Citation
Locus of I-E control expectancy and expectancy changes of disadvantaged mothers

Material Information

Title:
Locus of I-E control expectancy and expectancy changes of disadvantaged mothers
Added title page title:
I-E control
Creator:
Bilker, Larry Michael, 1940-
Publication Date:
Language:
English
Physical Description:
viii, 125 leaves : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Child psychology ( jstor )
Children ( jstor )
Control groups ( jstor )
Infants ( jstor )
Locus of control ( jstor )
Mothering ( jstor )
Mothers ( jstor )
Parent education ( jstor )
Parents ( jstor )
Social psychology ( jstor )
Attitude change ( lcsh )
Dissertations, Academic -- Psychology -- UF ( lcsh )
Psychology thesis Ph. D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis - University of Florida.
Bibliography:
Bibliography: leaves 121-125.
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
This item is presumed in the public domain according to the terms of the Retrospective Dissertation Scanning (RDS) policy, which may be viewed at http://ufdc.ufl.edu/AA00007596/00001. The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator(ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
022272229 ( ALEPH )
13578433 ( OCLC )

Downloads

This item has the following downloads:


Full Text











LOCUS OF I-E CONTROL EXPECTANCY

AND EXPECTANCY CHANGES OF

DISADVANTAGED MOTHERS















By
LARRY M. BILKER














A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY













UNIVERSITY OF FLORIDA 1970













ACKN OWLFDGEMENTS


This research was made possible through the cooperation of Dr. Ira J. Gordon, principal investigator of the Parent Education Project,* and his staff, particularly Dr. J. Ronald Lally, project field director. I am indebted to them for their ideas and their assistance in collecting and programming the data. Thanks are due also to Mr. Toni Williams for his excellent computer programming and consultation. I would like to thank my Supervisory Committee, particularly Dr. Louis D. Cohen, my chair-nan, for their intensive, comprehensive, and enlightened examination of this research and for their insightful suggestions. They are particuilarly to be commended for accomplishing much of this task at a distance of 1,200 miles frorni the writer. I would also like to thank Dr. Anthony Davids for his generous help with some of the practical matters involved in completing this work. The encouragement and forbearance of mny wife, Marilyn, were of inestimable value.


*The Parent Education Project was initially funded by the Fund for the Advancement of Education. The continuation was partially funded by the Children's Bureau, Department of Health, Education and Welfare.











TABLE OF CONTENTS


Acknowledgements

List of Tables

Abstract

Chapter 1 -Problem Chapter 2 - The Parent Education Project Chapter 3 - Review~ of Related Research Chapter 4 - Study I: Hypotheses Chapter 5 -Study I: M'Iethods Chapter 6 - Study I: Results and Discussion Chapter 7 - Study I: Conclusions Chapter 8 - Introduction to Study TI Chapter 9 - Hypotheses for Study TI Chapter 10 -Study TI: Methods Chapter 11 - Study TI: Results and Discussion Chapter 12 - Discussion Chapter 13 - Summary Appendix A - 'Modif ied I-E Scale ("RI) Appendix B - Parent Edtzcator We,'ekly Report (PI", T and the
teaching behavior indi ces (Index of Positive Verbal Interaction, Index of Ne! ative Verbal Interaction, Index of Attitude Towaard T'arent
Education Project) derived from it. Appendix C -- Series Test Bib liography


iii.


Pa',e

ii iv vi


1

5

9

23 26 37 58

60 62

64

67 88 96

103


109 118

121









LIST OF TABLES


Page


T reatment Plan for the Parent Education Project


A Comparison of Three Populations With Subjects in Groups 1 and 2 in the Present Study on I-E Control


Table 2 Tab le 3 Table 4 Tab le 5 Table 6 Table 7 Table 8 Tab le 9 Table 10. Table 11.


Table 12. .**Differences in I-E Control Expectancy Scores (SRI)
Between Mothers Rated High and Low in Project
Involvement (PEP)

Table 13 . . . . A Comparison of Three Different Populations With
Project Mothers on I-E Scores

Table 14 . . . . Differences Between Black and White Mothers in
Initial and Final I-E Control Scores


7


. . . Correlations Between Age and Parity and Initial
(pre), 'Final (post), and Changes in I-E Control
Scores for 32 Subjects

* . Motherst Initial (pre), Final (post), and Changes
in I-E Control Scores (SRI) Grouped by Level of
Verbal Interaction and Education

. . . Analyses of Variance of Initial (pre), Final (post),
and Changes in I-E Control Scores (SRI) by Mothers'
Verbal Interaction Level. and Education

. . . Changes in Children's Series Test Scores (6-24
months) Grouped by Their Initial Score and Their
Mother's Initial Expectancy

. . Changes in Children's Series Test Scores (12-24
months)

* .Differences Between Success Scores of Children
Grouped According to I-E Scores of Their Mothers
at the end of 24 Months

*..Changes in Mothers' Expectancy Scores (I-E)
Scores (3-21 months) Grouped by Their Initial
Score and Their Child's Achievement at 6 months

. . . Differences Between I-E Scores (Final SRI and
Changes Over 15 Months) of Mothers According
to Achievement of Their Children

. . . Differences Between Treatment Groups in Initial,
Final and Changes in Expectancy Scores (SRI)


55 68 68


iv


Table 1


39 39



43-44 45-46 51 52 53 53



54 55












Table 15 . . . . Correlations Between Education, Age, Parity and
Initial (pre), Final (post), & Changes in I-E
Scores (N=28)

Table 15.1 ..Analysis of Variance of the Multiple Correlation of
Age, Parity, and Education With Final I-E Control
Scores

Table 16 .. . Differences in Mean Initial and Final I-E Scores
Between Mothers Above and Below the Median on
Behavior Indices

Table 17 . . . . Analysis of Variance of Changes in Mothers' I-E
Scores Effected by Positive Verbal Interaction
(VI+) and Positive Attitude Toward Project (PEP)

Table 18. ...Analysis of Variance of Changes in Mothers'
Expectancy Effected by Negative VI & Positive
Attitude

Table 19. ..*Differences Between Achievement Scores of Children
Grouped by Median Split of Mother's Initial (pre),
Final (post), and Changes in I-E Control (SRI)
Scores

Table 20 . . . . Differences Between Initial (pre), Final (post), and
Changes in I-E (SRI) Scores of Mothers Grouped by
Median Split of Children's Achievement Scores
(6 and 12 Month Series Tests, and Griffiths Test)

Table 21 . . . . Differences in Initial (pre), Final (post), and
Changes in I-E Control Expectancy Scores (SRI)
Between Experimental (E) and Control (C) Groups

Table 22 . . . . Differences in Initial (pre), Final (post), and
Changes in I-E Control Expectancy Scores (SRI)
Between the Two Experimental Groups (El+E2) and
the Control (C) Group


v


Page


69 69 71 72 73 80 82





84 84





Abstract of IDiss( rc atiori T'resc:;t -d to t-'
Graduate Council of the University of Florida in Partial Fulfillment
of the Requirernents for the Degree of Doctor of Philosophy



LOCUS OF I-E CONTROL EXPECTATICY ANID EXPTECTANICY CEHAN1GES OF DTSADVANT'-ACED MOTHEP



by

Larry 11. Bilker

June, 1970

Chairman: Dr. Louis D. Cohen
Major Deoartnent: Psychology

This research investigated existing internal vs external control

of reinforcement expectancies (I-F Control) among indigent mothers, and changes in expectancy as a function of partici-pating- in an educational program. The two studies renorted attempted 1) to validate theory and research indicating that the lo-w.er sociccononic class individual is one who tcnds to perceive events as externally controlled,, 2) to explore the relationship between specific maternal behaviors and I-B Control, 3) to identify indi-vidual differences within a population samole from the lower socioeconomic class on the -dimonsion of I-F Control. oxnectancy, 4) to examine the interrelationship between a mother's locus of control expectancy and her child's achievement in a prograr, designed to enhance development, and 5) to extend previous research findings on correlations between internal I-F Control and participation in or affiliation with social action programs. It was -oredictcd that narticinatlion in a social action noa by Jowcr class mothers would chanreo their expectancies in Pn intcrna-. direction overti.

The Subjects were indigent inothers a~nd their infants who nartici.Dated in a nrog-ranm desir-ne-d to imr.,1ove fllctherin, skills. The nrop-ram was based


vi






on research in child development wh',ich' was fouvnJ to be related to the cognitive growth of children. In Study I and Study II which used better controlled procedures, families participated or were assessed over the child's first year of life.

Maternal expectancy was assessed at the beginning and end of the programr by a version of Rotter's I-1-7 Control Scale modified for a fourth grade reading level. The revision was found to be reliable over a one month interval for an indigent population. Maternal behaviors were observed and recorded weekly by parent educators, indigenous non-professionals who taught the mothers in their or hones. Infant achievement was assessed by progress on the program materials developed out of Piagetan theory and research, and two standardized tests of infant achievement.

Thne sig-nificant finding-s were:

1. Indigent others were moore external than subje-cts from more advantaged populations reported by TRotter.

2. Indigent -others were more external- than indigenous nomn-professionals from the sam- background.

3. Black mothers were more external than w hites in Study I but the finding-s were non-significant in Study II.

4. Mother's frequency of positive verbal interaction with her child was significantly- re.latcd to internal1 control in Study I but not in Study I. 5. Mothers who partIcipated in the programing for 9 months became significantly more internnl than brothers who did not p-artici-nate.

6. Mothers who were taughbt by parent educators who desir-no:. thcir 0,7f prog-ram bcca.7:', significantly i.:ore inter,, J.. tin control g-roup rothcers, MotherFs taught by educatcrz; vho usc-0 an alrcn.0y designed progr,4M beca.-e rore internal bizt t .e findinf-s 'were not sig;nliciant


vii






Mother's I-P Control was not significantly related to her level of negative verbal interaction, her degree of involvement in the program, her age, parity and education or her child's achievement.

The finding that participation in an educational nroj7-ram, Darticularly the one in which the teachers, former indigents themselves, designed their own program, was effective in changing indigent mothers' locus of control expectancy in an internal direction was discussed in terms of its implication for positive social action. Suggestions for future research were made on the basis of methodologlical and theoretical shortcomings of the studies. The possibilities of using an I-F Control measure more reliable over time, more sensitive at the internal end and more specific to the subjects' situational context wvere considered. Possible intervening variables between mother's expectancy and child's achievement were suggested.
















CHAPTER 1 PIROBLE"I


The purposes of this research are to investigate locus of control

expectanciJes amo-.ng lower socioeconomic class mothers and its relationship to selected maternal behaviors and to their children's achievernent; and to assess the effects of an attempt to change existing control expectancies.

Expectancy is defined by Rotter (1954) as "the probability held by an individual that a particular reinforcement will occur as a functicon of a specific behavior on his part in a specific situation or situationss" Internal vs external control of reinforcements describes a generalized expectancy which determines to what extent certain

outcomes of behavior will be categorized as within the individual's personal control and understanding. A person who generally categorizes situations as internally controlled tends to expect that it is an individual's own characteristics and skills rather than externals which influence what reinforcements he receives. On the other hand, a person who generally categorizes events as externally controlled tends to expect that chance, fate, powerful others, or an incomprehensible complexity has the greater influence over what reinforcements he receives (Rotter, Seeman, and Liverant, 1962).

Rotter (1966) summarized findings by himself and others which indicate that internal vs external control expectancy is a personality


- I





I


characteristic which has predictive value in relation to other behaviors of an individual. More specifically, the findings show that an individual who tends to categorize events as internally controlled is more likely to be alert to aspects of the environment which provide useful information for future behavior, to take steps to improve his environmental conditions, and to place greater reinforcement value on acQuisition of skills. Simmons (1959) also found internal control exrlectancy related to organizing, planning, and realistic goal-setting abilities.

Many social psychologists and sociologists have described the

lower socioeconomic class individual, particularly the lower class

Negro, as a person who feels powerless and alienated (Cohen and Hodges, 1963; Dean, 1961; Irelan, 1966; Reiff, 1966). A low expectancy that an individual can control his reinforcements is often referred to as powerlessness or alienation (Neal and Seeman, 1964). The empirical findings of Coleman (1966), Lefcourt and Ladwig (1965b), and Battle and Rotter (1963) corroborate this description. Thus, lower class individuals are more likely to believe in external control and lack just the personality characteristics described above which they need to improve their situation.

Other researchers have emphasized the importance of individual differences within socioeconomic classes. Bell (1965), e.g., has

pointed out that different subcultures can be distinguished within the lower class continuum. One would expect to find that within,









and cutting across,* subcultuires there are individual differences on a behavior continum such as internal vs external control expecta-ncy, and that these are related to differences in other behavior.

The first purpose of this research is to investigate lo,,er class mothers' expectancy of internal vs eternal control and a) the direct relationship of that expecta-ncy to, 1) their &an'ount and kind of verbal activity, 2) their conmitrnent to a program designed to enhance their mothering skills and, 3) the indirect relationship between control expectncy a-Ld the intellectual development of their infants.

The second purpose is to evaluate an attempt to change thosec

expectancies through participation in a parent education project. It is important for all people as well as disadvantaged people to become aware of the degree to which success or failure is contingent upon the acquisition of certain skills. Although it has been shown that members of social improvement groups are higher in the belief in internal control than non-members (Neal P-nd Seeman, 1964;

Strickland, 1965), the question of whether participation in such a group, e. g. the parent educatio-., project described in the next chapter, will chane expectancy to a more internal direction has not been previously investigated.

It seems~ particularly important for kvocor class mothers to develop the internal expectation thcat their rmothering abilities and the abilities t-rhey c 7n el.vevlop in their yor*ngz-ters ca~n load to Flleviation of povZ:tv, conditions and a inore successful life. PAccording, to a reccznt resf 7,rch rE!D0rt (Gi~ctberc., 1969), Head Sta:rt p:rorai we~re









only successful to the extent that the mothers became involved and developed a more positive attitude.














CHAPTER 2 THE PARENT EDUCATION PROJECT


The subjects for the present study came from the ongoing Parent Education Project begun in June, 1966. In this program, a sy-all group of disadvantaged women were instructed and arranged to visit other disadvantaged homes to teach mothers to stimulate the activities of their infants. The methods of stimulation used had been shown to be related to cognitive development.

In addition to being trained to teach stimulation techniques, the parent educators were also taught to make and record observations objectively in the ho,: e. This teaching was done by means of presenting films and verbal descriptions of home situations and asking the group to judge and record these until they could all agree on various kinds of observations. In order to reduce interference with the natural environment, the parent educators were the sole observers in the hone. Thus, no reliability check on their observations was possible. However, during weekly individual supervisory hours, each educator went over her observations with a supervisor. Periodically at six nionth intervals, visits to the homes wt re made by both the parent educator anud her supervisor. Both made independent observations and then che. eed then out. There wzre few discrepancies.

The parent educators were 15 (12 Iegrocs, 3 Caucasion) high

school gradvalk-es whcsze aveira-e f&aufliy incc'-- b--fc-re b.-_ing hired was $280 -oer rlonth for the support of on av,rar- of four Dcople.


- 5-





- 6-


Thus, most of the parent educators were better off economically than the indigent population they were to work with but were Still within the same socioeconomic class. All but one of the parent educators were mothers with an average of 2 to 3 children. Seven were married. Negro educators worked with Negro families; Caucasian educators with Caucasian families.

The project treatment plan is outlined in Table (Gordon, 1969).

Series means weekly visits by parent educators to instruct the mother in materials develoD,:d by the project staff. Group C3 mothers were taught by another set of parent educators from backgrounds similar to those of the original parent educators. These new parent educators were former Head Start aid nursery school workers who developed their own pattern of stimulation.

For all groups mothers and their infants were indentified at the birth of the latter by the obstetrics staff of the J. Hillis IMiller Health Center Teaching Hosoital of the University of Florida. Only

families with an economy ic coide of indigent on the hospital acir-liSsionl form and residence in Ala~chua and eleven other nearby counties were selected. In addition, the obstetrics staff used the following criteria: single birth, no breech or Caesarian delivery, no co, qlications

to mother or infant, no gross evidenc of inftint's mental retardation, and no evidence of mother's mental illnes,7s. These criteria were selected because of their -Dossible relationship to normj.-al intellectual potential. TFhe birth rate at the hosnital was such that, be inninga June 15, 1.9(6, about 30 babies a- month were Padded to the sample through October 31) 101(7 (attrition on narticipation in this st'iidy was ,,,Out





7 -


TAB T' 1

Treatment Plan for the Parent TBdue-ation Project





Group Firnal NI IP fants Sinfants
(E=-Eperimental; Treatr-rnt front Treatrment from
C=Controi) Age 3 mo-l year age 1 year-?
years



Original grouns began 9/15/66

1. El 36 Series Series

2. E/C 36 Series No Treatmnent

3. C2/C 36 No Treatment No Treatment



New groups begun 7/1/67

)4. E2 21 Series

5. C3 22 Stirnulation Program
Desi[gned by Parent
Educators

6. c14 26 No Treatment



30%) (Gordon, 19690).
Brief note on, the. Use of the indigenousro-rf inJ

reviewed in the literature is added here. Riessimn(V) and Levinson and Schiller (19r5) repoi't that the use o2' no, -professionals from the sae3c)o. socioeeacno te;t!.:ri group nrce cornsiuication effectiveness with the population beinc- stuclicd. A training pr-oremr (aftcr i-which the Parent Fh3ucation Project's was





-8


modeled) emphasizing inter-personal relationships, commnicative skills, professional confidentiality and responsibility, continuous interaction with professionals and personal identification resulted in a high level of competence (Levinson and Schiller, 1965).














CHAPTjER 3 REVILW OF RELATED RSSEA1XOF'


I-E Control and Attempts of Peonle to Better Their Life Conditions

A person who feels powerless to better his life conditions was

described in the introduction as one who has a low expectancy that people can control their own reinforcements. Such a person, in fact, does less to improve his circumstances than one who has an internal expectancy. In experimcc.ntal tests of the above assumption using7 Rotter's Internal vs Ext.-ernal (I-E) Control of Reinforcer-ents Scale (Rotter, 19656), Seeman and Evans foun-d that the more internally controlled TB patients knew., more about their own condition, qyc'stionel the doctors and nurses -ore,, expressed less satisfaction with infoemation feedback, and were rated as better patients by the staff;

Seenann (1963) found indep--ndcnt of intelligence (correlation of Beta I.Q. with I-E scale was .03) that the more internally controlled refoirI-J-ory inmates knev Tnore about hio.,r the refomoln--to ry was run, parole, and pertinent infornation aloiit ecc-io~xiic ojp )ortunitics for post-refoxTmatory living. Kiehlbauch (1968) found that prisoners who attainod a work release status, i.e.* were allowed to work in the community towards th3 end of their incarceration, did not show the rise in extmraa-0ity e r uont'h prior to release that a matched samonle of prisoeir, diJ. co.-nd Pottcr (11Y 3) founr willinancss to noake a coaimiitr.int to social action on civil ri-ghts was related to internal control w;nga ground Of 'TgCL_ enrolled in a southern clce


- 9 -





- 10 -


and Strickland (19165) extended the latter finding to actual. participation in social action. Students actively involved in the civil rIghts movement were nore internal thian a matched control group of non-participoants. Male smokers who quit after the Surr-eon General's report were found to be more internal than those who believed the report but did not quit smok-inrg (Jamires, '-, oodruff, and. Werner, 195).

The predictions in this study are that mothers using7 more mothering skills, more specifically verbal interaction, would be more internal, and that those who saw the noroject in a positive light and cooperated would also be more intern~al..

T-F Contro) andLearning

Potter's (1966) "... basic hypothesis is that if a person perceives a reinforcement as contingent upon his own behavior, then the occurrence of either a positive or negative reinforcement will. strengthen or weaken potential for that behavior to recur in the

same or similar situation" (Page 5). In a series of studies which contrasted learning under experimr~n tally defined condition ons of chance and skill, subjects made more FMprop~riate responses to positive and ngterinocmnsicraein nositivelr reinforced behaviors, decref:se in nr-.7gatively reinforced ones), generalized their learning- to a significantly greater degree, and *were more resistant to extinction under skill conditions than under chance conditions (Phares, 1957; Jar ,es, 1957; James arnd Potter, lQ5,9; Potter, Liverant, and Crowne, lQ('l).

A few studies have contracted individuals who tended to see

events as skill controlled vs those that saw then -vs cht.rncc controlled. Pharcos (1957), us-ingr the first precursor of the I-EP Control Scale, fotind a tenlre tcr- for suh jccts scoring' hir&h-r on in-





- 11 -


ternality to give more appropriate responses to reinforcement. These findings approached significance. Using the james-Phares Scale from which the I-E Scale is derived, James (1957) found significant relationships between internality and all three learning variables mentioned above. Thus, the greater tendency an individual has to perceive situations as internal, the more he will learn from that situation;

the more he will generalize that learning to related situations and the more he will remerber over time. Coleman (1966) surveyed students and school conditions in all twelve grades across the country. They found that differences in school conditions such as facilities, curriculum, teaching quality and availability were insufficient to explain differences in verbal and non-verbal ability, reading comprehension, mathematical achievement, and general information about the natural sciences, social studies, humanities, and practical artS as assessed by comprehensive tests administered through the local schools. On the other hand, they report that a Negro child's achievement is highly correlated with his feeling that he can control his own destiny.

There is also evidence that need for achievement is related to

I-E Control. In a study using the Intellectual Achievement Responsibility Scale, Crandall (1962) found that for boys in the first three grades, achievement motivation as measured by free play achievement behavior and achievement test scores was related to more

internal scores. The findings did not hold for girls. Franklin (1963) found 15 of 17 kinds of reported evidences of achievement motivation correlated with scores on the I-E Scale in a national stratified sample of 1,000 high school students. Using latency of





- 12 -


decision on a matching task as a measure of achievement motivation, Rotter and Muiry (1965) found that when unsel-ected subjects were divided at the median I-E score into internals and externals, the former showed greater need for achievement. In summary, I-E Control is related to amount, generalizability, durability of learning, and need for achievement. Applying this logic to the present study, mothers who learn more and use this increased knowledge on their children should be more internal.

Personality Correlates of I-E Control

Although a more pre Cise relationship between exact score on the I-E Scale and specific personality attributes has not been validated, Rotter (1966) suggests that the extremes may be related to adjustment problems. That is, an extremely high external score may be an indication of defense against failure while an extremely low one may be associated with the assumption of an unrealistically high amount of responsibility and consequent guilt for personal actions. Effran's 1963 study of high school students indicated a relationship between the tendency to repress failures versus successes, and scores towards the internal end of the dimension. This was interpreted as an indication that inter-nals feel a greater need to defend against failure while externals already have a convenient rationalization (Rotter, 1966). Butterfield's study (in Lefcourt, 1966) indicated that internals depict themselves as goal directed workers who strive to overcome hardships while externals portray themselves as suffering, anxious, and less concerned with achievement than with their emotional response to failure.





- 13 -


Potter (1966) repor,:s that st - xa_,ninng rE'1ation. hips between measures of adjustment and the 1_- scale are siu7gestive but inconelus ive. Potter and Rafferty (in ,ctter, 1966) cornared the scale with the Potter Incomnlete S)entetnces 1TKI'rik, (T', ) for several Sampies of college students. Generally licrcorreations were insignificant. So!me curvilinear relationsnhip between extreme scores on the I-F Scale and maladjustment scores on the Potter TSB_ approached significance. The extreme scores were less well adjusted. Simmons (1959) similarly found a positive but nonsignificant correlation ratio (eta) between the two measures. He also found no reliable cor-relation between the Potter ISB dependency score and the T-E Scale for college males. Kiehibauch (1966B) reportsod vs poor adjustment grourns of prisoners did not differ on I-F scores.

Ware (Rotter, 2266) found a significant correlation of .2L

between externality and the Taylor Manifest anxiety scale. Yffran, using the swne measures, found a correlation of .00 (Rotter, 19166). Cromwell, Rosenthal, Shnlkow and Kahn found schizophrenics higher in externality than normals (Lefcourt, 19r66). Overall, the results suggest some relationship between extreme scores and maladjustment with nerha-ns greater variance corning from the external extreme. In test behavior internals seem, to have a more constructive responsre to failure.

Several studies (Jwn, s, 1957; Battle and Pottcr, l6;arid

Simmnons * 1959) have found the goal settling, of externals nore variable and unrealistic. The external subjects are nore pron.2 to the gamnbler's fallacy of exnectin- to win after a sre ~~:

in skill situations. Sirrions (1059) fiurtherynore fonrvJd cxtern~1ity









correlated positively with other mfaladaptive goal setting patterns, and negatively with Edwards Personal Preference Schedule needs for

order, nurturance and endurance for college women. He concluded that external females are miore disorganized, lack nlanning and realistic goal setting abilities and are fatalistic comnoarel to internal females. Several studies have found relationshins between I-E Control and self-reliance. Julian and Katz (1968) found on a

competitive task that internal college students preferred to rely on their own resources rather than those of their opponent even though they were shown by exterimental maninulation that their opponent was more comnretent. Odell reported a significant rel-ationshiD between I-F Control and Rarron 's indenendence of judgement scale. Crowrnie and Liverant found externals conform significantly more in an Ashe type conformity situation (Lefcourt, l-)66). Deever (1967) reports internals choose their personal reinforcement history rather than retorted per-formance of others as an index or cue to expectancy of personal succe!;s in the future.

In the present study successful. participation deTnended upon the mother's active use of what she learned in the parent educator's absence. This would seem to be related to her self-reliance and organization. Thus, it was predicted that a volunteer mother's

success both in ters,-s of her performance as a teacher of her child and the child's success would be related to 1-7 control.

I-E Control and Affiliation

N'eal. and Seei-nan (l9 5', usinc the I-E Control Scale, found that members of a work- based organization wore si~nific artly lower in





- 15 -


externality than controls matchedd for age, socioeconomnic level and education) who were non-members. Furthermore, activity within the union and general knov1edgpe of political events were positively related to internality. Differences were not related to generalized despair scores as .Iea rcd by SroJle 's Anomnia Scale. n.he authors concluded that internality was r ?Jated to affiL"liation with can organization that can better one' s life circiuzstances. The direction of influence wEs not assessed. Either z:-ore internal people affiliate, or affilir'.ticn engemle rs -, tern a].ity. The present study proposed to test the assumption that affiliation modifies expectancy in ran internal direction.

1?pectancy and De:no,,aphic V~ariables

'The population for this stvdy was selected from volunteers in

the Parent Education Project. Since they we)-e lover socioeconomic class mothers (6:1 ratio of blacks to whites), the primary concern here was the relationship between expectrancy and class and expectancy and race.

Race

Several studies have found that Yegc;es arc! significwntly more external in their control. exp:ectancies than Caucasians. Lefeourt and Ladvtig (l6j ndfattle D.9t6_2) using the I-E Scale found a small1 (a score of 9 co::iared *,o 8 out of a possible total score of 23) but significant diffe-reicc. Coleman.- (1966) found that in a colItry-"eid--aal of sefrol children a siGnificantly hig'I~c-r proportion of black than white children ans-crel three quzstea bout control expectancy in an externnlj~c~in Lcfcoirt (1966)





- 16 -


argues that blacks eo.s-ily perceive impediments in the way of' goal striving. Segregation and discrI1~lination are interpreted as neaning that their own effort will not pay off in reinforcements. On the other hand, Kiehibauich (1968) fond a small difference in the samae direction between black and white prisoners in their I-E scores which did not reach a significant probability level. Thus, every one of these four studies comparing black and white subjects has found blacks to be nore exterra1 than whites. The results in only one study were not significant. If Lefcourt's argumient is correct, end it seems to hold up country-wide in Colemnan's study, it should hold in a southern cozxnity. Consequently, it was predicted that Negro others would be more external than Caucasian mothers within this disadvantaged population. Socioecono-nic Level

For popu-lations in which social class differences are small no relationships between elkss and 1-E Control hr--lc br.'cn found. That is, there were no Clfference::s in I-E Scale c:*s by social class in Ohio State University classes (Rotter, 19,06), Florida A & M classes (Gore amd Rotter, 1963), or in a sample of prisoners (Le-fcourt and Ladwig, 1965b). Since subjects in the present st11udy wor1e frona % rcolant;.v21y h:.xiG soci"Icv~~ differcoce in I-E Scale scoi'e3 by socic co-cA.Ic level c2dnot be cxaa--ined. Differences have bcen fou-id in populations more lieterog!eneous for soeioecononic cl~'.Persons in the 0ctr c.s tend to scorch

more at the eytex-v71 end of the I-E Scale. One !:tuay of chi).,c 1(Battle ard Potter, 1-963) fo~ind a sir'rificcant diffcrcnoe be-tweeon social classes with race r-nd intelligence controlled. I-ost of the









variance wras accounted for by the: difference between lwrclans 11ebyro s and middle class Ilegroes and whites. Dean (1961) fcri ~cw (.10 to .23) but significant correlations between alienation end level occupation and income. His questionnaire included components of powerlessness, normlessness Clack of rules), and social introversion. The prediction here is that the present sanmi1e of lower class subjects will be more external than sanmples stratified by social class. Intelligence

Correlations of externality or powrerlessn-ess w-ith intelligence are generally lov and insignificant but there are exceptions. Strickland found no correlation between I.Q. and I-E scores with a samiple of Ohio State coeds. Seemon (1963) found no significant correlation between prison inmates Beta I.Q. scores and I-E scores. In the Kiehlbauch (1966) study there was no significant correlation between N-E aid I.Q. scores. On the other hand,, Ohio State University womens' external scores on the I-E dimension were found by Cardi. (Rotter, 19066) to be

correlated -.22,, and by Simmoils (1959) -.47 with intelligence. Since the balance of evidence is for no correlation between I.Q. and I-F Control, the present study did not exa-mine this relationship. Edu cati on

Evidence for the relationship between edueAVcn and I-B Control expectancy is equivoceal. Rotter (1966) reports unselected hi.rh school students score higher (nore external) on the I-EB Scale than colle-: bound students. Dean (1961) found a low.. but si__Pni*ficz.nt, inverne correlation betweazn alienat--n -!nO e eeuctional lcvrel. On the other hand, in thbsuisPoo~ ~ (96)s:::e the-' scored


- 3.7 -





. 18 -


of college students are ren L>_ic~rzally not different frm ri! On innmatcs with an eighth grade reading level. The present study predicted that since the educational level would probably be relatively homogeneous, there would not be any sig~nifltcnnt relationship between educational level and perceived loctis of control.



Age differences have not been found to relate to differences in control expectancy. AJ1-though the present sample includes very young (15 years ol.d) ard older (76 years old) mothers or smothering ones (see subjects,p. 21) the prediction is for no correlation between age and expectancy.

Effects of lMother-Child Interactio-n I-ECotrolMotherig and Infunt Corfnitivey Dzvelop:.ent

It is assumed here that a mother's expectancies and behaviors will be interrelated with her child's cognitive development. A mother internal in orientation, according to the foregoing research discussed, should tahe stejs to improve her mothering skills. Such inqprovemant is ex-pected to show up in increased cognitive development of her child. The focus here will be on verbal behavior. Conversely, if a child is successful cognitively, thus confiriirg his mother's ability, his other's expectancy shcntld chp:nre in can internal direction.

That maternal behaviors are a significant influcrice on a. child's cognitive develop:n 2nt is well knvxrn. Mternal Lz4-_, rations and expectr-tions, re-,rird -nd p'nz~tof her cnild's behavior, provisi I o_ r~il Iz'_ oth- opporV,.ties. tosiult he child's





- 19 -


development end policies for rearing of the- child have all been found in the child develocMn,:-.it literature to be re] Aed to the cognitive develo-pment of the pre--school child (Gray and M-iller, 1966). These are particularly crucial in disadvant-aged homes where, for example, Bayley (196D6) suCests there is a lack of stimulation which may be caustive of retardation. The lack of verbal stimulation has -oarticularly been emphasized. Hiunt (196,6) points out that verbal stimulation is lacking- in disadvantaged homes during the first two years of an infant's life. Bernstein (1961) points up the long range implications of this. He find that low,-er

class members experience a relatively narrow range of variations in language. Thus, they cone to rely on highly predictable inplicit utterances which poorly equip then to express themselves. The inflexibility of their verbal expression interferes with their ability to beco.Te actively responsible for their ow,.n behavior or learning.

The child influences the other as wTell. For example, a child's executive conetence has been found to influence the mother's

emotional reactivity (WJenar and Wenar,, 1964). In the present study a child's success is predicted to lead to his mothe-r's movement in an internal direction analogous to the subjects in the studies of Rotter, Liverant and Cro'.rnme,, Bernnion, and Blackman (Rotter, 1966) WhIo Wer-e li!ke ly to scce a sequence of reinfocc-,ent as skill controlled internall) whenl reinforced as right rro thani 505 of the tine. A child's success in the project is assumed to be reirnforcln,7 to the otr










Love Oriented vs Po,.ier Assrtive Di., inline

A more internal mother is expected to perceive the reinforcements her child gets as conting ent upon his characteristics or actions. She would seem more lky, then', to use discipline which engenders the development or internal standards. In a review of the literature on parental discinline, Becker (1964) refers to love ori ented

techniques such as reasoning with and pralsing the child. These are reflected in the positive verbal interaction index used in

this study by items such as: "Mother explains and, describes things when talking to the baby"; "listens to thv!b' wen the baby talks"; "her tone of voice sounds softu a.nd o4r.


On the other hand, a more external mother perceives reinforcefments that happen to an individual to be continent upon chance, fate, inconurehensible- comp lexity or the actions or chauracteristics of more powerful others. Thus, she would seem to be more likely to use discipline techniques which lead to externalized reactions to behaviors (e.g. fear o~f punishment, projected hostility). These techniques classified as power assertive would seem represented by the items in the negative verbal interaction index such as

"Mother orders or tells the baby to do or not to do things"; "her tone of voice sounds cross and anry,. i

The consensus of research Is that lov'e oriented techniques, most reliably reaso~ni-nrg and praise, have been found to be correlated with the occurrence of inter.-slized reactions to transrrresslons in the form,, of acceptance of self-res-consibility, while now.-er asse-rtive techniques corroate with externalized reactions (ecr,1064).






- 21 -


More specifically, Tolor and Jalowlec (1968) found that college students who perceived maternal attitudes of t.Uthoritarian control and hostility rejection as assessed by the Parent Attitude Research

Inventory (PAMI) were more external op, the I-TE Scale than those perceiving democratic attitud:-s. The foregoing studies deal with parent behaviors influencing child's control expectancy. The present study deals with the relationship between a mother's behaviors and her own control expectancy.

In suy~ary, for the women in this study, the maternal role wa~s considered a significant fo cus of the life situation. Therefore, motivation for self-improvement was assunied to include a desire to enhance the performance of the maternal role. Fuhxthcn,_ore, it was assumed that the opportunity to participate in the Perent Education Project was an opportunity to improve the performance of the maternal role. At least this is the wayr the parent educators perceived it when introduced to the progrvn.. It is also consistent with their informal descriptions of how mnost of the volunteer mothers perceived the prograua (Gordon, 1969). Thus, a mother's degree of participation and involvemcmt in the project (as assessed

by such behaviors as: demonstrating ability to perforri the skills taught, encouraging her child, keeping- vxpointrma nts with the parent educ ators) were considered a Pmanifestation of attempts to improve her life situation. One inportlrnt reflection of this irsprovement was assu:,med to be the child's success in the project, program. Thus, it was predicted tha t diffcronccs in the amount and quality of rnatex'no-l care, partl.C;~1nrly verbnl bch.aviror arld attitude toward. the project, would be related to I-E control and





- 22 -that this in turn would be interrelated with the cognitive develop7.,nt of' her infant.














CITAPTER 4 STUDY I: HYPOTHESES


The Parent Education Project described in Chanter 2 began as a

pilot study. Thus, the usual difficulties encountered in naturalistic research were added to by such nroblcm3 as the inexperience of the parent educators a-nd chan-ing plans as to the most appropriate or expedient tire to collect a particular datu'. Somre of these

problems were ironed. out by the time three new groups of subjects began participating in the project one yea-c la'.er. These new groups were used for a second study which was pla-nned to invest.igate the saeissues as Study Y. The hypotheses for the first study are listed belov.

Ijynot1 h ss

Class, Race, Are, Parity

1. The present sample of lower class mothers will be more external in I-E Control than previously studied samples that were not restricted to the lo-wer class.

2. Black mothers will be more external than ,7hite mothers.

3. There w-ill be no significant correlation betwTeen age or parity of the mother and N-E Control.

Maternal Verbal1 Interactioni Level

i.a) M'others higher in positive verbal i ritera icn with their child initially will be norc internal in I-E Control than mothers lcver in positive V':rbn Intr~in


- 23 -





- 4


b) Mothers higher in positive verbal intera-Iction wIth their child

will become more internal in I-E Control over a 15 month interval than mothers lower in positive verbal interve.tion.

5. a) Mothers higher in negative verbal intCeraction initially will be more external in I-E Control than mothers lower in negative verbal interaction.

b) Mothers higher in negative verbal interaction will become more external in I-E Control over a 15 month inter-val than mothers lower in negative verbal interaction.

6. Mother's educational level, individually or in interaction with verbal interaction, will not be related significantly to I-E Control.

Infant Cognitive Dcelov me-nt

7. There will be a significant two way relationship between mother's I-E Control and her child's cognitive dCvelopmont such that:

a) Children of mothers more internalI in I-E Control will learn

more eand be more highly develope-d by the end of their second year of life than those ofLL mothers moie externrtl -in I.-IE ControlJ

b) Mothers of children ,,ho are more highly developed cognitively will be nore internal in I-E Control than mothers of less well developed children.

c) These mothers will also inn:e ore internal over a 15 month

interval.

Effects of Participation

8. Mothers who pairticipate longer (21 tiont hs) in V. program to improve their mothering skills wIll bcco-- inore internal in l-E Control than thosc who -it-i~cipate for- a shorter pieriiod of tLe(9 monthss.






- 25



9. a) Mothers who make better use of the procnram as manifested in such behaviors as keeping appointment and mastering materials will be more inter-nal in I-E Control initially than mothers less involved.

b) Mothers who rnirL':e better use of the program will become more internal in I-E Control over a 15 month interval than mothers less involvedi.














CITAPTIL2, 5 SYUDY 1: I





The subjects haVe been described unlier "tr- I 'on Project in Chapter 2. 7hey are in treat.,,ent Crol.s 1-3 (sce T~-- 7), Thus, data on mothers arnd their children cover fro-m the thirk -to Vie twenty fourth month of the child-ren's lives. For this s'yonly -.:"nc or mother substitutes (e.g. granc'Mr)ther) who had the r'l&jor r.ib~t for the child's care were selected. The number of subjects for ticdifferent variables chan-es so,-:e-.what because of miUssing data.

Desif2n

The rittjor variable studied, 1-E Control, was the score atta-,incd by the mother on a modified version (SRI) of Rotter's I-E Scale (see In-struftirits). P.l. of the mothers studied. were given the SRI when their babies wcre 6 mntsold. This wac; conside-red the initial or pre SRI

score. Mothers were Yreeted when thcir babies were 21 rzonths old. This was considered the final or post SRI score. The difference between the post and pre SRI scores was calIled the SRI change score.

Except for the hypothesis in which infant achievement was the depcndent variable, thez- SRI score -was the d:-pendent varira.ble. To allows for initial d-iffic-rnccs und chnegenerAly the relationships between the in cerendcnt variable and all three SRI scores (pre, pesZt Vnd



The treatr-c-nt pltln to test the effect of' part Lm-.%,-ti7g in a prorcr.to improve nothering- shiJ.ls on a vno 1hr's P-E Control is; ch,-.Xted on


- 26 -





- 27



page 5, Chapter 2. The group 1 mothers partiAcipated for 21 months compared to 9 months only for group 2 and no participation for group 3. Insufficient data were available on group 3. This precluded assessing

the effects of participation vs non-nartlicination. Instead, the effect on I-E Control of the len,,th of narticination was tested by a t test of the difference between the mothers' expectancy scores of group

1 and group 2.

Mothers' verbal interaction levels (positive and negative) and mothers' degree of involvement in a progrwai to improve mothering skills were rated by the parent educator in weekly observations (seo PEWR in Instrument section) over a 21 month period for treatment group 1 and over a 9 month period for treaty w.nt grout 2. Tn assessing the relationship between positive or negative v rbal interaction level and I-E Control, first mothers were grouped as high or low on verbal level. Then they were split at the median on educational level. The hypotheses were then tested by a two way analysis of variance of the SRI scores.

In order to test the relationship between mothers' degree of involvement in the program and mothers I-F Control, first Piothers were grouped as high or low in involvement. Then th!ey were split again t-cording to 'Whether they were high or low in positive verbal intcracticn. This hypothesis was also te: ted by a two way analysis of variance.

The inter-rellationship bcttw -c n mothers' I-E Control and infant cognitive development was tested with the "short term prospective model" (Wenar and 7err 91) his model allows for consicreration

of the direction of inf-1.hvincc in stuiiie3 of ,-rent-child interaction1. Families were classified in t-,1.o ways, i.e. for I-E control tand infant





- 28-


intelligence. Fwamilies -were divided into high, and low, on the basis of mother's initial exterap.l scored and thc-n, urlr. on thce basis of the child's initial intelligclnne ratlirr as hig-h or low. Then the data were set up twice, once for intell.4gence cha nre scores an! once for expectancy change scores. When the child's cognitive prog-ress is the dependent var-iable, significant differences would be attributed to

the child's intellectual development. When changes in mother's external scores are plugged into the sarne design then significant differences are attributable to maternal expectancy.

The hypothesis concerning I-E Control and socioeconomic class was

assessed by a t test of the differences between the I-E scores of the present subjects nnd scores of sublJects from more advantaged populations. The relationship between race and I-E Control -was examined by testing differences in SRI scoy-es between black nd w hite subjects. Age, parity and I-E Ccntr.:4 scores were correlated to assess whether they were significantly related.

Non Standardized InstrUMenItS

Rtter's Internal vs F:~ :lControl of Reinforcev-ents Scale

This instrument was dzr; to tcsses, s the extent to which an individual categorized events as internally or externally controlled. It is a 29 itemi forced choice test which includes six filler items and is entitled 'Social eacio Invento2ryl' to raethe purpc,3e of thf2 test somcwlat -n,uoil

The test uarsto beitLl' ceiA;A r-1l-ale r2l ~

sional1, and har; good discrirmin,,L-t and external validity-.. The item biserial corrc2 'ticns fc2 n,~:ol of II00 ww. rD-!-~n range fro:t1 .1-1 to .48 with 13 out of' 23 itcens groaterthn 23 Tne Split





-- 20, -


reliability ranges from .69 to .73. This is below desirability. However, the items are additive rather than equivalent. Rotter (1966)

reports one month test-retest reliabilities in the .70's. Kiehlbauch's study (1968) showed a three month reliability of .75. With 6 and 9 month intervals the coefficients are nonsignificant: .39 and .16 respectively. Factor analysis revealed that a general factor accounted for 53% of the variance with the rest being accounted for by several factors of a few items anLd a small degree of variance each. Correlations with the Marlowe-Growne Social Desirability Scale range from -.35 to -.07 when administered by authority figures in various institutions. The highest correlation can be explained by the fact that the test was administered with a battery of classification tests given to new prison inmates (Rotter, 1966). Low correlations with intelligence and tests of external validity have been described in the literature review.

Modification of Rotter's I-E Scale for Below an Eighth Grade Reading Level (SRI)

The I-E Scale- has previously been found satisfactory for people

with at least an-. eighth grade reading ability. Since it was suspected that many indigent mothers in the project might not meet this criterion, an attempt was made to adapt the test for this population. The parent educators took the original test. Afterwards, they were asked to pick

out words and phrases they thought their poorest reading volunteer mother might not understand. They were asked to rephrase these in terms they would use in expressing themselves to these mothers. The revisions were then collated and edited for correct grammar and consistency of meaning with the original version. The final revision was






- 30 -


edited for words inappropriate for higher thrin third grade~ reading according to the Thorndike-Lorg-e (1944~) word list. In a few cases fourth grade words had to be included. A test-retest reliability coefficient of .78 and with an of 35 was obtained for a two week interval on a population comparable to the subjects of this study (Freijo, Gordon, and Bilker, 1968).

The above revision followed an attempt to adapt Bialer's (1961) Children's Locus of Control Scale. The wording of the latter was modified slightly to make it appropriate for adults (e.g., in "Do you think a kid can be whatever he wants to be," wom-an was substituted for kid and she for he). Then, this modified scale was given to the 15 parent educators. It correInted -.20 with the I-E Scale arid so was abandoned.

It was thought furth,2r that the administration of the scale might

present other problems such as those encountered by Radin and Glasser (1965) in their use of the PARI with disadvantaged mothers. In addition to liguage complexity, they found that interruptions in the home, reluctance to disclose information to a person of a different

background and insufficient rapport with the examiner to tell if the subjects understood were nxzoblemzs. Althoug-,h many of these problems are mitigated by the use of indigenous workers, in anti cipation of any that might remTain, the parent educators wcere instructed to read the items to each mother while the r-othcr ~:her choices on hey, own copy, to make it clear that there were no right or wrong answers and that their responses were confidential, and to p:u~for interrv:tl ): and even return e-rnother day if circun;istanes~c'::3 to beifl'c a volunteer's state of nind. Thu,., the PAdinistreation of th-e, I-J,





- 31 -


Scale used here eliminated langa-( conoplexSty and word connotation differences between backgrounds, facilitated the good rapport, and allowed the subjects to tanke the scale under natural conditions. The Parent Educator Wzeakly Reyrort taid Derived Behavior Indices

As was described earlier, each mother was visited once a week by

the sazne parent educator. The mother had stimulation materials presented to her and did the previous week's activities with the baby for the parent, educator's evaluation. The parent educator made an appointment for the next visit before she left. Irmedlately after leaving the honf , the parent educator filled out an observation report, the PEWR (sec Appendix B).

The PEWR (Parent Educator Weekly Report) is an omnibus fori-i prepared by the project staff (cornpozcd of educato: Z, psychologists and nurses) to include d7:.tn on hone conditions presirmed to be ralevant to the learning of the baby and tlie rather. It includes surggestions from Rheingold (1960 ) and Stedn-n2. * Inter-obscrver ag-reement between supervisor and parent educators o;- b1--nal visits was high. A formal relia'bility test of repea,-ted obser-v .ttiodns was riot done in order to ir~ini:;iz:-e the presence of the profession'ti. staff in the horn . It was felt that the presence of an unfamiliar authority figure would interfere with the ob:;ervatioj~ns of nn--tural on cc:-ditions --nd interfere with the rapport bet,.ccn a rnothc-r ttnd her parent eduoutor. Tteris from-, the PEWR were orgl-ized into three indices: the indCex of Pos3itive Verbal Interaction (VI+); the indcx o-f negative verb-al intieaction (VI-); and the index of attitu:- toward Parent Education Project (rPiT). The itc-z which Ere citv rt t1hj!.. of cyv . i J1


*Dr. Donald J. Stedna, 1in U
pecrzojn~t1 co ,ncni~





- 32 -


were chosen on a face validity basis. E.g., a typical item for the VI+ index (in the mother column since the subjects for this study were mother or mothering one's) was "Tone of voice sounds soft and loving." A sample VI- item was "In a few words, order or tell the baby to do or not to do things." An illustrative item of the PEP index scored positive was "How did the mothering one react to your instructions?"

1. "Looked at you while you were talking, and/or asked questions." A sample item scored negative was "When you finally got to see the mothering one: 1. She said nothing about missing her appointment."

After the Parent Education Project was well under way a validity check was made on the index of positive attitude toward the project (PEP).

The parent educators were asked to rank each of their volunteer mothers on positive attitude toward the project. The mothers were then ranked independently according LO the PEP index. The ranks for the index correlated better than .80 with the parent educator's ranks according to Gordon.* (Appendix B lists the specific items and manner of scoring for

each index.)

Further validity tests are provided in studies by Gordon (1969) and

Gordon, Herman, and Jester (1968). Gordon found mothers' VI index positively related to their children's achievement. Gordon et al. found a reliable difference between achievement of children of mothers whose attitudes were rated positive compared to those not so positive on the PEP index.

Stimulation Series Test

The stimulation series, samples of which appear in Appendix C,


*personal communication, 1970





- 33 -


was designed by the project staff to provide infants with the kind of experience through which, according to Piagetan principles, adaptation through accommodation will lead to greater modification of development

and greater cognitive organization than would otherwise be expected from growing up in a culturally deprived environment (Gordon, 1969). Sigel (1964) points out the possibility that adequate stimulation in the first two years of life may be necessary for a rate of intellectual development which will allow for acquisition of necessary knowledge later on. He suggests that one reason why children from disadvantaged homes have difficulty in kindergarten and first grade is that there was a stimulation decrement in the first two years.

The assessment of progress on the series was carried out by the five

staff members who taught the material to the parent educators. Parent educators and their caseloads were randomly assigned to each supervisor. Thus individual differences in scoring of the series were randomly distributed with respect to other variables in this study.

Although the series test is not independent of participation in the program, the use of it does not violate any assumptions inherent in the hypotheses of this study. Each of the children being compared has

an equal opportunity to progress on the material.

The derivation of the series is not the main focus of the present study and will be sketched briefly here. The reader is referred to Gordon (1969). The exercises were chosen for ease of teaching and evaluating. They were derived from Uzgiris and Hunt, 1966; Bayley (1933), Gesell (1949), and Gattell norms (1947); Hess et al. (1965); and Bernstein, 1960 (in Gordon, 1969). The materials are organized





- 3h4 -


so that they were introduced to the infant before the behavior should occur, according to the aforementioned norms and studies. Order of presentation also took developmental norms for pos-ition into account so that, for examznple, the first set of exercises have the baby prone or supine,while a later set has him creeping (Gordon, 1969).

StandardizedInstrume nts

Griffiths Mental Develornment Scale

The test was developed by the author (Griffitlhs, 1954) to measure rental development during the first 2 years of life. The scale consists of 260 items selected from a larer pool of items fro M other infant scales after extensive pretesting and observation of infants. On the basis of pretesting of infants the items were also grouped into five subscales and arrant-ed in oricr of increasing difficulty. There are an equal number of items in each of the five subscales for each age period. The subscales are: Locomotor, Personal-Social, Hearing anid Speech, Eye and Hand, and Performance.

The scale was standardized on 571 British children largely from infant welfare clinics and day nurseries. Like the Stanjfoi.'d-Binet, this test is an ayge scale. Thrce itcms passed represents one week of credit in the first year. Two items passed counts as one week of credit in the second year. The mean general quotient (GQ) for the standardization sample was 99.7 with a standard deviatlion of 1.2.1.

The retest reliability cozefficient over an interval of an average of 30 weeks (a rzxi.- of from 7 to 70 weclks) onl 60 casc~s was .87. Correlations with Stanford-B~inet IQ's at age 5 were higher than those reported for other infant tests. TheGrffth GQ at 6 m-onths orc9ted .32 with the Binet IQ at aD-e 5. 1Ii. 28 ;tstP C .40 with the Binet IQ at ar, o 5.





- 35 -


To insure adequate familiarity of the exctminer with the test, the author permits testing materials to be sold only to persons who have been trained by her in their usa (Griffiths, 19514). 1,leyMent Ll Development Scale

This test was desig-ned to measure devzlopmenttl progress from birth to two and one half years (Bayley, 1969). The Mental Develop~ant Scale was desi.gned to assess early learning, problem solving, commL-unication and abstract thinking- capacities. The items (163) were derrived mainly from the California Preschool Me-ntal Scale and the

Califor-nia First Year VMntal Scale. The iteiis were selected on the basis of data on 1560 children.

The standardization sample consisted of 1,262 children from 2 to

30 months old. The sample was selected to renprsernt the U.S. population (1960 census) within this ag-e range. The are placement of iters was the age at which 50% of the children passed a given item. The

mean standard score of the scale (called the Mental Developicent Index) was 100 with a standard deviation of 16. The split half rliability coefficients rpnged from .81 to .93 with a median of .88. The mean percentage of retest nre-_:-ent over a one week interval. was 76.4 (S.D. = 13.7). The Mental Developm-ient Index (IMI) and Stanford-Binet (Fori.A L-1,1) IQ were found to correlate .57 (Bayley, 1969).

Data Collection

The parent educr~tor .e!-,ly report observations were filled out for' eve-ry ho. e visit be,-,~h 2.Q vin each baby was three months old. Thiu,, if there were no missed visits, there would be 13 reports by the tinic: the baby was 6 mernths cMd. Thei I-E SecLe (Lcr etonTvno'










was a&.iin-4ste.-rcd by the parent edcctor whm~ eolch mother's baby was

6 mo~nths. old rxnd acain at 21 mo-tsis. Staf ' r.-hrbcrs adrinis-tered the Series Test at 6, 12, 18 and 24 nmon-thls of the baby's a.Lc, the Griffith's Mental De-v~loprient Scale at age one, and the Bayley Mental Development Scale at n-e two.
















CHAPTER 6 STUDY I: RESULTS AND DISCUSSION


Class, Race, Age and Parity


Hypothesis 1 was that the present sample of disadvantaged mothers

would be more external than previously studied samples not restricted

to the lower class. Table 2 shows a comparison of the external scores of the indigent mothers in the project when their infants were 6 months old with the scores of two samples presented by Rotter (1966) and those 'of a group of women from lower class backgrounds chosen for Project Follow Through which was modeled after Project Head Start.

Hypothesis 1 is confirnTed. Southern, indigent mothers, mostly

black (13:1), were significantly more external than samples of people from all socioeconomic levels. This confirms Battle and Rotter's

(1963) finding of the relationship between class and control expectancy and extends the findings to adults. It al-so confirms Dean's (1961) results of correlations between scores on a scale of powerlessness, and low socioeconomic level. It is interesting to note that the Project Follow Through educators, most of whom came from the ghetto area, were significantly less external than the project mothers. While

this may reflect the influence of being chosen for a relatively high status job), it may also be an indication of different sub-cultures within the lower class. As will be seen in the next section,


- 37 -




- 38 -


education was not a significant variable. The Project Follow Through educators obviously v-ere alert to infornation relevant to their life situation and took steps to im-rioo; tlieir life situation in securing, the job. Thus, this firidln, rnv bhe trCken as consistent with the studies presented in the literature review, indicating persons who know more about their surroundings and make conimitments to social action are more internal in orientation.

Hypothesis 2 was that black mothers would be more external than

whites. The mean SR1 score after 21 months of poartici-pation for 11 white mothers in the project (Grouns 1-3) ,!as 7.18 with a standard deviation of 3.63. The mean for 514 black volunteers was 10.25 with a standard deaviation of 3.511. The difference of 3.07 yielded a t of 5.18 significant at less than .001. Thus, hypothesis 2 is confirmed. As sugg-ested by previous research and theory, lower class ae: rae less likely to see reinforcements as contingent unDon an individual's skills and characteristics than the "opulation at large. Furthe-rmore, within our lower class Sami~ile, the percention of an individual as Powerless or externally controll ed was more extreme among black than white subjects.

Hypothesis 3 pi-edicted no significant relationships beft7een geor

Parity of mothers nd their I-E scores. Ta--ble 3 show.-s tho correlations between age, parity, and I-B' Control af ter 3 (Iore and 15 (-post) months project rt i c oT n.LI n c tln~ diffc r ncabtw e the t'7o sco-_2:



Since anl r pj;L __ tec1 tUi037wwr3h nCz.L: o-siefi:

results at the .05 levol, there no reliable relationship betc~l,1c:





- 39 -


Table 2

A Comarison of Three Populations UIitii Subjects in Groups 1 and 2 in the Present Study on I-E Control


Scale Uscd


Sex


N X


Prisoners, 18-26, 8th ! rade
reading (Ladwiz, 1963)

National Stratified Sam,.ple
10th, 11th, 12th grades
(Frankr-lin, 1963)

Follow Through Parent
Educators (Gordon, 1969)

Parent Education Project
Mothers-,groups 1 & 2


I-E I-E


SRI SRI


M 80 7.72 3.65


M&F 1, 000


8.50 3.74


F 40 7.21 3.58 F 42 11.50 3.14


*Si.-nificant at p less than 0001.




Table 3


correlations Bcte~en Agle and Parity and Initial (pro),
Final (post), & Changes in I-E Control Scores for 32 Subjects SRI


VariablePre (X=11.50
S. D. =3. l4)


Post (X=ll.0
S.D. =3. 14)


Change (X=0.50
S.D.=3.88)


Age (X--25.9:, S.D .=]lO Parity (X=3.41, S.D.=2.20)


40.27


+0. 32


+0.18 +0.31


Sample


S.D.


t


5 .4;9 * 5. 28*


5. 4 3


-0.02

-0.02


Va-riable-










MaternU_ Behaviors and Attitudes

Positive Verbal Interaction (VI+),

Mothers were grouped as either high or low depending on whether their positive verbal interaction index,calculated to stv-iarl-xze 21 months of visits by a parent educatLor, fe2.l above or be-low the median index of all others. Then they were split at the median again according to education. SRI (Miodif'ied I-E Control Scale) scores were then analyzed again at an initial point and near the end of the project. Change scores were also analyzed.

Hypothesis 4a predicted that mothers higher in positive verbal

interaction would have a more internal control expectancy. Hypothesis

4~b predicted these sa,,e mothers would become more internal by the end of a 15 month interval. Tables 4.1 through 4.4 summarize the mean I-E scores (SRI) of subjects grouped by positive verbal interaction and education. Tables 5.1., 5.2, and 5.3 show the analyses of variance of I-E scores at the beginning (pre), and the end (post) of the project and the change in I-E scores over the 15 month interval.

While there is no significant reJ ations;hip bctiacn positive verbal interaction tand initial locus of control, others with a higher level of verbal interaction are more internal in orientation cand have changed significantly more in an internf:l direction after a 21 month interval than those mothers with a lo-,r-r level of verl it-er-a-ction. There is no confir '-.tion of hypothesis a.Mothers higher in positi-vverbalizations were not more intern-dJ at the beginning of the Project.. The data suggest that level of verbal 5-y-teraetion is not directly related to locus of control. Ho~~~,it may be relvatcd indircuc-hap~~~~~~~~~~~~~~ tf.Oj C' r~' v''c> oILrU1f C':vc.~~






- 41 -


is, in ccnfir-.iation of hypothesis 4b a brother's levCl of positive verbal interaction with her chi]O predicted her moeatin control orientation. If one accepts thr assumption that the verbal interaction index reflects an important maternal skill, then these results conform to those of previous research. Those mothers more active verbally in mothering changed more in an internal direction. There is some evidence to suggest that this was an important skill for mothers in the project. Gordon (1969) reports sone significant correlations between amount of adult (mostly others) positive verbal interaction and children's cognitive develop:.,ent as ziiasured by the Bayley elid Griffiths Scales. One could speculate that mothers, seeing their efforts reinforced by their children's progress, beca!.me more internal in generalized control expectation. Negative Verbal Interaction NvI-)

To test hypotheses 5a and 5b which dealt with the relationship

between rierative verbal interaction (VI-) and I--E Control, the mothers were grouped in the sama way as for the dcIlta on positive verbal interaction with a Y.sedirn split on their VI- ind-;x Lcl ztrgain on education. Hypothesis 5a prodictcd thar2t rtr.other , with a higher frequency of negative verbal vr~t~m ith their child would be more external in

initial I-E Control. Hypothesis 5b was that the1 saT.1e r:oCheis vou.ld beconie more vxen, ~c 15 inonluhrtca.

Inspection of the! nzan I-Ta zcc~rcz:i ~h~~1~. d48a ef

the analyses of the variances in Tabic-1 5.4 to 5.6 reveals no sirnificant relationsi7hipo between lev 21 o.,f ncI-:v~~9 :~~~cc~ control expectanL-cy. H~pte~~5.: and 5) ; .~ ..1

it is interesting to note thel ic t~ of -11E:fnln~i opcc





- 42 -


of what was predicted. Rather, the effect of negative verbal interaction for the pre, post, or Change reasure o^r I-E Control is in the same direction as that of positive verbal interaction but it is nonsignificant. Perhaps the word negative is a misnomer here. Statements which go into the VI- index such as: mother "in a few words

directs the child to do or not to do something" might be construed as reflecting the mothering skill of setting limits. In that light, the division- of mothers on this variable might result in two groups of disciplinarians, one relatively active and one passive. Consequently, the prediction would be for those mothers higher in VI- to be and become more internal. The findings do not support that prediction either but suggest that replication on another sample would

be worthwhnile.






- 43

Table 4


Mothers' Initial (pre) , Final (post) , and Clir.na-'es in I-E Control Scores (SRI) Crouned by Level of Verbal Interaction and Education Positive Verbal. Interaction (VI+)


Frequency





Low VI+
High


Table 4.1 of Mothers per Cell Education Low High

9 7

7 9


Table 4.3 Post SRI Cell Means

Education

Low High

Low 12.22 12.29 V1+
High 9.00 10. 33 Column 10.81 11.19


Row,7

12.25 9.75


Table 4.2

Pre SRI Cell I-eans

Education

Low

Low 12.33
VI+
High 11.57

Columln 12.00


Vi+


Table 4.4 Chan-ges in SRI Cell

Education

Low

Low -0.11

High -2.57

Column -1.19


High 9.43 12.22 11.00


Pow

11.06 1.1. 94


M,,eans (Post-Pre)


Hig~h

2.86

-1.89

0.J. 9


,Rote7

1.19

-2.19






-~ !,4

Table 4, (continued)

Negative Verbal Interaction (VI-)


Table 4.5

Frequency of Notlhers per Cell

Education

Low High

Low 10 7
vi
High 7 0


Table 4. 7 Post SRI Cell M1eans

Education

Low Hi gh

Low 11.40 11.57 vi
High 10.29 10.62 C o 1u,, 10.44 11.07


Row' 11.47 10.46


Table 4. 6 Pre SRI Cell Means

Education


Low vi
High1

C o1Un


Low 12.50 11.29

12.00


H-Iigh 9.86 11.88 10.93


Row 11.41 11.60


Table 4.8

Changes in SRI Cell M11eans (Post-Pre)

Education


Low&

Low -1. 10
VI
High --1.00 Column -1.06


H ig(,h 1.71

-1.25

0.13


Row,, 0.06

-1.13






- 45

Table 5

Analyses of Varinnce of Initial (pre) , 'Final (post) , and Chanc~es in I-E Control Scores (SRI) by M4otherst Verbal Interaction Level and Education

Positive Verbal Interaction (VI+)

Table 5.1

Pre SRI Scores

Source Sums of Squares Dl? Mean Squares F

Rows (VI ) 8.127 1 8.127 0.86

Columns (Education) 10.003 1 10.003 1.05

Interaction 24.890 1 24.890 2.63

Within 264.985 28 9.464


Source

Rows Columns Interaction Within







Source

Rows Columns Interaction

Within


Table 5. 2

Post SRI Scores Sumis of Squares

52. 718

3.841 3.174

248.9005


DF 28


Table 5. 3

Chainaes in SRI Scores (Post-Pre)

Sums of Squares Dl?

102.240 1

26.240 1

10.286 1

330.349 28


Nean Squares

52.71S~

3.841 3.174

8.892







TMean Squares 102.240 26.2/40 10.286

12.034


0'~ 01 =- 7.


F

5. 93*

0.43 0.36










F

8 . 4 5 2.17 0.85






- 46 -


Table 5 (continued)

Negative Verbal Interaction

Table 5.4

Pre SRI Scores


Source Rows MV-) Columns (Education) Interaction

Within









Source

Rows Columns Interaction Within


Sums of Squares

0.28 9.07

20.99 275.80




Table 5.5

Post SRI Scores Sums of Squares

8.03

0.12

4.30

297.36


DF 28









DF 28


Table 5. 6

Chano~cs in SUI Scores (Post-Pro)

Sums of Squares DF


Source Rows Colu-mns


1

1


11.37 11.33

21.47 441. 84


Interacti on


Wi thin


1

2 F


Mean Squares

0.28 9.07

20.99

9.85









Mean Squares

8.03

0.12 4.30 10.62


11can Squpres

11.37 11.33 21. 17 15.78


F

0.03 0.92 2.13











F

0.76 0.01

0.40


F


0.72 1. 36










Education V2erbal Interactioo, and I-E Control

Hypothesis 6 stated that other's edticvtional1 leve'-l would not

significantly effect I-E Control and that education e2nd level of verbal interaction either positive or negative would not have a significant interaction effect on I-E Control. Table 2; shows small differences between I-E scores of mothc-rs high on educational level ai)J those low in education. The analyses of varianice in Table 5 indicate thecdifferences are not significant.

The data confirm hypothesis 6 and previous research that within a relatively homone,,,ect,3 pop'.fiwion there is no relationship betwaCe level of education and intern-3 vs external control expectancy. Edcation also did not interact siEgr'ificrntly with level of either

positive or negative intersection to affect expuect.,-ncy scores.

Infant C iiv:Dov lo-r-ent

Influence of Mother on Child

Hypothesis 7a predicted. thut children of i:others more internal in control. expectancy would learn rore and be ii ore highly developed cognitively by the tiri - the,,/ reached their second birthd-y. Tables 6, 7', anld 8 sho2,. the result,,. The asse,.srrcnt of the effect of mother's control ex D-ctancy was done in several. ways. First, to control for differences in the child's initial achieve:'ent, the child's achiev.ment scores werce plu.-Sed into the short terin prospective model (e~ and 1,en-r, l9i) h hlc rr2 grouped first accord-ing to

whether theIr i-hcr: wrc abOve o-r belowr the md~nin initial etc tancy, then nccrigto iwheither they werc- hig-h oir 3.o-., in in-iti-,:l


- 47 -






- )18 -


In apparant contradiction to the hypotheses, the children of initially more external mothers learned more althou-h the &.*fference does

not reach a significant level of probabili-ty. However, this is only true when the amount of achievement is calculated from the child's 6 month score. Table 7 shows that there is no significant relationship bet~jeen the child's initial and later achievement and little difference between children of external vs internal mothers in the child's achievement during his second year of life. In fact, because of a ceiling effect there is a reversal. Children of more internal mothers learn more in the second year. This suggests, since the series test had a finite number of items and there was a significant inverse relationship between initial achievem.-nt scor-2 and what was left to learn, that children learned about the sa-me number of items over t,,;o years regardless of their mother's expectancy but the children of the more external mothers learned more by the tine they were 6 months old. They then learned less than children of more internal mothers over the next 18 months. This is supported by the findings in Table 8. There is a difference of less than one iten achieved by chiildren of external v3 internal mothers at either 18 or 24 r.-onth of age. Thus, mother's I-E control expec'u -nc:y is nut rclk-tizd to hrcr chi-Id's progress on the series test.

Table 8 further indiatez; no si[gnifcnt1 Ylationship bct-,;OO1 YLztc,:rnal expectancy and children's scoirEs on the stand-U.kdized cognitive development scaics. Although these results are in the pi-edicted dirrction, the hypothe-sisc that brother's control ex>zi~yil~~czhor infant's cocmitive acevor:-n is ro- r itc7




- i49 -


Influence of Child on Vother

Hypothesis 7c predicted that a child's achievement would significantly affect his mother's I-F Control expecta-ncy. The hypothesis stated that, the children who achieved more or were more highly developed cog~nitively would effect changes in their mothers' expectancy in an internal direction. The results were analyzed the same way as for the assessment of the influence of mother on child. In this procedure changes in mothers' expectancy scores were p].uggad into the model(Table 9). There were no significant relationships between child's initial. achievement or mother's initial expectancy and changes in maternal expectancy.

Hypothesis 7b predicted that mothers of more highly develoncd children would be more internal in final expectancy. Children's succ,2ss over the course of the project as measured by their final Series Test scores and

the stand-a rdized tests was- not found to be sig-nificantly related to a mother's final expectancy or changes in expectancy. (See Table 10). Again, although h r:iost of the results were in the predicted direction, hypothesis Tb is not confir-med.

At the end of the project mothers of children who achieved more were

more internal than the others of less successful children. This is not true of changes in expectancy over the duration of the project. With one achieveme-nt mc-a:;ure (Griffiths) mothers of higher scoring children became more internal. With a second (Rayley) there was no differc-nce a-nd with a thirc, (Secries Tc~it) they becwrti ore external.

While the2rp is sone eviJdencLe that mothers' control. expect ancy and their childrens' achievement .!ffcct eac_,h other, the results were not reliable. However, the results wee5v'~t''aiwere c e. f:hrwt

a now sve! in StUr,'- TT.









Effects of Participation

Axnount ofPrtiain

Hypothesis 8: Mothers who participate: longer in a program (about

21 months) to imnrove their rothering sk1ills will bcoeroeintcrnial in I-E Control than tho3 l wo rarticiu!at e a shorter noint of tirme (9 months).

Insufficient datua weeavaila'ble to test diffe,.cnc ?s bet ;cen mothers

who did versus those who did not participate in the project. The parent educators were not able to reach enough mothers in treatment groups 3 to administer the SRI. There were 21 mothers each in groun l, and 2 who were given the SRI when their babies were 6 months old and again 15 months later. The tests betwee7cn the:;c groups which r-ppear in Table 11 assess the effects of m~aternal expectancy and expectancy chl-nces of 9 vs 21 months of participation in the project.

The results do not support the hypothesis that lon-er participation

changes control expectancy in an internal. direction. %'Others who parti-cipated lonr-er were m.ore extterna!l anf! charn-,ed more in an e-xternal direction than m.,others ,:ho did not participate in the second year of the project. The change for all 4~2 mothers was a m.,ean of 0.02 which sug -,est7, the project did not effect caesin control exineet,:ncy. IFo--ever, these findings are not too raaninr,,ful in the absence of E, control. gro,. A no trect~ment control group was available for Study II. Quality of Particinat-ion

Hypothesis 9a stated that mothers who make better u!se of tLh rcy

as manifested in sue>bd,'jiaw as, Tke:.-in[- trpoint!ments a-nd nasterinrterialIS !,ill *b yoiitm in initial T-7,' Control . 'uo,- h."r i 9b






- 51- -


Table 6




Chang-es in Children's Series Test Scores (6--24 months)
Grouped by Their Initial Score and Their 'Mother's Tnitial Expectancy


Child's 6




Mother's Internal (N=19)
Initial
Expectancy External (N=16)

Column M.eans


b.

df


Sre

Rows Columns Interact icia With in


*np = < .001


1

1


1 31


a. Means

Month Series Score

Low (N=16) Hiah (N=19)

35.33 27.08

35.50 K 3-1-.33'

35.44 28.42


Analysis of Variance

Mean Squares

38.34

302.45

32.80

14.07


Row 14eans

29.68

33.94


F

2.73 21.5*

2.33






- 52 -


Table 7




Changes in Children's Series Test Scores (12-24 months)


Mother's Initial Expectancy


a. Means

Child's 6 Month Series Score

Low (N=16) H~igh (K5)Row Means

Internal (~16) 11.17 8.50 9.63

External (14=15) 9.10 6.80 8.33

Column Means 9.88 8.07


Source

Rows Colunms Interaction

Within


b. Analysis of Variance DF Mean Squares

1 27.77

1 40.30

1 0.03

27 15.36


F

1.81 2.51 0.00






- 53

Table 8

Differences Between Succes.3 Scores of Children Grouped
According to I-E Scores of Their ',others
at the FEnd of 24 Months Mother' s Expecta-acy 'nt '.Iaure Internal Exter


Bayley Mental Development Scale

Griffiths Mental Development Scale

Series Test (24 months)


83.55 112.79


44.17


nal


80.50 109.56


43.53


t


0.97 1.18 0.56


Table 9

Changces in MNothers' Expectancy Scores (I-E) Scores (3-21 months) Grouped by Their Initial Score and Their Child's Achievement at 6 Months

a. 'Means

Child's 6 Month Series Score


Mother' s Initial Expectancy







Source

Rows Columns

In tera ct ion Within


Low High Row Neans

Internal (N='21) 1.14 0.36 0.62

External (N=21) -0.29 -1.14 -0.57

Column Means 0.19 J -0.14


b. Analysis of Variance

df Mean Squares F

1 20.012 1.30

1 6.2930 0.41

1. 0.012 0.00

3 5- 15. Z,1 5


Anlil ~'V&~f9C





- 54


Table 10


Differences Between I-- Scores (Finol SRI arnd- Chan-es Oviir 15 monthshs)
of Mothers According to Achiievoment of Their Children


Achievement 'Measure


Final (Post) SRI Score


N


Bayliey


Low 16


11. 13


9.94

Changes in SRI Score (Post-Pre)


0.37 0.37


0.00


Post SRI


t


10.94 10.63


0.36


Change SRI


24 11onth Series


Low 17 High 16






N

Low 16


10.11 Chang-e SJKT

-0. 44


tic~I . 106


t


1.7


1.03


High Low high


16 17


Crif fiths


Low 17 High 16


1.29

-0.63


1.47


Post SRI


x


10.83


1 0
u


t


0.59


hich Low


1 .10


x






- 55 -


Table 1.1


Differences Between Treatment Groups in Initial, Final and
Changes in Expectancy Scores (SRI) Treatm-ient Group


Mother's Expectancy


Initial


Final


Change


1 (21 mo) (N= 21)


10.43 11.24 0.81


2 (9 ma) (N=~21)


11.86 11.10

-0.76


t


-1.39

0.14 1.34


Table 1Differences in I-E Control Expectancy Scores (SRI) Between
Mothers Rated Eigh & Low in Project Involvement (PEP)


Mothers' Positive Attitude Index (PEP)


Mothers' Expcctancy Initial


Final Change


Low (N=20)

11 .30


11.45 0.15


Hfigh (N=22)

11.00 10.93.


-0.09 0.28


t


0.77




.56 -


stated that such mothers would also beoemo-Le internal in I--E Cont Lrol over a 15 month1 in1terval. In odrto test these hypotheses ohr

above and below the median on the PIT? index wer3 ccm~ared on their exnectancy scores at the beginning and end of the pro--Ject.

Table 12 reveals no significa- nt differences in expact.)ncy between mothers high on the PEP index Nerzus mothers low on the index. The two hypotheses are rejected. ?,other's &da,,ree of involvement in a program to improve ber mother'Lnhg skills vas, not significantly related to her I-E Control expectancy or her changes in expectzaicy over the course of participa-tion.

The reasoning for this hypothesis was in line with otIlher studies: those subjects who knew -lore about their role, whora1 a greater commitment to it and vho actually did perform., the-,-ir role2 in the pOroject better would be those- who were m.-ore internal in control ext.ectancy. It would seem that the r.mothers did not perceive their involvement in the project as crucial to their general life situation or at least not in th,, i-wa!y thet the items r~srdit. Some of the items, e.g. those dealin,- with the ntirber and z11nci11uaJAty of a-r-.noint-nments ken' ray reflect a middle class bins. Furthermore, it is Dossible, that the projcut wais not comletely succeofv~l in getting the other to identify with the role of her child's teache- dso the

heavy emphasis by the narent educators (Gordon, l9r(9) of the importance! of the mother in influencing hier child's cco-nitive gro;;th.

Perhaps a more fLruitf-ul w to conce-e of this variable is E-s an index of maternal activity whic-h in inteira:,ction w-ith ¬lhcr tenK






- 57



activity variable such as verbal level ihthave an effect on maternal expectancy. This was attempted in Study I.















CPTER 7 SUD : Co:C)' si:


As expected and in line with Previous research, the present smple of lower class mothers was found to be more external in locus of control than samples not restricted t~o the lower class. Within this leT)-er class sample, black moth-ers w-ere more external than whites. The age, number of children or education of the mother were not significantly

related to control expectancy.

The number of nositive verbaliz~tirie, of a mother to her child was related to her I-E CoDntrol with those miothcrs verbalizing more rore internal in control. On the other hand, ne-ative verbalizations w,,ere not related to I-E Co'ntrol; behaviors preniuried to riarif c-t nothers'I involvemi-ent in the project and the teaching role were not related to I-E Cntrol.

The prediction of a t'.To w-ay relationship between maternal expectancy and her child's achieveme- nt wa~s not bo-:ne out. aTe findings were suggestive of a rnelationshin but were not reliable.

In an naturalistic stud0y such as this with srjva-riables beyond the experimenter's control, th.- reliability of the data increases as procedures irnrove with the benefit of practice and hnsgh. For exearnnle, in Study I time needed to solidify data collection jpl-_s precluded the availability of a nurorua'e of the mriother's initi-I expectancy. 7n at i s, shbce en tcc:d th ie pr oj ect when1 h er b n:~tw


~2






- 59



months of age but was not given the SRI until three months later. Thus, possible changes in expectancy during- the first three months of participation might have been lo:st. Furthemore, missing data prevented the avail-ability of a control group to contrast with a

participation group to assess the effect of participation on control expectancy. Study II was planned with more adequately control-led procedures to get a more reliable assessment of the variables investigated in Study I.
















CHAFiP IU'9~0~iCIU1,0 1'-"JD 11


Three new groups of mothers were available for investigation of' the relationships between their level of verbal interaction, their degree of involvement in the project, the effect of their participation in a prog-ram--. to enhance their bothering skills and their I-E Control. The two way relationship between mother's I-E Control and her child's achievement and the effect of class, race, education and parity of the mother and her I-E Control were also re-investigated

using this new sample.

Better controlled proc-edures werc used in Study II. The parent

educators wore rore experienced researchers. That is, their obser-vz,tion and recording skills had improved. Entry scorc2s irnstkcaj of entry plus 3 month scores on the I-E Control Scale were available as the measure of initial expectancy. The re--test on the scale was given at the end of the mothers' part cipation in the progru. ,- instead of 3 rzionths before the end of the project as the! measure of final expectancy. Thus, chan~jcs in I.-E Control s'.z(final-initial1 scores) more accurately refJlectced chan,7es in expectancy as effected by project participation, and rmaternal behaviors ru2:?n, du thc course of the project. Fvrthcrnore, the test-rutc,:t 5.-r-tcvrJ v7;0r. shorter (9 months). Con-:t~ucnt1 y, the rctez~t raliabil;ty of the I-IB Sc-ale whiell teont,t; 'cc) ~~>i9~o:I< :~


- 6 r - -


U_





- 6 j. --


than 6 months (Kiehibauch, 1963) was increased. Finally, tw,.o new control groups were used: a non-participant control Croup ',Lcs available to best the effect of participation on I-E Control and a control group that, participated in a r~odified program to toest the effect of participating in a specific type of prcgrax- on I-E Control.

Some conceptual changes were made in line with su-r-e-tive findings

from Study I. The negative verbal interaction (VI-~) and positive attitude (PEP) indi-ces -we2'e conceived of as measures of maternal activities which singly or in interaction would be related positively to internal

control. When these r:esures did not relate sig-nificantly to I-E Control in Study I, the items vere re-exm- .ned. The author felt that these items perhaps reflecte-.d the mother's level of activity and P'nouilt of involvenie-nt with the role of her child's teacher in the sane wsy3, that the itemi which make up the positive attitude index (VI+) do. In fact, the VI- index varied with I-E Control scores in the same direction as but with slightly less nagni-tude, than the VI+ index. In addition, the author was also interested in investigating whether certain combinations of these variables would interact to effect I-E Control.

Level of mother's education showed no reliable tendency to affect mother's I-E Control and also showed no significant interaction with verbal level in its effect on I-E Control in Study I. Consequc:ntly, the prediction in Study II was that education would be iiinfcml correlated with I-E Control.




















The hypotheses for the second study a:-e the sare~ as for Study I except for the changes discussed in Chapter 8. Demoraphic Variables

1. The present sample of lower class rothers will be more ex-terna. in I-E Control thaxi previously studied sa-mplen that irer.3 not rcZ!tricted to the lower class.

2. Blatck mothers will be more external than -rhite others.

3. There will be no siginificazit correlations betWeen r:othe,re Vge, parity or eductation -,nd I-h Co-Atrol. Maternal h2.or

14. Mother3 who have a higher frequency of behaviors r.21ated to their role as child's teacher: a. Positive vcerbalite:tn(:) b. NegEative verbal interaction (I-) c. Positive -Utti!.0~etcnr. a progr7Am to enhance her teaching skills (PEP), will be rmore interna:L in I-E Control than mothers with a lower freque .ncy of these behaviorS. 5. Mother with a higher frequeanclr of behaviors li-.t1ed in hypothesis ha, b, and c will also b Yco-~eore it:nlin T-E Control over a 9

month iritervai.

6. There will be a significtant interaction effect between rotheri' positive verbz .I inte-raction and poziti\ve attlitu"1 : acp;zz"rr.s a to enhfa:eri, her- tre:? cl-il-fw z~ljo (tY c.' - cAo iCti~






- 63 -


7. There will be a significant interaction between mothers' negative verbal interaction frequency and positive attitude to-wards a teaching prorra (PEP) on I-E Control expectancy changes. Infant Comitive Development

8. There will be a significant two way relationshin between mother's I-E Control and her child's cccnitive dcvel opn-nt such that: a) Children of others more internal in I-E Control will learn more and be more highly developed by the end of their first year of life than children of mothers more external in I-E Control.

b) Mothers of children who aro more highly developed co-nitively will be more internal in I-E Control than mothers of less well developed children. These mothers will also become more internal over a 9 ronth interval.

Effects of Particibation

9. Mothers who participate in a program- to improve their mothering skills will become more internal in I-E Control than brothers who do not.

10. Mothers who particip: te in tw~o different programs to improve their smothering skills will not differ in I-E Control score chang-es over the prog ram 's duration.

















CIU.PTEB, 10 STUD)Y II: 1.4LTH0DS


Samnipe

The subjects were selected usinC the samez criteria as for Stud3y I.

Thus, they were fro. i indigent families. Infants with complications possibly related to coEgnitive development were screened out. These 54 mothers snd their infants began participating in the Parent Education Project (see Chapter 2) one year af'tcr tho subjec-ts in Study 1.

Design

The treat!-ent pltn for the groups in this study is outlined in Table I (pagec 5). In that table they are referred to as groups 5, and 6. The first experimental group (here called El) received weekly visits from a parent educator from the tire the baby was 3 months old until his first birthdz~y. The parent educator used teaching materials designc-d by the profh esi on,-,. staff of the Parent Educa.tion Project (see page 26, Stimul-ation Series) in working with the mother. Another exoer incntel group (E2) also received weekly visits froma parent educator. Flo;:veir, theso ej~ucators, also indigenous non-professionals, pILnned their own tc nchinr prograi., beL c:, thcir c experiences working, in Head Sta-rt and nursery school zc'r:. supervised by a resorrch za sistc-..t on the Staff of the Parent Mrucatic,-: Project but one who h-2d no specific kno-wled-o of the Series Test. A control groiw ( c) of -~o",ers in .J ~ ~~v1retc~ t~


- C ": -,









tested.

All three groups of mothers iw.ere given the I-E Scale (SPI) at

their entry into the project (in%.',:nt is 3 months old) and ar-.'-in at the end (infant is one year old). r~h infants werc given. the Series Test at 6 months old and at thcir first birthday. They were also given the Griffiths Test at one year old. Except for the control group families who were not visited, the parent educators recorded their observations after each wekyvisit on the PFJ.-T, Data Analyses

The measures used are identical with the instruments and indices

used in Study I. The hypotheses concerning class and race were tested by t tests of differences botcen rgroups on their TE score!3. The remaining demogranhic variables w.erc tcstcd ca-lculatinr a Pearsonian r between each independent variable and I-E Control scores and a iP2n]tiple r between all the indenenclent variaiblos and I-F scores.

The hypotheses concerning maternal behaviors were tesztcd by t test of mothers' initial (pre), and final. (post) I-E scores by an analysis of variance of chiancc3 in ifothcrs' I-E Control scores grouped by median s-olits on the three maternal behavior indices.

The relationrhiTn between maternal T-E Control and child coj-nitive develonn)_ent was tested first by srlittnr, rntesat the mc~ci~an on their I-E scores ail,! c'1ein a t for the difference between their children's ac-ve>2nt scoresi., then by sul.itlting children at the median of their achievement scores andl doin- a esIf h di fr


ence between their notbers' I-P C .cs The effects o r p tiint7:cvc a cc s tcu~ f4>C2

ences between eyx: i~retal arv! cT-a rrsonte~-fn1 17'





-66



and I-E score changes. The di~'~2 ewcnthe two exnri%.ental

approaches was as,:essed b a t test of the differzmcas betw..een groups El and E2 in final I-E score adI-E score chan,-es.















CHAPTER 1.1 STUDY II: RESULTS ATD DISCUSSION


Demogra-phic Variables

Class

Hypothesis 1 states that the present sample of lower clasc. mothers would be more external in I-E Control than previously studied samples not restricted to the lower class. The results in Table 13 replicate

those of Study I. The National Stratified samp.,Ile and the prisoners (from varied socioeconomic backgrotunds) were significantly more internal than the present sz-peof indi_ gent mothers. Thus, hypothesis

1 is confirmed.

The comparison of the Project Follow Through educators wvith the disadvantaged brothers again indicates within class differences in I-E Control. Again on:e can specul2ate that the vomen hired for Follc,, Through have received reinforcement for their personal~ characteristics rnd skills. This is consistent with Rotter's (1966) theory that co,)sistent reinforcements for personal qualities and abilities leads to adoption of the internal control expectancy. Race

Inspoctiona of the data in Table III reveals that the h!-,;_T thesis for c racial difference in I--E Control svp-,ort .Cd by th(2 Stl.,Cy I findings is not supported here. LlthcgV Thro rothcrs are ra ( crn thani Caucasivn notbers both at the beginning ard end of their pro~leet pa~rti~fcAior , hz~ di f~~n x~noat '.T asirieiticA


__ (), I " -






-- 68 -


Table 13


A Comparison of Three Different Populations
with Projcct '-others on I-E Scores


Scale


N X


National Stratified Sample 10th, 11th, 12th grades (Franklin, 1963)

Prisoners, 18-26, Sth ,-rade plus reading (Ladwi,-, 1963)

Follow Through Educators (Gordon, 1969)

Project Mothers


*


43 1 i


I-E I-E


SR'I


SRI


1,000 8.50




80 7.72 40 7.21 54 10.04


13.99 13.32 12.82


14.59


2. 91* 3. 610-3


.01 level of significance .001 level of significance







Table 14


Differences Between Black and JHhite ',others
in Initial and -Final I-E Control. Scores


Initial Expectancy


N


Blacks Whites


43 11


X

10.26 9.1.8


t

0.84


Final Expectancy


Blacks
1,4hit e s


9.02 7.61:


1.15


Group


S


t







- 69


Table 15


Correlations Between Education, A e Parity and
Initial (pre), Finial (post), and Chan-ges in I-E Scores (1,123)


SRI Score Education WE


Pre


Pos t


Change


.00


-0.13

-0.08


Agc (A)

-0.22

-0.30

-0.13


Parity (P)

0.20 0.25 0.01


Multiple SRI EAP

0.26 0.38

-0.14


10. 21 years


19.25 years 2.61 children


Table 15. 1


Analysis of Variance of the Multiple Correlation of A-ge,
and Education with Final I-E Control Scores


Parity,


Nul1tiplIe
r


Source


0.30 Regression
'Residual

0.37 Regression
Residual

0. 38 Re.-ression
Residual


Mean
df Squares


1 26

2 25

3
24


29.37
11.54

23.63 11.29

15. 80 11.75


Means p=. 05 =0. 37


Variable Entered


Age


Parity


Ed ucation


F Ratio

2.55 2.09


1.35


p-=.05

4.22 3. 3! 3.01





- -j0( -


It should be kept in mind, hcer, that the mean difference of 1 between blacks trnd whites is typical of previo-us stud ies. Defco"Irt and Ladwig (1965)) found P. siTlr differc-nce which held up as significant on a larger IN (120). Kichlbauch found a slightly smaller difference in the sam e direction on P-n N of 80 which was also non-significant. If Study I is included, ell four studies con2bined show an average difference of more tha-n one. While the question bears further investigation, a review of all the studies strong-ly sujGgests there is a racial difference in I-E Control albeit small as measured by Eotter's scale.

Age, Parity, and Education

Hypothesis 3 predicted that there would be no significant correlation

bet-ieen mother's age, parity and education and I-E Control. Table 15 shows the correlations of these three variables with initial (prec), and final (post) exp _ctancy (I-E Control. scores) aridl che-s in expectancy. The correlations are s7.-all (o to 0.25) and ins i gni ficaint. Table 15.1 indicates the degree to which the variables predict expectancy scores if weighted and combined multiplee r). The varian1ce indicates prediction of I-E Control scores front , parity and educaltionfal level of the subjects is non-significant.

Hypothesis 3 is confirmed. As predicted, age, parity, and ed'uczt1ic-n of the mother was not sitg.nificantly related to I-E Control. Even in combination these three variables could. not pre dict expctc- y changes significa-ntly better tha -n ehlane.

Materrn1 Th'

Pos iti ye Vei1 Intc!-a t ion (V .,)

Hypoci ' ttt~~ th1 Cc; S;t) _21 hS r of






- 71. -


Table 16


Differences in Mean Initial & Final I-L Scores
Between Mothers Above & Below the Median on Behavior Indices


Initial Expectancy


Final Expectancy


Behavior Index Positive Verbal
Interaction (VI+) Negative Verbal Interaction (VT-) Positive Attitude Toward Project (PEP)


N

High 14 Low 14


x

10.43 9.86


High 14 8.64
Low 14 11.64

High 14 9.57
Low 14 10.71


*p< .05


X


t


t

0.39


2. 21 0.84


8.86
7.43

7.93
8.36

7.29 9. 00


-1.03


0.33


1 .30






- 72 -


Table 17



Analysis of Variance of Chan-es in '!others' I-E Scores
Effected by Positive Verbal Interaction (VI+) and
Positive Attitude Tow,,ard Project (PEP)


a. Means




VI+ Low 01=14)

High (N=14)

Column Means


Positive Attitude (PEP) Low (N14) fHigh (N=14)

-1.33 -4.40

-2.40 -1.11

-1.71 -2.29


b. Analysis of Variance


Source


Rows Columns Interaction Within


df








24


Mean Squares

7.936 5.079

30.489

16. 887


Row :*eans

-2.43

-1.57


f

0.47 0.30

1.81





- 73 -


Table 18


Analysis of Variane of Chan es in MToi.hers' Expectancy
Effected by Ne; ;ative VI & Positive Attitud-e Positive Attitude


a. Means


VI- LOW (14)

High (14)

Column 'Means


LOW7 (14)

-2.00

-1.33

-1.71


HigQh (14)

-5.00

-0.25

-2.29


b. Analysis of Variance Source

Rows

Columns Interaction 11ith irt


df








24


Mean Squares

50. 293

6.298

28.583 15. 201


Row , Means

-3.29

-0.71


f

3.31

0./il 1.88










positive verbal interaction would be more internal in I-E Control than mothers with a lower frecquency of positive verbal interaction. Hypothesis 5a was that those maothers w,,ould. also become more internal ovrr a 9 month interval. Tables 16 nnd 17 sur=.iarize the findir.-.s.

Inspection of the rea!ns and t SCo2re in Table 16 reveals that mothers above and below,, the median on the VI+ index do not differ significantly in initial expectancy. This finding is consistent with that in Study I. However, contrary to the finding in Study I mothers high and low on this index do not differ significantly in final I-E scores. In addition, inspection of the mean change I-E scores and the ro,7 variance in Table 17 reveals these mothers do not differ significantly in expectancy changes. In fact, these findings tend to be in the opposite direction. That is, mothers with a higher frequency of positive verbal interaction are more external after a, 9 month interval anid have changed maore in an external1 direction. Hypotheses 5a and 6a weare not confirmed. The significant findings on positive verbal interaction and I-E Control in Study I were not replicated. Negative Verbal Interaction

Hypothesis 4b stated that mothers with a higher frequency of neg7ative verbal interaction (VI-) would be more internal in I-E Control than mothers with a lower frequency of negative verbal interaction. Hypothesis 5b predicted that these mothers would also become more internal over a 9 month interval.

Inspection of the r.eans and t score in Ta ble 16 reveals that mothers above and below thle niedian on the VI- index dif.Cer significantly inl I-E Contfol.. AS orcJeer)ar:hir~ n ~S ~ v ~II~5]


- 74 --





- 75 -


more internal. However, after a 9 month interval there is no longer a significant difference between roth-ert hig.h and low on this index in their I-E scores. Inspection of rcan I-E change scores in Table 18 shows that mothers low on VT- change considerably more in an internal direction. While the change is relatively large, analysis of the row variance sho-,,7s that it is not significant and the find-ings tend to be in the opposite direction of Study I in which mothers higher in the VI- index became more internal. Differences between these findings in Study I and Study II will be discussed, at the end. of the section on maternal behaviors.

The hypotheses are generally not supported in Study II. However, one wonders about the relationship between the VI- index and initial expectancy in Study II. Mothers who use raore negative verbalizations

initially are significantly more internal than those who use less. Over the duration of the project mothers lower in negative verbalizations change more in an internal direction. Thus, at the end of the project the difference between high and low negative verbailizers has becoieL insignificant. Perhaps participation in the project has a leveling effect on I-E Contz-ol which obscures other variables. A

control. group of non-participants on whom behavior indices were available would be necessary to test the siprnificance of this effect. It also must be keut. in mind, that the scale h,- a floor of zero. At a mean of 8.64~, the scores of the high VI- mothers were not as free to vary dow nward as the scores Of the low. VI- others. This s s more Optimal instrur.-nt, for this variable ,would be one wc~JT~iU






-- 76 -


Positive Attitude Towards a Projrat-i to Enhance Mothers' Teaching Skills (PEP)

Hypothesis 4c stated that mothers with a higher frequency of behaviors which reflect their positive attitude towards a program to enhance their teaching skills (PEP) would be more internal than mothers with a lower frequency of such behaviors. Hypothesis 5c predicted that these mothers would also become more internal in I-E Control over a 9 month interval during which time they participated in this program.

Inspection of the means and t scores in Table 16 and of the analyses of the column variances in Tables 17 and 18 reveals small but non-significant differences in I-E scores at the beginning or end of their participation in the program and in changes in I-E scores between mothers high and low on the PEP index. This is consistent with the Study I findings. Mother's attitude toward the program she is participating in as measured by the PEP index was not found to have any significant, main effect on her I-E Control Expectancy.

Positive Attitude and Positive Verbal Interaction

Hypothesis 6 predicted an interaction effect between positive verbal interaction and positive attitude towards the project(C PEP) on I-E Control expectancy changes. Inspection of the interaction variance in Table 17 reveals no significant findings. Hypothesis 6 is not supported.

Positive Attitude and Negative Verbal Interaction

Hypothesis 7 predicted an interaction effect between negative verbal interaction and positive attitude (PEP) on I-E Control expectancy

changes. Analysis of the chaig score variances in Table 18 reveals no significant interaction -zrm. Hypothesis 7 is not supported.





- 77 -


Positive attitude towards or active involvement with a program to

improve child teaching skills as assessed by the PEP index did not have significant main effects or interaction effects with level of verbal interaction (positive or negative) on mothers' I-E Control expectancy. However, it is interesting to note (see Tables 17 and 18) that mothers lower in verbal activity (either positive or negative) and higher in project involvement became the most internal (a change of 4-5 scale points). This may suggest incompatibility between the two maternal behaviors. The observations which go into the behavior indices are made when the parent educator is working with the mothers. Those mothers who got the most out of the project may have spent more time listening to the materials presented than talking to their babies.

Maternal Behaviors Summarized

Of the three behaviors studied (VI+, VI-, PEP) only negative verbalizations were related to mother's control expectancy assessed when

her baby was three months old. None of the behavior indices were related reliably to final expectancy, neither were they singly or in interaction significantly related to expectancy changes. Final expectancies were in the predicted direction for positive attitude and level of negative verbal interaction but not for level of positive verbal interaction. That is, mothers higher in frequency of verbal commands and negatively toned verbalizations during their child's first year of life were more internal while those higher in frequency of verbalizations of tender, loving care were more external. The findings on positive verbalizations are opposite those of Study I, and Study II involved babies one year younger. Perhaps the mothers found negative






- 78 -


verbalizations more effective in coping, witb younger infants, and posi-tive verbalizations more effective in coping with oldT.r on~es. Possibly this greater or lesser degree of effcctivcn.-;,zs was slightly rel,1it_-d to I-E Control.

Inconsistences and non-reliable findlings in the two studies on the maternal behavior variables End I-E Control may also be a fu-nction of

the limitations of the measures used. While validity studies of the behavior indices have been reported (p23) no refinement of the internal consistency of the items or investigation of the factorial composition of thle indices ha!- been attempted. Quite possibly individual items or factors from these behavior indices might provide a better test of the relationship between the variable each index is purported to measure and I-E Control. As presently constituted each index is a rough measure of its respective variable.

Theni too, the I-E Control Scalc, as was me,_ntioned in an earlier section, has a low re-tcezt reliability over an intc-cval beyond 6 months. If changEs in epcty o';--zr tim,- aire to br' related to other-I Variables such as maternal behaviors in future research, it seems necessary to us-e a control group for which measures on such Naritibles are aveaileable and/or to develop a nore stable measure of I-E Control.



Hypothesis 8 predicted a significant two way relationship betiweecn mother's I-E Control and her child's cognitive development. As in Study I there were no siGnific-ant relationships in either direc2tiou%. The findings in Table 19 indica te no( si
I-E Control on her child's rcw o : n th, , eic Toet at either C






- 79 -


(Table 19.1) or 12 months (Table 19.2) or on his overall intellectual growth as measured by the Griffiths Test (Table 19.3).

Inspection across Tables 19.1, 19.2, and 19.3 reveals that the

children of initially vore external (higher scores) mothers progress more on the series by the tire they are 6 months old. Then perhaps (because of missing data they are not exactly the same subjects) there is a reversal. The children of initially more internal (lover scores) mothers are slir-htly ahead on the series by their first birthday. These same (with two exceptions) children are also more highly developed in intellectual growth as measured by the standardized (Griffiths)

test.

At 12 months old (after 9 months' participation in the Parent Educa-tion Project.) children of -.cAhers ,rho ar- .ro acre internal at the end of their p tiiic hxelearned scn.ht ore on the series and

score higher on the Griffiths Test. However, none of these trends are significant.

P).rthernore the result: a 'e c~eled by the reversed relationship between chan-es in r'~ther's I-E Control and her child's achievement on the Series Test and the Griffiths at 12 i.,onths. That is, the children of mothers who change more in an internal direction are less developed cognitively. Since- mothers who were more intern9,.l at the pre and post test of the SRI had chil-dren that were mocre developed cognitively the rcevcrsal of this relationship for the SIRI ch-tngre measure suggests different mothers are grouped as interc-al for each of the SRI measures. As in the preceding section of this paper, t1his arfaill ciuesticns h re-test reliability of the PRSae

The f inli n~ in Ube20 indietel no sgifc fetOf aci'






- 80 -


Table 19

Differences Pet..re~n Achieve ,1 nt Scores of Children
Grouped by M~edian Split of T-othe-r's Triitlal (Pre), Final (Post)
and Changes in I-F Control (SRI) Scores Tabhle 19.1

Series Test Scores (6 months)


PreSPI


t


low (iT(8) 9.13 high (t'=6) 10.33


Table 19.2

Series Test Scores (P. months)


t


PostSRI


low (N=15) high (N=13)


314.60
33.62


0.59


low (,,=14) 34.17 high (Ir14) 33.41


Change SR' low.. (N=16) high (NI=12)


32.67 35.11


Table 190.3 Griffiths Test Scores

t Post SIT


low (N=l4) 112.00
high ('11l2) 107.0


loW (N=13) 111.77
hig-h (>d)107.r,9


Change SRI lo, (N=:l4) 1.09.0 hirgh ( T'= 1 2,) lj)h. 25


Pre S91f


t


0.50


t

-1.35


Pre SRI


t


1.05


t

-0.91'









achievement on his mother's I-E Control. Prorezss on the Series Test at 6 months vas not significantly related to mother's initial expectaney (Table 20.1). Children vho learn more initially have more external. mothers but by the time they are a. year old children who have progressed more on the series have mothers with the same expectancy as children who have oro,-res!:cd less (Table 20.2). Children who scored higher on the standardized test of infant mco-nitivc growth hav~e -iothers who are more internal in final expectancy.

Table 20.3 indicates a surprising- but insignificant trend. othersrz of children who score lower on the achievement tests chran:e more in a-n internal direction. This is opposite from what one -vould expect from previous rcsearch indicating, an inverse relationship between failure and internality as surzmari7.ed in the literature review, of this paper.

The hypothesis of mutual influence between maternal T-1- Control and child achievement must 'be rejected. A mother's cx-ncctancy about events in the world is too general to be influenced by a specific achievement of her child. This is perlhais too small a sample of her infant4-'s behavior after too shcrt a tirae in the infant's life to chrnre her overall e)M-ctancies of internal control. It probably takes a wider range of her own and her child's lifc activities and cicm~ecsto influence this orieritation. Even within the area of possible influence on her child's aChiC-veC ment, especially considering the variability of intelligence tests at this age, the effect may be reduceO by the -lackz of reliebiDity of these tests as indieattor, of the child's overvIl. co,.nJii-Jve gro-.ith. The lack: of reliability of this rel1ationshiin is f~hrrde. yt :civ

of many intervenir,7m variables sueh as3 ther c -'iCIclt <' under which the: attitude and intelle2cttual. tc- ts :r 4 e






- 82 -


Table 20

Differences Between, Initial (Pre), Final (lost), and Chrrcnes in
I-E (SRI) Scores of M4others Grouped by U*edian Split of
Children's Achievement Scores
(6 and 12 Month Series Tests, and Griffiths Test)


Table 20.1

Pre SRI Scores


Series Test
(6 month)

low (1-47) 10. 10 high (N=7) 10. 10


t


0.00


Table 20.2

Post SRI Scores


Series Test (6 month)

low (N=3) 8.25 high (N=9) 9.44


t


-0.74


Series Test


low (:1)9. 30 high (N=12) 9.30


CriFfiths Test low (N.=12) 8.75
high (,N=14) 7.43


Table 20.3

SRI Score Chang(es


Series Test
(6 mon01th)

low (N=7) -3.00 high (N=7) -2.14


t


-0.40


Series Test


low (i--1) -2.50 hig-h (=1)-07


Grif fiths -Testlo,. (1=12) -2. 92


t


0.00


t

0.97


t


-1. 06


t

-1.03










(Gordon, 391.9) and the low re-test reliability of the I-E Scale for an interval greater than b months.

Variables als o inte-vcin-_ tctwcan mother'z expectancy of I-E Control and child achieve: -it -vhich reduce the strength of relationship between the two variables. As was seen in the irnnidiately preceding section of this study, behaviors (herein called PEP and VI- respectively) which researchers have found related to children's achievement, such as mother's --nount of participation, and frequency of imperative statements without rationale (Grotberg, 1969) were not related to control expectancy.

Effects of Particination

Hypothesis 9: M~others who participate in a program to improve their mothering- skills will becze r2 ore interna-l in I-E Control than mothers who do not.The relevant results are su.-mar~ized in Table 21. Hypothesis 9 is

mostly supported. While thcre are no significant differences between the experimental and control groups in pre or post test means, changes in expectancy approach the .05 level of significance. The &actual probability level, is .07 for a two tailed test. This seenn close enough11 to the conventional level to consider the results significant. Thus, mothers who participate in the project become significa-ntly more internal in control orientation than thc'sc who do not. This is consistent with research findings (Neal &-nd Seen.Ln, 190"L; ScernLn and Evans, 1962; ~ 1063; Kielhlbv.u h) 19168; Stricklncnd, 1965) that

persons affiliint,,cl ;:-ith so~ioal action group.is or involved in atvte to Tiromote i -irovix--cnt in tlncir JAifa cv -;t1mc 1 -0i"'- ! than h:c POrc c.


- 83 -






- 84 -


Table 21

Dif fer, nces in Initial (pre) , Final (post) , and Changes in
I-E Control Expectancy Scores (SR.I) Between
Experimental (F) and Control (C) Groups


Pre SRI X

10. 14 9.92


t

0.21.


Post SRI X

8.14 9.73


t

1.45


Cliang-e SRI X

-2.00

-0.19


Table 22

Differences in Initial (pre) , Final (post), and Chan7,es in
I-E Control Expectancy Scores (SRI) Between
the Two Exoerirnantal Groups (E1+E2) and the Control (C) Group Group Pre SRI X t Post SRI X t Change SRI

El (N=14) vs 10.07 -0.10 8.93 1.20 -1.14

E2 (N=14) 10.21 7.36 -2.86

El vs 10.07 0.311 0.93 06

C (N=26) 9.92 9.73 -0.19

E2 vs 10.21 0.24 7.36 -2.10*k -2.86

C 9.92 9.73 -0.9


* .07 probability level S.05 probability level


IV


Group E (N-28) C (N=26)


t

1. 87


t

1.12




-0.75










In the Head Start Program (Kitano, 1967 in Grotbe20, 1969) designed to assess the effects of participation in a parent education program on parents, the author found that particiao uas not significantly

related to differences in powerlessness as naasured by the UCLA Alienation Scale. In view of this finding, despite the fact that a different instruments w;as used to meas-ure locus of' control, the lack of positive findings between the two treatment groups in Study I and the marginally significant finding in the present study, further investigation of the issue seems ncessary.

Hypothesis 10: others who participate in two different programs to improve their mothering skills will no differ in I-E Control score changes over the program's duration.

Table 22 suzrzsthe relevant, findingrs. Hypothesis 10 is supported. There are no significant diffcrenccs betwe.-en experimental groups in expectancy. However, the more imn,,orttant finding is that while there are no significant differences in final expoc tancy and expectancy changes between rneirei group 3- and the control group,

there are between experimental group 2 and t-he control group.

This suggests that the parent educators (whlo hadl cxpei:e in le'ad Start anid nurzoery program_ Is) who part ic i ated in Clcsiarling thr? s-iuulation materials -nd irothers' education program had L significant effect in chri-rn- ~t~s extcctncics. Thre parent eductors ho u'sed materials zrid pro-r&:t 3 designed by the professional. staff rather than designing, their o7,n did no' si EBi j cam i 3y a.^-" ct thec c :ngin.expectanicies oP the mothers wit~n ,hc,_- thc -'v:~d

This finding- was siauricing c. drn h erx xa nbt


- 85 -





- 86


treatment programs, on imbuing th other wit," thle feeling thavt her involvement in the program was crucial -to her chiild's cognitive growth (Gordon, 1969). Perhaps the educators who planned their own program developed or already had a belief analagous to generalized internal. control: That is that pa-t-ticipation in their progreni rather than external variables would result in the others' improvement in child teaching skills. Perhaps they communicated this belief to the volunteer mothers they worked with.

Since the two exjperimnental groups of mothers and children did not differ significantly in success as measured by the children's final achieve-mnent scores (Gordon, 19069) it seems reasonable to assume that characteristics of the parent educators had a significant effect. Previous research in social psychology suggests that the parent educators who

planned their own progre-,,n would be more involved in and committed to their jobs.

Studies by Lewin (1958), Pelz (1958) and their colleagues found participation in a group discu~szioni in whiIch a decision was made Twas significantly more effective in changing the social. conduct of mothers than lectures, individual instruction or group discussion without group decision. Group discussion and decision were felt to be necessary for effecting incased involvement in and co Mlitment to an action.

It is interesting to note that the tcchniquc., used to teach the parent educators who worked with the experintal group 1 Trrothers included lectures, individual supvrviszio! and pro-,-; dliscussion. The professional staff was senzitive to the educatorr. ineliv~i-ual rieeds and

pr~hcm cla Xv o i: svr~edter J~sfol into the procedures.

Howevo-,0ny the cXZ20Swho work -ed vith the nothers in enrmna










group 2 decided as a group what the exact nature of their prog-ram would be. Consistent with the Lewinian studies then, these educators would be more actively involved in and cc:.ii-tted to their program. This greater involvement could have rubbed off on the mothers in experir~nLal groi~p 2 resulting in the mothers' developing a stronger belief in an~ individual's ability to influence her own life.

There are, of course, other possible differences between the educators of the two groups which may have tiude them more effective in altering existing expectencies. Most obviously, Group E2 parent educators probably had more experience in working with disadvantaged families. Assesshi-nt of differences in characteristics between the parent educators related to their effectiveness in giving nothers a belief that persons can control their o~ndestinies could be the subject of another study. In fact, the whole area of the etioJLoLgY of control exoectcencies is relatively virginal research territory. This author was only able to find one stud~y. Tolor and Jalo'J.,ec (1968) found that college students who rated thei-'r others higher in authoritarian attitudes on the PARI were more external than those that did not.


- &Y
















CHAPTER 12 DISCUSSION


The major objectives of the tw,,o studies in the present research were to investigate internal vs external control expectancy among disadvantaged mothers, and examine its relationship to maternal behaviors,

to their children's cognitive growth and to assess the effectiveness of participation in a parent education program in changing expectancies in an internal direction. Presumably if this change occurred improvement would also occur in the children's intellectual development.

Existing expectancies among the predominantly black indigent population sampled were found to be significantly more external than sarnples reported in the literature which included a wider social class range. Blacks were more external than whites although in Study II the difference was insufficient for the number in the sample to reach the conventional probability level. However, the difference was of the same magnitude and direction as those studies reported by Rotter (1966).

The importance of these findings is suggested by Lefcourt (1966)

summarizing the arguments about social class, race, and expectancy. Individuals within the lower socioeconomic class, particularly blacks, early perceive impediments in the way of goal striving. For example, segregation and discrimination convey to blacks that their own efforts will lead to no reinforcements. Thus, they come to disbelieve that

the efforts of an individual pay off.


- 88 -




- 8) -


This i~s not true for all disadvII-anta-,ed people, as the inclusion of the I-E sco-.es of Project Follow; Through (proad !Id after Head Start) educators in tha Drenent sui dntad.Although from the saxme

backgroimd of th-- indint, voluate~r m,,ot-hcr6, they were significan-rtly more internal. One can speculate on the variables which eih xplain this. The Follo-.. Throu.-I educators were high school. graduates, self selected by th-cir a-leitness to the availability of c. good job and staff selected for skills in relating and for apparent integrity. Thus, they had already received valuable reinforcements for their efforts, their skills, and their personal characteristics.

The results of the Follow Through1 educators ara similar to thozce of Holmes as reported by Grotberg (1969). Holmies found differences among, parents who referred their children themselves to Head Start as conpared to those iose children wzre sou-.'ht out by Head Start personnel. The self -ref erred group Of parents were more like middle class parents in the-ir higher aspirations for themselves and their children.

Thre-e sets of tiaternal behaviors which seemed to reflect mastery of or vattenrts to inprove upon : oth-rinc skills related to infant cognitive gro;,th were select,d, for study: frequency of positive verbalizations, f:ca-uency of (presuiiabJly non-growth producing) negative verhalizatio,-c such as o:.xQswithor- tio ,ll rd bahy talk, v7' d frequency of-1 beha !Iors - -,''K to --no-, i Ce, wthadco ct

to a mothers' educLticon prorrcn. TTo relatiMonshins. Were found to be relia-bl#e betn.;ucn nratcrnal hca1~zshow,,ing deg-ree of involvement in the program as; moasured by thne positive attitudec iniexY (PEP) 3and interfaJ 11lro xeea a.Te results en positive.' md "negrative-"





- 90 -


verbal interaction ser-m equivocz.l in the tWo studies. However, consideration of the age differences betwdcon the t-vo samples may help clarifyj this.

In the first study which involved infants 3 months to two years old, positive or potentially growth producing verbalizations such as explaiingand describing7 thinric to tecild by the mother were significantly related to internal expectancy and changes in that direction. Interestingly enough Gordon (1969) reports mothers who were more verbal on the sarime index had children who achieved more in the program.

Still another study with much older children (pre-schoolers) by Hess and S3hipc :an found "openness of mother to her child's questions" and "infrequcncy of irrocrat ive statements to child without rationale" were the b"c't poredictors of a child's achiovocient. Othcnr good predictoi-st which were; related to I-E Control were the parents ' vlipirations (Grotberg, 1969).

In Study IT which- involv~.d younrper children (3-12 n-cnths) posiLi-ve

verbal. intexacti~on was not related to expectancy o: exc:p-eatancy changes. So called negratiye verbalizations were related to internal expectancy of mother r.-surJ ,,hon her infa:,nt was 3 months old, but not when he was one year old. Taken together the results fro-M the two studies lead to the tentative- conclusion that frequency of mother's verbal interaction with her ch-ild is related -to internal control if the VI indcx is based o.- observation of vorballis-ations appropriate to the halcy's a.O

The inconsistencies discussed abocrc also sei to involve itocooi

cal proble-,s,. What period of -;sc--cvation en riothar sho-u,1Ld be related to what tinie at ,.hich her control c::pccta oe11y I imeasured.? In these






- 91~ -


studies the observations weroc tabulated raero _s thE. entire~ duration of the project while expectancy was rraured at the beginninS and the end. Thus, possible syste:n~tic variation of the t-w.o treasures over time could not be assessed. Perhaps this could be overcome in a future

study which used and. corpared several. measurements of the two variables over time or one in which the VI measure was elaborated to give more information et dilfferent tim~es.

That the positive attitude (PEP) index did not predict final expectancy or expectancy changes may also suggest that the behaviors selected reflected too much of a middle class bia,,s, or that involvem,--t in the project was not highly valued enoughl to influence the mothers'

perceptions or rcinforce-,z-nt contingencies. IRelated to the former, the index was made up of such itens as the rnuber of delays and missed appointments a mother had which may not be relevant in an indigent population where rany reality factors influence punctuality. Related to the latter, participation did seem. to be valued enough to

change expectancies in an internal direction at least in Study II.

Perhaps the weak relationship between the PET inda:x and I-E Control indicates an. oversight of more potent variables. Rotter (3.966) Suggests in his rliscuL.-sion of the origin of T-E Control ex pcctancies that these expectancies vould6 be influancod by the actual. reinforcement contin-cncies in an individJual's life cirmi!zt-nees. In the case of the disaI. nt.: d vc-Lothers stu-died t-r::4e contingencies iould seeri to bc th availability of, ti~rc, funawz;, friend,-,, etc. , rzsu's the tuz.OLII of de r~ lnaced upon ho r. Th:*for Yx~.~ oi rjo,

Gordon and Rilizor (1%'8) found tha[t uohr infi is uurt~ b





- 92 -


income other than welfare were significantly more internal tha families on welfare. T-his could be interna-eted to mean greater availability of funds was related to inteinal-ity. In addition, mothers in families where there was less illness observed iwere moCre internal -than fam-.,ilies in which there was more illness (Preijo, et aW , 96S). This could. be interpreted to nean that brothers less oa' ru.,helined by denand.- were more internal.

There were no sienifica-nt relationship,, between moth!:-rs' locus of control and ehildrer's achie'fenent. Pcsibly these t\;o variables are at the extremes of a chain of causal relationshi-os. The intermediary links of the chain need to be irvestig-ated be-fore th'e extremes lend themselves to greater predictability. Fruitful lines of research have already been mentioned. The question of whether parental behaviors influence achievement has been investigated in T'ead Start progrvitis. For example, Hess and Shinmnn (in Grotberg, 1969) found that, although a few,, behaviors predicted chiJldrEn's achievement, for the most part parent behaviors were not found to predict children's achievement. On the other ha ncl, scenei studies fund parent ptarticios.tion predicted children's succc ss in the itrogr-,m (Grotbzrg, l16)9). The whole area of disad-vantac-ed parental characteristics and their relationship to children's achie-vene-nt is a vide o-nean research territory.

Other important intcr::.:!diarv links would sento be the rela-!tic'nship) between locu.s of control a~ogthe children the-:se--lves anqd their ach ieverne:nt; between parental eaontrol expeottvncies and parental achieveme,-_nt; betwt.,.n p!.rem-tsi aci'ac.-n :i~ : le ~ and between parental control exieetancy -nd chin r sc t" :;. 'Iy




Full Text

PAGE 1

LOCUS OF l-E CONTROL EXPECTANCY AND EXPECTANCY CHANGES OF DISADVANTAGED MOTHERS By LARRY M. BILKER A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1970

PAGE 2

ACKNOWLEDGEMENTS This research was made possible through the cooperation of Dr. Ira J. Gordon, principal investigator of the Parent Education Project,* and his staff, particularly Dr. J. Ronald Lally, project field director. I am indebted to them for their ideas and their assistance in collecting and programming the data. Thanks are due also to Mr. Tom Williams for his excellent computer programming and consultation. I would like to thank my Supervisory Committee, particularly Dr. Louis D. Cohen, my chairman, for their intensive, comprehensive, and enlightened examination of this research and for their insightful suggestions. They are particularly to be commended for accomplishing much of this task at a distance of 1,200 miles from the writer. I would also like to thank Dr. Anthony Davids for his generous help with some of the practical matters involved in completing this work. The encouragement and forbearance of my wife, Marilyn, were of inestimable value. *The Parent Education Project was initially funded by the Fund for the Advancement of Education. The continuation was partially funded by the Children's Bureau, Department of Health, Education and Welfare.

PAGE 3

TABLE OF CONTENTS Acknowledgements List of Tables Abstract Chapter 1 Problem 1 Chapter 2 The Parent Education Project 5 Chapter 3 Review of Related Research 9 Chapter A Study I: Hypotheses 23 Chapter 5 Study I: Methods 26 Chapter 6 Study I: Results and Discussion 37 Chapter 7 Study I: Conclusions 58 Chapter 8 Introduction to Study II 60 Chapter 9 Hypotheses for Study II 62 Chapter 10 Study II: Methods 64 Chapter 11 Study II: Results and Discussion 67 Chapter 12 Discussion 88 Chapter 13 Summary 96 Appendix A Modified I-E Scale (SRI) 103 Appendix B Parent Educator Weekly Report (PEWR) and the teaching behavior indices (Index of Positive Verbal Interaction, Index of Nej^ative Verbal Interaction, Index of Attitude Toward Parent Education Project) derived from it. 109 Appendix C Series Test 118 Bibliography 121 iii

PAGE 4

LIST OF TABLES Page Table 1 . . . . Treatment Plan for the Parent Education Project 7 Table 2 .... A Comparison of Three Populations With Subjects in 39 Groups 1 and 2 in the Present Study on I-E Control Table 3 . . . . Correlations Between Age and Parity and Initial 39 (pre). Final (post), and Changes in I-E Control Scores for 32 Subjects Table 4 . . . . Mothers' Initial (pre). Final (post), and Changes 43-44 in I-E Control Scores (SRI) Grouped by Level of Verbal Interaction and Education Table 5 . . . . Analyses of Variance of Initial (pre). Final (post), 45-46 and Changes in I-E Control Scores (SRI) by Mothers' Verbal Interaction Level and Education Table 6 . . . . Changes in Children's Series Test Scores (6-24 51 months) Grouped by Their Initial Score and Their Mother's Initial Expectancy Table 7 . . . . Changes in Children's Series Test Scores (12-24 52 months) Table 8 . . . . Differences Bet^^7een Success Scores of Children 53 Grouped According to I-E Scores of Their Mothers at the end of 24 Months Table 9 . . . . Changes in Mothers' Expectancy Scores (I-E) 53 Scores (3-21 months) Grouped by Their Initial Score and Their Child's Achievement at 6 months Table 10. . . . Differences Between I-E Scores (Final SRI and 54 Changes Over 15 Months) of Mothers According to Achievement of Their Children Table 11. . . . Differences Between Treatment Groups in Initial, 55 Final and Changes in Expectancy Scores (SRI) Table 12 . . . . Differences in I-E Control Expectancy Scores (SRI) 55 Between Mothers Rated High and Low in Project Involvement (PEP) Table 13. ... A Comparison of Three Different Populations With 68 Project Mothers on I-E Scores Table 14. . . . Differences Between Black and White Mothers in 68 Initial and Final I-E Control Scores iv

PAGE 5

Page Table 15. . . . Correlations Bebjeen Education, Age, Parity and 69 Initial (pre), Final (post), & Changes in I-E Scores (N=28) Table 15.1 . . Analysis of Variance of the Multiple Correlation of 69 Age, Parity, and Education With Final I-E Control Scores Table 16. . . . Differences in Mean Initial and Final I-E Scores 71 Between Mothers Above and Below the Median on Behavior Indices Table 17. . . . Analysis of Variance of Changes in Mothers' I-E 72 Scores Effected by Positive Verbal Interaction (VI+) and Positive Attitude Toi^jard Project (PEP) Table 18. . • . Analysis of Variance of Changes in Mothers' 73 Expectancy Effected by Negative VI & Positive Attitude Table 19. . . . Differences Be&Jeen Achievement Scores of Children 80 Grouped by Median Split of Mother's Initial (pre). Final (post), and Changes in I-E Control (SRI) Scores Table 20. . . . Differences Between Initial (pre). Final (post), and 82 Changes in I-E (SRI) Scores of Mothers Grouped by Median Split of Children's Achievement Scores (6 and 12 Month Series Tests, and Griffiths Test) Table 21. . . . Differences in Initial (pre). Final (post), and 84 Changes in I-E Control Expectancy Scores (SRI) Between Experimental (E) and Control (C) Groups Table 22. . . . Differences in Initial (pre). Final (post), and 84 Changes in I-E Control Expectancy Scores (SRI) Between the Two Experimental Groups (E1+E2) and the Control (C) Group V

PAGE 6

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 LOCUS OF I-E CONTROL EXPECTANCY AND EXPECTANCY CHMIGES OF DISADVANTAGED MOTHERS ty Larry H. Bilker June, 1970 Chairman: Dr. louis D. Cohen Major Denartment : Psychology This research investigated existing internal vs_ external control of reinforcement expectancies (l-E Control) among indigent mothers, and changes in expectancy as a function of participating in an educational program. The two studies renorted attempted l) to validate theory and research indicating that the lover socioeconomic class individual is one who tends to perceive events as externally controlled, 2) to explore the relationship "between specific maternal behaviors and I-E Control, 3) to Identify individual differences within a population sample from the lovrer socioeconomic class on the dimension of I-E Control expectancy, h) to examine the interrelationship between a mother's locus of control expectancy and her child's achievement in a program designed to enhance development, and 5) to extend previous research findings on correlations between internal I-E Control and participation in or affiliation with social action programs. It was predicted that participation in a social action program by lower cla^js mothers would change their expectancies in an internal direction over time. The subjects were indigent mothers and their infants who narticicated in a program designed to improve motherin?,' skills. The nrogram was based vi

PAGE 7

on research in child development which vas found to be related to the cognitive growth of children. In Stud;/I and Study II which used better controlled procedures , families participated or were assessed over the child's first year of life. Maternal expectancy was assessed at the beginning and end of the program by a version of Rotter's I-E Control Scale modified for a fourth grade reading level. The revision was found to be reliable over a one month interval for an indigent population. Maternal behaviors were observed and recorded weekly by parent educators, indigenous non-professionals who taught the mothers in their a^m hones. Infant achievement was assessed by progress on the program materials developed out of Piagetan theory and research, and two standardized tests of infant achievement. The significant findings were: 1. Indigent mothers were more external than subjects from more advantaged popixLations reported by Hotter. 2. Indigent mothers were more external than indigenous non-professionals from the same background. 3. Black mothers were more external than whites in Study I but the findings were non-significant in Study II. h. Mother's frequency of positive verbal interaction with her child was significantly related to internal control in Study I but not in Study II. 5. Mothers who participated in the program for 9 months became significantly more internal than mothers who did not participate. 6. Mothers who vrero taught by pai-ent educators who designed their o^m program becav.o significantly more internal than control group mothers. Mothers taught by edxicators who used an already designed program became more internal but the findings were not significant. vii

PAGE 8

Mother's I-E Control vas not significantly related to her level of negative verbal interaction, her degree of involvement in the program, her age, parity and education or her child's achievement. The finding that participation in an educational program, particularly the one in which the teachers, former indigents themselves, designed their own program, was effective in changing indigent mothers' locus of control expectancy in an internal direction was discussed in terms of its implication for positive social action. Suggestions for future research vrere made on the basis of methodological and theoretical shortcomings of the studies. The possibilities of using an I-E Control measure more reliable over time, more sensitive at the internal end and more specific to the subjects' situational context were considered. Possible intervening variables between mother's expectancy and child's achievement were suggested. viii

PAGE 9

CHAPTER 1 PROBLEM The purposes of this research are to investigate locus of control expectancies among lower socioeconomic class mothers and its relationship to selected maternal behaviors and to their children's achievement; and to assess the effects of an attempt to change existing control expectancies. Expectancy is defined by Rotter (1954) as "the probability held by an individual that a particular reinforcement will occur as a function of a specific behavior on his part in a specific situation or situations." Internal vs_ external control of reinforcements describes a generalized expectancy which determines to what extent certain outcomes of behavior will be categorized as within the individual's personal control and understanding. A person who generally categorizes situations as internally controlled tends to expect that it is an individual's own characteristics and skills rather than externals which influence what reinforcements he receives. On the other hand, a person who generally categorizes events as externally controlled tends to expect that chance, fate, powerful others, or an incomprehensible complexity has the greater influence over what reinforcements he receives (Rotter, Seeman, and Liverant, 1962). Rotter (1966) summarized findings by himself and others which indicate that internal vs^ external control expectancy is a personality 1 -

PAGE 10

2 characteristic which has predictive value in relation to other behaviors of an individual. More specifically, the findings show that an individual who tends to categorize events as internally controlled is more likely to be alert to aspects of the environment which provide useful information for future behavior, to take steps to improve his environmental conditions, and to place greater reinforcement value on acquisition of skills. Simmons (1959) also found internal control expectancy related to organizing, planning, and realistic goal-setting abilities. Many social psychologists and sociologists have described the lower socioeconomic class individual, particularly the lower class Negro, as a person who feels powerless and alienated (Cohen and Hodges, 1963; Dean, 1961; Irelan, 1966; Reiff, 1966). A low expectancy that an individual can control his reinforcements is often referred to as powerlessness or alienation (Neal and Seeman, 1964). The empirical findings of Coleman (1966), Lef court and Lad^^?ig (1965b), and Battle and Rotter (1963) corroborate this description. Thus, lower class individuals are more likely to believe in external control and lack just the personality characteristics described above which they need to improve their situation. Other researchers have emphasized the importance of individual differences within socioeconomic classes. Bell (1965), e.g., has pointed out that different subcultures can be distinguished within the lower class continuum. One would expect to find that within.

PAGE 11

3 and cutting across, subcultures there are individual differences on a behavior continuum such as internal vs external control expectancy, and that these are related to differences in other behavior. The first purpose of this research is to investigate lover class mothers' expectancy of internal vs external control and a) the direct relationship of that expectancy to, l) their astount and kind of verbal activity, 2) their coiTrnitnent to a program designed to enhance their mothering skills and, 3) the indirect relationship between control expectancy and the intellectual development of their infants . The second purpose is to evaluate an attempt to change those expectancies through participation in a parent education project. It is important for all people as veil as disadvantaged people to become aware of the degree to which success or failure is contingent upon the acquisition of certain skills. Although it has been shown that members of social improvement groups are higher in the belief in internal control than non-members (Neal end Seenan, I96U4 Strickland, I965). the question of whether participation in such a group, e. g. the parent education project described in the next chapter, will change expectancy to a more internal direction has not been previously investigated. It seems particularly important for lower class p.others to develop the internal expectation that their mothering abilities and the abilities they can develop in their youngsters can load to alleviation of povci-^y conditions end a more successful life. According to a recer.t research report (Grotberg, I969), Head Stixrt programs were

PAGE 12

1» only successful to the extent that the mothers became involved and developed a more positive attitude.

PAGE 13

CHAPTER 2 THE PARENT EDUCATION PROJECT The subjects for the present study came from the ongoing Parent Education Project begun in June, 1966. In this program, a small group of disadvantaged women were instructed and arranged to visit other disadvantaged homes to teach mothers to stimulate the activities of their infants. The methods of stimulation used had been shown to be related to cognitive development. In addition to being trained to teach stimulation techniques, the parent educators were also taught to make and record observations objectively in the home. This teaching was done by means of present ing films and verbal descriptions of home situations and asking the group to Judge and record these until they could all agree on various kinds of observations. In order to reduce interference with the natureil environment, the parent educators were the sole observers in the home. Thus, no reliability check on their observations was possible. However, during weekly individual supervisory hours, each educator went over her observations with a supervisor. Periodically at six month intervals, visits to the homes were made by both the parent educator and her supervisor. Both made independent observations and then checked them out. There were few discrepancies . The parent educators were 15 (12 Negroes, 3 Caucasion) high school graduates whose average family incojTia before being hired was $280 per month for the support of an average of four people.

PAGE 14

6 Thus , most of the parent educators vere "better off economically than the indigent population they were to work with but vere still within the same socioeconomic class. All but one of the parent educators were mothers with an average of 2 to 3 children. Seven were married. Negro educators worked with Negro families; Caucasian educators with Caucasian fejnilies. The project treatrtent plan is outlined in Table (Gordon, 1%$) . j Series means weelcly visits by parent educators to instruct the mother in materials developed by the project staff. Group C3 mothers were taught by another set of parent educators from backgrounds similar to those of the original parent educators. These new parent educators were former Head Start and nursery school workers who developed their own pattern of stimulation. For all groups mothers and their infants were indentified at the birth of the latter by the obstetrics staff of the J. Hillis Miller Health Center Teaching Hospital of the University of Florida. Only families with an economic code of indigent on the hospital admission form and residence in Alachua and eleven other nearby counties were selected. In addition, the obstetrics staff used the following criteria: single birth, no breech or Caesarian delivery, no complications to mother or infant, no gross evidence of infant's mental retardation, and no evidence of mother's mental illness. These criteria were selected because of their possible relationship to noraal intellectual potential. The birth rate at the hospital was such that, beginning June 15, 1966, about 30 babies a month were added to the sample through October 31, 19^7 (attrition on participation in this study was about

PAGE 15

1 TABLE 1 Treatment Plan for the Parent Education Project Group (E=Experiraental ; C=Control ) Final N Original groups began 9/15/66 1. EI 36 2. E/C 36 3. C2/C 36 New groups begun 7/1/67 k . E2 21 5. C3 22 26 Infants Treatment fro.Ti Age 3 no-1 year Infants Treatment from age 1 year-2 years Series Series No Treatment Series No Treatment No Treatment Series Stirau].ation Program Designed by Parent Educators No Treatment 30^) (Gordon, I969). A brief note on the use of the indigenous non-professional as revievred in the literature is added here. Riessinsn (1966) and Levinson and Schiller (1965) report that the use of non-professionals from the same low socioeconomic treatment group increases communication effectiveness with the population being studied. A training program (after which the Parent Education Project's was

PAGE 16

8 modeled) emphasizing interpersonal relationships, communicative skills, professional confidentiality and responsibility, continuous interaction with professionals and personal identification resulted in a high level of competence (Levinson and Schiller, 1965) .

PAGE 17

4 CHAPTi® 3 REVIEW OF RELATED RESEARCH I-E Control and Attempts of People to Better l.'heir Life Conditions A person who feels powerless to better his life conditions was described in the introduction as one who has a low expectancy that people can control their own reinforcements. Such a person, in fact, does less to improve his circximstances than one who has an internal expectancy. In experimcintal tests of the above assaniption using Rotter's Internal vs_ External (I-E) Control of Reinforcements Scale (Rotter, 1966), Seensji and Evans foujid that the more internalljr controlled TB patients knew more about their own condition, questioned the doctors and nurses more, expressed less satisfaction with infoj.'mation feedback, and were rated as better patients by the staff; Seenan (1963) fovmd independent of intelligence (correlation of Beta I.Q. with I-E scale was .03) that the raore internally controlled re format orj' inmates knew more about how the refoimatory was run, parole, and pertinent infonnation about economic opportunities for post-refoimatory living. Kiehlbauch (1968) foxmd that prisoners who attained a work release status, i.e. were allowed to work in the community towards the end of their incarceration, did not show the rise in externality one month prior to i-eleaso that a matched sample of prisoners did. Gore and Rotter (1963) found willingness to make a corcniitrnent to social action on civil rights was related to internal conti'ol araong a group of Ilogroes enrolled in a southern cones'?, 9 _

PAGE 18

10 and Strickland (19d5) extended the latter finding to actual participation in social action. Students actively involved in the civil rights movement were more internal than a matched control group of non-participants. Male smokers vho quit after the Surp;eon General's report were found to he m.ore internal than those vho believed the report hut did not quit smoking (Jcrnes, Woodruff, and Werner, 19^5). The predictions in this study are that mothers usin;^ more mothering skills, more specifically verbal interaction, vould be more internal, and that those vho saw the project in a positive light and cooperated would also be more internal. I-E Co ntrol and J.,earnin g Rotter's (1966) "... basic hypothesis is that if a person perceives a reinforcement as contingent upon his ovn behavior, then the occurrence of either a positive or negative reinforcement will strengthen or weaken potential for that behavior to recur in the same or similar situation" (page 5). In a series of studies which contrasted learning under experimentally defined conditions of chance and skill, subjects made more appropriate responses to positive and negative reinforcements (increase in positively reinforced behaviors, decrease in negatively reinforced ones), generalized their learning to a significantly greater degree, and vei^e m.ore resistant to extinction under skill conditions than under chance conditions (Phares, 1957; James, 1957; James and Rotter, 1958; Rotter, Liverant, and Crowne, 19^1 ). A few studies have contrasted individuals who tended to see events as skill controlled vs_ those that saw them as chance controlled. Phares (1957), using the first precursor of the I-E Control Scale, found a tendency for subjects scoring higher on in-

PAGE 19

11 temality to give more appropriate responses to reinforcement. Itiese findings approached significance. Using the James-Phares Scale from which the I-E Scale is derived, James (1957) found significant relationships between in temality and all three learning variab]es m.entioned above. Thus, the greater tendency an individual has to perceive situations as internal, the more he will learn from that situation; the more he will generalize that learning to related situations and the more he will remenfoer over tiir.e. Coleman (1966) surveyed students and school conditions in all twelve grades across the country. They found that differences in school conditions such as facilities, curriculum, teaching quality and availability were insufficient to explain differences in verbal and non-verbal ability, reading comprehension, mathematical achievement, and general information about the natural sciences, social sttjdies, humanities, and practical arts as assessed by comprehensive tests administered through the local schools. On the other hand, they report that a Negro child's achievement is highly correlated with his feeling that he can control his own destiny. There is also evidence that need for achievement is related to I-E Control. In a study using the Intellectual Achievement Responsibility Scale, Crandall (1962) found that for boys in the first three grades, achievement motivation as measured by free play achievement behavior and achievement test scores was related to more internal scores. The findings did not hold for girls. Franklin (1963) found 15 of 17 kinds of reported evidences of achievement motivation correlated with scores on the I-E Scale in a national stratified sample of 1,000 high school students. Using latency of

PAGE 20

12 decision on a matching task as a measure of achievement motivation. Rotter and Mulry (1965) found that when unselected subjects were divided at the median I-E score into internals and externals, the former showed greater need for achievement. In summary, I-E Control is related to amount, generalizability , durability of learning, and need for achievement. Applying this logic to the present study, mothers who learn more and use this increased kna/ledge on their children should be more internal. Personality Correlates of I-E Control Although a more precise relationship between exact score on the I-E Scale and specific personality attributes has not been validated. Rotter (1966) suggests that the extremes may be related to adjustment problems. That is, an extremely high external score may be an indication of defense against failure while an extremely low one may be associated with the assum.ption of an unrealis tically high amount of responsibility and consequent guilt for personal actions. Effran's 1963 study of high school students indicated a relationship between the tendency to repress failures verstos successes, and scores towards the internal end of the dimension. This was interpreted as an indication that internals feel a greater need to defend against failure while externals already have a convenient rationalization (Rotter, 1966). Butterf ield' s study (in Lefcourt, 1966) Indicated that internals depict themselves as goal directed workers who strive to overcome hardships while externals portray themselves as suffering, anxious, and less concerned with achievement than with their emotional response to failure.

PAGE 21

13 Rotter (1966) reports that studies examining relationships between measures of adjustment and the I-E scr.le are suggestive but inconclusive. Rotter and Rafferty (in Rotter, I966) compared the scale with the Rotter Incomplete fJentences Blank (ISB) for several sajnples of college students. Generally linear correlations were insignificant. Some curvilinear relationship between extreme scores on the I-I-^ Scale and maladjustment scores on the Rotter ISB approached significance. The extreme scores were less well adjusted. Simmons (1959) similarly found a positive but nonsignificant correlation ratio (eta) between the two measures. Ke also found no reliable correlation between the Rotter ISB dependency score and the I-E Scale for college m.ales. Kiehlbauch (1968) reports good vs poor adjustment groups of prisoners did not differ on I-E scores. Ware (Rotter, I966) found a significant correlation of .21 between externality and the Taylor Manifest anxiety scale. Effran, using the same measures, found a correlation of .00 (Rotter, I966) . Cromwell, Rosenthal, ShsJww and Kahn found schizophrenics higher in externality than normals (Lef court, I966). Overall, the results suggest some relationship between extremes scores and maladjustment with perhaps greater variance coming from the external extreme. In test behavior internals seem to have a more constructive response to failure. Several studies (James, 1957; Battle and Potter, I963; and Simmons, 1959) have found the goal setting of externals more variable and unrealistic. Trie external subjects are more prone to the gambler's fallacy of expecting to win after a series of failures in skill situations. SinMs (1959) furthermore found externality

PAGE 22

Ik correlated positively vith other maladaptive goal setting patterns, and negatively vith Edwards Personal Preference Schedule needs for order, nurturanee and endurance for college vomen. He concluded that external females are nore disorganized, lack -ol&nning and realistic goal setting abilities and are fatalistic comt)arei to internal females. Several studies have found relationships betveen I-E Control and self-reliance. Julian and Katz (1968) found on a competitive task that internal college students preferred to rely on their own resources rather than those of their opponent even though they were shown by experimental manipulation that their opponent was more competent. Odell reported a significant relationship between I-E Control and Barron's independence of judgement scale. Crowne and Liverant found externals conform significantly more in an Ashe type conformity situation (Lefcourt, I966). Deever (I967) reports internals choose their personal reinforcement history rather than reported performance of others as an index or cue to expectancy of personal success in the future. In the present studj' successful, participation depended upon the mother's active use of what she learned in the parent educator's absence. This would seem to be related to her self-reliance and organization. Thus, it was predicted that a volunteer mother's success both in terms of her performance as a teacher of her child and the child's success would be related to 1-E Control. I-E Control and Affiliation Neal and Seeman (lOoJj ), using the I-E Control Scale, found that members of a work based organization were significantly lower in

PAGE 23

15 externality than controls (matched for age, socioeconomic level and education) who were non-memhers . Furthen.iore , activity within the union and general knowledge of political events were positively related to intemality. Differencor. were not related to generalized despair scores as measured by Srolo's Anoinia Scale. The authors concluded that internality was related to affiliation with an organization that can better one's life circiunstances . The direction of influence w&s not assessed. Either s'.ore internal people affiliate, or affilieticn engenders internality. The present study proposed to test the assumption that affiliation modifies expectancy in an internal direction. ^P-g.9!t ^-^cy &"<^ Deno^raphic Variables Tne popiilation for this study was selected from volunteers in the Parent Education Project. Since they were lower socioeconomic class mothers (6:1 ratio of blacks to whites), the primary concern here was the relationship between expectancy and class and expectancy and race. Race Several studies have found that Negroes are significantly more external in their control expectancies than Caucasians. Lefcourt and Lidvig (1965b) and Battle {1962) using the I-E ScjQe found a small (a score of 9 compared v.o 6 out of a possible total score of 23) but significo-nt difi'erence. Co1ck::ii (1966) found that in a co;uitry-vide sample of school children a significantly higher proportion of black than white children answered three questions about control expectancy in eii external direction. Lefcourt (I966)

PAGE 24

16 argues that blacks easily perceive impeditnents in the way of goal striving. Segregation and discrinination are interpreted as meaning that their ovn effort vill not pay off in reinf orcensnts . On the other hcind, Kiehlbavich (1968) found a small difference in the same direction between black and white prisoners in their I-E scores which did not reach a significant probability level. Thus, every one of these four studies conparing black end white subjects has found blacks to be more external than whites. The results in only one study were not significant. If Lefcourt's argtfflient is correct, and it seems to hold up country -wide in Coleraan's study, it should hold in a southern coinmunity. Consequently, it was predicted that Negro Eothers would be more external than Caucasian mothers within this disadvantaged population. Sociosconordc Level , For populations in which social class differences are small no relationships between class and I-E Control have been found. Ihat is, there were no differences in I-E Scale scores by social class in Ohio State University classes (Rotter, 19o6), Florida A & M classes (Gore and Rotter, I963), or in a sample of prisoners (Lefcourt and Ladwig, 196^?^) . Since subjects in the present study were from a relatively homogeneous social class, differences in I-E Scale scores by socioeoonoaic level could not be examined. Differences have been found in populations more lieterogensous for socioeconomic class. Persons in the lover class tend to score more at the extemr.1 end of the I-E Scale. One study of children (Battle vvA Rotter, I963) found e significciit difference between social classes with race vxid intelligence controlled. Most of the

PAGE 25

17 variance vas accounted for "by the difference between lower class Negroes and middle class Negroes and vhites. Dean (l9ol) found lovr (.10 to .23) but significant correlations betveen alienation end level occupation and income. His questionnaire included components of powerlessness , normlessness (lack of rules), and social introversion. The prediction here is that the present sample of lover class subjects will be more external than samples stratified by social class. Intelligence Correlations of externality or powerlessness with intelligence are generally low and insignificant but there are exceptions. Strickland foimd no correlation between I.Q. and I-E scores with a sample of Ohio State coeds. Seemaji (1963) found no significant correlation between prison inmates Beta I.Q. scores and I-E scores. In the Kiehlbauch (1968) study there was no significant correlation between I-E a:id I.Q. scores. On the other hand, Ohio State University womens ' external scores on the I-E dimension were fotmd by Cardi (Rotter, I966) to be correlated -.22, and by Simnons (1959) -M with intelligence. Since the balance of evidence is for no correlation between I.Q. and I-P Control, the present study did not examine this relationship. Education Evidence for the relationship between education and I-E Control expecta-ncy is equivocal. Rotter (1966) reports unselected high school students score higher (more external) on the I-E Scale than college bound students. Dean (1961) found a low but significant inverse correlation between alienation and educational level. On the other hand, in the studies Rotter (19S6) suir:marized the scores

PAGE 26

18 of college students are generally not different from prison inmates with an eighth grade reading level. The present study predicted that since the educational level vould probably be relatively homogeneous, there voxad not be any significant relationship between educational level and perceived locus of control. Age Age differences have not been found to relate to differences in control expectancy. A.1. though the present sample includes very young (15 years old) and older (76 years old) mothers or mothering ones (see subjects, p. 2l) the prediction is for no correlation betveen age and expectancy. Effects of Mother-Child Interaction I-E Control, Mothering, and Infant Cop;nitivo pevelopT~ent It is assviraed here that a mother's expectancies and behaviors will be inteiTelated vith her child's cognitive development. A mother intenial in orientation, according to the foregoing research discussed, should take steps to improve her mothering skills. Such improvement is expected to shov up in increased cognitive development of her child. The focus here will be on verbal behavior. Conversely, if a child is successful cognitively, thus confirming his mother's ability, his mother's expectancy should change in soi internal direction. Tliat m.aternal behaviors are a significant influence on a child's cognitive development is well kno'/n. Maternal aspirations and expectations, reward end pujiishirient of her child's behavior, provision of materials and other opportunities to stimulate the child's

PAGE 27

development and policies for rearing of the child have all been foxind in the child development literature to "be related to the cognitive development of the pre-school child (Gray and Miller, 1966). These are particularly crucial in disadvantaged homes where, for example, Bay ley (1966) suggests there is a lack of stimulation which may be causative of retardation. The lack of verbal stimulation has particularly been emphasized. Hunt (19^6) points out that verbal stimulation is lacking in disadvantaged homes during the first tvo years of an infant's life. Bernstein (1961) points up the long range implications of this. He find that lover class members experience a relatively narrow range of variations in language. Thus, they come to rely on highly predictable implicit utterances vhich poorly equip them to express themselves. The inflexibility of their verbal expression interferes vith their ability to become actively responsible for their own behavior or learning. The child influences the mother as well. For example, a child's executive competence has been found to influence the mother's emotional reactivity (Vfenar and Vfenar, 196k), In the present study a child's success is predicted to lead to his mother's movement in an internal direction analogous to the subjects in the studies of Rotter, Liverant and Cro'.me, Bennion, and Blackman (Rotter, 1966) who were likely to see a sequence of reinforcement as skill controlled (internal) when reinforced as right more than 50/' of the time. A child's success in the project is assumed to be reinforcing to the mother.

PAGE 28

20 Love Oriente d vs Power Assertive Dic sin line A more internal mother is expected to perceive the reinforcements her child gets as contingent upon his characteristics or actions. She would seem more likely, then, to use discipline vhich engenders the development of internal standards . In a reviev of the literature on parental discipline, Becker {l96k) refers to love oriented techniques such as reasoning vith and praising the child. These are reflected in the positive verbal interaction index used in this study by items such as: "Mother explains and describes things vhen talking to the baby"; "listens to th'S baby vhen the baby talks"; "her tone of voice sounds soft and lov.lnr;." On the other hand, a more external mother perceives reinforcements that happen to an individual to be contingent upon chance, fate, incomprehensible complexity or the actions or characteristics of more powerful others. Thus, she would seem to be more likely to use discipline techniques which lead to externalized reactions to behaviors (e.g. fear of punishment, projected hostility). These techniques classified as power assertive would seem represented by the items in the negative verbal interaction index such as "Mother orders or tells the baby to do or not to do things"; "her tone of voice sounds cross and angry." The consensus of research is that love oriented techniques, most reliably reasoning and praise, have been found to be correlated with the occurrence of internalized reactions to transgressions in the form of acceptance of self-responsibility, while power assertive techniques correlate with externalized reactions (Recker, 196h) .

PAGE 29

21 More specifically, Tolor and Jalowiec (1968) found that college students vho perceived maternal attitudes of authoritsrian control and hostility rejection as assessed by the Parent Attitude Research Inventory (PARI) vere more external on the I-E Scale than those perceiving democratic attitudes. The foregoing studies deal with parent behaviors influencing child's control expectancy. The present study deals with the relationship between a mother's behaviors EOid her own control expectancy. In summary, for the women in this study, the maternal role was considered a significant focus of the life situation. Therefore, motivation for self -improvement was assumed to include a desire to enhance the performance of the maternal role. Furthermore, it was assumed that the opportunity to participate in the Parent Education Project was an opportunity to improve the performance of the maternal role. At least this is the way the parent educators perceived it when introduced to the program. It is also consistent with their informal descriptions of how most of the volvmteer mothers perceived the program (Gordon, 2.969). Thus, a mother's degree of participation and involvement in the project (as assessed by such behaviors as: demonstrating ability to perform the skills taught, encouraging her child, keeping appointments with the parent educators) were considered a manifestation of attempts to improve her life situation. One important reflection of this improvement was assu.nied to be the child's success in the project program. Tlius, it was predicted that differences in the amount and quality of maternal care, particularly verbal behavior and attitude toward the project, would be related to I-E control and

PAGE 30

22 that this in turn vould be interrelated with ths cognitive development of her infant.

PAGE 31

CHAPTER k STUDY I: HYPOTHESES The Parent Education Project described in Chapter 2 began as a pilot study. Thus, the usual difficulties encountered in naturalistic research vere added to by such problems as the inexperience of the parent educators and changing plans as to the most appropriate or expedient tine to collect a particular datura. Some of these problems vrere ironed out by the time three new groups of subjects began participating in the project one year later. Tliese new groups were used for a second study which was plsnned to investigate the same issues as Study 1". The hypotheses for the first study are listed below. ]fynothej>es_ Class, Ra ce ,^gg_,_Parity 1. The present sample of lower class mothers will be more external in I-E Control than previously studied sa;nples that were not restricted to the lower class. 2. Black mothers will be more external than white mothers. 3. There will be no significant correlation between age or parity of the mother and I-E Control. Matern al Verbal Intei-aet ipn_Leve_l h. a) Mothers higher in positive verbal interaction with their child initially will be more internal in I-E Control thaii mothers lower in positive verbal interaction. 23 -

PAGE 32

2h b) Mothers higher in positive verbal interaction with their child will become more internal in I-E Control over a 15 month interval than mothers lover in positive verbal interaction. 5. a) Mothers higher in negative verbal interaction initially will be more external in I-E Control th.m mothers lower in negative verbal interaction. b) Mothers higher in negative verbal interaction will become more external in I-E Control over a 15 month intem'al than mothers lower in negative verbal interaction. 6. Mother's educational level, individually or in interaction with verbal interaction, will not be related significantly to I-E Control. Infant Cognitive Developme nt 7. There will be a significant two wsy relationship between mother's I-E Control and her child's cognitive development such that: a) Children of mothers more internel in I-E Control will learn more and be more highly developed by the end of their second year of life than those of mothers more external in I-E Control . b) Mothers of children who are more highly developed cognitively will be more internal in I-E Control than mothers of less well developed children. c) These mothers will also become more internal over a 15 month interval . Effects of Particiriation 6. Mothers who participate longer (21 r^onths) in a program to improve their mothering skills will become more internal in I-E Control than those who pai-ticipate for a shorter period of time (9 months).

PAGE 33

25 9. a) Mothers vho make better use of the program as manifested in such behaviors as keeping appointments and mastering materials will be more internal in I-E Control initially than mothers less involved. b) Mothers who make better use of the program vill become more internal in I-E Control over a 15 month interval than mothers less involved.

PAGE 34

CHAPTER 5 STUDY I: METHODS Subjects The subjects have been described under Parent Education Project in Chapter 2. Ibey are in treatment groups 1-3 (see page 7), Thus, data on mothers and their children cover from the third to t:ie twenty fourth month of the children's lives. For this study only luothers or mother substitutes (e.g. grandnother) vho had the najor responsibility for the child's care vere selected. The number of subjects for thtdifferent variables changes somewhat because of nissing data. Design The major variable studied, 1-E Control, was the score attained by the Eother on a modified version (SRi) of Rotter's I-E Scale (see Instruraents ) . Al l of the mothers studied were given the SRI when their babies were 6 months old. ^nis was considered the initial or pre SRI score. Mothers were retested when their babies were 21 months old. This was considered the final or post SRI score. The difference between the post and pre SRI scores was called the SRI change score. Except for the hypothesis in which infant achievement was the dependent variable, the SRI score was the dependent variable. To allow for initial differences and changes, generally the relationships between the independent variable and all three SRI scores (pre, post and change) were exa,rained. The treatment plan to test the effect of participatirg in a program to improve mothering skills on a mother's I-E Control is chro-ted on 26 -

PAGE 35

I 21 page 5, Chapter 2. The group 1 mothers participated for 21 months compsLred to 9 months only for group 2 and no participation for group 3. Insufficient data vers available on group 3. This precluded assessing the effects of participation vs^ non-participation. Instead, the effect on I-E Control of the length of participation was tested by a t test of the difference between the mothers' expectancy scores of group 1 and group 2. Mothers' verbal interaction levels (positive and negative) and mothers ' degree of involvement in a program to improve mothering skills were rated by the parent educator in weekly observations (see PEWR in Instrvment section) over a 21 month period for treatment group 1 and over a 9 month period for treatment group 2. In assessing the relationship between positive or negative verbal interaction level and I-E Control, first mothers were grouped as high or low on verbal level. Then they were split at the median on educational level. The hypotheses were then tested by a two way analysis of variance of the SRI scores. In order to test the relationship between mothers ' degree of involvement in the program and m.others I-E Control, first mothers were grouped as high or low in involvement. Then they were split again according to whether they vrere high or low in positive verbal interaction. This hypothesis was also tested by a two way analysis of variance. Tlie inter-relationship between m-others' I-E Control and infant cognitive development was tested with the "short term prospective model" (Wenar and Uenar, 196h) . mis model allovrs for consideration of the direction of influence in studies of parent-child interaction. Families were classified in two ways, i.e. for I-E conti'ol and infant

PAGE 36

28 intelligence. Families vere divided into high and lov on the basis of mother's initial external score and then, again, on the basis of the child's initial intelligence rating as high or lev. Then the data vere set up twice, once for intelligence change scores and once for expectancy change scores. When the child's cognitive progress is the dependent variable, significant differences would be attributed to the child's intellectual development. VJhen changes in mother's external scores are plugged into the same design then significant differences are attributable to maternal expect^cy. The hypothesis concerning I-E Control and socioeconomic class was assessed by a t test of the differences between the I-E scores of the present subjects and scores of subjects from more advantaged populations. The relationship between race and I-E Control was examined by testing differences in SRI scores between black and white subjects. Age, parity and I-E Control scores were correlated to assess whether they were significantly related. Non Standardized Instru ments ^j-l^jl's Internal vs External Control of Reinforcement s Scale Ihis instrument was designed to assess the extent to which an individual categorized events as internally or externally controlled. It is a 29 item forced choice test which includes six filler items and is entitled "Social Reaction Inventory" to make the purpose of the test somewhat ambiguous . The test appears to be internally consistent, reliable, unidimensional, and has good discrimina.'Jt and external validity. The item bisorial correlations for a sample of IfOO men emd vor-en range from .11 to .1(8 with 13 out of 23 items greater than .25. The split

PAGE 37

29 reliability ranges from .69 to .73. This is below desirability. However, the items are additive rather than equivalent. Rotter (1966) reports one month test-retest reliabilities in the .70's. Kiehlbauch's study (1968) shoijed a three month reliability of .75. With 6 and 9 month intervals the coefficients are nonsignificant: .39 and .16 respectively. Factor analysis revealed that a general factor accounted for 53% of the variance with the rest being accounted for by several factors of a few items and a small degree of variance each. Correlations with the Marlowe-Crowne Social Desirability Scale range from -.35 to -.07 when administered by authority figures in various institutions. The highest correlation can be explained by the fact that the test was administered with a battery of classification tests given to new prison inmates (Rotter, 1966). Low correlations with intelligence and tests of external validity have been described in the literature review. Modi fication of R otter' s I-E Scale for Belo^^ an Eighth Grade Readin g Level (SRI) The I-E Scale has previously been found satisfactory for people with at least an eighth grade reading ability. Since it was suspected that many indigent mothers in the project might not meet this criterion, an attempt was made to adapt the test for this population. The parent educators took the original test. Afterwards, they were asked to pick out words and phrases they thought their poorest reading volunteer mother might not understand. They were asked to rephrase these in terms they would use in expressing themselves to these mothers. The revisions were then collated and edited for correct grammar and consistency of meaning with the original version. The final revision was

PAGE 38

30 edited for words inappropriate for higher than third grade reading according to the Thomdike-Lorge {l9^h) word list. In a few cases fourth grade words had to be included. A test-retest reliability coefficient of .78 and with an of 35 was obtained for a two week interval on a population comparable to the subjects of this study (Freijo, Gordon, and Bilker, I968). The above revision followed an attempt to adapt Bialer's (1961) Children's Locus of Control Scale. Ihe wording of the latter was modified slightly to make it appropriate for adults (e.g., in "Do you think a kid can be whatever he wants to be," woman was substituted for kid and she for he). Then, this modified scale was given to the 15 parent educators. It correlated -.20 with the I-E Scale and so was abandoned. It was thought further that the administration of the scale might present other problems such as those encountered by Radin and Glasser (1965) in their use of the PARI with disadvantaged mothers. In addition to language complexity, they found that interruptions in the home, reluctance to disclose inforaation to a person of a different background and insufficient rapport with the examiner to tell if the subjects understood were problems. Although many of these problems are mitigated by the use of indigenous workers, in anticipation of any that might remain, the parent educators were instructed to read the iterrs to each mother while the mother made her choices on her own copy, to make it clear that there were no right or wrong answers and that their responses were confidential, and to pause for interruptions and even ret^u•n another day if circmr^tances* seemed to be influencing a volunteer's state of mind. Thu., the administration of the I-E

PAGE 39

31 Scale used here elinsinated language complexity and word connotation differences between "backgrounds, facilitated the good rapport, and allowed the subjects to take the scale under natural conditions. The Parent Educator W eekly ReDort nnrl Derived Behavior Indices As was described earlier, each mother was visited once a week by the sarae parent educator. The mother had stimulation materials presented to her and did the previous week's activities with the baby for the parent educator's evaluation. The parent educator made an appointment for the next visit before she left. Immediately after leaving the home, the parent educator filled out an observation report, the PEWR (see Appendix B). The PEV/R (Parent Educator Weekly Report) is an omnibus form prepared by the project staff (composed of educators, psychologists and nurses) to include data on home conditions presumed to be relevant to the learning of the baby end the mother. It includes suggestions from Rheingold (196O ) and Stedman.* Inter-observer agreement between supervisor and parent educator on bi-.xinual visits was high. A formal reliability test of repeated obsein-ations was not done in order to minimise the presence of the professional staff in the home. It was felt that the presence of an unfamiliar authority figure would interfere with the observations of natural hone conditions and interfere with the rapport between a mother and her parent educator. Items from the PETO were orgmiized into three indices: the index of Positive Verbal Interaction (VH-); the index of negative verbal interaction (VI-); and the index of attitude toward Parent Education Project (PEP). Tne ite:.s which are additive rather thr^n of e^u:.! weight **Dr. Donald J. Stediran, Duke University, personal conununicaticn , I966

PAGE 40

32 were chosen on a face validity basis. E.g., a typical item for the VI+ index (in the mother column since the subjects for this study were mother or mothering one's) was "Tone of voice sounds soft and loving." A sample VIitem was "In a few words, order or tell the baby to do or not to do things." An illustrative item of the PEP index scored positive was "How did the mothering one react to your instructions?" 1. "Looked at you while you were talking, and/or asked questions." A sample item scored negative was "When you finally got to see the mothering one: 1. She said nothing about missing her appointment." After the Parent Education Project was well under way a validity check was made on the index of positive attitude toward the project (PEP) . The parent educators were asked to rank each of their volunteer mothers on positive attitude toward the project. The mothers were then ranked independently according to the PEP index. The ranks for the index correlated better than .80 with the parent educator's ranks according to Gordon.* (Appendix B lists the specific items and manner of scoring for each index.) Further validity tests are provided in studies by Gordon (1969) and Gordon, Herman, and Jester (1968). Gordon found mothers' VHindex positively related to their children's achievement. Gordon ejt al. found a reliable difference between achievement of children of mothers whose attitudes were rated positive compared to those not so positive on the PEP index. Stimulation Series Test The stimulation series, samples of which appear in Appendix C, *personal communication, 1970

PAGE 41

33 was designed by the project staff to provide infants with the kind of experience through which, according to Piagetan principles, adaptation through accommodation will lead to greater modification of development and greater cognitive organization than would otherwise be expected from growing up in a culturally deprived environment (Gordon, 1969). Sigel (1964) points out the possibility that adequate stimulation in the first two years of life may be necessary for a rate of intellectual development which will allow for acquisition of necessary knowledge later on. He suggests that one reason why children from disadvantaged homes have difficulty in kindergarten and first grade is that there was a stimulation decrement in the first two years. The assessment of progress on the series was carried out by the five staff members who taught the material to the parent educators. Parent educators and their caseloads were randomly assigned to each supervisor. Thus individual differences in scoring of the series were randomly distributed with respect to other variables in this study. Although the series test is not independent of participation in the program, the use of it does not violate any assumptions inherent in the hypotheses of this study. Each of the children being compared has an equal opportunity to progress on the material. The derivation of the series is not the main focus of the present study and will be sketched briefly here. The reader is referred to Gordon (1969) . The exercises were chosen for ease of teaching and evaluating. They were derived from Uzgiris and Hunt, 1966; Bay ley (1933), Gesell (1949), and Cattell norms (1947); Hess et al. (1965); and Bernstein, 1960 (in Gordon, 1969). The materials are organized *

PAGE 42

3'4 so that they were introduced to the infant before the behavior shoiild occur, according to the aforementioned norms and studies. Order of presentation also took developmental norms for position into account so that, for example, the first set of exercises have the baby prone or supine, while a later set has him creeping (Gordon, I969). Standar d ized Instrume nts Griffiths Men tal Developnie nt Scale The test was developed by the author (Griffiths, 195^+) to measure mental development during the first 2 years of life. The scale consists of 260 items selected from a larger pool of items from other infant scales after extensive pretesting and observation of infants. On the basis of pretesting of infants the items were also grouped into five subscales and arranged in order of increasing difficulty. There are a:i equal number of items in each of the five subscales for each age period. The subscales are: Locomotor, Personal-Social, Hearing and Speech, Eye and Hand, and Performance. The scale was standardized on 571 British children largely from infant welfare clinics and day nurseries. Like the St an ford-Bine t , this test is an age scale. Thi-ee items passed represents one week of credit in the first year. Two items passed counts as one week of credit in the second year. The mean general quotient (GQ) for the standardization sample was 99.7 with a standard deviation of 12.1. The retest reliability coefficient over an interval of an average of 30 weeks (a r^Jige of from 7 to 70 weeks) on 60 cases was .87. Correlations with Stanford-Binet IQ's at age 5 vere higher than those reported for other infant tests. The Griffiths GQ at 6 months correlated .32 with the Binet IQ at age 5. At I8 months the GQ correlated .UO with the Binet IQ at a^o 5.

PAGE 43

35 To insure adequate familiarity of the examiner with the test, the author peinnits testing materials to be sold only to persons who have been trained by her in their use (Griffiths, I95U). Bayley Ment al DevelopTTient Scale This test was designed to measure developmental progress from birth to two and one half years (Bayley, I969). The Mental Development Scale was designed to assess early learning, problem solving, communication and abstract thinking capacities. The items (163) were derrived mainly from the California Preschool Mental Scale and the California First Year Mental Scale. The items were selected on the basis of data on 1560 children. The standardization sample consisted of 1,26? children from 2 to 30 months old. The sample was selected to represent the U.S. population (i960 census) within this age range. The age placement of items was the age at which 50% of the children passed a given item. The mean standard score of the scale (called the Mental Development Index) was 100 with a standard deviation of I6. The split haJ-f reliability coefficients ranged from .81 to .93 with a median of .88. The mean percentage of retest agreement over a one week interval was 76.1* (S.D. = 13.7). The Mental Development Index (MDI) and Stsmford-Binet {Fovhi L-M) IQ were foimd to correlate .57 (Bayley, 1969). Data Colloction The parent educator weekly report observations were filled out for every ho.-ne visit beginning when each baby was three months old. Thus, if there were no missed visits, there would be 13 reports by the time the baby was 6 months old. The I-E Scale (Social Reaction Inventory)

PAGE 44

vas adxflinistered by the parent educator when each mother's baby was 6 months old and again at 21 months. Stcff members adninistered the Series Test at 6, 12, 18 and 2k months of the baby's ags, the Griffith' Mental Developnient Scale at age one, and the Eay]ey Mental Development Scale at age two.

PAGE 45

CHAPTER 6 STUDY I: RESULTS AND DISCUSSION Class, Race, Age and Parity Hypothesis 1 was that the present sample of disadvantaged mothers would be more external than previously studied samples not restricted to the lower class. Table 2 shows a comparison of the external scores of the indigent mothers in the project when their infants were 6 months old with the scores of two samples presented by Rotter (1966) and those of a group of women from lotjer class backgrounds chosen for Project Follow Through which was modeled after Project Head Start. Hypothesis 1 is confirmed. Southern, indigent mothers, mostly black (13:1), were significantly more external than samples of people from all socioeconomic levels. This confirms Battle and Rotter's (1963) finding of the relationship between class and control expectancy and extends the findings to adults. It also confirms Dean's (1961) results of correlations between scores on a scale of pOT\'erlessness, and low socioeconomic level. It is interesting to note that the Project Follow Through educators, most of whom came from the ghetto area, were significantly less external than the project mothers. While this may reflect the influence of being chosen for a relatively high status job, it may also be an indication of different sub-cultures within the lower class. As will be seen in the next section, 37 -

PAGE 46

education was not a significant varia>)le. The Project Follow Through educators obviously vere alert to information relevant to their life situation and took steps to improve their life situation in securing the job. Tiius , this finding may be taken as consistent vith the studies presented in the literature review indicating persons who know more about their surroundings and make commitments to social action are more internal in orientation. Hypothesis 2 was that black mothers would be more external than whites. The mean SRI score after 21 months of participation for 11 white mothers in the project (Groups 1-3) was T.l8 with a standard deviation of 3.63. The mean for 5^ black volunteers was 10.25 with a standard deviation of 3.5^*. The difference of 3.07 yielded a t of 5.l8 significant at less than .001. Thus, hypothesis 2 is confirmed. As suggested by previous research and theory, lovrer class members are less likely to see reinforcements as contingent upon an individual's skills and characteristics than the population at large. Furthermore, within our lower class sample, the perception of an individual as povrerless or externally controlled was more extreme among black than white subjects. Hypothesis 3 predicted no significant relationships between age or parity of mothers and their I-E scores. Table 3 shows the correlations between age, parity, and I-E Control after 3 (pre) and 15 (post) months project participation tnd. the difference between the two scores (change). Since an r greater than 0.37 would be nece3Sar:v' for significr^nt results at the .05 level, there if; no reliable relationship between maternal age or parity emd I-R Control. Hypothesis 3 is confirmed.

PAGE 47

39 Table 2 A Comparison of Three Populations Uith Subjects in Groups 1 and 2 in the Present Study on I-E Control Sample Scale Used Sex N X S.D. t Prisoners, 18-26, 8th c-rade I-E M 80 7.72 3.65 5Ji9* reading (Ladwig, 1963) National Stratified Sample I-E M&F 1.000 8.50 3.74 5.2810th, 11th, 12th grades (Franklin, 1963) Follow Throuf^h Parent SRI F 40 7.21 3.58 5.43=^Educators (Gordon, 1969) Parent Education Project SRI F A2 11.50 3.14 Mothers-groups 1 & 2 *Significant at p less than ,0001. Table 3 Correlations Between Age and Parity and Initial (pre). Final (post), & Changes in I-E Control Scores for 32 Subjects SRI Variable Pre (X=11.50 Post (X=11.0 Change (X=0.50 S.D. =3. 14) S.D. =3. 14) S.D. =3. 88) Age (X=25.94, S.D. =11. 10) +0.27 +0.32 -0.02 Parity (X=3.41, S.D. =2. 20) +0.18 +0.31 -0.02

PAGE 48

-l;0 Maternal Behaviors and Attitudes Positive Verbal Interaction (VH -) Mothers were grouped as either high or low depending on whether their positive verbal interaction index, calculated to si!r=ir.arize 21 months of visits by a parent educator, fell above or below the median index of all mothers. Then they were split at the median again according to education. SRI (Modified I-E Control Scale) scores were then analyzed again at an initial point and near the end of the project. Change scores were also analyzed. Hypothesis l+a predicted that mothers higher in positive verbal interaction would have a more internal control expectancy. Hypothesis kb predicted these same mothers would become more interned by the end of a 15 month interval. Tables k.l through k.k summarize the mean I-E scores (SRi) of subjects grouped by positive verbal interaction and education. Tables 5.1, 5.2, and 5.3 show the analyses of variance of I-E scores at the beginning (pre), and the end (post) of the project and the chejige in I-E scores over the 15 month interval. While there is no significant ' relationship between positive verbal Interaction and initial locus of control, mothers with a higher level of verbal interaction are more internal in orientation and have changed significasitly more in an internal direction after a 21 month interval than those mothers with a lower level of verbal interaction. There is no coafin^ation of hypothesis lia. Mothers higher in positive verbalizations were not more internal at the beginning of the project. The data suggest that level of verbal interaction is not directly related to locus of control. Ho-/ever, it may be related indirectly perhaps through an intervening variable of potential for change. That

PAGE 49

-Ill is, in confirmation of hypothesis k'h a mother's level of positive verbal interaction vith her child predicted her movement in control orientation. If one accepts the assumption that the verbal interaction index reflects an important maternal skill, then these results confora to those of previous research. Those mothers more active verbally in mothering changed more in an internal direction. There is some evidence to suggest that this was an important skill for mothers in the project. Gordon (1969) reports some significant correlations between amount of adult (mostly mothers*) positive verbal interaction and children's cognitive development as measui-ed by the Bayley and Griffiths Scales. One co-old speculate that mothers, seeing their efforts reinforced by their children's progress, became more internal in genera].ized control expectation. Neg ative Verbal Interaction (VI-) To test hypotheses 5a and 5b which dealt with the relationship between negative verbal interaction (VI-) and I-E Control, the mothers were groupsd in the same way as for the data on positive verbal interaction with a median split on their VIindex and again on education. Hypothesis 5a predicted that mothers with a higher frequency of negative verbal interactions with their child would be more ejrtemal in initial I-E Control. Hypothesis 5b was that the same mothers would become more external ove-r a 15 month interval. Inspection of the mean I-E scores in Tables Ij. 5 end 1;. 8 and of the analyses of the varitmces in Tables -,.h to 5.6 reveals no significant relationships between level of negative verbal interaction e-jid control expectancy. H:^otheses 5a and 5b ai-e not supported. In fact, it is interesting to note the direction of the findings is opposite

PAGE 50

k2 of what was predicted. Rather, the effect of negative verbal interaction for the pre, post, or change measure of I-E Control is in the same direction as that of positive verhal interaction but it is nonsignificant. Perhaps the word negative is a misnomer here. Statements which go into the VIindex such as: mother "in a few words directs the child to do or not to do something" might be construed as reflecting the mothering skill of setting limits. In that light, the division of mothers on this variable might result in two groups of disciplinarians, one relatively active and one passive. Consequently, the prediction would be for those mothers higher in VIto be and become more internal. The findings do not support that prediction either but suggest that replication on another sample would be worthwhile.

PAGE 51

43 Table 4 Mothers' Initial (pre), Final (post), and Changes in I-E Control Scores (SRI) Grouped by Level of Verbal Interaction and Education Positive Verbal Interaction (VI+) Table 4.1 Frequency of Mothers per Cell Education VI+ Low Low 9 High 7 High 7 Table 4.3 Post SRI Cell Means Education Low High Low 12.22 12.29 High 9.00 10.33 Column 10.81 11.19 VI+ Row. 12.25 9.75 Table 4.2 Pre SRI Cell Means Education Low High Row Low X/ . J J ±1 . Ub VI+ X z. • ^ ^ J. J, . Col iiinn — — 12. 00 XX • \J\J Table 4.4 — — ^ Changes in SRI Cell Means (Post-Pre Education Low High Row Low -0.11 2.86 1.19 VI+ High -2.57 -1.89 -2.19 Column -1.19 0.19

PAGE 52

44 Table 4 (continued) Negative Verbal Interaction (VI-) Table 4.5 Frequency of Mothers per Cell Education Low High Lov7 10 7 VIHigh Table 4.7 Post SRI Cell Means Education Low High Low 11.40 11.57 High 10.29 10.62 Column 10.44 11.07 VI8 Row 11.47 10.46 Table 4.6 Pre SRI Cell Means Education Low High Low 12.50 9.86 High 11.29 11.88 Column 12.00 10.93 VIRow 11.41 11.60 Table 4.8 Changes in SRI Cell Means (Post-Pre) Education Low High Row Low -1.10 1.71 0.06 High -1.00 -1.25 -1.13 Column -1.06 0.13 VI-

PAGE 53

45 , Table 5 Analyses of Variance of Initial (pre), Final (post), and Changes in I-E Control Scores (SRI) by Mothers' Verbal Interaction Level and Education Positive Verbal Interaction (VI+) Table 5.1 Pre SRI Scores Source Sums of Squares DF Mean Squares F Rows (VI+) 8.127 1 8.127 0.86 Columns (Educa tion) 10-003 1 10.003 1.05 Interaction 24.890 1 24.890 2.63 Within 264.985 28 9.464 Table 5.2 Post SRI Scores Source Sums of Squares DF Mean Squares F Rows 52.718 1 52.718 5.93-" Columns 3.841 1 3.841 0.43 Interaction 3.174 1 3.174 0.36 Within 248.935 28 8.892 Table 5.3 Changes in SRI Scores (Po st-Pre) Source Sums of Squares DF Mean Squares F Rows 102.240 1 102.240 8.45* Columns 26.240 1 26.240 2.17 Interaction 10.286 1 10.286 0.85 Within 330.349 28 12.084 *p .05 =4.18; '•'••'^p .01 = 7.64

PAGE 54

46 Table 5 (continued) Negative Verbal Interaction Table 5.4 Pre SRI Scores Source Suras of Squares DF Mean Squares F Rows 0.28 1 0.28 0.03 Columns (Education) 9.07 1 9.07 0.92 Interaction 20.99 1 20.99 2.13 Within 275.80 28 9.85 Table 5 S Post SRI Scores Source Sums of Squares DF Mean Squares F Rows 8.03 1 8.03 0.76 Columns 0.12 1 0.12 0.01 Interaction 4.30 1 4.30 0.40 Within 297.36 28 10.62 Table 5.6 Chan gcs in SRI Scores (Post-Pre) Source Sums of Squares DF Mean Squares F R0X7S 11.37 1 11.37 0.72 Columns 11.33 1 11.33 0.72 Interaction 21.47 1 21.47 1.36 Within 441.84 28 15.78

PAGE 55

hi Edu cation, Verbal Inte raction, and T-E Co n-tyol Hypothesis 6 stated that mother's educational level would not significantly effect I-E Control and that education and level of verbal interaction either positive or negative wo'old not have a significant interaction effect on I-E Control. Table k shows small differences between I-E scores of mothers high on educational level a^.d those low in education. The analyses of variance in Table 5 indicate these differences are not significant. The data confirm hypothesis 6 and previous research that within a relatively horaogeneous population there is no relationship between level of education and internsl vs external control expectancy. Education also did not interact significantly with level of either positive or negative intersiction to affect expectancy scores. Infant Cog nitive Deyelcroynt Influ ence of Mother on Child Hypothesis 7a predicted that children of nothers more internal in control expectancy would learn more and be novo highly developed cognitively by the tir,o they reached their second birthday. Tables 6, 7. and 8 show the results. The assessment of the effect of mother's control expectancy was done in several wa^-s. First, to control for differences in the child's initial achiever.ent , the child's achievement scores were plugged into the short tera prospective model (Uenar and Wenar, 196h) . The children were grouped first according to whether their r.tothers were above or below the medi.^n in initial expectancy, then accordi.ns to whether they were high or low in initial achievement.

PAGE 56

Ji8 In apparant contradiction to the hypotheses, the children of initially more external mothers learned more although the difference does not reach a significant level of probability. However, this is only true when the amount of achievement is calculated from the child's 6 month score. Table 7 shows that there is no significant relationship between the child's initial and later achievement and little difference between children of external vs internal mothers in the child's achievement during his second year of life. In fact, because of a ceiling effect there is a reversal. Children of more internal mothers learn more in the second year. This suggests, since the series test had a finite number of items and there was a significant inverse relationship between initial achievement score and what was left to learn, that children learned about the smne number of items over two years regardless of their mother's expectancy but the children of the more external mothers learned more by the time they were 6 months old. They then leai-ned less than children of more internal mothers over the next l8 months. This is supported by the findings in Table 8. There is a difference of less than one item achieved by children of external vs internal mothers, at either 18 or 2h months of age. Ihus, mother's I-E control e:tpectancy is not related to her child's progress on the series test. Table 8 further indicates no significant relationship between maternal expectancy and children's scores on the standardized cognitive development scales. Although these results are in the predicted direction, the hypothesis that mother's control expectancy influences her infant's cognitive achievement is not supported.

PAGE 57

I;9 Inf]Aaen^^_Ch Hypothesis 7c predicted that a child's achievement vould significantly effect his mother's I-E Control expectancy. The hypothesis stated that the children vho achieved raore or vere more highly developed cognitively vould effect changes in their mothers ' expectancy in an internal direction. The results were analyzed the same way as for the assessment of the influence of mother on child. In this procedure changes in mothers' expectancy scores were plugged into the model (Table 9). There vere no significant relationships between child's initial achievement or mother's initial expectancy and changes in maternal expectancy. Hypothesis Tb predicted that mothers of more highly developed children would be m.ore internal in final expectancy. Children's success over the course of the project as measured by their final Series Test scores and the standardized tests was not found to be significantly related to a mother's final expectancy or changes in expecto.ncy. (See Table 10). Again, although most of the results were in the predicted direction, hypothesis 7b is not confirmed. At the end of the project mothers of children who achieved more were more internal than the mothers of less successful children. This is not true of changes in expectancy over the duration of the project. With one achievement measure (Griffiths) mothers of higher scoring children became more internal. With a second (Bayley) there was no difference p-nd with a third (Series Tost) they becarn.e more external. V/hile there is sons evidence that mother.' control expectancy and their childrens' achievement .ffoct each other, the results vere not reliable. / However, the results vere .ng.e.tivc and vere investigated further with a nev saraple in Study IT.

PAGE 58

50 Ef fects o f Far-ticipatioji Amount o f Pa rticipation Hypothesis 8: Mothers who participate longer in a progran (about 21 months) to improve their mothering skills will hecome more internal in I-E Control than those who participate a shorter amount of time (9 months ) . Insufficient data were available to test differences "between mothers who did versus those who did not participate in the project. The parent educators were not able to reach enough mothers in treatment groups 3 to administer the SRI. There were 21 mothers each in group 1 and 2 who were given the SRI when their babies were 6 months old and again 15 months later. The tests between these groups which appear in Table 11 assess the effects of maternal expectancy and expectancy changes of 9 vs 21 months of participation in the project. The results do not support the hypothesis that longer participation changes control expectancy in eji internal direction. Mothers who participated longer were more external and changed more in an ex-ternal direction than mothers who did not participate in the second year of the project. The change for all k2 mothers was a mean of 0.02 which suggests the project did not effect changes in control expectaiicy. However, these findings are not too meaningful in the absence of a control group. A no treatment control group was available for Study II. Hypothesis 9a stated that mothers who make better use of the program as manifested in such behaviors as keeping appointments and mastering materials will be more internal in initial I-?I Control. Hypothesis 9b

PAGE 59

51 Table 6 Changes in Children's Series Test Scores (6-24 months) Grouped by Their Initial Score and Their Mother's Initial Expectancy a . Means Child's 6 Month Series Score Low (N=16) High (N=19) Row Means Mother's Initial Expectancy Internal (N=19) 35.33 27. 08 External (K=16) 35.50 31. 33 Column Means 35.44 i 28.42 b. Analysis of Variance Source df Mean Squares F Rows 1 38 34 2.73 Columns 1 302. 45 21.50*-"Interaction 1 32. 80 2.33 Within 31 14. 07 29.68 33.94 *"P < .001

PAGE 60

52 Table 7 Changes in Children's Series Test Scores (12-24 months) Mother's Internal (N=16) Initial — — „ Expectancy External (N=15) a. Means Child's 6 Month Series Score Lov7 (N=16) High (N=15) 8.50 11.17 9.10 Column Means 9.88 6.80 8.07 Row Means 9.63 8.33 b, Anal^'sis of Variance Source np Mean Squares F ^ 27.77 1.81 Columns i • , 1 40.10 2.51 . 1 0.05 0.00 Within 27 15.36

PAGE 61

53 Table 8 Differences Betv7een Success Scores of Children Grouped According to I-E Scores of Their llothers at the End of 24 Months Achievement M easure Bayley Mental Development Scale Griffiths Mental Development Scale Series Test (24 months) Mo ther ' s _Eig)_ect aiicy Internal 83.55 112.79 44.17 External 80.50 109.56 43.53 0.97 I.IS 0.56 Table 9 Changes in Mothers' Expectancy Scores (I-E) Scores (3-21 months) Grouped by Their Initial Score and Their Child's Achievement at 6 Months a . Means Child's 6 Month Series Score Lovr High Mother' s Initial Internal (N=21) 1.14 0.36 External (N=21) -0.29 _^ ^l.l^ Column Means 0.19 -0.14 Row Means 0.62 -0.57 b. Analysis of Variance Source df Mean Squares F R0V7S 1 20.012 1.30 Columns 1 6.298 0.41 Interaction 1 0.012 0.00 Within 38 15.415

PAGE 62

54 Table 10 Differences Between I-E Scores (Final SRI and ChangerOver 15 N.onths) of Mothers According to AchiovaiTient of Their Children Achievement Measure Bayley L0X^? jnigh I J ) L i Low I High Final (Post) SRI Score N X t 16 11.13 1.08 17 9.94 Changes in SRI Score (Post-Pre) 16 0.37 0.00 17 0.37 Griffiths 1 Post SRI I N X i Low 17 , 10.94 .High 16 10.63 i i Change SRI Low 17 1.29 iHigh 16 -0.63 t 0.36 1.47 24 Month Series Low High I j Low ! I ; High Post SRI N X 16 10.88 18 . 10.11 Change SPII 16 -0.44 18 1.06 t 0.59 1.10

PAGE 63

55 Table 11 Differences Betv;een Treatment Groups in Initial, Final and Chanties in Expectancy Scores (SRI) Treatment Group Mother's Expectancy 1 (21 mo) 2 (9 mo) t (N=21) (K=21) Initial 10.43 ll.Srj -I.39 ^i^al 11.24 11.10 0.14 Change 0.81 -0.76 1.34 Table 12 Differences in I-E Control Expectancy Scores (SRI) Between Mothers Rated High & Low in Project Involvement (PEP) Mothers' Positive Attitude Index (PEP) Mothers' Expectancy Low (N=20) Hiah (N=22) t ^^^^^^^ 11.30 11.00 0.44 Fin^l 11.45 10.91 0.77 ^^^"^^ 0.15 -0.09. 0.28

PAGE 64

stated that such mothers vould also become more internal in I--S Control over a 15 month interval. In order to test these hypotheses mothers above and below the median on the PEP index vers compared on their expectancy scoi^es at the beginning and end of the project. Table 12 reveals no significant differences in expectancy betveen mothers high on the PEP index versus mothers low on the index. The tvo hypotheses are rejected. Mother's degree of involvement in a program to improve her mothering skilli; vas not significantly related to her I-E Control expectancy or her changes in expecta:icy over the course of participation. The reasoning for this hypothesis vas in line vith other studies: those subjects vho knew more about their role, who made a greater coraiiiitment to it and vho actually did perfon^i their role in the project better would be those vho vere more internal in control expectancy. It vould seem that the mothers did not perceive their involvement in the project as crucial to their general life situation or at least not in the vay that the items measured it. Some of the items, e.g. those dealing vith the number and punctuality of appointments kept, may reflect a middle class bias. Furthermore, it is possible that the project was not completely successful in getting the mother to identify with the role of her child's teacher despite the heavy emphasis by the parent educators (Gordon, 1969) of the importance of the mother in influencing her child's cognitive growth. Perhaps a r,iore fruitful way to conceive of this variable is as an index of maternal activity which in interaction vith another maternal

PAGE 65

5T activity variable such as verbal level might have an effect on maternal expectancy. This vas attempted in Study II.

PAGE 66

CHAPTER 7 STUDY T: CONCLUSIONS As expected and in line vith previous research, the present sample of lover class mothers vas found to be more external in locus of control than sa-nples not restricted to the lower class. Within this lover class sample, black mothers vere more external than whites. The age, number of children or education of the mother were not significantly related to control expectancy. The number of positive verbalizations of a mother to her child was related to her I-E Control vith those mothers verbalizing more more internal in control. On the other hand, negative verbalizations were not related to I-E Control; behaviors presumed to manifest mothers' involvement in the project and the teaching role were not related to I-E Qjntrol. The prediction of a two way relationship between m.aternal expectgjicy and her child's achievement was not bo:ne out. Ti^e findings were suggestive of a relationship but were not reliable. In an naturalistic study such as this with several variables beyond the experimenter's control, the reliability of the data increases as procedures improve vith the benefit of practice and hindsight. For example, in Study I tim.e needed to solidify data collection plans precluded the availability of a purer measure of the mother's initial expectancy. TT^at is, she entered the project when her baby vas three 58 -

PAGE 67

59 months of age but was not given the SRI until three months later. Thus, possible ch?jiges in expectancy during the first three months of participation might have been lost. Furthermore, missing data prevented the availability of a control group to contrast with a participation group to assess the effect of participation on control expectancy. Study II was planned with more adequately controlled proced^ares to get a more reliable assessment of the variables investigated in Studj' I.

PAGE 68

CHAPTER 8 INTRODUCTION TO STUDY II Three new groups of mothers vere available for investigation of the relationships between their level of verbal interaction, their degree of involvement in the project, the effect of their participation in a program to enhance their mothering skills and their I-E Control. The two way relationship between mother's I-E Control and her child's achieve:nent and the effect of class, race, education and parity of the mother and her I-E Control were also re-investigated using this new sample. Better controlled procedures were used in Study II. The parent educators were more experienced researchers. Tliat is, their observation and recording skills had improved. Entry scores instead of entry plus 3 month scores on the I-E Control Scale were available as the measure of initial expectancy. Tha re-test on the scale was given at the end of the mothers' participation in the program instead of 3 months before the end of the project as the measure of final expectaaicy. Thus, changes in I-E Control scores (final-initial scores) more accu^-ately reflected changes in expectancy as effected by project participation, and maternal behaviors measured during the course of the project. Furthermore, the test-retest interval was shorter (9 months). Consequently, the retest reliabil^y of the I-E Scale which tends to diminish significantly over intervals greater 60 -

PAGE 69

61 than 6 months (Kiehlbauch, 1968) was increased. Finally, two new control groups were used: a non-participant control group was available to test the effect of participation on I-E Control and a control group that participated in a modified program to test the effect of participating in a specific type of progran on I-E Control. Some conceptual changes were made in line with suggestive findings from Study I. The negative verbal interaction (VI-) and positive attitude (PEP) indices were conceived of as measures of maternal activities which singly or in interaction would be related positively to internal control. When these measures did not relate significantly to I-E Control in Study I, the items were re-examined. The author felt that these items perhaps reflected the mother's level of actix-ity and amount of involvement with the role of her child's teacher in the same way that the items which make up the positive attitude index (VI+) do. In fact, the VIindex varied with I-E Control scores in the sar.e direction as but with slightly less magnitude than the VI+ index. In addition, the author was also interested in investigating whether certain combinations of these variables would interact to effect I-E Control. Level of mother's education showed no reliable tendency to affect mother's I-E Control end also showed no significant interaction with verbal level in its effect on I-E Control in Study I. Consequently, the prediction in Study II was that education would be insignificantly correlated with I-E Control. '

PAGE 70

CHAPTER 9 HYPOTHESES FOR STUDY II The hypotheses for the second study are the same as for Study I except for the changes discussed in Chapter 8. Demographic Variables 1. The present saicple of lover class mothers will be more ertemal in I-E Control than previously studied samples that vrere not restricted to the lover class. 2. Black mothers will be more external than white mothers. 3. There will be no significant correlations between nother's age, parity or education and 1-E Control. Maternal Be haviors k. Mothers who have a higher frequency of behaviors related to their role as child's teacher: a. Positive verbal interaction (VI+) b. Negative verbal interaction (VI-) c. Positive attitude towards a program to enhance her teaching skills (PEP), will be more internal in I-E Control than mothers with a lower frequency of these beha^/iors . 5. Mothers with a higher frequency of behaviors listed in hypothesis Ita, b, and c will also become more internal in I-E Control over a 9 month interval. 6. There will be a significant Interaction effect between mothers' positive verbal interaction and positive attitude towards a program to enhance her teaching skills (pep) on I-E Control e>:pectancy changes. 62 ^

PAGE 71

63 7. There vill be a significant interaction between mothers' negative verbal interaction frequency and positive attitude towards a teaching progi^aa (PEP) on I-E Control expectancy changes. Infant Cognitive Development 8. There vill be a significant two way relationship between mother's I-E Control and her child's cognitive developnant such that: a) Children of mothers more internal in I-E Control will learn more and be more highly developed by the end of their first year of life than children of mothers more external in I-E Control. b) Mothers of children who are more highly developed cognitively vill be more internal in I-E Control than mothers of less well developed children. These mothers will also become more internal over a 9 month interval. Effe cts of Participation 9. Mothers who participate in a program to improve their mothering skills will become more internal in I-E Control than mothers who do not . 10. Mothers who participate in two different programs to improve their mothering skills will not differ in I-E Control score changes over the prograon's duration.

PAGE 72

CHAPTER 10 STUDY II: J-STHODS SaTTip le The subjects vere selected using the same criteria as for Study I. Thus, they vere from indigent families. Infants vith complications possibly related to cognitive development vere screened out. These 5h mothers and their infants began participating in the Parent Education Project (see Chapter 2) one year after the subjects in Study I. Design The treatment plc.-i for the groups in this study is outlined in Table I (page 5). In that table they are referred to as groups If, 5, aiad 6. The first experimental group (here called El) received veekly visits from a parent educator from the time the baby was 3 months old until his first birthday. T^ie parent educator used teaching materials designed by the professional staff of the Parent Education Project (see page 26, Stimulation Series) in working with the mother. Another experimental group (E2) also received weekly visits from a parent educator. However, these educators, also indigenous non-professionals , planned their o-.ai teaching progra.1 based on their own experiences working in Head Start and nursery school programs. Thoy were supervised by a research assistant on the Staff of the Parent Education Project but one who had no specific knowledge of the Series Test. A control group (c) of nothers and inf...ts received no treatment but was 6h -

PAGE 73

65 tested. All three groups of mothers vere given the I-E Scale (SRi) at their entry into the project (infant is 3 months old) and again at the end (infant is one year old). The infants vore given the Series Test at 6 months old and at their first birthday. They were also given the Griffiths Test at one year old. Except for the control group families who were not visited, the parent educators recorded their observations after each weekly visit on the PEI-m. Data A nalyses The measures used are identical with the instruments and indices used in Study I. The hypotheses concerning class and race were tested by t tests of differences between groups on their I-E scores. The remaining demographic variables were tested calculating a Pearsonian r between each independent variable and I-E Control scores and a multiple r between all the independent variables and I-E scores. The hypotheses concerning maternal behaviors were tested by t tests of mothers' initial (pre), and final (post) I-E scores by an aiaalysis of variance of changes in mothers' I-E Control scores grouped by median splits on the three maternal behavior indices. The relationship between maternal I-E Control and child cognitive development was tested first by splitting mothers at the median on their I-E scores and calculating a t for the difference between their children's achievement scores, then by splitting children at the median of their achievement scores and doing a t test of the difference between their mothers' I-E scores. The effects of participation vore assessed by t tests of the differences between experimental and control groups on their final I-E score

PAGE 74

66 and I-E score changes. Tlie difference between the tvo experimental approaches was assessed hy a t test of the differences between groups El and E2 in final I-E score and I-E score changes.

PAGE 75

CHAPTER 11 STUDY II: RESULTS AND DISCUSSION Demographic Variables Class Hjrpothesis 1 states that the present sample of lower class mothers vould be more external in I-E Control than previously studied sairiples not restricted to the lower class. The results in Table 13 replicate those of Study I. The National Stratified sejnple and the prisoners (from varied socioeconomic backgrounds) were significantly more internal than the present seaple of indigent mothers. Thus, hypothesis 1 is confinaed. The comparison of the Project Follow Through educators with the disadvantaged mothers again indicates within class differences in I-E Control. Again one can speculate that the vomen hired for Follow Through have received reinforcement for their personal characteristics and skills. This is consistent with Rotter's (1966) theory that consistent reinforcements for personal qualities and abilities leads to adoption of the internal control expectancy. Race Inspection of the data in Table ill reveals that the hypothesis for a racial difference in I-E Control supported by the Study I findings is not supported here. Although Negro mothers are more e>:ternal than Caucasian mothers both at the beginning and end of their project participation, the difference does not reach a significant level. 67 -

PAGE 76

68 Table 13 A Comparison of Three Different Populations with Project Mothers on I-E Scores Group National Stratified Sample 10th, 11th, 12th grades (Franklin, 1963) Prisoners, 18-26, 8th grade plus reading (Ladwig, 1963) Follow Through Educators (Gordon, 1969) Project Mothers Scale N X I-E 1,000 8.50 I-E SRI SRI 80 7.72 40 7.21 54 10.04 s t 13.99 2.91* 13.32 12.82 14.59 3.68** * .01 level of significance ** .001 level of significance Table 14 Differences 35etv7een Black and VJhite Mothers in Initial and Final I-E Control Scores Initial Expectancy N X t Blacks 43 10.26 0.84 Whites 11 9,18 Final ExpectancyBlacks 43 9,02 1.15 Ifliites 11 7,5/,

PAGE 77

69 Table 15 Initial Correlations Between Education, Af>e, Parity and (pre), Final (post), ana Changes m I-E Scores i SRI Score Education (E) Age (A) Parity (P) Multiple^^^ • EAP Pre .00 -0.22 0.20 0.26 Post -0.13 -0.30 0.25 0.38 Change -U.uo -U.±j U.UX Means 10.21 years 19.25 years 2.61 children p=.05 =0.37 Table 15.1 Analysis of Variance of the Multiple Correlation of Ar^e and Education with Final I-E Control Scores , Parity, Variable Entered Multiple Mean r Source df Squares F Ratio P-.05 Age 0.30 Regression 1 29.37 Residual 26 11.54 2.55 4.22 Parity 0.37 Regression 2 23.63 Residual 25 11.29 2.09 3.38 Education 0.38 Regression 3 15.80 Residual 24 11.75 1.35 3.01

PAGE 78

70 It should be kept in mind, however, that the moan difference of 1 between blacks and wttites is typical of previous studies. Lefcourt and Ladwig (19633) found a simiJar difference which held up as significant on a larger N (120 ). Kiehlbauch found a slightly smaller difference in the same direction on an N of 80 which was also non-significant. If Study I is included, all foiir studies combined show an average difference of more than one. While the question bears fiirther investigation, & review of all the studies strongly suggests there is a racial difference in I-E Control albeit small as Measured by Rotter's scale. A ge, Parity, and Education Hypothesis 3 predicted that there would be no significant correlation between mother's age, parity and education and I-E Control, Table 15 shows the correlations of these three variables with initial (pre), and final (post) expectancy (l-E Control scores) and changes in expectancy. The correlations are STnall (O to 0.25) and insignificant. Table 15.1 indicates the degree to which the variables predict expectancy scores if weighted and combined (multiple r). The variance indicates prediction of I-E Control scores from age, parity and educational level of the subjects is non-significant. Hypothesis 3 is confirmed. As predicted, age, parity, and education of the mother was not significantly related to I-E Control. Even in combination these three variables could not predict expectancy changes significantly better thaii chance. Mate rn al Beh av i or s ^°^-PJ-Jf2-X^2)r^l^lSIB^l2-J>}} (VI+) Hypothesis Ija stated that mothers with a higher frequency of

PAGE 79

71 Table 16 Differences in Mean Initial & Final I-E Scores Betv/een Mothers Above & Below the Median on Behavior Indices Behavior Index N Positive Verbal His^h 14 Interaction (VI+) Lev; 14 Nef»ative Verbal Hi^h 14 Interaction (VT-) Low 14 Positive Attitude Hi^h 14 Toward Project (PEP) Low 14 Initial Expectancy X t 10.43 -0.39 9.86 8.64 2.21''11.64 9.57 0.84 10. 71 Final Expectancy X t 8.86 -1.08 7.43 7.93 8.36 7.29 9.00 0.33 1.30 *P< .05

PAGE 80

72 Table 17 Analysis of Variance of Changes in Mothers' I-E Scores Effected by Positive Verbal Interaction (VI+) and Positive Attitude Toward Project (PEP) a. Means Positive Attitude (PEP) Low (N=14) High (N=14) Rov; Means VI+ Low (N=14) -1.33 -A. 40 -2.43 High (N=14) -2.40 -1.11 -1.57 Column Means -1.71 -2.29 b. Analysis of Variance Source df PvOWS 1 Columns 1 Interaction 1 Within 24 Mean Squares f 7.936 0.47 5.079 0.30 30.489 1.81 16.887

PAGE 81

73 Table 18 Analysis of Variance of Chan<^es in "others' Expectancy Effected by Negative VI fi Positive Attitude Positive Attitude a. Means Lov7 (14) High (14) Row ^'eans VILow (14) -2.00 -5.00 -3.29 High (14) -1.33 -0.25 -0.71 Column Means -1.71 -2.29 b. Analysis of Variance Source df R0V7S 1 Columns 1 Interaction 1 Within 24 Mean Squares 50.298 6.298 28.583 15.201 f 3.31 0.41 1.88

PAGE 82

positive verbal interaction would be more internal in I-E Control than mothers with a lower frequency of positive verbal interaction. Hypothesis 5a vas that these mothers vould also become more internal over a 9 month interval. Tables l6 and 17 summarize the findings. Inspection of the means and t score in Table l6 reveals that mothers above and below the median on the VI+ index do not differ significantly in initial expectancy. This finding is consistent with that in Study I. However, contrary to the finding in Study I mothers high and low on this index do not differ significantly in final I-E scores. In addition, inspection of the mean change I-E scores and the row variance In Table 17 reveals these mothers do not differ significantly in expectancy changes. In fact, these findings tend to be in the opposite direction. That is, mothers with a higher frequency of positive verbal interaction are more external after a 9 month interval and have changed more in an external direction. Hypotheses 5a and 6a vrere not confii-med. The significant findings on positive verbal interaction and I-E Control in Study I were not replicated. ^ggMye Verbal Interactio n Hypothesis kh stated that mothers with a higher frequency of negative verbal interaction (VI-) vould be more internal in I-E Control than mothers with a lower frequency of negative verbal interaction. Hypothesis 5b predicted that these m^others would also become more internal over a 9 month interval. Inspection of the means and t score in Table l6 reve£.J.s that mothers above a.nd below the median on the VIindex differ significantly in I-E Control. As predicted, mothers higher on this index were initially

PAGE 83

75 more internal. Hovever, after a 9 nonth interval there is no longer a significant difference "between Eothers high and low on this index in their I-E scores. Inspection of nean I-E change scores in Tahle 18 shows that mothers low on VIchange considerably more in an internal direction. While the change is relatively large, analysis of the row variance shows that it is not significant and the findings tend to be in the opposite direction of Study I in which mothers higher in the VIindex became more internal. Differences between these findings in Stu^y I and Study II will be discussed at the end of the section on maternal behaviors . The hypotheses are generally not supported in Study II. However, one wonders about the relationship between the VIindex and initial expectancy in Study II. Mothers who use more negative verbalizations initially are significantly more internal than those who use less. Over the duration of the project mothers lower in negative verbalizations change more in sxi internal direction. Thus, at the end of the project the difference between high and low negative verbalizers has become insignificant. Perhaps participation in the project has a leveling effect on I-E Control which obscures other variables. A control group of non-participants on whom behavior indices were available would be necessary to test the significance of this effect. It also must be kept in mind that the scale has a floor of zero. At a mean of 8.61* , the scores of the high VImothers were not as free to vary do-^ward as the scores of the low VIraothers . This suggests a more optimal instrument for this variable would be one which discriminated better at the more internal extreise. *

PAGE 84

76 Positive Attitude Towards a Prog ram to Enhance Mothers' Teaching Skills (PEP) Hypothesis 4c stated that mothers with a higher frequency of behaviors which reflect their positive attitude towards a program to enhance their teaching skills (PEP) would be more internal than mothers with a lov;er frequency of such behaviors. Hypothesis 5c predicted that these mothers would also become more internal in I-E Control over a 9 month interval during which time they participated in this program. Inspection of the means and t scores in Table 16 and of the analyses of the column variances in Tables 17 and 18 reveals small but non-significant differences in I-E scores at the beginning or end of their participation in the program and in changes in I-E scores between mothers high and low on the PEP index. This is consistent with the Study I findings. Mother's attitude tavjard the program she is participating in as m.easured by the PEP index was not found to have any significant, main effect on her I-E Control Expectancy. Positive Attitude and Positive Verbal Interaction Hypothesis 6 predicted an interaction effect betx^een positive verbal interaction and positive attitude towards the project C PEP) on I-E Control expectancy changes. Inspection of the interaction variance in Table 17 reveals no significant findings. Hypothesis 6 is not supported. Positive Attitude and Negative Verbal Interaction Hypothesis 7 predicted an interaction effect between negative verbal Interaction and positive attitude (PEP) on I-E Control expectancy changes. Analysis of the change score variances in Table 18 reveals no significant interaction term. Hypothesis 7 is not supported.

PAGE 85

77 Positive attitude towards or active involvement with a program to improve child teaching skills as assessed by the PEP index did not have significjint main effects or interaction effects with level of verbal interaction (positive or negative) on mothers' I-E Control expectancy. However, it is interesting to note (see Tables 17 and 18) that mothers lower in verbal activity (either positive or negative) and higher in project involvement became the most internal (a change of 4-5 scale points). This may suggest incompatibility between the two maternal behaviors. The observations which go into the behavior indices are made when the parent educator is working with the mothers. Those mothers who got the most out of the project may have spent more time listening to the materials presented than talking to their babies. Maternal Behaviors Summarized Of the three behaviors studied (VI+, VI-, PEP) only negative verbalizations were related to mother's control expectancy assessed when her baby was three months old. None of the behavior indices were related reliably to final expectancy, neither were they singly or in interaction significantly related to expectancy changes. Final expectancies were in the predicted direction for positive attitude and level of negative verbal interaction but not for level of positive verbal interaction. That is, mothers higher in frequency of verbal commands and negatively toned verbalizations during their child's first year of life were more internal while those higher in frequency of verbalizations of tender, loving care were more external. The findings on positive verbalizations are opposite those of Study I, and Study II involved babies one year younger. Perhaps the mothers found negative

PAGE 86

78 verbalizations more effective in coping with younger infants, and positive verbalizations more effective in coping with oldsr ones. Possibly this greater or lesser degree of effectiveness was slightly related to I-E Control. Inconsistences eJid non-reliable findings in the two studies on the maternal behavior variables and I-E Control may also be a function of the limitations of the measures used. While validity studies of the behavior indices have been reported (p23) no refinement of the internal consistency of the items or investigation of the factorial composition of the indices hab been attempted. Quite possibly individual items or factors from these behavior indices might provide a better test of the relationship between the variable each index is purported to measure and I-E Control. As presently constituted each index is a rough measure of its respective variable. Then too, the I-E Control Scale, as was mentioned in an earlier section, has a low re-test reliability over an interval beyond 6 months. If changes in expectancy over time are to bo related to other variables such es matei-nal behaviors in future research, it seems necessary to vise a control group for which measures on such variables are available and/or to develop a more stable Eeasm-e of I-E Control. Infajit Cognitive Growth Hypothesis 8 predicted a significant two way relationship between mother's I-E Control and her child's cognitive development. As in Study I there were no significant relationships in either direction. The findings in Table 19 indicate no significaiit effect of mother's I-E Control on her child's progress ou the Series Test at cither 6

PAGE 87

79 (Table 19.1) or 12 months (Table 19.2) or on his overall intellectual growth es measured by the Griffiths Test (Table 19.3). Inspection across Tables 19.1, 19.2, and 19.3 reveals that the children of initially Kore external (higher scores) mothers progress more on the series by the time they are 6 months old. Then perhaps (because of missing data they are not exactly the same subjects) there is a reversal. The children of initially more internal (lovrer scores) mothers are slightly ahead on the series by their first birthday. These same (with tvo exceptions) children are also more highly developed in intellectual growth as measured by the standardized (Griffiths) test. At 12 months old (after 9 months' participation in the Parent Education Project) children of mothers vho are more internal at the end of their participation have learned somewhat more on the series and score higher on the Griffiths Test. However, none of these trends are significant . p-uorthermore the results are weakened by the reversed relationship between changes in mother's I-E Control and her child's achievement on the Series Test tmd the Griffiths at 12 months. That is, the children of mothers who change more in an internal direction are less developed cognitively. Since mothers who were more internal at the pre and post test of the SRI had children that were mere developed cognitively the reversal of this relationship for the SRI change measure suggests different mothers are grouped as internal for each of the SRI measures. As in the preceding section of this paper, this again questions the re-tcst reliability of the I-E Scale. The findings in Table 20 indicate no significant effect of a child's

PAGE 88

80 Table 19 Differences Eetvcen Achieveraent Scores of Children Grouoe'd by Median Split of Mother's Initial (Pre), Final (P( and Changes in I-E Control (SRi) Scores Table 19.1 Series Test Scores (6 months) Pre SPI t low (II-8) 9.13 -0.62 high (N=6) 10.33 Table 19.2 Series Test Scores (l2 months) Pre SRI t Post SRI t low (N=15) 3!*.60 0.59 low (N=lli) 3^.11 0.50 high (N=13) 33.62 high (N=in) 33.1*1 ^ Chan gi e SR I low (N-l6) 32.6? high (N=12) 35.11 -1.35 Pre SRI low (N=ll;) 112.00 high (11=12) 107.08 Table 19.3 Griffiths Test Scores t Post SRI low (N=13) lll.TT high (TI-13) 107.69 C_hp.nge _SRI low (N=llO 109.80 high (r^l2) 111*. 25 -0.91

PAGE 89

81 achievement on his mother's I-E Control. Progress on the Series Test at 6 months vas not sifjnif icantly related to mother's initial expectancy (Table 20.1). Children vho learn more initially have more external, mothers but by the time they are a year old children who have progressed more on the series have mothers vith the same expectancy as children who have progressed less (Table 20.2). Children who scored higher on the standardized test of infant cognitive growth have ^aothers who are more internal in final expectancy. Table 20.3 indicates a surprising but insignificant trend. Mothers of children who score lover on the achievement tests change more in an internal direction. This is opposite from \;hat one would expect from previous research indicating an inverse relationship between failure and intemality as siysnarized in the literature review of this paper. The hypothesis of mutual influence between maternal I-E Conti-ol and child achievement must be rejected. A mother's expectancy about events in the world is too general to be influenced by a specific achievement of her child. Ihis is perhaps too small a sample of her infant's behavior after too short a time in the infant's life to change her overall expectancies of internal control. It probably takes a wider renge of her o\m and her child's life activities and circumstances to influence this orientation. Even within the area of possible influence on her child's achievement, especially considering the variability of intelligence tests at this age, the effect ma,y be reduced by the lack of reliability of these tests as indicator;of the child's overell cognitive growth. The lack of reliability of this relationship is further reduced by the presence of many intervening variables such as the difficult field conditions under which the attitude and intellectual tests were administered

PAGE 90

82 Table 20 Differences Between Initial (Pre), Final (^ost), and Changes in I-E (SRI) Scores of >;others Grouped by Median Split of Children's Achievement Scores (6 and 12 Month Series Tests, and Griffiths Test) Table 20.1 Pre SRI Scores S eries Test t (6 month) low (N=7) 10.10 0.00 high (N=7) 10.10 Table 20.2 Post SRI Scores Series Test t Series Test t (6 m onth) (12 month) low (N=8) 8.25 -0.74 low (!I=1A) 9.30 0.00 high (N-9) 9.A4 high (N=12) 9.30 Gr i ffjitjn s T e st t low (N=12) 8.75 0.97 high (K=14) 7.43 Table 20.3 SRI Score Changes Series Test t Series__Test t (6 month) jCjXi?k'2th)_ low (N=7) -3. CO -0.40 low (ii=12) -2.50 -1.06 high (N=7) -2.14 high (N=12) -0.75 Griffiths Test t low (N=12) -2.92 -1.03 high (i\=14) -1.29

PAGE 91

83 (Gordon, I969) and the lov re~test reliability of the I-E Scale for an interval greater than 6 months. Variables also intervene between mother '3 expectancy of I-E Control and child achieveraent vhich reduce the strength of relationship between the two variables. As was seen in the immediately preceding section of this study, behaviors (herein called PEP and VIrespectively) which researchers have found related to children's achievement, such as mother's ar.ount of participation, and frequency of imperative statements without rationale (Grotberg, I969) were not related to control expectancy. Effects o f Participation Hypothesis 9: Mothers who participate in a program to improve their mothering skills will become more internal in I-E Control than mothers who do not. The relevant results are summarized in Table 21. Hypothesis 9 is mostly supported. V/hile there are no significant differences between the experimental and control groups in pre or post test means , changes in expectancy approach the .05 level of significance. The actual probability level is .0? for a two tailed test. This seems close enough to the conventional level to consider the results significajit . Thus, mothers who participate in the project become significantly m.oro internal in control orientation than those who do not. Tliis is consistent with research findings (Neal and Seem.an, 1961^; Seemiin and Evans, 1962; Seenan, 1963; Kiehlbauch, I968; Strickland, 1965) that persons affiliated with social action groups or involved in activities to promote iriproYc:.ent in their life circumstances are more internal than those who ere not.

PAGE 92

84 Table 21 Differences In Initial (pre) , Final (post) , and Changes in I-E Control Expectancy Scores (SRI) Between Experimental (E) and Control (C) Groups Group Pre SRI X t Post SRI X t Change SRI X t E (N=28) 10.14 0.21 8.14 1.45 -2.00 1.87* C (N=26) 9.92 9.73 -0.19 Table 22 Differences in Initial (pre) , Final (post) , and Changes in I-E Control Expectancy Scores (SRI) Between the Two Exnerimcntal Groups (E1+E2) and the Control (C) Group Group Pre SRI X t Post SRI X t Change SRI X El (N=14) vs 10.07 -0.10 8.93 1.20 -1.14 E2 (N=14) 10.21 7.36 -2.86 El vs 10.07 0.11 8.93 -0.66 "1.14 , C (N=26) 9.92 9.73 -0.19 E2 vs 10.21 0.24 7.36 -2.10--* -2.86 C 9.92 9.73 -0.19 t 1.12 -0.75 -2.24* * .07 probability level .05 probability level '9

PAGE 93

85 In the Head Start Program (Kitano, 19^7 in Grotberg, I969) designed to assess the effects of participation in a parent education program on parents, the author found that participation i
PAGE 94

86 treatment programs on imbuing the mother with the feeling that her involvement in the program was crucial to her child's cognitive grovth (Gordon, I969). Perhaps the educators who planned their own program developed or alrearly had a belief analagous to generalized internal, control: That is that participation in their program rather than external variables would result in the mothers ' improvement in child teaching skills. Perhaps they communicated this belief to the volunteer mothers they worked with. Since the two experimental groups of mothers and children did not differ significantly in success as measured by the children's final achievement scores (Gordon, 1969) it seems reasonable to assume that characteristics of the parent educators had a significant effect. Previous research in social psychology suggests that the parent educators who planned their own program would be more involved in and committed to their jobs. Studies by Lewin (1958), Pelz (1958) and their colleagues found participation in a group discussion in which a decision was made was significantly more effective in changing the social conduct of mothers than lectures, individxial instruction or group discussion without group decision. Group discussion and decision were felt to be necessary for effecting increased involvement in and coramitment to an action. It is interesting to note that the techniques used to teach the parent educators who worked with the experimental group 1 mothers included lectures, individual supervision and group discussion. The professional staff was sensitive to 'the educator;^ individual needs and problems and often incorporated their sugsoKtlons into the procedures. However, only the educators who worked with the mothers in experimental

PAGE 95

er group 2 decided as a group what the exact nature of their program would be. Consistent with the Levinian' studies then, these educators would be more actively involved in tund ccmitted to their program. This greater involvement could have rubbed off on the mothers in experimental group 2 resulting in the mothers ' developing a stronger belief in an individual's ability to influence her ovn life. There are, of course, other possible differences between the educators of the two groups which may have made them more effective in altering existing expectancies. Most obviously, Group E2 parent educators probably had more experience in working with disadvantaged families. Assessment of differences in characteristics between the parent educators related to their effectiveness in giving mothers a belief that persons can control their o-m destinies could be the subject of another study. In fact, the whole area of the etiology of control expectancies is relatively virginal research territory. This author was only able to find one study. Tolor and Jalo-^iec (1968) found that college studeiats who rated their mothers higher in authoritarian attitudes on the PARI were more external than those that did not.

PAGE 96

CHAPTER 12 DISCUSSION The major objectives of the two studies in the present research were to investigate internal ys_ external control expectancy among disadvantaged mothers, and exand.ne its relationship to maternal behaviors, to their children's cognitive growth and to assess the effectiveness of participation in a parent education program in changing expectancies in an internal direction. Presumably if this change occurred Improvement would also occur in the children's intellectual development. Existing expectancies among the predominantly black indigent population sampled were found to be significantly more externa] than samples reported in the literature which included a wider social class range. Blacks were more external than whites although in Study II the difference was insufficient for the number in the sample to reach the conventional probability level. Hot/ever, the difference was of the same magnitude and direction as those studies reported by Rotter (1966). The importance of these findings is suggested by Lef court (1966) summarizing the arguments about social class, race, and expectancy. Individuals within the Iw-zer socioeconomic class, particularly blacks, early perceive impediments in the way of goal striving. For example, segregation and discrimination convey to blacks that their own efforts will lead to no reinforcements. Thus, they come to disbelieve that the efforts of an individual pay off. 88 -

PAGE 97

69 This is not "true for all disadvantaged people, as the inclusion of the I-E scores of Project Follov Through (modaled after Head Start) educators in the present studies de^nonstrated. Although from the same background of the indigent, volunteer mothers, they were significcjitly more internal. One can speculate on the variables which might explain this. The Follov Through educators vere high school graduates, self selected by their alertness to the availability of a good Job and staff selected for skills in relating and for apparent integrity. Thus, they had already received valuable reinforcements for their efforts, their skills, and their personal characteristics. The results of the Follov Through educators are sinilar to those of Holmes as reported by Grotberg (1969). Holmes found differences among parents who referred their children themselves to Head Start as compared to those whose children vrere sought out by Head Start personnel. The self-referred group of parents vere more like middle class parents in their higher aspirations for ther.iselves and their children. Three sets of maternal behaviors which seemed to reflect mastery of or attempts to improve upon .niothcring skills related to infant cognitive gro'.rth vere selected for study: frequency of positive verbalizations, frequency of (rjresunably non-grovrth producing) negative verbalizations such as cor.iQiandc without rationale and baby talk, and frequency of behavior;; presuT.ed to show involvement with a;id commitment to a mothers' education program. No relationships were found to be reliable betvcen rcaternal behaviors shoving degree of involvement in the program a-j mca'sured by the positive attitude index (PEP) and interna], control expectancy. The results on positive end "negative"

PAGE 98

90 verbal interaction seem equivocal in the two studies. However, consideration of the age differences between the two samples may help clarify this. In the first study vhich involved infants 3 months to two years old, positive or potentially growth producing verbalizations such as explaining and describing things to the child by the mother were significantly related to internal expectancy and changes in that direction. Interestingly enough Gordon (1969) reports mothers who were more verbal on the same index had children who achieved more in the program. Still another study with much older children (pre-schoolers) by Jless and Shipaan found "openness of mother to her child's questions" and "infrequency of imperative statements to child without rationale" were the best predictors of a child's achievement. Other good predictors vhich were related to I-E Control were the parent^'' aspirations (Grotberg, 1969)In Study II vhich involved younger children (3-12 months) positive verbal interaction was not related to expectancy or e^^eotancy changes. So called negative verbalizations were related to internal exi)ectajicy of mother measured when her infant was 3 months old, but not when he was one year old. Taicen together the results from the two studies lead to the tentative conclusion that frequency of mother's verbal interaction with her child is related to internal control if the VI index is based on observation of verbalisations appropriate to the baby's age. The inconsistencies discussed above also seem to involve methodological problems . Vtoat period of observation cn mother should be related to what time at which her control expectancy' is measured? In these

PAGE 99

91 studies the observations woro tabu-lated across the entire duration of the project while expactancy vas measured at the beginning and the end. Thus, possible systematic variation of the tvo measures over time could not be assessed. Perhaps this could be overcome in a future study which used and eospai'ed several measurements of the two variables over time or one in which the VI measure was elaborated to give more information at different times. That the positive attitude (PEP) index did not predict final expectancy or expectancy changes may also suggest that the behaviors selected reflected too much of a middle class biets , or that involvement in the project was not highly valued enough to influence the mothers' perceptions or reinforcer.;ent contingencies. Related to the former, the index was made up of such items as the nujnber of delays and missed appointments a mother had which ms:y not be relevant in an indigent population where m.any reality factors influence punctuality. Related to the latter, participation did seem to be valued enough to change expectancies in an internal direction at least in Study II. Perhaps the weak relationship between the PEP index and I-E Control indicates an oversight of more potent variables. Rotter (1966) suggests in his discussion of the origin of I-E Control expectancies that these expectaiicies would be influanced by the actual reinforcement contingencies in sn individual'? life circt^.n-.stances . In the case of the disadvmitaged mothers studied these contingencies would seem to be the availability of tin^e, funds, friends, etc.. versus the ariioujit of demcaids placed upon her. Thuo , for example Freijo, Gordon ond Bilker (1968) four.c that mothers in families supported by

PAGE 100

_ 92 income other thou welfare were significantly more internal than families on welfare. This could be interpreted to mean greater availability of funds was related to internality. In addition, mothers in fa'nilies where there was less illness observed were iriore internal than families in which there was more illness (Freijo, et^ aL , 1968). Ihis could be interpreted to mean that mothers less over'./helEed by demands were more internal. There were no significant relationships between mothers' locus of control and children's achievement. Possibly these two variables are at the extremes of a chain of csiisal relationships. The intermediary links of the chain need to be investigated before the extremes lend themselves to greater predictability. Fruitful lines of research have already been mentioned. The question of whether parental behaviors influence achievement has been investigated in Kead Start programs. For example, Hess and Shipnan (in Grotfcerg, I969) found that, although a few behaviors predicted children's achievement, for the most part parent behaviors were not fovmd to predict children's achievement. On the other hand, some studies foimd parent participation predicted children's success in the program (Grotberg, I969). The whole area of disadvantaged parental characteristics and their relationship to children's achievement is a vide open research territory. Other important intermediary links would seem to be the relationship between locus of control sj^ong the chiJ.drcn themselves and their achievement; between parental control expectancies and parental achievement; between parental achievement andchildren 's 'achievement ; and between parental control expectancy and children's control expectancy.

PAGE 101

93 Ongoing rc-soarch has already suggested Eoruo interesting a!aswers to these questions. Freijo, et_ a3.. (196B) , for example, foiind that whether or not a family was on welfare v&s related to mother's control expectancy. Welfare families were more extei-nal perhaps suggesting a relationship between achievement and expectancy. Stabler and Johnson found that ccnpared to an external orientation , children who perceived events as internally controlled retained inforaaation longer (Grctberg, 1969). Thus, there seems to be a relationship between expectancy of I-E control and achievement vithin the sarne individual but interacting influences between parent and child, teacher and pupil are complex and as yet undetermined. The most inportsiit conclusion of this research was that the relatively external expectancies of disadvantaged mothers were susceptible to change significantly in an internal direction. VHiile this did not turn out to be true for the mothers in Study I there are factors allowing for giving less weight to that finding. That is, findings were contaiainated by the lack of entry scores. No appropriate control group was available and the parent educators lacked the greater experience and confidence they shovred subsequently when working with the new groups. Perhaps the most worrisoae characteristic? of externa.lly oriented persons is their failure-avoidant skill accfaisition pattern. As Lef court and Lad;^ig (l965a) point oy.t, this is particularly true of blacks in biracial, coripciitive achievement tasks. It see,r.s that much of life in Black America involves just such tasks. Competition for achiever.snt, a fact of current life, iv> particularly difficu3.,t for both blacks and whites in dinadvantagod popul atioios . If the disadvan-

PAGE 102

tagcd continue to expect failure and withdi*av fro.'n achievement situations, it vill be difficult to break the poverty cycle. If other maladjusted persons feel po^rerless to imrirove their lives, community mental health cannot be improved. Thus, it is important to find a vay of changing expectancies. If one considers the significant change in the internality of the experimental mother's I-E Control as representing success in enhancing the volunteer mother's identification with the role of mother and child teacher, then the findings of Lef court and Ladvig (I9n5a) seem relevant here. VJhen these authors were successful via •orestige manipulations in creating identification with the role of jazz musicians among black prison inmates, the inmates persisted without failure-avoidance in a biracial competitive achievement task (in which failure was experimentally prearranged) much longer than inmates who weren't given this prestige manipulation. By analogy, one would expect the present sample (Study II) of participants to persist in enhancing and encouraging skills in theiaselves and their children even among the larger, biracial population longer than the control group of non-participants. It would be interesting to test this notion. Of even greater interest would be the strength and duration of this change in expectancy. Although it remains to be investigated, the findings on changes in expectancy may generalize to other areas besides achievement and to other populations besides the lower socioecononac class. For examnle, characteristics related to I-E Control seem important in psychothera-oy . Knowledge of self, knowledge about one 'spresent life situation and

PAGE 103

95 motivation for improvement certainly are airas of psychological treatment. Research on expectancies of patients and their therapists vould seem profitably extended to include I-E control expectancies. Finally, a cominent on the possible implications of the fact that mothers worked vith by pja-eut educators vho made up their own program became the most internal. More research on characteristics of any person whose job it is to influence exiother person is needed. However, at present, the possibility that the greater selfregulation of the teacher lead to the more internal control expectancy should be of interest to educators, teacher educators, parents, therapists, or others involved in teaching and helping others .

PAGE 104

CHAPTER 13 SWB-IARY Internal ys_ external control of reinforcements describes a generalized expectancy vhich deteri/iines to what extent certain outcomes of behavior will be categorized as within the individual's personal control and understanding. A person who tends to categorize events as internally controlled tends to expect that is is an individual's own characteristics and skills which influence what reinforcements he receives, in contrast to a parson who categorizes events as externally controlled and tends to expect that chance, fate, powerful others or an incomprehensible complexity has the greater influence over the reinforcements a person receives . The two studies reported here attempted (l) to validate theoi-y and research indicating that the lower socioeconomic class individual tends to perceive events as externally conti-olled or perceives himself as alienated and powerless (to use related sociological tenns); (2) to extend research findings on the relationship between improvement of one's life situation aiid internal I-E Control to the relationship between specific maternal behaviors and I-E Control; (3) to identify individual differences within the lower class on the dimension of I-E 'Control expectsjicy; (k) to ex-^mine the direct interrelationship between mothers' locus of control expect5.ncy and children's achievem.ent and (5) to extend previous research findings on correlations between 96 "

PAGE 105

97 internal I-E Control ajid socicl &cticn and affiliation to predict that participation in a social action program by lover class mothers would change their expectancies in an internal direction over time. The mothers' I-E Control expectancy vas assessed by a questionnaire originated by Rotter and others and found to have construct validity and reliability. The questions were revised for a fourth grade reading level. The revision vas found to be a reliable one for an indigent population. Maternal behaviors were assessed by the observations made by indigenous non-professionals , the parent educators who made weekly visits to the subjects' homes. Tliese were combined into three mothers' behavior indices: positive verbal interaction, negative verbal interaction, and involvement in the project. Infant achievenient was assessed by progress on the prograia snaterir.ls developed out of Piagetan theory and research, and by tvro standardized tests of infant achievement . In Stu.dy I, the subjects were 65 indigent mothers and their infants (5^ black and 11 white) who volunteered to participate in the Parent Education Project. V/eekly observations were made by parent educators during Mother training visits from the time the infant was 3 months old until he vas ?. yeai's old for one group of 21 mothers and until he was 1 year old for another group of 21 mothers. Mothers in both experimental groups were given a modified I-E Scale (called here SRI) when their infants were 6 months old and again when they were 2 years old. Tlie 23 mothers and their infants used as controls were not visited weekly but were tested at the sane intervals as the ex-perimontal mothers . '

PAGE 106

93 The major findings in Study I were: 1. Mothers were more external than samplep not restricted to the lower cltiss . 2. Mothers were more external thzsi a sample of .lower class, indigenous non-prof easionals . 3. Black mothers were more external than white mothers. k. I-E scores did not correlate significantly with mothers' age and parity. 5. Mothers higlier in positive verbal interaction (VI+) with their child were more internal. 6. Mother's education alone or in interaction with verbal level did not affect I-E Cor.trol scores. 7. Negative verbalizations (VI-) in in.others' interaction with their children were not related to mothers' locuii of control. 8. !Ihere were suggestive but insignificant relationships between mothers' expectancy and childi-en's achievement. 9. Mothers' degree of involver.:ent in the project (PEP) was unrelated to expectancy changes. 10. Length of participation was not significantly related to expectancy changes . A second study was planned to attempt to replicate the findings and to tighten up sor-.e of the procedures, in particular the possibly contaminated entry scores on the SRI m^d the unavailability of a ncnparticipa^it control group. The basic plan and objectives of Study II were the same as the fir3t study.

PAGE 107

99 The sample consisted of three groups of indigent mothers and their infants: One group of ih mothers vho were visited veekly by the now experienced parent educators; another group of visited by nev parent educators vho vere former Head Start and nursery school workers vho designed their own teaching program; and a control group of ?.6 who vere assessed on maternal expectancy and infant achievement but not visited weekly for teaching and observation purposes. The major findings were: 1. Mothers vere more ex-ternal than saraples not restricted to the lower class . 2. Mothers were more external than a sample of indigenous non-professionals . 3. Black mothers were more external than whites of a magnitude similar to previous studies but not statistically significant for this N. h. Mothers' I-E scores correlated insignificantly with age, parity, and education. A multiple correlation of the 3 variables with I-E Control Score was also non-significant. 5. The three maternal teaching behavior indices (VI+, VI-, PEP) vere not related to initial or final control expectancy or expectancy changes, except that mothers higher in negative verbalizations (VI-) vere more interncl at the beginning of the project. 6. There vere no significant interactions between each verbal index (positive and negative) and the project involvement index in their effect on I-E Control. 7. There vere no sigraficant interrelationships between children's achievement and mothers ' control expectancy.

PAGE 108

100 8. Participatinf^ mothers changed nore in an internal direction than the control group of non-particinatinf; mothers at a .07 probability level. 9. The two groups of experinental mothers worked with weekly in the program were not significantly different in control expectancy or expectancy changes but the group worked with by educators who designed their o".m independent program were significantly Rore internal at the end of the project than the control group. The mothers teught by parent educators who used the "nrografii designed by the professional project staff were not significantly more internal and had not changed significantly more in an internal direction by the end of the project than the control group. The most important reliable finding was the change in I-E Control among the experimental mothers in Study II. Mothers participating in a mother and infant education project became significantly more internal than a control group of non-participating mothers. The lack of cross validation of this finding in Study I was considered a function of methodological flaws in Study I. When each of the two groups of experimental mothers in Study II was compared individually with the non-participant control e:roux), one experimental group was significantly more internal than the controls while the other v.'as not. The experimental groun of mothers whose parent educators designed their o\m program vrere significantly more internal than control groun mothers. This finding seemed to have implications for the effect of group discussion and group decision on commitment to social action.

PAGE 109

101 other stable findings in both studies were: (l) greater external I-E Control scores of disadvantaged mothers conpared to scores of socially stratified populations reported by Rotter (1966), (2) greater external I-E scores of disadvantaged mothers compared to scores of indigenous non-professionals from the same ghetto area and (3) nonsignificant correlations between I-E Control arid the subject's age, parity, or education. Suggestions for future research were made on the basis of methodological and theoretical shortcomings of the present studies; e.g., the degree of internal consistency of the behavior indices, the appropriateness of the behavior indices to the class of the family and age of the infant in the mother-child interaction vere considered important variables for designing future research. It vas considered necessary to examine the relationships between mothers' expectancy and achievement, children's expectancy and achievement, and the effects of maternal behaviors and expectancy on children's expectancy of control, areas presently insufficiently researched, before examining further the relationships between maternal expectancy and children's achievement. The timing of the measures of the relevant parent and child variables vas also discussed. The possibilities of using an instrument with a wider range of scores at the internal end, more reliable over time and m.oro specific to the subjects situational context were considered. Trie duration of expectancy changes was considered an important question for future research. Practical and heuristic implications of the findings were discussed, particularly the importance of changing failure avoidance 'ar.ong disad-

PAGE 110

102 vantaged Negroes. The effectiveness of indigenous nov.-protcEsinnals vho plsnned their ovm indapendcnt program in chor.gititthe relatively external expectancies of disadvantaged mothers in a more internal direction vas also emphasized. This seoraed to have important implications for onsoing social action programs such as Head Start, for present ejid futiire programs designed to educate parents and children, and for mental health programs. The possible generalizability of the findings to non-indigent populations was discussed.

PAGE 111

Appendix A Modified I-E Scale (SRi)

PAGE 112

10k Name Age Usual Job School Grade CcTipleted Tlunber of Children^ Trainer SOCIAL REACTIOI'I Iir/EITTOKY Instructions This is a questionnaire to find out the vay in ^rhich certain events in our society affect different people. Each question has tvo choices, called a or b. Please choose the one of each pair (and only one) which you more strongly believe to be the case as far as you are concerned. Be sure to select the one you actually believe to be more true rather thaii the one you think you should choose or the one you would like to be true. This is a measure of personal belief; obviously there are no right or wrong answers. For each question, after I read both remarks to you, put a circle around a if you believe remark a more strongly; put a circle around b if you believe remark b more strongly. After each question tell ins when you have made your choice. Then I will read the next one. Feel free to ask me to read any question over again. Be sure to print your name and other information asked for at the top of the page. Please do this now. In some instances you may discover that you believe both remarks or neither one. In such cases, be sure to select the on£ you more strongly believe to be the case as far as you are concerned. Also try to respond to one question at a time when making your choice; do not be influenced by your previous choices. RKMEIffiER, in each case, choose the remark which you gersonjllyjb true.

PAGE 113

105 I more Strongly Believe That : 1. a. Children get into trouble because their parents punish them too much. b. The trouble with most children (today)* is that their parents are too easy with them. 2. a. Many of the unhappy things in people's lives are partly due to bad luck. b. People's (troubles) result from the mistakes they make. 3. a. One of the (biggest) reasons xjhy we have wars is because people don't take enough interest in (government). b. There will always be wars, no matter how hard people try to prevent them. 4. a. In the long run people get the respect they deserve in this world. b. (It is the sad truth that) an individual's worth often passes (without being recognized) no matter how hard he tries. 5. a. The idea that teachers are unfair to students is ("hot air") • b. Most students don't realize (how much) their grades are influenced by (accident or chance) . 6. a. Without the right breaks one cannot be a (good and able) leader. b. (Able) people who fail to become leaders have not taken advantage of their opportunities. 7. a. No matter how hard you try, some people just don't like you. b. People who can't get others to like them don't understand how to get along with others. 8. a. (What a person is born with) plays the (biggest part) in determining (what they're like). b. It is one's experiences in life which determine what they're like. ) indicate a change from Rotter's I-E Scale

PAGE 114

io6 9. a. I have often found that what is going to happen will happen. b. (Patting trust in) fate has never turned out as veil for me as making a (pl?Ji) to take a (certain) course of action. 10. a. In the case of the vail prepared student there is (hai'dly ever) such a thing as an unfair test. b. Many tires (test) questions tend to be so (different from class vork) that studying is really a (waste of tine).. 11. a. Becoraing a success is a matter of hard work, luck has little or nothing to do with it. b. Getting a good job depends mainly on being in the right place at the right time. 12. a. The average citizen can have an influence in Government (plans)* b. This v:orld is run by the few people in power, and there is not much the little guy can do about it. 13. a. \
PAGE 115

107 18. a. Most people don't realize the (point) to vrhica their lives are controlled by (accident and chance.) h. There is really no such thinf^ as "luck"'. 19. a. One should alvays be villinp; to admit his inistakes. b. It is usually best to cover ut> one's mistakes. 20. a. It is hard to know vhether or not a person really likes you. b. How many friends you have depends upon how nice a person you are, 21. a. In the lon^ run the bad things that happen to us are (made ud for ) by the good ones . b. Most (troubles) are the result of lack of (know-how, lack of knowledge, being lazy,) or all three. 22. a. With enough effort ve can (clean up dirty government.) b. It is difficult for people to have much control over tlie things (government leaders) do in office. 23. a. Sometimes I can't understand how teachers arrive at the grades they give. b. (ihe harder) I study (the better grades) I get. 2h. a. A good leader expects people to decide for themselves what they should do. b. A good leader makes it clear to everybody what their Jobs are. 25. a. Many times I feci that I have little influence over the things that happen to mo. b. It is impossible for me to believe that chance of luck plays an important (part) in my life. 26. a. People are lonely because they don't try to be friendly. b. There's not much use in trying too hard to nlease people. If they like you, they like you.

PAGE 116

108 27. a. There is too much (importance placed) on (tearii sports) in high school. b. Team sports are an excellent way to build chai*acter. 28. a. What happens to me is Eiy cm doing. b. Sometimes I feel that I don't have enough control over the direction my life is taking. 29. a. Most of the time I can't understand why (government leaders) behave the way they do. b. In the long run the people are (at fault) for bad government on a national as well as on a local level.

PAGE 117

Appendix B Parent Educator Weekly Report and the teaching "behavior indices (index of Positive Verbal Interaction, Index of Negative Verbal Interaction, Index of Attitude Toward Parent Education Pi-oject) derived from it.

PAGE 118

liO Weekly Report Parent Educator Koma Visit ( PH7R) Observer Mother Date Time in Mnutes Date Last Vif:it_ Mother's Number Visit Number Was this a TEST visit? Yes_ If yes , vhich one? No 6~inonth , 18 month, CODE: 1. M=Mother, 2. F=Father, 3. S=Sibling (Brother end/or Sister), k, GM=Grandmother, 5. A=Aunt, 6. BS=Baby sitter, ?. G=Other, 8. Nobody 1. People in the home: A) With vhoin did you work? MPS GM A BS Other B) Is this the person you usually vork vith in this home? Yes 2. No C) Is this the person vho cares for the baby most of the tine? 1. Yes 2. No _12 month, 2k month D) Ho'.r many adults vere present at least part of the time in the room in which you vorked (besidec-? the person vith vhoa you worked)? E) Ho'.r many children vere present et least part of the time in the rooa in vhich you vrorked (besides the baby)? 2. General Information: A) Hov much activity was in the room in vhich you presented the exercises? 1. Nothing was going on besides the training 2. Other activities were going on but did not^attracF the attention of the baby 3. Other activities in the room ol^erTpu'lle d the baby's attention away from the training h. Tnere was such a great deal of activity~irr'the room that it made it difficult to present the exercises Revised 2/21/68

PAGE 119

Ill 3. Series Information: A) How did the mothering one react to your instructions? 1. Lool:ed at ycu vhile you were talking, and/or asked questions 2. Did other things vhile you vore shoving her how to do the exercise (exai.iples of other things: straightened baby's clothes, looked around the rooia, did houstvork) 3. Walked out of the room vhile you were explaining things to her it. Refused to do an exercise 5. Laughed at and/or scoffed at instructions 6. Other VThat? B) Mothering one's ability to repeat exercises: 1. Could repeat exercises the trainer had explained to her 2. Could do part of the exercise by herself but needed the trainer's help 3. Couldn't repeat exercises the trainer had explained to her C) V('hat vas the child's response to objects used? 1. Bid not look at or any vray indicate interest in the objects 2. Glanced at^ and held objnTts*Triefly but did not explore them 3. Played vith rnaterials vhen presented, but lest interest in them after a brief reaction Kept up interest in each itea presented 5. Didn't vrant to give up materials ~ D) V7hen the mothering one goes over last veek's exerciser vith her child she:. 1. Doesn't kno'.r vhat she's doing 2, Knovs vhat she's doing ' *' E) When the mothering one goes over last veek's exercises vith her child she: 1. Tries them on the child more th;m once if ii doesn't go veil the first time 2. Gets discouraged or is satisfied after doing them once even if it doesn't go veil the firit tiras 3. Does them inore than onec even if it goes very veil" the first time F) Hov many interruptions vero there during training that made the mothering one stop thrj exercise for a tine? None .1 . ? h

PAGE 120

112 ~ G) VOiat kinds of interruptions vere there? 1. Mothering one had to care for ojiother child 2. An adult wanted something 3. Ilie phone rang k. Visitors carrie 5. The bahy had to fed 6. The baby vent to sleep 7. Other ^_ 8. None H) VJhat other type? of activities vere presented by the trainer to the mothering one? 1. Songs k. Rhythm Gases 2. Imrsery Bhymes 5. Other Jr/hat? 3. Toy Making 6. None I) Check if you observed: 1. Honemade toys around the house 2. Mobiles hanging by baby's bed_ 3. Mothering one using songs or games you showed her k, Other_ 5. None of the above k. Baby's Health and Developirent : A) Did the mothering one say the baby was sick? ]. . She said the baby was sick 2. She said the baby was not sick 3. She did not soy whether the baby was sick or not If the mothering one said the baby was sick, explain: B) Did you think the baby was sick? 1. Yes 2. No_ Explain if you have a different idea than the mothering one: C) Vrnat has the baby learned to do since you saw him last in addition to the series? 1. Rolls from side to side 2. Sits alone for a short time withouFTupport 3. Crawls (creeps on hands and Imees ) ~ h. Walks alone 5. Climbs on low chair 6. Runs or jUi;ips__ ^ZT" 7 . Climbs to a strai'd on chaTi"" 8. None of the above ~ D) How many clear words do3s t])e baby u-;e? 1, Ma^.es sounds, but no c3ear worils 2. Babbles, but no clct.i•'..''ords

PAGE 121

113 3. 1 word h. 2 or 3 words 5. h cr 5 voris ; 6. 6 to 9 vords 7. 10 to I'l vorJs 8. 15 to 20 wordr, 9. More than 20 vord.5. Social Inforration: a) When the mothering one is in the room the child: 1. Watches her 2. Tries to get to her 3. Goes on as if mothering one ^•rasn't in the room._ It. Tries to get her attention 5. Other ^ B) When the mothering one coraes near the child he; 1. Frovms ' 5. Smiles 2. Watches her 6. Vocalizes 3. Laughs_ T. Reaches for her_ h. Cries 3. Ignores her 9 . Other _yria.t ?_ Verhal In formation: A) To vhat extent do people talk to the baby? 1. No one talks to the baby 2. The one vorking with the baby talks to the baby about things with which the;^ are working 3. Tl\c one xrorking with the baby talks to the baby about things besid es those with which they are working U, People other than the one working with the baby talk to the baby M_ L_L^-^il_/ Other Nobody B) Vao talks to the baby most of ' ' the tir^e (r.iore thaji half of the' > » < • ' » • tine ) I I I I 1 I I I Kow people talk to or about the baby: C ) L()ok_di r^^^ctl:^ Jrito _hi s_fac:e__ D) Talk about Mm as though "he were not__thierp E) Talk sovmds rather than words G) Tncir tone of voice sou.'ids

PAGE 122

llh K) 0) H) Their tone of voice pounds soft and loyina; I) Use the taby's nsjneTor nicknajT.e) when sresking t_o_hi)n j) Repeat sonnds the baby nakes in a questioning way ^ K) Interpret to others what the baby_ E_ays L) Listen to the baby when the baby talks M) In a few words, order or tell the baby to do or not to do thinF;s Explain ajid describe thingc vhen_ talkin.o; to the baby How many words are there in most of the sentences spoken to the baby by the mothering one? 0123l+56789_^ Give two sentences used by mothering one while talking to the baby Missed Appointments and Delays: a) \-Ja.s the beginning of training delaj'ed today? Yes " No B) If yes, why? Because the mothering one wanted to: eat^ 6. finish talking with 2. feed the baby friends or relatives 3. do housework 7. care for older k. dress the baby children 5. get dressed herself 8. let the baby"sleep~^ 9. Other 21 C) How many trips did you make before you got to see the mothering one for this visit? If you made more than one v'i s i t Im^wer the"l^llowing : D) Did the mothering one leave a message for you on any of the trips? Yes IIo E) Wien you finally got to see the mothering one: 1. She said nothing about missing her appointments 2. She gave a confusing explanation 3. She gave an vmderstandable explanation

PAGE 123

115 Index of Positi ve Verbal Interact i_on Pos. VI = tally of positive indicators 11.0 X no. of visits Items tallied if checked 1. 6 C 2. 6 F 3. 6 G k. 6 I 5. 6 J 6. 6 k 7. 6 L 8. 6 N 6 0 if sentence length 1 2 or 3 tally 1 H 5 or 6 tally 2 7 8 or 9 tally 3 Total possible tallies per PEra = 11 Limits on index 0 pos. VI 1 Index of Negative Verbal Interaction Neg. VI = tan^_qf_nej^ati^ no. of visits Items tallied negative if checked 1. 6 D 2. 6 E 3. 6 II h. 6 M Total possible tallies per PEITK = It Litnits on index 0 neg. VI 1

PAGE 124

116 Index of A ttituiie Tovard Parent Edu cators Project I'alLy) There are 5 possible tallies either positive or negative on the PirvrR. A totally positive mother vould score 5 positive tallies on every visit — with a resulting 0 negative tally. The calculated index for a totally positive mother would be +1.00. On the other hand, a totally negative mother would score 5 negative tallies, with resulting 0 positive tiillies — thus her index vrould be -1.00. Tne resulting range of the index is : -1.00 attitude index +1.00 Computation of Tallies : If item 3A is scored 1 tally 1 positive othewise If item 3A is scored 3 tally 1 negative other-wise If item 3A is scored ? and item 3B is scored 1 or 2 tally 1 positive othervrise If itein 3A is scored 2 with item 3B scored 3 but item 3F is scored>l and item 3G is not scored 6 tally 1 positive othenrise tally 1 negative This will result in 1 and only 1 tally— either positive or negative . If item 3B is scored 1 or 2 tally 1 positive otherwise If item 3B is scored 3 tally 1 negative If ite.ni 3D is. scored 2 — tally 1 positive otherwise If iteiii 3D is scored 1 tally 1 negative

PAGE 125

117 If item 3S is scored 1 or 3 otherwise If item 3E is scored 2 If item YC is scored 1 taJ-ly 1 positive otheivise If iten 7C is scored 3 or more tally 1 negative othervrise If iten TC is scored 2 aiid itera is scored yes or If item 7C is scored 2 and item 7E is scored 3 tally 1 positive otherwise tally 1 negative — tally 1 positive — tally 1 negative

PAGE 126

Appendix C Series Test

PAGE 127

119 ~ SAI^PLES OF STINULATION MATERI.ftXS SERIES I POSITION 1. Baby's position: 3yino: on his back on the floor, bed, crib or sofa. 2. Mother's position: kneeling at the side of the baby. ACTION 1. \Ihen the baby makes any sound usable in speech, have mother tickle his tummy and sr.-^ile. Ihese sounds are not crying, fussing or noisy breathing. 2. Have mother repeat the sounds her baby iiiake?.* AIM Baby responds by making similar sounds. PURPOSE To help the child cor-':!miicate vei-bally with other people. *Alyay_s encourage mother to use the baby's naine vhon zhc talks to him.

PAGE 128

120 SERIES IV POSITION Varied ACTION 1. Point to and naine parts of the baby's body, such as, arm, eye, nose, r.outh, head, neck, ear, foot, toe, and hand. 2. Continue to nar.e and explore pictures of objects in magazinoo and objects around the house vith the baby. AIM OF THE GA:-IE The baby vocally responds after you name the object, PURPOSE To give names to objects.

PAGE 129

BIBLIOGRAPHY Battle, Esther S. "The relationship of social class and ethnic group to the attitudes of internal vs. external control of reinforcement for children." Master's thesis, Ohio State University, 1962. Battle, Esther S. and Rotter, J. B. "Children's feelings of personal control as related to social class and ethnic groups." Journal of P ersonali ty, 1963, 31, 482-490. Bayley, Nancy. "Ttie trvo year old: Is this a critical age for intellectual development?" Lecture delivered at Diike University, May 5, 1966. Bayley, Nancy. Manual for _the Bayley Scales of Infant Develo pment. New York: Psychological Corp., 1969. Becker, W. C. "Consequences of different kinds of parental discipline." In: Hoffman and Hoffman (Eds.), Review of C hild Development Re search . New York: Russell Sage, 1964, 169-209. Bell, R. R. "Lower class Negro mothers' aspirations for their children. Soc ial Forces , 1965, 43, 49 3-501. Bernstein, B. "Social class and linguistic development: A theory of social learning." In: A. Halsey, J. Floud and C. Anderson (Eds.), Education , Economy and Society . New York: Glencoe Free Press, 1961, 288-314. Bialer, I. "Conceptualization of success and failure in mentally retarded and normal children." Journal of Personality , 1961, 29 , 303-320. Cohen, A. K. and Hodges, H. M. , Jr. "Characteristics of the lov/er blue collar class." Social Problems , 1963, 303-334. Coleman, J. S. Equality of Educational Opportunities . Washington: United States Printing Office, 1966. Crandall, V. J., Katkovsky, W. , and Preston, Ann. "Motivational and ability determinants of young children's intellectual achievement behaviors." Child Develop ment, 1962, 643-661. Dean, D. G. "Alienation: Its meaning and measurem.ent. " American Sociological Review , 1961, 753. 121 -

PAGE 130

122 Deever, Stanley G. "Ratings of task oriented expectancy for success as a function of internal control and field independence." Doctoral dissertation, University of Florida, 1967. Franklin, R. "Youth's expectancies about internal versus external control of reinforcement related to N variab les . " Doctoral dissertation, Purdue University, 1963. Freijo, T. D. , Gordon, I. J., and Bilker, L. "Internal-external control of expectancy of reinforcement and family characteristics in disadvantaged homes." Unpublished manuscript. University of Florida, 1968. Gordon, I. J. "Early child stimulation through parent education." Unpublished manuscript, University of Florida, 1969. Gordon, I. J., Herman, Susan and Jester, R. E. "The relationship between maternal attitude toward stimulation and child performance. Unpublished manuscript. Institute for the Development of Human Resources, University of Florida, 1968. Gore, Pearl M. , and Rotter, J. B. "A personality correlate of social action." Journal of Personality , 1963, 58-64. Gray, Susan W. and Miller, J. 0. "The training program for mothers of deprived children — study I." Unpublished progress report, George Peabody College, 1966. Griffiths, Ruth. The Abilities of Babies . London: McGraw-Hill, 1954. Grotberg, Edith H. Review of Research 1965 to 1969 . Washington, D. C. : Research and Evaluation Office Project Head Start, Office of Economic Opportunity Pamphlet 6108-13, 1969. Hunt, J. McVicker. "Toward a theory of guided learning in development." In: R. Ojem.ann and K. Pritchett (Eds.), Giving Emphasis to Guided Learning . Cleveland: Educational Research Council of Greater Cleveland, 1966, 9 8-146. Irelan, Lola M. Low Income Life S tyles . Welfare Administration publication #14, 1966. James, W. H. "Internal vs. external control of reinforcement as a basic variable in learning theory." Doctoral dissertation, Ohio State University, 1957. James, W. and Rotter, J. "Partial and 100% reinforcement under chance and skill conditions." Journal of Experimental Psychology , 1958, 55 , 39 740 3. James, W., Woodruff, A. and Werner, W. "Effect of internal and external control upon changes in smoking behavior." Journal of Consulting Psychology, 1965, 29, 184-186.

PAGE 131

123 Julian, James W. and Katz, Stuart B. "Internal vs. external control and the value of reinforcement." Jour, of Pers onali ty an d Social Psychology . 19 6 8 , 8 , 89 94. Kiehlbauch, J. B. "Selected changes over time in internal-external control expectancies in a reformatory population." Doctoral dissertation, Kansas State University, 1968. Ladwig, G. W. "Personal situational and social determinants of preference for delayed reinforcement." Unpublished doctoral dissertation, Ohio State University, 1963. Lefcourt, H. M. and Ladwig, G. W. "The effect of reference group upon Negroes' task persistence in a biracial competitive game." Jour , of Personality and Social Psychology , 1965a, _1, 668-671. Lefcourt, H. M. and LaAjig, G. W. "The American Negro: A problem in expectancies." Jour, of Person ality and Social Psychology , 1965b, 1, 377-380. Levinson, P. and Schiller, J. "The indigenous non-professional." Res earch Issues Research paper #6, Division of Research, Welfare Administration, March, 1965. Lewin, K. "Group decision and social change." In: Maccoby, Newcomb and Hartley (Eds.), Readin gs in Socia l P sycholo gy. New York: Henry Holt, 1958, 197-212. Neal, A. G. and Seeman, M. "Organizations and powerlessness : A test of the mediation hypotheses." American Sociological R eview , 1964, 29 , 216-226. Pelz, Edith B. "Some factors in Group Decision." In: Maccoby, Newcomb and Hartley (Eds.), Readings in Social Psycholog y. New York: Henry Holt, 1958, 212-219. Phares , E. J. "Expectancy changes in skill and chance situations." Jour, of Abno rmal and Social Psychology , 1957, 5U^, 339-342. Phares, E. J. "I-E Control as a determinant of amount of expected social influence." Jour, of Personality and Social Psychology , 1965, 2_, 642-646. Radin, Norma and Glasser, P. A. "The use of parental attitudes questionnaires with culturally disadvantaged families." Jour, of Marriage and the Family , 1965, 373-382. Reiff, R. "The ideological and technological implications of clinical psychology." In Comraunity Psychology ; A report of the Boston Conference on the Education of Psychologists for Community Mental Healtli, Edited by the Conference Committee, Boston, 1966, 51-64.

PAGE 132

124 Hhelngold, Harriet L. "The measurement of maternal care," Child Development , 1960, _31, 565-573. Riessman, F. "The role of the indigenous non-professional in a community mental health neighborhood service center program." Paper delivered at American Orthopsychiatry Association Meeting, 1966. Rotter, J. B. Social Learning and Clinical Psycholog y. Englewood Cliffs: Prentice-Hall, 1954. Rotter, J. B., Liverant, S. and Crowne, D. P. "The growth and extinction of expectancies in chance controlled and skill tests." Jour. of Psychology , 1961, 52 , 161-177. Rotter, J. B., and Mulry, R. C. "Internal vs. external control of reinforcement and decision time." Jour, of Personality and Social Psyphology , 1965, 2, 59 8-604. Rotter, J. B., Seeman, M. , and Liverant, S. "Internal versus external control of reinforcements: A major variable in behavior theory." In: N. F. Washbume (Ed.), Decisions , Values and Groups , Vol. 2. London: Pergamon Press, 1962, 473-516. Seeman, M. "Alienation and social learning in a reformatory." Am.er. Jour, of Sociology , 1963, 69., 270-284. Seeman, M. and Evans, J. W. "Alienation and learning in a hospital setting." Amer. Sociolo gical Review , 1962, 2^, 772-782. Sigel, I. "The attainment of concepts." In: Hoffman, M. L. and Hoffman, Lois W. (Eds.), Review;? of Child Development Research . New York: Russell Sage, 1964, 209-248. Simmons, W. L. "Personality correlates of the James-Phares Scale." Master's thesis, Ohio State University, 1959. Strickland, Bonnie R. "The prediction of social action from a dimension of internal-external control." Jour, of Social Psychology , 1965, 66, 353-358. Thonidike, E. L. and Lorge, I. The Teacher's Wordbook of 30,000 Words . New York: Teacher's College, 1944. Tolor, Alexander and Jalowiec, John E. "Body boundary, parental attitudes, and internal-external expectancy." Jour, of Consulting and Clin. Psychology , 1968, 32(2) . 206-209.

PAGE 133

125 Wenar, C. and V/enar, S. "The short term prospective model, the illusion of time and the tabula rasa child." Child Development , 1964, 34(3) 6 79 70 3. Yarrow, L. J. "Separation from parents during early childhood." In: Hoffman, M. L. and Hoffman, L. W. (Eds.), Review of Child Development Research. New York: Russell Sage, 1964, 89-136.

PAGE 134

BIOGRAPHICAL SKETCH Larry M. Bilker was bora in Brooklyn, New York, on April 16, 1940. He attended public schools in Brooklyn and upon his hi^h school graduation entered Brooklyn Colle;?e from v/hich he received a B.A. degree in psychology in June, 1960, He entered the University of Florida graduate school in the Department of Psychology in September, 1963, and vzas awarded the degree of Master of Arts in April, 1966. Subsequently, he was a research associate for the Parent Education Project at the University of Florida. He is currently a staff psycliologist at Bradley Hospital, Riverside, Rhode Island. He is also a consulting psychologist at the Providence Scb.ool Clinic for Educationally and E-aotionally Disadvantaged Children. His major professional interests are in child and community psychology and in family dynamics. He is a member of the PJiode Island Psychological Association. Mr. Bilker's professional career has also included positions as a social v/orker for the New York Department of Welfare, a playground director for the New York Department of Parks, and a teacher of biology in the New York City high schools. Mr. Bilker c-nd his wife, Marilyn, a native New Yorker and a talented professional artist, currently reside in Providence with their two children, Jaben and Eric.

PAGE 135

This dissertation was prepared under the direction of the chairman of the candidate's supervisory con-.mittce and has been approved by all members of that committee. It v/as submitted to the Dean of the Collec',e of Arts and Sciences and to the Graduate Council, and was approved as partial fulfillment of the rcquirepients for the dec^ree of Doctor of Philosophy . June, 1970 / Dean, Graduate School Supervisory Committee: Chairman 6 V c