenefits of the new

system to annual costs is 0.:'0 'Fe'.eral otherr techniqes for analzin

cc t and benefit dat. are currc-nt I, used Arnmcng these, t'e coI t 'ber-~fit

rItio anri paybar' period for the ir..e: ttmenlt in the coiput-r system i.ere

corp j ted.

Tahlr- F,-1

PDP 11/20 C.y, tem iith [lat;, .n r..ne I. i r;tan to IEl 3 ,'6

Lost 'Pencf-it Analv\ i

(Annriru l Fa: iT)

Benefi ts:
Pr-ductiorn PPiss innrP nrt of rlricral t.Aff
(Peduction in staff needed t m, ea'ur-.
phyv;iologicl v.aria blec.)
(Peduction in I, .r.uri.:h costs) p ,

Pecponsi'.enets Fastr.r r',t.a Transfer
(ability to npot errors while u iec:ts
.nri r_-iuir.irrent i confi',uratlionr, are st ll
,. ai l..i.e) 1 rir,r

Accuracy: Increise in Peliabi il ity of Data Sari.li rng
(''lue of increased accuracy reduc,,d need
to 'eperformr e, erimnint; i n:reasie in confi-
drnce in r ge-n ratirn report etc.) 1,0rij
CoirprehenrS i ,en .i es CLFpa ,i 1 it. for Increased
Scope of Data .Acquisition
(valuE cof additional inforrration: ..alue of
new Droic.ts won because of oparide-d
capabilities) 5,CO

Total Est ir.cated Anr. ua r, l Bene; fit r 15,

Cost/ renefit Estiir'i te.r Deri 'ed from teta-Distributed
Estimate-: Und-er Uir.:ertaint :

0 + 4Jr-L + F

where P = retsilnistic
0 = .pt rimistic
ML = ro't likelv

Table 6b-,
PDP 11/20 5..stenm th DatalChone LinnkagR to IE.lI 36.0/6'
Cost/Errnefit Anal;, i
( nn,.al Ea: 1 s i

LOs ts
Jonrl ur rr 1 in
C F .. Furcrhasl
-t 1 6-b1 t r,;-d,'a"-rit' memory
F' rcira ;iiIm r corn:1-cle
Easic mounting bo.:
Po r..' r supril,'
ASP-33 tele .tipe :n control rc. ,00
I i16 l it rcad -orite core 3.1 .li
'. to Di conv'erters, 6 channrr el 2,ri"l
Da t phorir 1 rO
Si te-rr irs :. lotion -debu. i ring 1 ,000

<, 1'9,30
5.-'ear iife innrual Basis 3.FO0

Pe: prr in.]
[ataphone I .o.Jern Frit l I. 1O 'l/ni. ) 1 ,'
S., tfrii pri'O i'arrmi ri ?' 000
3y:terii ri,:iintenjrrar 1 ,0I"l
S;sterim s;.iprirl is-: 500
CFI.I t ill.e,' ,stem :..C, 5,ClO

Toutjl rstirratgd annual o:(st I ..60

Cc,:t hdl' idc-d amrron three ,proje t: 1 ,e. 20

'Co:t'Eeref-t Estim these Derie..d Frorm E.c-t-Di:itr .uted
Est.im ts U-s i r L, Incerta i n.,'
P Ji!L + 0
wh r- FP = rsimi t i ; rl = .or in siist i; 'L : mos t 1 i lely

Sou'r:e: 'ioi t l E u- ipnr. nt Cor rF:,'tioin FPDP 11,' 0 F'rice List. rio,..1970

The payback period is defined as:

; Cl
C( t CC it T
3 j

where T = expected life of the systemm
B = the ith benefit in time period t
CC = the ith recurring cost in time period t
CI = the kth initial cost
P = the time value of money or cc.st of capital

For the data described in Tables '.-1 and 6-2, the pab.a:k periodd is

1.SS y,ears. These data indicate that the ..:;tem ,will pay for itself

in appro,.imi tell, onr, and one-half earsr. Since p 5-..ear life was assumed,

it appears that these results indicate that investment in the syvten is


The ccst,'benefit ratio is computed from the formula:

r (I3 z CC)
t I Cl
(0 + n)

The data yield a ccst-btnefit ratio of 0.A2-1, indicating that over

the 5-year period the costs associated wih th the system, e'en ..hen con-

sidering the time-valu- of ;none.. invested, are less than orne-half of

the generated benefits. These results also support arnuMients fc.r acqui-

sit ion of the syite.n.


As is Cobvious, the data presentf-J in Taible 6-1 represent *.'ery

unrcrtrin estimates of benefit t- Hcucver, ev.r if we car assign oly

a Drr.ot.bility oi f C.5,E to the c y.nected occurE-nc of these benefits the

cc.st benefitt ratio .till i les; than 1.0. This result allows scon:e

degree of confidence in the statement that .4 sste.,i carn t-e justified.

In surr:m r.,. it aDpears that researchers should in'.estiq, te Dos-

sitilities cornputer s.,cstens offer as scOurces of moree ad.' arinced E -cerii-

rinrital control 3nr. an l. 'sis. ith-riou h "-pecific cost anrid b nefit figures

depend or inidi:'idu.5l project requi renmentz, the d.ta prir.eri.td above

suggest thit, in some c. aes inestmcrnts in these s, 'i.st' s can I.e eas.il




The results of this studj raise several questions ahout the

direction of future research on pupillometrlcs. If we accept the hyoo-

thesis that the pupil does respond differentially to stimuli and that

it could be a useful indicator of emotional reactions, then .'e are faced

with the apparent need for much ojre basic and much more intensive work

than has been previously reported As with this dissertation, for

ever, "positive" finding reported in the pupillometric li terature,

there seems to bte a corrolar. "negati-e" one. Until this tyoe of

research reaches a stage where general uniformity of results is accom-

plished, it will nct be accepted as a useful techrinue for more advanced


Attitude Theory

Given the complexity of design necessary for evaluating attitude:

with physiological variables, an effort should be made to determine the

additional information this d:sicn ;ields over other, more Lonventioral,

measures of attitudes. it appears that paper-and-pencil instruments

such as the semantic differential, the Likert technique, or the Tliurtone

technique would yield morE information per input unit of work e''ended

that uould the physiolro ical techniques. At the present state of the art

in pupill -ri etr icL, it al o appears that the e scale: would d yield a

cqrc: ter absolute r..ai ni tude of infi:rnat; t hLC'i h : uld rupil studies.

The generally r'eportd rel iai lity and validity coefficients associ-

a ited \,ith the r. aer-ar.ii-penrcil r.methiods .are higher than tho~e ass oci-

ated with pupillnimetrics (Wloodmarsee, i 'j sc it rppears that for

mou: t asseiL srS'rent probli-er.-, the converitionafl r,,aS ur rinenrt iins trurrer it

imay b-e more appr'opri.ate.

Conmpar tiv e studjie ailing this liner could defi; n- ;ri me r*elEc.ait

attitude doman'in tco he .arimlied.,I thern je..'-iloor aszesrment ins trument: cf

b.-tli t.,:pe roL.tin r the co. t: : f d IS~ oprient in t irne, dollars nranpor, ier' ,

anI other r urcei e..r.ended. Peliahi.li ,, nd ,al-i.it., crefrfi: iir nts

:ul.d he cc.'.ipuled us inr, data g3thEe*' d from "I rne., in group.'. Per unit

c:ontri'LuuticrO and a sol ute contritution'i of infjorrcatior, could h.e cal ulated

ftor" each technique.

PFscearch o n 'tiriulu: Pre nritction

PerhaFrp one of the most limiting fac.to's ir pupiilomer.ict s ii

the mer t:r.d ,C:f joper.atiornalizling independent variab'les throu'. h visual Stipriuli.

1, 1:,O c' tac t he re.- ressior mode l de.e loped in Cihaptps V derr.on t s' tr.

tlis pi ic.tlen fI ny of the fac.:tr. is affecting pupil size .ar';- .a ocia te

with the e.,e a 'r. input medium if r :timuli. Thie near vi ic.r, i refle.., the

oi i'ieting refie., color adaprts tion, i.iht intens'i stv (h.oth -..ot .nd

ScrarTiiled), :tinilu u' ccriple.. ity, .nd the phy;icai locationr of the .viual

s.tinuulu. all cause char,.es in pLpipi 1 ize. if it W.ere po',ss Bl to de'eloCD

a different t,.pe of stimulus inrut, such as auditory stimulation, the

laqrg number of contributors to puioil change :ould be reduced, and

perhaps the effects due to ttimuUlus content could be better isolated.

Beyond this problem, there is the difficulty of projecting ac-

titude objects through the use of visual stimuli. As several subjects

in the present study noted \a ery larce differen..e t is t beti'enr an

attacking police dog and a picture of an qtt-ckinq police dog. ESii-

larly, a picture of a Nlegro rin, not ev'l.e the same t.oe of response that

a living being does.

A very difficult p'e blem must be ovecor.rme before pupill metric

techniques can be applied in areas such as the rcO'isonriel furictiior. To

date. most successful studies hae involved fairly simple, basic atti-

tudes. Host of these could have been c.'alujted equjll / ,ell using ccn-

.'entional techliriiuss. The hope is that some method for presenting comple.

emotional stimuli can be de'velope-d. Of all stimulus t;pes, ccmple,.

visual stimruli are the least preferred, since the comiple. ity requires

greater coping behavior from the I..upil and cause: greater Ee search

of the stimulus, both of whichh confound pupil change data. Smnie mc thod

which does not cause confounding interactions Pith the dependent variable

must be developed for presenting nor.r couple. stir.iul .

Field Fcsearch

In the Iarleting area, in addition to research in stimulus pre-

sentation, much work needs to be done o.n equipim.Pnt aind mohilit,'. P 'r erit

techniques are still v'ry "obtrusive'" and iriLer activ.'E itl. ti resnorn-

dent. The dev',-elopmrnt of a norn-int ict'i -e reasuririg inst ur7in for

research on consumer's re acti r's to p..a'a'e e si no and other promoti onral

techniques would be a imiajor ad/vanr ce.

If r-arletirigg esIarchers ps1 r, to us- Fr pi l m1 ii etrii c a useful

stud, could bt r made using Fortable physiolco ical recordiri. de.ices.

A phy.:si cloqic,: recordl-er, f'or inst arce c, uld tirecord several tr.ac : o, f

data simultaneously, during a iar'l:etling .tijd., The data could th!rn

be tale-n bed to- the labor1oortcr,' arnd rad directly, fro:i the reccrd-'r

into computer. Th.- laq timri'- between data collection, and anal ,sis

c',:uld tbe .ignificintl,,- i educea using this- tec iilque PResearcher. could

u.,,al u.te the tra d .cff '- bet. ei. r u si ng st 1 il -f raTe '.er us ,u i c: i a lrars,

and could de eiop ciit/b Snefit ana l se'. of the irnt uI ri rt ul in th


L ?bo rator',' Fese r:ar'ch

[luIchi r ri. still n eo se Lc. be done:- in l he are.? cof insti urn.ntatic n

in rpuFil 1lometric res: i'earjclh. 1 1. appears t*hat th.1 re is .'ry little question,

other than relati *,e co ts about the c-rcira] super iori t,' of video s stemss

such as t t 1he television pupilloin- ter, e r.ecia li,' ,hnr such U s steriis car,

tbe interfaced rith a computer. whetherer the Sir. n ficanrt addi tiora costs

of thr tele'visicin s-ystesri are lusti fied ,rineriends upon the v'alLe of the

inrforl, nlion 1r. be gather red and the -.a i labi li t,' of s-:pr tin e o. nic 'q en t.

Finally, applications of the mini-ccmputer in physiological

research car now be evaluated. .-'ork needs to be don? on the compatibility

of the small computer systems with larger systems. The information

processing, storage, and input/outout requirements of physiological

data processing systems need to be investigated. With a.dances in

computer techr,olgr;y continuin.- at the current rate, researchers who dc.

not recognize the potential of the computer for their projects may

soon be left with obsolete methods and inferior results.


1.11.i^ f I:,D CCi!CLU .IO1iNS

Gc 1 ls

Th& goal: of this dis er-tttion were to, repr licate fini.ings that

pupil chrqe is a useful measure ocf &emot.l: na ri e:tic''. to .isual sLiiul i,

de.'elop a lin rear nmouel e1 pli niir,g tIhe cnl-tri bit tctrs to r-.u~ l charge de.el op

a cci.puter sariplin r[slhoi i and c'e.,ine the fe i,-i ilit:,' of its use i Ii pupil

change studies, and studio the o their phy,,-iolc. c.ai Co.ariates in pupil

change. It was hor,=d that the recc.rdir n of chanQ-..s in r.otihe r 'a,";ar il

al ni wi r.th pupi 1 il l r. e ~ oulr il ed to i mer ns of 1di -r irii nati ng t et..ceer

pupil dii at ia or.s cause, i b.: p I c ur tt.e stii:uli and d iaii li urt .iur--:l t.,

at iet, -ar .u-. irq s- tii. ul i.


Fourteen -ales e.eA..ed as indeperdent rate s cf 57 l ide stirii' .

Of the 57, 16 were .-ho:en for use in the stud,. The 1. were divided into

-1 -isual stimulus .c teoj ries: ple -saiit, nreutr l1, iunc'leisant, rand anr iety-

c.r' feair-':,,i u i nj. The mediar te t wai used- tc. test f r differences

tleteern the sets of stimuli. Each set wa; a s inificarnt'i different from

the cther -et. at the .01 level. It appears, h.or,'.*ver, that the sl de

used as an an- iety-aroui in stirmulus ,.' .actu 11, not t capable c.f afc'u .irngI

anxiety in the subjects use:. Two startle stimuli--s buzzer and a drop

of several inches in the chair subjects sat iri--rere also included in

the an.:iety set.

The study tool place in a laboratory ..'here environmental light

and sound levels could be controlled. Subjects were seated in 3 dentist's

chair, and looked into an apparatus similar to a Hess ['.o.. Slides were

presented in random order on ; screen at the end of the bo" orpo, site

the subject. The tuo startle stimuli always occurred as stimuli number

17 and 18.

A television pupillo:meter measured change in pupil sire. In

addition, heartrate, blood pressure, and slin potential %ere measured.

These variables, tcaetheI with tile liiht interisi t, of the slide screen,

were recorded with a Dyrlrgraph recorder, which Droduced charts of the

data for each subject.

The Eunl:er-Pa computer as coutennercEted with analog-to-digi tal

(onverters to the D.nograph recorder and recorded second-by-second

digital readings for each of the ,.'ariabls, Ft the end of each session

with a subject, a paper tape record of the 2,5GO readings taken on that

subject iwas produced

A Eunler-Playmo process control computer wal connected to to

iodal carousel '50 sliie p--ojectors and controlled the timing of the

slide presentations. Each stimulus slide was preceded by a 20-second

blani sli e. The E:unl.er-Pay:'-. by initiating a change in the projector

containing the stimulus slide before iremToving the controls slire, reduced

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