A multivariate investigation of professed and inferred self concepts of fifth and sixth grade students


Material Information

A multivariate investigation of professed and inferred self concepts of fifth and sixth grade students
Physical Description:
v, 58 leaves. : ; 28 cm.
Graves, William Heniger, 1939-
Publication Date:


Subjects / Keywords:
Self-perception   ( lcsh )
Child development   ( lcsh )
Child psychology   ( lcsh )
Foundations of Education thesis Ph. D   ( lcsh )
Dissertations, Academic -- Foundations of Education -- UF   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis--University of Florida.
Bibliography: leaves 52-57.
Additional Physical Form:
Also available online.
Statement of Responsibility:
by William H. Graves.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 022663105
oclc - 13958940
System ID:

Full Text

A .Hulivaxiate InvestU' ,. ori of Professed and Inferred
Sellf Concepts of :'.th and S3.xth Grade Stud:nts


WILLIAi I. GRA',"-;, JR.

A D .:.: O,' ]':T V ''.;TD TO TE ,RDJATS, OO'UN,. OF
I-' UIvfSiTy 0op FT.,C".-; 1"P P.-,Ai ,.L-
] ..; .: OP TI F':::; '.,l,-. 71 '_ .- T'7.-
DOCO'<' OF .ilLG Y


r,-, ... cI of people who have influenced the author's thought is

long, an') 1 i o c.. i on. s with th'am are fondly remembered. The contri-

butions f ,o- Jc-n &o 3- .. as to demand xc'.clal reco-nition. The

author i1 !l',.. d to the chairman of his doctoral committee, Dr.

,', V ', for his personal warmth and for his exemplary devo-

ti5on to 1ttr:'; edue tion for children. Dr. Vynce A. Hines' initial

and conti:':.! 3r -I ort .an a catalyst which will always be deeply appre-

ciatd, l)r;, .. e and CGuillermo F. Hascaro have unveiled many

mystoeriocr -. l' o.-n the author to uork with them closely in demanding

tasks. Toe t i... ;.- of Drs. *'-,on H. Guertin and C-arlesI M. Bridges,

as roll as th ir t..:; a.nd interest, were vcry influential. The children

and ctrf of tho co.hools involved in this study are thanked for efforts

without vhic- inis study would not have been possible.

Mrco o,;.l L, Stevens, who typed this dissertation under very

trying conditi. inr, is h.;jhly commended. The author finally wishes to

thank his faii),; ,ho have borne with him throughout his studies.

A,'C,, ,, I_ -"Du .... T

TAC11 ' IU D-T, I'- 7; 3
LTAI.. '' .J .'

ACK O*'IL* *i T* ..,.. .. .9 9 49. 99.. ,, 9. 9.... 4,4* .... ..*.... ii
LIST OF' TATLrY1 .,, , . o .......,,.,.9.94449,.,9 49 9~*9 .. 999 .. iv

A Sh f R1 T ................ .................................... v

I Statement of the Prob!V ........................... 1
;*-!.. ouid of tlhe . 9 . . .. 2
purposes of the Stu-iy ........................... 5

II Review of the Ltratu r ....................... ..9 7
Azpe c .'-, of Sc32f .. .. .. ..... ... .. .. . 7
Developmental Tp.ic:",>: of Sclf I .., ..... 8
Self Concept ad Sc: '> /1c;',.R.i'nt i..... .,.. 10
Measurement Thor',y .' pc'.1cptT.3. Psc '.V ,.. 13
Perceptual ,c ;:. ,:. 0 f. 0.: i .u .s ,4 .t. ff, .a 16
pde'PC:! c:!^' t-t;,.. -l.atJ 0. l so ... 22

III Design and Method .....*. .... C ,. 9,, 9. .4 ...... 25
Measures Used in th. ; *i' ....... ......... 25
Description of Chilr v: .:i Schools...,..... 27
Design and '`.Uhod. .....,,t ...,.no,., Hypotheses . ....., .... .*......4 .,. 29

IV Analysis of the D-ita ....... ,9..............,, 30
Relationships betwe.en : 1- .jues .,,,,,.,.,,.., 30
Comparisons between Schools ..... ,,,,, 34

V Summary and Conclus ions,..-, ....,,,,..,,,,.,,. 41
Summary ..... ,.,, ............. ... #. *.... . .. .. 41
Comparisons between Schools and Sexes ..,,,, 45
Conclusion .9.,.,...,.. ,.,, 9,.,.,,,,***** 46

A Florida Key .............,......9..9.......... 49

B Pupil Florida Key ........ ....,,.. ,.... 50

C Self-Esteem Inventory ..........9.,............. 51

REFERENCES .......................... ,.............. 52

BIOGRAPHICAL SKETCH ................................ .....*... 58


1 Rotated Factor Loadin!r for Flor5.(dL. y ............ 26

2 Iten Crons-Corrslatiors of -1oria Kory, PuilT
Florida K. v, and Self"7 io, T:'c ............. 31

3 Intercorrelations of Interred p.nA }'i';;ed
Learner Self Concept and Scf2Fst:c; Scores .,,...... 33

4 Curvilinear Comiponents oZ Thr.e: 11rf hc'xoort VAziIbles
Predict:i.- Inferred Sielf Ccnrp alal!s .......... 35

5 Summary of F-Ratios and CPAll r or Afo ly:es
of Variance of FK Factors .................... 36

6 Summary of F-Ratios and Cell M;ans for An.lyI.ys
of Variance of PFK Facor; ........................ 37

7 Summary of F-Rat.ios and Cell ,.c, for Analyses
of Variance of c.:I Scores .... ........ .... ,. ... 38

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



William H. Graves, Jr.

March, 1972

Cliairr.ar,: Dr. William Watson Purkey
;'OL Depax-tIent: Foundations of Education

Fifth and sixth grade pupils frc:i two North F.lc-'rl- elementary

schools having different practices were a:irilr:,tered the Pupil Florida

Key (PPK) and the Coopersmith Self-- st'cr- Inventory, tw-o self report

measures. Their teachers rated them on the Florida Key (FK), a learner

self concept inventory whose items paralleled those of the PFK. Cor-

relations calculated between parallel items on the FK and PFK showed

little coammonality in scoring, as did a subsequent analysis using

factor scores from these inventories./ The conclusion was drawn that

self report indices should generally be considered distinct from
inferred self concept measures./ i

Factorial analyses of variance using the above indices were per-

formed with school, sex, and their interaction identified as the sources

of variability. Girls tended to have higher indices, as did the more

innovative school.


StAe;i.ent of th:- Probsien

I. ..,-r years which hav- followed th- re-emergence of erceptua

.y-hzo Py as a significant force in psycholcjy and ecucation, thr r.

has !be#rn a pliferation :'f ethod:i and appro-ches for systematic study

of the individual within a .'. eno- o'o;oi,: framework. This should

cone as no s.:rLri'e within ar a-pproach izhich has as a .ijor concern the

explorat5nr of the uniqueness of the individual's perceptions. Ya'7-t'Jr-

mor-e, percePTual psychology inherited a long histc-'y of phi1lo.ochh.Ic

concern about the problem of mind and self.

Dig-ory (19r6) traces the history of theoretical development o_'

xJK.t about self fro- Lorner through Descartes, who initiated the con-

cern of trl interacticn of two aspects of man previously considered

independent: body and mini. Serious theoretical questions about mind

were follcwei frcm Descartes through the British Rmpiricists to vnliaa

J Les, who'je work -arked the beginnln!s of systematic study of the rer-

sonar. internal world of the experiencing individual. Concern at that

pcint in history '.e an to shift to aspects of self such as the develop.-

-t. an-d differentiation of self, the I-me relationship, and the

ques.on s to whether self is divisible, and, if so, how to best

categoriz'; aspects of self.

ft.-r a lull in in terest in self theory on the part of American

--*:-.-,,-.. of a .....iy' of this cnttry, a lull eneraly 7.3$S-

. ..... ... ...... te eh.ioV L -st school he repzsen.


students of the self began to reappear. In this renaissance of inter-

est, there was an aspect not prevalent earlier, empirical research

aimed at validating and making practical. the theory. At least two

forces are seen to have been working in concert to foster this develop-

ment. The first force was that the experimental, cause-seekin;,

orientation of behaviorism was still valued, even if the early aims of

behaviorism were not. A second factor was the co'nuiriuir. development

of research tools for behavioral science such as desl-.ri, statistics,

and measurement theory. Field theorists such as Koffka (1935) hplprdl

link the various schools by specifying that the self, at the core of

the ego, represented acts pertinent to real needs of people, and that

the early behaviorists had been concerned with quasi-needs; thus more

global and individually determined concepts were needed to account for


Self, or, more objectively, phenomenal self, has been defined by

Combs and Snygg (1959, p. 58) as "all those parts of the phenomenal

field which the individual experlencc as part or characteristic of

himself," where the phenomenal field is the experience of the per-

ceiver. Within the phenomenal self, they define self concept as in-

cluding "those parts of the phenomenal field which the individual has

differentiated as definite and fairly stable characteristics of himself"

(p. 112).

The usefulness of measures of self concept has been defended by

Hilgard (1949), who has stated the necessity of using self concept to

understand defense mechanisms. Hebb (1960) postulated the necessity of

using self concept measures for studying thinking. Numerous studies

(Cowen et al., 19551 Perkins, 1958a; Bruck and Bodwin, 1962t Brookover

et al., 1964; Cline et al., 19641 Purkey, 1970), which will be reviewed

later, have demonstrated relationships of self concept variables with

social achievement.

Wylie (1961) published a revlcw of the research in perceptual

psychology, in which she critically assessed the "state of the art" at

that time. In this review, -.le larnter the lack of standardization

of terminology and i, t'ua:.ntation, .d i p L-cilly critical of re-

search efforts which devise a srcl;%l instrument which is only used

that one time. Work such as that undertaken by Fiedler, -cO.fe, Jones

and Hutchins (1958), in which they develop indices of relitlh"ip

between instruments, is seen as very valuable by providing neans to re-

late findings between studies which have used different measures. This

approach could spark more intensive theoretical analysis if variables

previously considered similar were to show little or no re .t!onship

experimentally, or conversely, if attributes previously considered dis-

tinct and independent were to show correlation. The lack of widely

accepted criteria a-;ainst which to compare self concept measures is a

major difficulty.

Background of the Study

This study will explore the interrelations of the two major ap-

proaches in determining an index of self concept. One approach used is

to rely on the subject to overtly provide information about himself.

The intent of the instrument may be either obvious or disguised, global

or particular. Instruments of this nature will be referred to here-

after as self reports. The use of the self report is supported by

statements such as the following by Moustakas (1965, p. 45): "The

individual's perception of his own feelings, attitudes, and ideas is

aiorc V2.11d in*- jy _oUL;ld_ !i3..=i3 n be"T-hs ae_%,_rta:,ain

..-, the '2l r'. ~z-t ?xr that it car pro'z.de .qe6surez frcO a V -: -

n.. bi of subjects on numerous items rel.tced to ",he ,henor!,inov' under

investigation, and dro so rith 9. relatively low expenditure 'c" ns ad


"'Te se!or.d na -" aasurlng technique is to ask for an j-otv:

ratinr- of the subject. by saaeone else, gener,-Jly a tr/aned oo.-- ,

The Indices of self concept so garnered will be referred to as i.. .

self concept scores, Advanta.es to this nethod incl..e re action of
ideosyncratic variation due to semantic problems of ive (i,. '"ell

adjusted" nirht- mean to one :r-on what "adjusted." ri s to T,

or intetrprtation of terms. Reliance on an observer a];io vxuce ai-

ability attributable to individual tendencies to "fAke coxi" or "fae

bad," and allows for evaluation, of characteristics which a:7e not

apparent to the individual himself, i.:.ny measures of self concept, 'by

nature of b*!r-. concerned with items of obvious pez-cn2.i neiir[, car

be characterized as "reactive" as defined by C&rapbell and Stanley

(1963). Inferred self concept scores avoid this ger-iEal]2,; undesirable

source of variation, Practical difficulties arise using this procedure

if a i.ar-e number of subjects are desired, since stable aieasures i*.-

observations require time proportional to the complexity of the attri.-

bute in its koreetua-.iattcn and occurrence. It should bo e.iphaaized

that these tio methods of collecting data about perceptual phenomena

are not opposing methods, A *elf report instrument can provide vala-

able information to an observer, as an adjunct to his observation.,,

Self -3 cr-tr, on the other hand, are suspect when vsed with people with

clinical p hopathKol~ies for reasons listed by Com-'. So'c, and

CoUrson (I1:`3) utless the deviance is considered as a se.-ait variable.

The aspects Combs et al. felt to be necessary in the individual

for self report measures to yield valid indic;es of self concept areas

1. The clarity of the individual's a,,zreness
2. Thc availability of .q'l.... .. for expression
3. The willingness of the intlividual to cooperate
4. The social expectancy
5. The individual's fecclr'i; of I". -...-
6. His fcc-lin, of freedom frci. threat.

In support of their position, Co--,J et ccrrelateud responses of stu-

dents on 18 items such as "I feel worth while I don't amount to much"

with evaluations fro-" trained observers of thote students on the same

items. Correlations were predomiii:-intl/ low ad positive; however,

there was no attempt to determine the r, ity of the self 1,:po.rt

instrument and the range of possible stwa'ii ..ronses as limited with-

in these one-item scales.

Proes of the Stul.-

There is a continued widespread s of the term ".cif concept"

in studies using both inferred self concept and saif report procedures

despite the injunctions of Wylie (1961) zd Coibz et al. (1963). In

view of this and the methcdoloL;,.: .1 limitations of the Combs et al.

study, a reexamination of the relationship between these measures was

proposed. One purpose of this study was to e::- -iDine the multiple rela-

tionships between two self report instruments and one instrument in

which specific aspects of learner self concept are infcrrcd. Two of

these instruments used very specific terminology to reduce variability

due to differences in the interpretation of the rieanings of the items,

and restricted their emphasis to areas of learner self concept. All

three instruments were comprised of numerous related items to increase

reliability (Thorndike and Hagan, 1969, p. 183). High positive

relationships between the procedures could indicate that much of what

is knoyn from research uslni one proceledre Is a-ppPcable to researx-h

using, the other pr)ced'jrf, I.ov relattc.z3h;'. in oerta'i di-nsi'n., on.

the other hand, could mean that it would bes unsa fe tc ener lize re-

suits froi one +,e-.nlque to proble-,s in which tht ocher technique would

be nor0 rtrr,-),ri.te nmtil furthBr evidence coulr support the .7:erali-

zaticn. Lz'- relaltionships on certain dimensions could also ii .icate a

basic difference between self-perceptions and teacher t-:.. pons even

on frequency of observable behaviors. In the letter cie, th- major

implication for teacher trailn:i.- would be to mace :utz-re tea.cherz a-;t.r

of tha possible lack of cnrou! :e of perceptions, so that the' ihz

become more r-srceptive, both cf tho students' behavior P.nd cf t.:e

students' self-porceptions.

The second purpose of the study was to continue the ivaftize-

tion of the effects of deliberate and intensive efforts to .%-rharce

pupil self concept (Purkey, Craves, and Zeilner, 1970) by conpari-xn two

schools with different emphases in the area of self concept develo,..-t...

Onr school had a deliberate, public and planned effor-t for this purpose,

and the ot-her school was much less systematic ir this area.


Pevkew of the Literature

The need for a rhcio onlo21zcal system of psychology was out-

lined by Snygg (1941), who emphasized that the frame of reference from

which measurements are taken is a critical aspect in any field of

science. In p'i:.irolojictl p-ycholcgy, this point of r,-for-i-ce is at

the level of the individual's perceptions and experiences, rather than

at the level of the external stimuli which surround him. T;.c most

basic perceptions in terms of their centrality and durability are those

which reflect the person's ideas about himself. Rogers (1951) main-

tained that self is central to thc individrur. and thzft cnh1acc.ent of

self is a high priority issue, perhaps takin_ precedence over or':n-

ismic concerns. This position is related to that of Conbs and S.ynr

(1959) in which the role of the self is seen as the determiner of

behavior. A summary of the important characteristics of the self was

made by Purkey (1970), who described the self as organized, dynamic, and


Aspects of Self

There have been numerous attempts to subdivide the self into

aspects which are distinct and amenable to identification and which have

differential effects within the person. One major system for categori-

zation of aspects of self has been to divide the self in two parts. One

aspect of self is seen by this system as that part or process which views

and evaluates itself, the notion of the self-as-object. The other

aspect of self within this system does not have this component of self-

awareness, but is the prime determiner of behavior, the self-as-

procecs. Cordon and Cej:rfcn (1968) outlined some implications of this

formulation of the self in the area of self-understanding. Rentz and

,,ii.te (1';7) have shown that these two concepts, both defined u;ing self

report instruni:ents, have some shared aspects as well as some unique

aspects,. It would seem, however, that the self report is more aripro-

priate whon considering the self-as-object, and the procedure of infer-

ring self concept is more related to the self-as-process.

The development of an individual self is seen by many (Snygg,

1941; Allpoi-t, 196l1 Jersild, 19SOr Lecky, 1969) to be through the

mechanis-i of l ,_:'L:n, wherein aspects of the phenomenal field are dif-

ferentiatcd which wore previously undifferentiated. In the development

of the self ">L'-.:'.-I learnin murphy (1947) and All port (1955) have

exaninci the special importance attached to the condition of striving

with -r.,--d to the differentiation of the phenomenal field, especially

in the area of self-knowledge, Combs (1952) outlined a procedure

through which i,'tCL-lli -tice could be considered dependent upon an ade-

quate self concept, showing the reciprocal nature of the internal and

external world of the individual.

Developmental Inplications of Self Concept

The role of "significant others," such as parents, in the

"learning-of-self" process is of importance to educators and psycholo-

gists in order that they might better understand human development.

Such understanding can be used when these professionals have to interact

with the "significant others" in connection with the children and,

therefore, must know the ramifications of this psychological dynamic in

order to coordinate with and advise parents to maximize the child's

development. Furthermore, teachers and other professionals become

"significant others" to children because of their roles as agents of

the society having opportunity and obligations to effect chan2-r in the

children. Thus, it is imperative that they understand the implications

of this develop.:ert.'il process. For example, when working with children,

it is Arnort.-int. to consider that new prceptions, if they Irtrr.ere

with or threaten the dynamics of the self, can be rejected even when

such rejection is both logically inconceivable to the outside observer

and unpredictable from an ex-iination of the external stimuli (Tecly,

1969, p. 117). The more central a concept is to self, the more stable

and resistant to the above change.

Some degree of stability of the self concept is a necessary con-

dition if relationships of i.,,Leno.enoloeical variables to human behaviors

are to be explored. Research in different areas at different a,;e levels

employirLT different instruments has demonstrated a fairly stable re-

lationship within those instruments and age levels (Brownfain, 1952;

Engel, 1959). On the other hand, changes in the way pupils think about

and see themselves have been investigated by Yeatts (1967), who reported

differences in spatial relationships between self report variables ob-

tained from children at different grade levels. Gergen (1968) reviewed

the literature and theory of consistency and indicated that very high

levels of psychological consistency are not desirable in humans. This

is in keeping with an organism so dependent on adaptation to a changing


In brief, if the self be viewed as the internally consistent,

dynamic determiner of behavior, dependent on a supportive environment

.cr M'.&airnaz jrwth =d ff':nezs, the rsl&ti-nzhi' cf Soif theory

tith the scc.1 it: .... ,.ion f education is s :r:nr- -.d ob'vi'.s,

"eL Conceot and School Achievement

r:iJ'-rcus studies, us1l! a rmultitude- of instrwaents anri oriez:-.a-

tionz, have shown consistent relationships bet;esn selli c..onep+t, ai

school .chiev..nt nfror- kindergarten t.ro :h college. In scme of the

earliest work in this area, Sears (1940) used a discrecancy score pro-

cedure and found that children's rc-,rels of aspiration were related to

their academic r-cifornance. Work by Combs and Soper (`163) and related

research by La.y (1965) indicated that measures of inferr.1 d self cov-

cept, *j .. o-,- th observer-as-in ,!ti-ent technique they pioneered, were

related to ac,-adeenic rerfor-.ance in kirder -arteners, Stabilt'+, tvf this

relationr.h.p ira.s established in this research by the cC.-rrsro.den.e of

predictions .iade ,-5 Com'zs a-nd Sone.r with ratin-s =a-lc rn these c&lZrisn

the next y-r by their new teachers. Sex-based differences were

checked on ?. kindergarten population by Ozehcskey and Clark (1970), in

response to work with older children (fourth and sixth grades) which

showed girls with hIgher self concepts than boys (Bledsoe, 1964). Ai-

.hocuh COzehoskey and Clark found no differences associated with sex at

this earlier aqe, they did find inmrrssive relationships between self

concept meaztunres and both the Metropolitan Readiness Test and the

childrens gr rade-point averse.

Bledsce (1964) also found significant correlations of achievement

measures with his own self report instrument (modified froia Bills' IAV,

1951) for boys but not for girls. Underachievers were found by Bruck

and ).'dw-in (l!-'.) to have lower self concept scores, -s inferred, from

the ,a-a--rson -.:, -, e ,. zot-rolin for differences in

intelligence. Fink (1962) found male lower achie';ers in the ninth grade

to be rated lower than higher -.chievinj males on ;adequacy of self con-

cept ra.tin-; ssirrei by three ex%.rierlced raters review-:.' nruerous

-sych.ilo-rlcal protocols, Females in this study vere not found to dif-

fer sr.-nifircantly on this basis. In a study encompassir.g o-'v

secondary grade levels, Brookover, 1ho:mas, and Paterson (l?64) f .und

significant correlations of self ratings of general ability and. 'lity

in specific academic areas with academic achievement, controlling for

the influence of intc liqerce. A conclusion was drawn fru7 this

research thai a feeli-- of self confidence of ability may well be a

necessary (but not sufficient) .'o- -'tion: for achievement.

The rtla-'ior~hlp of self concept with achievement sees to re-

main even in higher education. In a study by Roth (1959), college

students who attended a re:iedial rea.l.-" clinic and were successful

tended to rate themselves hi-her on reality-oriented dimensions of a

Q sort for self concept, ideal self concept and "ideal-actual" self

discrepancy than those who did not succeed or who dropped out of the

progr,. The latter group tended to rate themselves higher on some

dimensions related to ideal self. Within a women's college, Flaherty

and Rautzel (1965) found female high achievers rating themselves higher

on such California Personality Inventory categories as responsiblee,"

"self-accaptinz," "self-assurant," and "poised." Low achievers in

colle_ ge, it should be remembered, still have a history of relatively

hi,-h .achi'vezment compared to the general population, and yet their

self-:omparioons are aprarently based on their perceived peer group

(I.e., college students).

The possibility exists that social class is a factor strongly

related to both achievement and self concept, and that many of the

above findin-s could be more pa2mrimon:ou-?.y and reliably attributed to

social class differs. *:... Shares anil Scares (1969) found, however,

low socioeconomic class black children having higher self report scores

than middle class white c}Il'irern. Davidson and Lang (1960) also exam-

ined this possibility and concluded that, even though lower social

class children reported lower self concepts, when social class was held

constant statistically, a distinct relationship between self reports

and academic performance still obtained. AnoLhcor finding by Davidson

and Lang was that girls in the fourth through sixth ,Tadcs reported

higher self concept scores than boys and also saw their teachers as

being more favorable toimrdz them in comparison to boys' perceptions of

teacher favorability. Girls of th!s age were also fo .. by Perkins

(1958a) to show higher "ideal-actual" self conx..uenco than boys.

The interactive effect of teachers on the self-picture of pupils

was examined by Staines (1?53) and it was found that teachers who deli-

berately attempted to instill and maintain a realistic level of aspira-

tion seem to effect a more realistic self concept in their students.

Work in the same area by Battle (1957) and Davidson and Lang (1960)

indicated that children whose values coincided with those of their

teacher tended to achieve higher. Perkins (1958b), Purkey (1967, 1970),

and Combs, Avila, and Purkey (1971) have written extensively about

methods and attitudes conducive to the enhancement of learners' self


Measurement Theory and. percanbal Psvch ".

Allport (1955) summariz'i the or-anlzation of observato.i and

interpretation in phenomenologfical psycs'-'l.y, starter.! with tnrae

broad rate-ories of constitutZional d.ar.osis, sociolulcal tkc cl....ues

and "ych.l'"-ical techniques. The psychological technique. h
divided into personal documents, self-appraisal, conduct s -n.-1.n" ,

ratings, scales and tests, projective techniques, depth analyses, and

expressive behavior.

Two basic categories of vheionoeno)o-ical variables have been

mentioned previously, self report and inferred self concept. Although

all techniques above could conceivably be used to get an index withlr

either category, some techniques and methods are more appropriate for a

given category. ?cr exya7ie, the depth analysis would be more a!-.oo

privately used( to infer self concept, while self appraisal best yields a

self report. Some techniques, such as scales, are commonly used to give

a.n index within either category.

Other dimensions will be discussed which may be superimposed on

the universe of perceptual variables. Consideration of these dimnenions

is useful in that they illustrate sources of variability between mea-

suring devices that can cause concern about reliability and validity of

the indices thus obtained. These concerns are especially important to

perceptual measurement because of the problem of validation. Ideally,

validation is accomplished by relatlnr a measure of a construct to an

objective criterion acceptable to experts within the discipline. The

l-icl. criterion for self concept measures would seem to be the ade-

quacy ;_ni effectivene's of the individual's l.fe-style in .cco1.lisMi.-.

his piros, but this data i noot irkly to become available for

validation purposes, so validity has to be established by indirect

methods, such as expert judg-elt or the relationships of self concept

measures with partially relat+.ed, confounded criteria. The indirect

approach dictates more concern with initial Instrument conrtructibn

since inaDn'rc:rlate aspects connot be as readily eliminated.

One dimension which is a source of variability iz over4:-:.% tlt

dei-re to which the true purpose of an instrument is explicit or ob-

vious to .he person supplying the information. This dimension, althoujr

it can be varied in either self report or inferred self concept tsch-

niques, is most commonly varied in the former. In nakLr-- the decision

of how much overtness to engineer into an instrument, the InterTlqy o0

two considerations, social desirability and validity, creates a dilea-a.

The systematic variations which can he attributed to the social desir-

ability of pcssessin? an attribute, rather than the degree of that

attribute actually possessed, is an area explored in detail by Crown-

and 'Marlowe (1964). Social desirability is of particular relcvar.ce to

humanistic psycholo:-y where the attributes of concern are, virtually by

definition, mratters of genera], social and personal importance, and social

d-s-zabhlnity can be expected to influence overt self reports and even

'uar.neri.-i and behaviors from which self concept measures are inferred.

Covert measures attempt to circumvent the variability due to
social desirability by stressing items which, though apparently innocu-

ous, reflect deeper traits. Accurate use of covert measures, however,

is dependent upon the degree of relationship between the directly mea-

sured attribute and the target trait for which the first measure provides

an inj-rect irdex. As an exaapie, Y-..nrber of cigarettes smoked may be

inferred :io.,Ti yelioq stains rrn fin:ers, but the index lacks so-in

validity as yellow stains do r.cI have to coat fro. cigarette r.t oking.

Furthermoret, the stairs can bt avoided or eliminated without char.in

sTkonUr frequency. Accuracy in covert me.sures is further dependent on

the reliability of the measurement of the measured. variable, 3och of

these considerations are critical in humanistic r5:ychofjc-- as the

degree of relationship between variables is generally neither known

empirically nor agreed upon theoretically.

Although a great t deal of inference is obvious in covert measures

of self report, this inference is a priori and, therefore, these mea-

sures are not to be corf'used with inferred self concept indices. Ir

the former .rcc-.ure, two people who con-letp protocols in the same

fashion trill obtain thn szcnt scowe, whereas in the latter prcc.i,'re,

this is not necessarily tz-je, as interpretation by the scorer is indi-

viduUlised for each subject and the scorer is free to use wat 5rdifier

(195) referred to az the "nascent surplus" of information which ths

more aechanistic scoring procedure cannot tap.

Another impcrtant dimension of measurement is the generality of

the attribute being measured. If all other considerations wire equal,

the More global measures would better satisfy the criterion of p-rsimonIy,

To tha extent, however, that a more global measure lacks operational

definition end, therefore, loses interrater reliability, positive results

from replication and generalizability suffer. Wylie (1961) concluded

that the comparatively specific explorations into particular aspects of

self have proven more fruitful than research using global measures. This

would seea a l:.-;i.al conclusion if the self is conceptualized, as pre-

viousl'v ".nt.cne'-., ?s the rrT-.,-t anrA prc.c..sa cf continual differentia-

tion, ,:. ,l is thoLrht of as h:vPn nr erous independent aspects,

this could serve as a further *xr-lqraton for the general lack of suc-

cess of research using g'l.oI,) measures since a Tartic.ular part of self

could be one-atlnj- effectively on a target behavior and vet an overall.

measure could show low effectiveness.

Some degree of generality must be present if research efforts

are to extend the riomol)o-ical development of the tbeory. It should also

be remembered that the differentia'tion of the self concept Is unique to

the individual and, within that, unique to the sitt7tlo.-i, uo that sore

degree of generality must be present, or no one could kr-.ow anyort else.

Further, self knowledge would be even more difficult than it now is due

to an inability to :eneralize and categorize from others.

Perceptual "easurement Techniques

One frequently used method of obtaining a self report is the 9

technique developed by Stephenson (1953). In this technique, a number

of items (usually 70-120) descriptive of personality traits are pre..

sented to individuals, who are asked to assign each of these items to

one of several (usually 5-9) discrete categories, with a fixed number

of responses falling into each category to approximate a normal distri-

bution. One of two procedures may then be followed to determine an

index of self report. One procedure is to have the individuals score

the same items twice (or more) accr-.in1r to differing sets of instruc-

tions which represent different aspects of the self (i.e., "sort these

as you see yourself now" actual self, or "sort these as you would

like to be" ideal self). 'he correlation between the sorts is con-

sidered to be index of congruence of the two attributes. Evidence

maz-shaled by Cole, Cettin-, and Hinkle (1967) indicated that either low

or high i2ac .nc?.s between "ideal" and "actual" self repor-s i.e

associated with clinical psyvhopathologies, and that a sample of non-

clinical people tended to have moderate levels of conrgruence. This

finding cast shadows on conclusions of earlier research which assumed

that high "ideal self actual self" congruence was the most d-sirable


The second procedure which may be followed in scoring the Q sort

is for the ceve.lopers to establish a priori wei.ights (,.atchiii., -to-

standard) for the items in order to determine the level of a particular

attribute. The correlation between those weights and the obtained dis-

tribution is taken to be the individual's index of that attribute.

Wylie (1961) raised the question of whether the latter procedure is in

fact pheno.L.icnolo-ical, the point beinz, that in the former procedure,

the order of selection of the variables is determined and judged within

the person, whereas in the latter procedure, an external criterion is

bein- imposed on people's selections.

The Q technique has been used with a variety of items, sorts

havirLn been made on words, phrases, and even short paragraphs. Items

have been selected from clinical protocols, exhaustive trait lists and

other sources. The type of item included and the source of items need

to be carefully considered for appropriateness to a particular research

problem. The "standard" list at present in phenomenological research,

from the frequency of use, is Butler and Halgh's (1954) list of 100

statements gleaned from clinical protocols.

A discrepancy procedure somewhat less extreme in psychological

distance than the "ideal self actual self" congruence technique has

been developed and used by Brownfain (1952). In this approach, people

are asked first to rate themselves as they think they are. Next they

are asked to rate themselves on the same scale but giving themselves the

benefit of every doubt, and on the third rating to be as self critical

as they truLhfutlly can.

Another popuilavr form of perceptual measuring device is the

raiMv':, scale which is used in both self report and inferential measure-

ment, In this technique, the individual is asked to measure the dcrce

to which he possesses or is represented by, an item prezumably related

to a more general attribute. Items can be selected which are either

lcjl:;.-1ly and/or functionally related to the attribute of concern.

Bills' Index of Adjustment of Values (IAV, 1951) is a scale of 49

traits which were selected from a longer list on the basis of their

)O:'.;Lu connection with adjustment, The Minnesota Nultlipha-ic Person-

ality Inventory ( ". i, Lathaway and McKinley, 1912) is an example of a

scale sometimnr used in pmcr:.-,.tual. research in which items were selected

for their correlation with psychological disorders or other attributes,

reardless of logical connection. Adjective or trait checklists are

subsets of rating scales and are generally dichotomous.

A variation on the theme of rating scales is the semantic differ-

ential appro-ch developed and reported by Osgood, Suci, and Tannenbaum

(1957), which is somewhat covert, and which assumes that most individ-

uals share connotative as well as denotative interpretation of numerous

adjectives which are common in the English language. In the semantic

differential procedure, an attribute is named and followed by a set

(3-20) of bipolar adjectives, each of which defines a continuum (i.e.,

hot cold). The extremities of each dimension are commonly divided by

a five-to-seven-point unlabeled scale which allows the individual to

select his own appraisal of the characteristic of the target attribute.

With the adjectives generally used, three independent characteristics

of attributes emerge froi a factor analytic procedut'e, evaluation,

potency, and activity. This indicates that people tend to categorize

attributes in terms of the perceived or, Z.-.A'tion ..d dynamics of those

attributes. It is interesting but futile to speculate whether people

tend to see the world in terms of c-L ;.;.. .-r:d dynamics because of

these characteristics of their ,clvw: or :hcthVcr olf theorists have

viewed the characteristics of self as they view the characteristics of

other phenomena.

A procedure which allows for individual selection of constructs

is the Role Construct Repertory T'e.t (Rep "'-. -t, Klly, 1955). In this

procedure a person generates his o0M category system and uses the

categories thus generated to rate r-.piTle significant to him. This al-

lows for a more individualistic analysis of the person's perceptual
field. A time-sharing, terminal-based c -ii.lLr pa orri for simplifying

the rather elaborate administration and scoring of the Rep Test was

developed by Guertin (1971) who also modified the administration to take

advantage of the Q sort methc1olo:y.

In an effort to combine individualized responses with self re-

port techniques, Bugenthal and Zelen (1950) instituted a procedure in

which they asked people to give three answers to the question of who

they thought they were. C. Gordon (1968) described his refinements of

this procedure, primarily by developing coding systems to organize the

wide-ranging responses into meaningful dimensions.

When a projective technique is administered and scored using pre-

arranged keys, a richer variety of information is available, in part

since numerous keys can be used to measure different aspects of

personality frc- one admIni.itration of the test, and in part since

grc-:.tcr flexibility in responding inherently provides the possibility

and probability of greater information. Nonetheless, projective tech-

j:.r1ues are anoii tho mor.e covert devices, and the latitude of response

is so var.'," tat valuable. information may well be lost when scoring

them "obj:ctiv')..," ". ,c as a stimulus in an observation period,

a projectivLw t..c:nque cac intensify the setting and provide some

structure and co. onai.i ty for the observer, as well as practically

i-. r,.ni..':'. sono 3{iUinru level of activity. 7zclosive use of this

technique is !': '.rally not indicated, however, due to the situation

gcn-rilly b inSr pr'-'eived as unnatural, thus allowing for situational

reactivity a.n! btc

Sor.. of tihe earliest work with use of expressive behavior as an

index in pnre<]et-v-+i e:'vch wa'; undertaken by IHuntley (1940), who had

subjects rcpond to *1.' -,,.graphs of mannerisms and gestures of body

parts. In order to counteract theoretical concerns about cross-cultural

mannerisms, or x-linked differences, Huntley provided his subjects

with unidentified cropped photographs of their own mannerisms for eval-

uation and had the subjects thus unwittingly rate themselves, providing

and index of self ideal discrepancy.

Inferences of self concept have been made from protocols such as

those already discussed and from systematic observation. The latter can

involve a prearranged formal situation, such as a situational test

(Cronbach, 1960), a less systematic but directed session, such as most

interviews, or it can be undertaken in a more natural setting, such as

the work by -qar,:'r (1968). In the work reported by Combs and Soper

(1963), all these procedures were combined to produce a technique known

as observer-a-instrument. An important consideration when accepting

inferences from one person about another is the possibility of system-

atic rater variability stcxnin'- from attributes other than those

relevant to the study. Bossom and Maslow (1957) explored this poisi-

bility with relation to raters' feelings of psycholo ira?. security.

Measuring self concept in children is a special case. It is a

truism that childhood is the period of most rapid development, psycho-

logically as well as physically. It follows, then, that self concept

is most malleable during this growth period, a conclusion vhich

Perkins' (195&-) study validated. In his chapter on "Growth of Self,"

?urlhny (1970) reviewed and expanded upon findings with implications to

early experiences and the growth of self concept. Evldencr.- of rela-

tionships between self concept measures and school achievement in the

early school years was presented by Shaw and McCuen (1960), Comibs and

Soper (1963), and Ozehoskey and Clark (1970).

Problems in measuring children's self concepts, especially self

reports, are numerous. In such a rapidly changing system, lack of

reliability can be attributed to "real" changes as well as to errors of

measurement such as faults within instrument design. Extraneous vari-

ables which could influence self report include: tendency to mimic

others' responses, inability to remember or comply with testing instruc-

tions, inability to discriminate shades of meanings within the scales,

or even inability to read with sufficient comprehension to understand

the items. These latter considerations have led to modifications such

as those found on the "I feel me feel" test (Yeatts and Bentley,

1971), in which scale points are indicated by faces which range from

very happy to very sad. Instructions and items can be read to pupils

to allow for some reading level variation.

Projective techniques, such as the Machover Draw-A-Person test

(Machover, 1949), are oc.ccwacnally used but the defense mechanisms

which the adult protective measures seek to circumvent are only devel-

oping in childhood (Sullivan, 1953) and the rationale for the use of

child protective techniques is thus "iJd. Although the individual's

suspicion and curiosity about the ulterior purpose of the test should

be less with younger groip)P, a counterbalancing consideration is that

child protective techniques arc *l:i:.p a more sul.erflcial and situa-

tionally oriented perceptual field, which may be further confounded

with motor skill development.

Relationships Betuen rprceTtu al Variables

Detailed and systematic conclusions are difficult to arrive at

in the research revio wed due, in part, to the abundance of measurement

techniques (Gordon and Co::,b, 19 Il Wylie, 1961; Diggory, 1966;discuss

this problem). Efforts to reaate different measurement techniques are

seen to be nt-ede.- to reduce the conceptual space presently in evidence

in phenomrenolog-lc.l research, to provide bases for comparing studies

which have used different instrumentation and thus to provide a feed-

back mechanism for theorists. In a study on college men, Calvin and

Holtzman (1953) obtained adjustment ratings on each man in a fraternity

from the MPI and also inferred adjustment scores from each man's com-

panions. Inference scores were taken from the companions in the form

of answers to rather global questions about the personality of the rated

person, such as his ability as a leader, and his tact and social under-

standing. Self report scores were taken in basically the same proce-

dure. The tendency to view self in an essentially negative way was

found to increase as adjustment decreased.

.'sin- adjustment techniques to arrive at a self concept index,

Storm, Child, and Frank (1956) found no relat on--hlp between Thematic

Apperception Test (TAT, Mlurray, 1943) scores and self ratir'- or child-

hood experiences. Low correlations wire found by Cowen et al. (1955)

between two standard self report instruments (Bills' IAV and Brownfain's

technique), both of which purported to measure self rc ,,... Cooper-

smith (1967) found low moderate correlaions between a self riort index

and a different, rather global, test he devised for teachers to infer

student self conrc-it-..

In a factor analytic investigation of slf r'.Ur.s; teacher

ratln-s and achievement scores which included some perceptually oriented

variables, Cline, Richards, and Needham: (1964) found internal consis-

tency within each rating technique but few high correlations of

parallel items from puTil to teacher.

Perkins (1958a) found correlations from 0.04 to 0.71 between

teachers and pupils on Q sorts of self referent statements for various

teachers ratin-; various pupils, no teacher rating the saiie student as

another teacher. Of the 50 items used, however, only five referred to

school activities, and only three of those five refr-rred to the pupil's

role as a learner, so the index could best be called a global self con-

cept rating.

In an attempt to experimentally confirm the strong theoretical

arguments in favor of the independence of self report and inferred self

concept indices, Combs, Soper, and Courson (1963) obtained indices on

59 sixth grade pupils from experienced observers on 18 items relating

to attributes of importance within perceptual psychology. The same 15

items were given to the sixth graders so that they might rate themselves.

Correlation coefficients were calculated for each item between the self

r", t'.,-' and the inferred rating:. Coefficients ranged from -0.20 to 0.34,

with an adjusted mean of 0.11 Indicating little self-other correspond-

ence withinn any one item. Single-item indices of personality are most

likely to iicli!' l.,,-e components of error variability, and means cal-

cul ted :f,-,: f correlations obt--in ed between such indices carry forward

th' error Instatd of balmancin; it out. SIelf concepLt indices obtained

fro, .-multliple unidirmensional samples of overt behavior are much less

lIl;y to include .-neo error variability.

A possible limitation to int..'cIrietation of the above results

could arisc if curvilinear relationships were to be found between in-

fe': c'i ratings and self r in's. An interaction of scores a3si-ned

by t chers and self assigned score was found to be a major source of

vaia.bidlity in a tXrnposed factor analysis of pupils when using both

self *, --. and inferred self concept measures (Graves, 1971?l). Curvi-

liirr:'.;r of these measures will be further explored in this study.


Denl-;n and Method

_______--.s Used in the Ltu,'v

The measures of self concept used in this study included the

Florida Key (FK), developed by Purkey, C.t. -., and Graves (l97?l); a form

of the sacae instrument modified for pupils (1.-7); the 25-item Cooper-

smith Self-Esteem Inventory, short for-m (SEI, 19607) and a three-item,

school-oriented subscale (SSEi) of the SIJ. Pi ?K is an instrument

developed to infer az-n:ct of leaxmer self concept (see At: ?rd.x A),

and is comprised of 18 observable behaviors rated by teachers zr.'nr3'r.n
to frequency of occurrence on a six-.point scale r..r.: r. co

(never) to five (very often). The 1T7 is sinlzr to the ... (see

AppendiJx B) except for rewording of the directions so that students can

give self reports of their perceived frequency of performing the same

18 behaviors. The S-I (see Appendix C) contains 25 statements which

the pupil is asked to evaluate as being "like me" or "unlike me," and

the 3SS; is the subset of SEI items 2, 17, and 23.

Previous work by Purkey et al, (1971) established that the FK

measured four areas of learner self concept, labeled relating, asserting,

investing, and coping (see Table 1). Interjudge and split-halves reli-

ability indices were high, averailn: 0.84 and 0.93, respectively.

Validity was judged high by a panel of educators and psychologists.

TAL2L' 1

Rotated Factor Loadings for Florida Key
values s below .400 omitted)

Item lRtinnr Assertir.2 Investing Coping





. 616









. 800

. 772






. 501

. 562


. 448

. 717




Description of Children and Schools

These instruments were administered to all children (N = 249) in

the fifth and sixth grades of two elementary schools in North Florida

in the spring of 1970. One school, hereafter called the innovative

school, was locatcd in a Northeast Florida suburban town. The 97

students there were predominantly white middle class. The depio.-rathic

breakdown of the 352 pupils at the other school, located in North Cen-

tral Florida and called the conventional school, was approximately 70

percent white and 30 percent black children, ranging, from, lover -iddle

to nJlc class. Innovations at the innovative school include team

teaching, flexible scheduling and a particularly heavy emphasis on the

personal development of each child, Classes in the conventional school

were copar-jqbcntalized by subject matter and ability level in the fifth

and sixth rc-a...

Deo -.-, and Method

Administration of the self report instruments (PFK and SET) wras

undertaken by the teachers of the innovative school, usin- the instruc-

tions of the investigator, who administered these instruments himself

at the conventional school.

The data analyzed included the 18 item scores and the four fac-

tor scores of both the FK and PFK, as well as the SSEI subtest score

and total score of the SET, giving a total of 46 scores for each stu-

dent. The factor scores for the FK were generated using the procedure

described in Cuertin and Bailey (1970). Factor scores for the PFK used

the factor score coefficients of the FK to assure logical comparability

of the FK and PFK measures.

To deocz;:-ine the relationship between the Iteas on the FK and

those on the FPK, particularly between tho3e items in which the only

difference is the person reporting the score (i.e., the items which are

paallel on the two tests), a cross-correlation of the 13 7K items with

he 18 Pa' items was 7erforned, usin-.: all subjects. In this way,

students' behaviors perceived in the same way by students and teachers

(perceptually con-ruent items) were identified. Perhaps more important,

itens on which students and teachers disaTreed as to the perceived fre-

quency of occurrence (perceptually discrepant items) were identified

and analyzed.

Factor scores, as more reliable measures of general aspects of

l1arrner elf concept, were corL.ired between the FK awd the PFK. The FK

factor scores were also correlated with the S2I scores. This analysis

deter-'.ned the decree of relationship between the measures of inferred

self concept as a learner and the self report measures.

A possible contaminant in the previous procedure could be present

if peroo-.s scorin- centrally on an FK scale were to, by some psyco-log-

ical i.y:- ic such as identification with a higher status group, score

lower o: the corr',.zcn..lin.c FK than pupils scoring either high or low

on the .1 scale. This ?o-obi!.ity was checked by mrnnitn.: a curvilinear

-re'r--or a-ralysis cn the parallel factor scores of the FK and FP.K,

and or',.e eiiiht F-3 combinations.
To explore possible differences between sexes at th- ^rade levels

I i.? a. well as differences between schools such as those found by

Pur:'::e, S'taves, anid Zellner (19?0), a, t,.:o-way factorial d kL-, analysis

cf v.,rl-.nce was performed on the ten indices of self concept (four

factor scores each on both the K." and PK, plus the SEI and S..: scores).


Thie folw r hypotheses we-r checked for ail tVn he-a!ures.

(-) A sinifica.nt difference will be Tfound between the means
.f the two schools used in this study.

(2) A ri'nlrica-nt difference will be found between the means
of x.oys and A-7rls in this study.

(3) 7?o significant Interaction will be found between sex
within school means.


Analysis of the Data

Relationships between Techniques

To ascertain the r;-?ree of relationship between the items of the

PK, the PFK, and the school and total S3I scores, a correlation matrix

was ccncjted. The P,! system 360/65 of the University of Florida

Co-r itinr- Center was used, in conjunction with the Education Evaluation

Library cc-nutcr programs. For interpretation of these data, the 99

percent level of confidence was chosen to afford some protection

against the commission of a Type I error (see Table 2).

Results indicated significant positive relationihlp3 on 11 of

the 18 variables of the FK and PFK. The highest relationship was

indicated by a correlation coefficient of 0.32 between teacher and pupil

ratings on variable five, pertaining to getting along with the teacher.

Seven variables, despite their common origin, had correlations which

could not be considered different from zero. In examining the correla-

tions between the school factor of the SEI with each of the FK items,

eleven coefficients were found which differed significantly from zero.

The highest of these (r = 0.29) was between the SSEI and teacher ratings

on variable 13, pertaining to speaking in front of the class. Comparing

the FK items with the total Self-Esteem Index, eight coefficients could

be considered larger than zero. The largest of these was the correlation

of the SEI.with variable seven, which asked the teacher if the student

reads in class.


Item Cross-Correlations of Florida Key,
Pupil Florida Key, and Self-Esteem Indices

z;,.ber .2:,iPFK ;Y,/SSEI FK/SEI PF'K/SSE! FK/T












1 8
~*p .01




. 22*











































S.16 -




















. 18"








. 18"









Relationships between the PFK and the school factor of the SEI

showed only three coefficients significantly different from zero, the

largest of which was again variable 13, the only variable which had

correlation significantly 1-irgor than zero in all five categories

cxaincd, C'.i:v ':?,.-.r-.r- that this latter analysis was between two self

repu rt indices, this low ,-3,.--rme of ..-lationship could be attributed to

a fun-'l:'ental difference between classroom behaviors and self-esteem

as oc ed, with school, or it could have been due to the limited amount

of qriablity available in the three-item school subscale. In com-

p.l.:.-... the tri.h the full SSI scale, fourteen coefficients appeared

Phh '.*-`- .'-_-- icantly ];r.err than zero. The largest of these was

vri.ble fouir, getting along with other students, which had a correl.a-

tion coefficient of 0.31 with the total SEI.

Factor scores were calculated for the four factors of the FM,

and the coefficients which were thus generated were used to generate

coni --':.. factor scores for the PFK. The SEI scales were carried for-

waxdt, ;'. these ten variables were intercorrelated,

nThe corri-Ilc-i..r' between the inferred self concept factors and

the profcesed self concept factors have been presented in the principal

diagonal of the matrix in Table 3. Two of the four factors showed

coefficients which could be considered significantly different from

zero. Thio magnitude of these correlations was sufficiently low to

indicate that these measures were scored differently by the teachers

and by the students. In order to check for a possible curvilinear

relationship between the inferred scales and the self report scales, a

polynomial fit was attempted between the dependent variable (inferred

self concept) and the independent variable (self report). Twelve


intercorrelations of Inferred and Professed Learner
Self Concept and Self-Esteem Scores

Inferred Inferred Inferred Inferred
Relating Asserting Investing Coping SSEI SEI

Relatin- .22* -.02 -.02 .02 .15* .29*

Asserting -,19* .19* .08 -.03 .17* .19*

Investing -.05 -.10 -.04 -.03 -.08 -.11

Dro:fe ssed.
Copin? .09 .05 .01 .14 .05' .0

SS3I -.041 .22* .14 .08 1.00 .59*
,SFT C9 .0
SE ,09 .10 .00 ,03 .59 1.00


analyses were thus perfored using student scores for all variables,

four with FK variables predicted from PFK variables and four each with

the FK variables predicted by the SIT and SSEI variables. Tlr-: of the

12 analyses indicated significantly nonlinear components in the :regres-

sion model (see Table 4).

Comparison between Schools

A set of analyses of variance, two-way factorial d'. vra, :as run

on the ten variables used within this study. Sex was identified as one

source of variability, school as the second source, and the school-sex

interaction as the third source. Summaries of the results of these

analyses were listed in Table 5 for the four factors of the Florida Key,

in Table 6 for the four factors of the Pupil Flt. di Key, and in Table

7 for the results of the two Self-Esteem Inventory indices. For these

analyses, the .05 level of significance wes employed, as fewer measures

were analyzed.

On the measure of teacher-perceived rr-latin-, differences slgnif-

icant at the .05 level were found for both sex and school. On this

measure, girls tended towards hl-h,.r scores as did students in the

innovative school, with girls in the innovative school scoring the

highest. Only a slight difference separated the mean of the innovative

school boys and the somewhat higher mean of the conventional school


When determining teachers' differential perceptions of asserting

in this population, no differences were found between sexes or schools.

On the dimension of perceived investing, however, a distinct difference

between the schools emerged, with the students in the innovative school

rated as having a stronger tendency to invest of themselves


Curvilinear Components of Threec Self Report Variables
Predicting Inferred Self Concept Variables

df SS

Teacher -e'.-tI-7 from Pupil
Linear Term
Quadratic Term

.a.' -r A -:rtir- from SEI
Lir,- ir .ern
Quadratic Term

Teacher >.ve-tinez from SEI
,:a'ecatr Ter-
Qi:adtcatic Term





eta = .26

eta .16

eta = .19

T'otAl deviation for all above analyses = ,

21.9, df = 248


Summary of F-Ratios and Cell Means
for Analyses of Variance of FK Factors

Source df Relatirn'- Asserting Investing Coping

F-Rat os

Sex 1 4.209* 0.870 0.689 12.045**
School 1 4.461* 0.001 9.519** 0.312
Sex X School 1 0.092 0.700 1.798 0.426

Error 245


G.C. 0.035 0.004 -0.121 0.272
G.I. 0.268 0.112 0.104 0.118
B.C. -0.269 -0.008 -0.186 -0.254
B.I. -0.043 -0.120 0.382 -0.242

Fchoo. Code: Conventional (C); Innovative (I)
Sex Code: Girls (c), Boys (B)

*p < .05
**p < .01


Summary of F-Ratios and Cell Means
for Analyses of Variance of PFK Factors





Sex X School



School Code:
Sex Code:







Conventional (C); Innovative (1)
Girls (G); Boys (B)

*p < .05
**p < .01




10. 346**




Summary of F-Rt.Uos and Cell Means
for Analyses of Variance of SEI Scores

Source df School Total

Sex 1 0.061 0.058
School 1 0.551 2.279
Sex X School 1 0.043 5.401*

Error 245

G.C. -0.066 -0.206
G.I. 0.059 0.289
B.C. -0.006 0.063
B.I. 0.064 -0. 042

School Code: Conventional (C); Innovative (I)
Sex Code: Girls (G); Boys (B)

*p < .05
df (1,245)

psychologically. :oys in the innovative school were the highest

scorers on this variable, but boys in the conventional school were the

lowest scorers, so no sex-linked differenc-cs were found.

SOn the question of how well the individual is perceived to cope

with his school environment, a strong sex-linked difference was found,

with girls perceived to be more efficient co-.^;. than ii,y-.

When asking the students to rate themselves on these measures,

girls in both schools rated themselves higher on ro?-.tir behaviors

than either group of boys, with girls in the innovative school showing

the highest self rating. A significant school-sex interaction was

found because the girls in the innovative school scored highecst cz-ao the

boys in the same school rated themselves lowest on this dimension.

With the dimension of asserting, no significant differences were

found between the sexes or between the schools, nor s a differential

effect found anong sexes between the schools. With self-perceived

investlir-, a sex-linked difference was found with boys scoring them-

selves higher on this dimension than girls.

In reporting on tendency to cope, students hoe.. a strong tend-

ency to differentiate on the basis of both sex and school, with girls

showing higher scores than boys and the pupils at the conventional

school perceiving themselves as more effective copers than those in the

innovative school. Girls in the conventional school scored highest,

followed by girls in the innovative school, who differed only slightly

from boys in the conventional school, with boys in the innovative

school indicating the lowest scores on this dimension.

In self-esteem measures, no differences were found on the school

factor of the SEI in any of the sources of variability examined. On

the total self-esteem index, a significant interaction term was found,

which resulted from the girls in the innovative school scoring highest,

while girls in the conve rtonal school scored lowest, and boys in the

innovative school scored somewhat lower than their counterparts in the

conventional school. There was a slirht, nor'-Ainificant tendency

towalds higher total self-esteem scores in the innovative school.

The first hypothesis, predicting significant differences in

scoring between the sexes, was accepted five times, with the means

favoring the girls four of the five tieos. The second !iypoth^cis, pre-

dicting significant differences of scores between the schools, was

accepted four times, with the mearns favoring the innovative school three

of the four times. The third hypothesis, predicting no siilfli'arit in-

teraction of scores due to sex within school considerations, was

accepted eight times.


Sunnarv and Conclusions

In examining the cross-correlations of inferred self concept

items and self report items taken respectively from the FK and the

?FK, eleven of 18 variables showed correlations between measurement

techniques significp.ntly different from zero (range: r = 0.15 to 0.32)

at the 99 percent confidence level (see Table 2). AS this far

exceeded chance probability of occurrence, common perceptions of pupil

actions by teachers and pupils seemed the most probable explanation,

The ,. t sE of these correlations, however, were so low &a to indi-

ate that, at best, only about ten percent of the variability in a

cr obtain Usn, one method could be accounteJ for b.- the score

_*--:v- t'-' oth methodl. L'his distinction between statistical siznifi-

cance z.n :;-nitude was of ilnortancp hezi: du o to wo factors, '7hS

;irst f'-cItcr ;as s at the itens answ.ered were essentially the same for

fstud .t teacher.', The second consideration -,as that the relative

,.becT=ivi., n.! overt classrooms behavior orientation of the scale would

e-. x;;ae 3es roon for viriatlon due to !ti- aibhguity. More

si't pt:, te indication is that tere was a 'Jtti,. s1.:nlarity and a

......... i. fference between the -,ay fifth aLd sixth trado children

: thls st s,; them salvos jn .c'hol-relatc.'ctivities a-id the way

theirr: t-. % [' L ;.-.'-;t:-a.. the-., 7-.,'- reofer- 'jtions of Co~ri Solor, and Courson

(1963) as to the lack of similarity of self report and inferred self

concept seem to have been well borne out by this study.

An internal examination of the seven ite;.s which did not corre-

late between techniques c .lio.ied two items ech in the areas of relating,

asserting, and cor,.,, and one item in the arcm of inwv
most con.'-ru rce in terms of the number of iter:i associaLed with a fac-

tor which show statistically significant c,:orla Lions between techni-

ques was within asserntir, where six of -even ite.s were correlated.

This would seem to imply that the area o.f a r'.i: is one in which

pupils and teachers can be aost i.- rf .i, an libpli-ton which will be

discussed later when factor indices are coae :;io The least congruence

in these terms was within the factor of ;s .. weve ner either item

correlated between techniques.

Eleven of the 18 items on the PlY e '.+', ;,',icar.tly with

the school self-esteem measure (. ), ;iti.h corelation coefficients

ranging from 0.15 to 0.29. No items within the factor of teacher-

perceived relating were correlated with the S5,2' Among the 11 items

which did correlate were items six t:.c.. "r. nin uhich relate to dis-

cussion within class. One of the three items of the W,'2 was "I find

it very hard to talk in front of the class," thus the correlations

could be seen to relate in part to the in-class verbalizations of the

student. It is interesting to note, however, that on the PFK, these

verbalization items did not correlate significantly with the SSEI.

Eight items of the FK correlated with the total self-esteem

score (SEI), with a much narrower range (r 0.15 to 0.20) and at a

lower level. No pattern could be discerned for the correlating or non-

correlating items, which was in keeping with the results of the factor

score correlations in this area to be discussed later.

'ithln the PFK, only three items showed any association with

the school factor of the self-esteem inventory (SSEI). The absence of

correlzaion of certain PK- items with the 3SEI factor, namely the

veralvisation items mentioned earlier, gave an indication of the lack

of reliability of self-reported single-item or limited-item indices.

A sizable increase in the number of correlations (14) between the PFK

iten a nd. ,thc total S21 score hinted at the increased stability of the

lo '. SI calo, Two of the four items which did not correlate signi-

ficantly with this scale were the items associated with invcstinr.

The conclution should not necessarily be drawn, however, that invest-

ment cf slf is unrelated to self-esteem, but rather that behaviors

seen V: teachers as indicative of self-investient may not be seen in
the eaao 1' ";'L \r', zF.n l ..

'actor "'orre ere calculated for all students on the basis of

the PK factor analv;-:'. Such scores were based on responses to all

itens m-ithin the test, but were most heavily weighted for those items

which .ecre most rQt'rr',ly associated with corresponding factors. This

was unlerthkcn both to increase reliability and to reduce dimensionality

while increasing cenccality. Cross-correlations of scores for the four

factors between the two techniques of measurement (self report and

inferred self concept) indicated that two of the factors, relating and

asserting, had statistically significant relationships (see Table 3).

The factor of relating, when subjected to a polynomial analysis to

determine whether pupil responses were curvilinearly related to teacher

responses, showed a significant quadratic effect, raising the correla-

tion index (eta = 0.26, Table 4). Further, the factor of coping,

although not meeting the stated criterion for significance, showed a

correlation coefficient of 0.14, significant at the 0.02 level.

In connection with the earlier discussion of item correlations,

althoug-h n-ore a-. '_- variables were correlated across techniques

than variables of the other three factors, the relating factor had the

highest correlation (r 0,22) across techniques. The factor of

az^tir_,, however, with a correlation coefficient of 0.19, was in the
same general ra as relt It was considered likely that these

two areas were seen as less central to self than the areas of coping

and irn-eetin:., and were, therefore, less susceptible to perceptual


The lack of correlation between teacher-perceived investing and

self-professed inv^stin- was indicative of either lower rellabi4ty due

to the number and relative loadin-s of the items w4ich ccirinitd' to

this factor or sone differentiation between teachers' and pupils' per-

ceptions of pupils' investing.

The relationship of teacher-inferred re!l. :-- to self-professed

azs3rti_ -howed a statistically significant -.&-ative correlation of

9,1), fi .ia-inz that those seen as high in relating behaviors by

weathers ndced, to a small extent, to see themselves as less a.sserting,

a... vice versa. trowever, when teachers evaluated assertion this was

>rs.l ated to pupils' self-oerceptions of rclatir--.

-: factor scores -'ere correlated with the S31I and the SSl,

only one correlation was found to be s-Latistically si-nificant, that of

te..c:-cr-inf ered a;.eti]n- and the SSEI (r 0.22). This indicated that

uc- :-.-tvit.es as offer.n- to answer questions in claz-s, and joining inr

'-1 ;l activities, were related to items such as beinm discouraged with

scol, an ire unset while there. Had -this correlation been sub-

-Luiallv >hi~~ier, or ija6 several factors correlated with this index, it

could have given a further indication of the dimension'ality of the

school self-esteem index (SSEIL). Curvilinear analyses of these data

showed the 371 to have significant quadratic components when used to

predict scores on the inferred asserting and investing factors of the

FK. The most plausible explanation of that finding was considered to

be that after a certain low in teacher opinion is reached, the influ-

ence of that teacher's opinion is discounted by the child, who seeks

other sources for maintenance and enhancement of self. The potential

for such a dynamic to generalize and transfer to other teachers or to

th- sch-c.ol should be obvious,

The self report factors of relating and acsertirn: froa the PFK

w-re s.-nificantly correlated with both the SSEI and the SEI (r 0.15

toc2. ? ,, 3), but professed i:..-estin- and coing_ were not. This

quite possibly reflects the absence of accomplish:-ent-criented items

in the :sI.

-air!, ma-nitudes of the correlation coefficients were only

.. 'u to account for, at most, five percent o the variability

of scores t.aken by one technique by considering scores from the other

* ::.-"-ue, w:.ch further confirmed the opinion that self r"p-ort aea-

suras do not measuree the sa.ne attributes that inferences of seif by


>'o~ iscs_ between Schcols ana. Sexes

T:. ,etei'n '.he difference.; between sexes, girls were per-

eid o-,bIy teachers and by themselves as higher in the areas of

IeloM[ : (see .r? ,l 5, 1, and ?). -1e nore advanced matu-

r ....l.... :ils at Itis a"-- level and th social expectancy fcr

; :t,, 'to of ",*iore social iv ox. ntpn. nobab?.l v'r-tr-b;ited. n i r oduc. .nz

this result, which was in concert with the findiln's of Perkins (1958a),

Davidson and Lan, (1960), and Bledsoe (1964). rc.- by school inter-

actions found in the areas of self-professed r',,1 -. lr', and on the SEI

indicated that factors associated with the innovative school were in

turn associated with higher scores for girls in that school as compared

to girls in the conventional school. Boys' scorc5 in the two schools

differed only slightly on these two variables, with theec in the con-

ventional school scoring hls;her by about onc-tenth of a standard score

in both cases.

Differences between schools were found in teacher-perceived

relating and investing and on pupil ratings of ),I t-' and c n'"

The innovative school teachers sa'r their puIpil. a1 po A ... ing. a higher

fzcji.c'!icy of csociallzin-1 behaviors, as measured by the T.'..-1i,-, index,

which perceptions were mirrored by the pupil; tIe'... el'., relative to

the conventional school. Pupils in the conventional school rated them-
selves higher on cping activities than the innovative school c'lldrc:n.

The present investigator felt that this probably reflected the lack

of a standard reference point more than any-tr,li- else, as he observed

more coping behaviors in the innovative school, albeit ".i :couJh informal



The absence of a high level of conformity between teachers' and

pupils' perceptions of the frequency of occurrence of overt, school-

based pupil activities can be attributed to distortions of the percep-

tions of these activities by teachers or pupils or both, or to pupils

playing different roles in different classrooms and the related situa-

tion of their responding to the questionnaire on the basis of an

.v.ri azsessnient of their school behavior, not necessarily repre-

-:-:3t oft their behavior within any one classroom. Another

interpretation could be differential interpretations of the scale

LitervI1s by different children, but the use of frequency assessment

wa!s chosen for its relative specificity.

If the explanation lies within the distortions of teachers'

ercc.ptiois:, the necessity for increased training, for sensitivity virthin

teacher education and in-service training. programs, such as that called

for by CoimbS, Avila, and Purkey (1971), is obvious. If, on the other

har.i, th distortions lie predominantly u-ithin the childrens' self-

percentionrs, then the sane conclusion nay be drawn, even though less

ob:i'iorly. The reasons for the later conclusion are twofold. More

accuraT.e- and effective self-awareness on the part of the pupil is

likely to cone only throu-h the intervention of the teachers, and thus

the teachers nust be cognizant of the fact of distortion, the factors

whichh a.re acscciated with the distortion, and effective attitudes and

e-thors to core with children's distortions of se]f-perception,

'ut-oire research will include a study to determine the extent of

ta.s:a*;s' -._-.ltivity to perceptual distortions of pupils, using a

..i .c..tio nf 7rownfain's technique (1952). Also to be examined is

.e o.:,st3cn of how well teacher .perceptions of pupils Qre related to

t*! z-pece-.L-onzs of trained observers, and whether these results hold at

(U'fers'.t 'ride levels and smith measures of different apects of the

self, r',e--.il,li;abiilty of these results to other populations of pui.'1.il

rc .cfi b- co: rciiser. by differences in "Ihe factors influencirg-

c'.i -:i'.rt pThaona.:.uities, Iotn school and, non-school, as well as factors

;-:1:2 : i; en? ...-"'.ls and teacht.r2' ocrceptions of school and scnool-

-' .":, d A hr /.V.oi s.

The value of the more relaxed and child-oriented atmoshpere of

the innovative school was strongly indicated here, although the design

of the study did not allow for conclusive statements on this point due

to the confo-nUr,-' of raters within schools. Since investing behavior

'., be seen as of a hi,.hnr order in terms of Maslow's (1943) hierarchy,

the large difference in inferred Jnvestlnr, behavior in favor of the

innovative school indicated a very pro-iisin- payoff for the emphasis

placed on efforts to enable the students of that school to inte-rnally

motivate and individualize their own learning.


Florida Key

This scale is to assist the teacher in evaluating how the
student perceives his or her "learner" self. Please select one
of the following answers and record the number in the blank space.

".". _0 [ .;"', A'"T.: 2 ALLY: 3 OFTEN: 4 (T'F: 5

Iaicf of ctu;.c.At to to cva.uxted'

Compared with other students his age, does this students

1. get along with other studc-nts? ____

2. get nio"-. ;ith the teacher?

3. keep cali when tVin-s go ironr?

4. say good things about his school?

5. tell the truth about his school work? ____

6. speak up for his own idea? _____

7. offer to speak in front of the class? ____

8. offer to answer qiurztions in class? ____

90 ask Trvrn-ful questions in class? _____

10. look Cpeople in the eye? ____

11. talk to others about his school work? ____

12. join in school activities? ____

13. seek out new things to do in school on his own? ____

14. offer to do extra work in school? ____

15. finish his school work? ____

16. pay attention to class activities? ____

17. do his school work carefully? ____

18. read in class?



,:!' Florida Key

This scale is to help you in thinking about yourself. Please
select one of the following answers and record the number in the
blank spaces.

1". 7 0 "r^':: 1 AWHILE: 2 ALLY: 3 O.FT'": 4 OFTEN, 5

Your :ui~e

.;--. d with other students my age, do I:

1. get along with other students?

2. get (.".r,- with the teacher?

3. keep calh when things go wrong?

4. say good things about my school?

5. tell the truth about my school work?

6. spe-ak up for my own ideas?

7. offer to speak in front of the class? __

8. '. :r to answer questions in class?

9. a3k meaningful questions in class?

10. look people in the ey-?

11, talk to others about my school work?

312. join in school activities?

13. seek out new things to do in school on my own?

14. offer to do extra work in school?

15. finish my school work?

16. pay attention to class activities?

17. do my school work carefully?

18. read in class?



Self-Est+. inventory (:.,)
Stanley Coopersmlth
University of Califorrjia, Davis

Please r.ark each statement in the following way:
If the statement descril'-es hw you usually
v-) in the colnn "LIKE 1 ..
If 'e statement dops not describe how you
a ceck (v) in the column "UNLIKE ME."
There are no r'.-',t. or wrong answers.
Ex .y T'rm a hard worker.

feel, put a check

usually feel, put

L -E E -7[ !j9

1. ~. cft

(7? i a t v'ery, hard to talk in front of the
3. There r- lot-s of thinn-s about myself I'd
a:.' if I could.
-- ca.":a.": v u 7y nind .-ithout too mUCn trouble.
S 'r lo- of fur. to be with.
SI-'t. *et easily at hone.
it taes .e a long time to iet used to anything

(. -.;*:T o-Is-- W'th' kis :riy own age,

C,,i.- In .er.' as!i .y,
] ...... s extec t oo n;uch cf me.
-. " -..... ._ !r. my -e,
Se s.al: follow -y ideas,

... -. ..I- .. ... ... ... / ..I... S' .V

S'; o', like than I an.

(-r '-r p-^eo



I __________________________
*----1- --



..lro-.-t, G. W. pattern and growth in personality. New York: Holt,
' er, R. C. environmental psycholoy. Stanford, Californiat
Stanford University Fress, 1968.

Battle, J. Relation between personal values and scholastic achieve-
nent. Journall of Experimental Education, 1957, 26, 27-41.

Dills, .r. 2., Vance, E. L., and McLean, 0. S. An index of adjustment
and values. Journal of Consulting Psycholoary, 1951, 15, 257-

?ledseo, j. C. Self concepts of children and their intelligence,
achievement, interests, and anxiety. Journal of Tndividual
> : c-,', 1954, 20, 55-58.

"1c;k, J. and Thomas, ?. Is satisfaction with self a measure of adjust-
men-.? Jou-naJ. of Abnornmal and Social Tsycholo, 1.:5, 51, 254-

:;so:(.'S, J. and ?asiow, A. H. Security of judges as a factor in i.-res-
sions of warmth In others. Journal. of Abnormal and Social
'Scholc-y, 1957, / 147-147c,.

1". 2., Thonas, S. B.- and Faterson, A. Slf concept of
ability and school achievement. Sociology of Education, 1964,
2?^ 271-278.

:rc" .fil.,I J. JT Sta'ilit' of the self concert as a dimension of
p'. 3onrlli:'.'. Journal oI Abnornal arnd .ocial -rs. ,-_-l-, 1952,
"-Z 597-606 .

., -.i-n, ,'. The relationship between self concept and the
z,-c.. or absence of scholastic underachievement. Journal of
,i.- -'al !;vhol o--, 1962, 18, 181-182. -

-.:, J. 3. SocJ.l psvckolaov d .d perception. In .. .'. Maccoby, T.
: e. re-1'"onb, and 7'. R,. !'y_ ,.l (rds.), Tenadtn'_ in social
z:, ch'o C'ew york; Holt, _3.

',' -::.], . 7T, ard. Tel-r, S. t fr...t.l.io.s into the self
*'.-e,,*' Jol.lrrl of Teosn!o1ity, lc..,O, 4 83-493.


Butler, J. M., and Haigh, C. V. Chianecs in the relation between self
concepts and ideal c.oc:npL. In C. R. Ro-'ers and R, F. Dymond
(Eds.), PsvchotheraDnv nd personality chranre. Chicago
University oi Unic o }rc!s, l'.".I

Calvin, A. D., and Hlolt.zman, h. H. Adjustment and the dircreanrcv
between self co,, and inferred self. Journal of Consultin
,Psycho!oY 1953, 1_, 39- 4.

Campbell, D. T., and ".L- ley, J. C. ri-.-1'-.nital and qu -1:i-experimental
designs for research in teaching. In N. L. Gage (Ed.), Handbook
of reach in tc,.;, Chicago: Rand ic7.Nlly, 1963.

Cline, V. D., ,ic'Iads, J, ;., and Nf L.i, W. Factor analysis of self
ratin-s, teacher ratings; and indices of achievement in high
school science, Jo'uirnal of Educctional Research, 1964, 58, 10-

Cole, C. I., OR.' 3. ., and Hinkle, J. Nonlinearity of self con-
cept discrepancy; the value dimension. olo-iP_-I ]'". 'mts,
19S7, 21, 58-60.

Comnbs, A. W. Intelli,-:-., fro-n a perceptual point of view. Journal of
Abnornmal and So .' ) o, 1952, 662-.673.

Combs, A. ., AvilJa, ,, T vnd P1r1ley, V, W .r .... i r gj
co ncrt. for t hc' eIC1 .c rofiaov Bc, t3.:,c. i',r,1, nl. IBacon,

Combs, A. W., and r.y-., 1). Individual behavior (rev. ed.). New Yorks
Harper, 1959.

Combs, A. W., and Soper, D. W, T7e relationrhin of child perceptions
to achievcr'ent aand behavior in inthe early school years.
Washington, D. C.: U. S. Depart;icnt of Health, Education and
Welfare, Office of P'ducation, Cooperative Research project no,
812, 1963.

Coabs, A. W., Soper, D. IT., and Courson, C. Measurement of self concept
and self report. Educational and T'ycholo-.lcal Measurement, 1963,
23, 493-500.

Coopersnith, S. A. The antecedents of self-esteem. San Francisco: W.
H. Freeman, 1967,

Cowen, E. L., Heilizer, F., and Axelrod, H. Self concept conflict
indicators and learn in-l'. Journal of Abnormal and Social Psychol-
5,1955, 51, 242-245.

Cronbach, L. J. Essentials of psychological testing (rev. ed.). New
York: Harnper, 19-60.

Crowne, D. P., and Marlowe, D. The aprovLtl motive: S'uails in
evaluative dependence. New Yor;: '.,Iley, 1964.

Davidson, H. H., and La.r-, G. Children's pTrcer't tons of their
teachers' feelinrvs towards then related to self perception,
school achievement and behavior. Journal of Exrn riirnit: I
Education, 1960, 29, 107-118.

Diggory, J. C. Self-evaluation: Coi,.niri and studies. New York: W(iley,

Eng-el, M. The stability of the self concept in adolescence. Journal
of Abniormal and Soc:l PsycbooTy, 1959, 58, 211-215.

Fiedler, F. E., Pc- 1, J., Jones, R., and Hutchins, :. Interrelations
a.non- neasui'es of personality adjustment in nonclinical popula-
tions. Journal of Abnoa.il and Social ? .ho].o~, l0-n, r,

Finkv M. 1. Self concept as it relates to academic underachieveennt.
California Journal of .,-: '1 Research, 1962, 1., 57-62.

Flaherty, H. R., and Reutzel, E. Personality traits of high awd low
achievers in college. Journal of ;.u.:a I Q Research, 19o5,
58, 409-411.

Gergen, K. J. Personal consistency and the presentation of self. In
C. Gordon and K. J. Gergen ( Lds.), The self in social. interaction.
Vol. 1. New York: Wiley, 1968,

Gordon, C. Self-conceptions: Con.ri-i:'.A rations of content. In C. Gordon
and K. J. Ckr!en (Eds.), 'The self in social interaction. Vol. 1.
New York: Wiley, 1968.

Gordon, I. J., and Combs, A. 1. The learner: Self and perceptions.
Review of Educational Research, 1958, 28, 433-444.

Graves, W. H. A transpose factor analysis of a learner self-referent
behavior inventory. Paper presented at annual convention of
Southeast Psychological Association, Miami, 1971.

Guertin, W. H. Transposed analysis of Kelly's personal construct pro-
ductions. Paper presented at annual convention of Southeast
Psychological Association, Miami, 1971.

Guertin, W. H., and Bailey, J. Introduction to modern factor analysis.
Ann Arbor, Michigan: Edwards, 1970.

Hathaway, S. R., and McKinley, J. C. A multiphasic personality schedule.
Journal of Psychology, 1942, 10, 249-254.

llebb, D. 0. The American revolution. The American Psychologist, 1960
l, 735-745.

Hilgard, E. R. Human motives and the concept of the self. The American
Psychologist, 1949, 4, 374-382.

Huntley, C. W. Jud-mnent of self based on records of expressive behavior.
Journal of Abnormal and Social Psychol o, 1940, 35, 398-427.

Jersild, A. T. Child r- vcholo. v. rnrle-1orj Cliffs, T'vew Jersey: Pren-
tice-Hall, 1I:0.O.

Kelly, G. A. The psychology of personal constructs. Vol. 1. A theory
,of* Yo), i:. Toxkon, 1955,
Koffka, K. Principles of r-stolt psycholoy'. New York: Harcourt, 1935.

Lamy, M. W. Relationship of self-perceptions of early primary children
to achievement in reading. In I. J. Gordon (Ed.), Human
Development. Chlca.-o: Scott, Foresananl, lio.

Lecky, P. Self-consistency (rev. ed.). Carden City, New York: Double-
day Anchor, 159.

Mach.cvcr, K. T'.'-onplty Projection in the draing of the huran fiLure.
Springfield, Illinois: Chax'lr!; Tho:ias, 749.

Maslow, A. H. A theory of human motivation, Psychol lc-1 Review.
1943, LO, 370-391.

Merenda, P. F. and Clarke, W. B. Differences in results of inferential
self-measurement in self concept analysis. Ferceptuatl and >tor
cvi1Ll, 1967, 25, 317-322.

Moustakas, C. E. True experience and the self. In D. E. HarLachek ("d.),
The self in growth, tetchin7, and learning. En-lewood Cliffs, New

Murphy, G. Personality, a biosocial interpretation of origins and
structure. New York: Harper, 97.

Murray, H. A. Thematic apperception test manual. Cambride, Massachu-
setts: Harvard University Press, 1943.

Neuringer, C. and Wandke, L. Interpersonal conflicts in persons of high
self concept and low self concept. Journal of Social Psychology,
1966, 68, 313-322.

Osgood, C. E., Suci, G. J., and Tannenbaum, P. The measurement of
meaning. Urbana, Illinois: University of Illinois Press, 1957.

Ozehoskey, R. J. and Clark, E. J. Children's self concepts and kinder-
garten. The Journal of Psycholagy, 1970, 7, 185-192.

Perkins, H. V. ?actors inf2.ercirc7 char.-e in children's self concepts.
Child Development, 19'5.9, 29 221-230 (a).

Perkins, H. v. Teachers' and peers' perceptions of children's self
concepts. Child Development, 1958, 29, 203-220 (b).

Purkey, W. 1., An independent std project for gifted underachievers:
"Project self discovery." Washington, D. C.: U. S. Department
of Health, Educztion and 'Welfare, Office of Education project no.
6-1734, 1967.
!' rke:-, 7. I' The search for self. FERDC Bulletin 4 ('2), Cainesville,
Florida. i'loridafaEd-ucaai-onal Research and Developnent Council,

Purkey, 7. I. Self concert and school achievement. 'nglewood Cliffs,
'"ew Jersey: Prentice-iall, 1970.

'urkey, *1 '.J., CaTe, J. *J., 8and3 Graves, 'W. The development of a pupil
..enavior invenrlry to infer learner self concept. Faper pre-
sented at annual convention of Anerican educational Research
Asso-iation, :ew York, 1971.

turkey ..,, rave, -'. -e., ard Zelner, .,. Celf perce-ptions of
.u..l. in xr ...n.... ele'.ent...z-' school. Ele:iertary School
JouI = l ^19?0, 71, 166-171.
.,., ,'. ., An. Thite, o. 2 3o rcc of the dlnennions of self as
o0' an-i._- self as process. The ,Jour.al of pIcholo0y, 1967, 67,
277' z ~5' ~

'-;5"erF, '];. r;,. "lij+_n11 ct-ce,_t therapy. 3oston; -ou-,.on-:ifflin, 1951.

?-0w.", '. ; role of self conceoDt in achleve-.ent. Journal of
Lxjerl-.ntal education, 1959, 27, 265-22..

.,'.t- :-, :. I[eels of aspiration in ?cu/lt''..cL'_x ucce. ."ul and unsuc-
.f...l... ch.i..ren. Journal of ASoroci1 an- Social 'sycholoy-, 194,0,
31, 45,}-536.

*"" .,, X nd McCufun, J. '. The onset cf academic underachievement in
L-ri&t children. Journal of Educational Psychologv, 1960, 51,

v' l. '-is need for a phenonenolo-ical s'-stem of psychology.
....s2 ',e~ w, 1 ..1, 48p 4V4_4.2.24.
-.,^l^.-^ ,eiw 19" Self ^-.^...
....", S., e oar- .. pe-rcepTxoe ons of socially disad-
. ...'A:t..", "... .ldr n. -.. ca . duca fioal Research J_.:rr. l. 196?,

Staines, J. 1. The self-pictture as a factor In the classroom. British
Journs! o.f _duca+tion-d Ts,-choloyv, 153, 2S, 97-111.

Stepi.enson, W, W. The study of behavior. Chicago: University of
Chicago Press, 1953.

Sallivan, H. S. The inter rsona. theory of 2 sychiatry. New York:
Norton, 1953.

Thorndike, P, L., and '-a-ar., S. Heasurcnent and evaluation in psychol-
o.y _lnd education (3rd. ed7. New York: wiley, 239.

Wylie, C. T7r self concept. Lincoln, Nebraska; University of
Nebraska Press, 1961.

Yeatts, p. p. Develcnnental clan-"es in the self report of iegro and
white children, ,-radies 3-12. FERDC -ullet 2),7 Gainesville,
Florida: 7lorida Educational Research and Development Council,

Yeatts, ., P., and 3entley, E. L. The development of a non-verbal
neasture to assess the self concept of youn- and non-verbal
chiildren. Paner presented at annual convention of American
educational Research Association, "ew York, 1971.


William H. Craves, Jr. was born in Dalton, Ceor,:la, on November

14, 1939. He moved to Daytona Bernch, Florida, where he attended elemen-

tary and secondary school. He attended Southern Methodist University,

Dallas, T'7.' in 1957-1958 and the University of the Americas, Mexico

City, Mexico, in 1958-1960, before enlisting in the United States Air

Force in 1950. During military service, he attended Indiana University

for one year. After discharge, he returned to the University of the

Americas, being c-ra-fted the Bachelor of Arts degree in secondary educa-

tion in 1965.

He taught mathematics for three years at Mainland Junior High

School in Daytona Beach and conversational Spanish at Daytona bLeach

Junior College during two of those years. In 1968, he enrolled in

Craduate work in psychological foundations of education in the Colle:e

of Education, University of Florida, Gainesville, Florida. He earned

the Master of Education degree in 1970.

He married the former Joan Lynn Dugand of Mexico City in 1966,

and they have two children. Professional affiliations include membership

in the American Educational Research Association, the Florida Educational

Research Association, and Phi Delta Kappa.

He is presently employed as Assistant Professor of Education,

University of Oklahoma, Norman, Oklahoma.

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

-illia.-7W. VFrey, Chairman ~
Professor of Education

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

T certify that 1 have read this study and that in ny opinion
It c" r-'.s to acce-ptable standards of scholarly presentation and is
f -' elqc>ate, in scope and quality, as a dissertation fcr the 1e-re?.
of Docto-o' cf 'hilosophy.

.rofzor A o education
Frofs-':or of Education

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

uillerno P. ascaro
Assistant Professor of


Thi; Ob L:,:;-"ion was submitted to the DT),an of the ClIr-ce of Fducatii',
1J t*cr. *;:'rdurfc Council, and was accepted as partial fulfillment of
IUe recm si< ts for the degree of Doctor of Philosophy.

Ki.ri h, 39??

Dean, Collegef education
Dea Y

Dean, Graduate School

~a',.. oF

N. I