Group Title: analysis of the content of oral language patterns of children
Title: An analysis of the content of oral language patterns of children
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00098400/00001
 Material Information
Title: An analysis of the content of oral language patterns of children
Alternate Title: Oral language patterns of children
Physical Description: ix, 99 leaves : illus. ; 28 cm.
Language: English
Creator: Pine, Shirley J. Robinson, 1930-
Publisher: University of Florida
Place of Publication: Gainesville
Gainesville
Publication Date: 1970
Copyright Date: 1970
 Subjects
Subject: Children -- Language   ( lcsh )
Speech   ( lcsh )
Speech thesis Ph. D   ( lcsh )
Dissertations, Academic -- Speech -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Bibliography: Bibliography: leaves 94-98.
General Note: Manuscript copy.
General Note: Thesis - University of Florida.
General Note: Vita.
 Record Information
Bibliographic ID: UF00098400
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000543087
oclc - 13097958
notis - ACW6792

Downloads

This item has the following downloads:

analysisofconten00pinerich ( PDF )


Full Text







AN ANALYSIS OF THE CONTENT OF ORAL

LANGUAGE PATTEIS OF CHILDREN














By
SHIRLEY J. R. PINE


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





































I .I













UNIVERSITY OF FLORIDA


3 1262 08552 3669














TABLE OF CONTENTS


LIST OF TABLES ................................. ........

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

ABSTRACT....................................... ..............

CHAPTER

I PURPOSE AND REVIEW OF THE LITERATURE.........

Purpose... ..... ....... ....... .... ....... ..

Review of the Literature.....................

II PROCEDURE...................................

Subjects...... .......... ......... ..... ........

Experimental Procedure......................

Data Analysis..............................

III RESULTS .......................... ...........

Analysis of Group I Subjects Comparing
Condition A and Condition B................

Analysis of Group II Subjects Comparing
Condition A and Condition B..................

Analysis of Group I and Group II Subjects
Under Condition A............................

Analysis of Group I and Group II Subjects
Under Condition B............................

Concept Index Score..........................


Pap e

v

vi

vii








2
1

1

2
18

19

22

27

37


38


42


45


48

48







Pagle

Leftover Words Not Processed.................. 55

Summary ..................................,... 57

IV DISCUSSION.................................. 58

Discussion of Results of Analyses as a
Function of Setting........................... 59

Discussion of Results of Analyses as a
Function of Age.............................. 68

Dictionary Revision........... ............... ... 72

Implications for Research.................... 74

V SUMMARY............ ........,................. 79

APPENDICES

A LETTERS SENT TO ADMINISTRATIVE STAFF OF
SCHOOLS AND TO PARENTS CONCERNING THE
INVESTIGATION................................. 83

B EXCERPTS FROM TWO TRANSCRIPTIONS AND EXAMPLES
OF PROCEDURES FOR TYPING, SEGMENTING AND
EDITING SPEECH SAMPLES ..........,........... 87

BIBLIOGRAPHY..... o ..... .. ... ............... 94

BIOGRAPHICAL SKETCH..................... .............. 99














LIST OF TABLES


Table Page

1. Ages and intelligence scores for
individual subjects in Group I and Group
II. The intelligence test was Peabody
Picture Vocabulary Test.................... 20

2. Harvard III Psychosociological Dictionary
content analysis categories. Examples of
entry words used by subjects in this
investigation are given within parentheses
following each category....................... 29

3. Summary of statistical analysis between
Condition A and Condition B, Group I
subjects. U 55; p 0.05.................. 39

4. Summary of statistical analysis between
Condition A and Condition B, Group II
subjects. U 55; p = 0.05.................. 43

5. Summary of statistical analysis between
Group I and Group II subjects, Condition A.
U = 55; p = 0.05............... ......... 46

6. Summary of statistical analysis between
Group I and Group II subjects, Condition B.
U 55; p 0.05........................... 49

7. Summary of Concept Index Scores for
Experimental Condition A and Experimental
Condition B for Group I and Group II subjects 51

8. Leftover list of words not entered in the
Harvard III Psychosociological Dictionary
spoken by subjects in this investigation..... 56

9. Summary table of statistically significant
differences between Condition A and Condition
B. Dependent variables are located in
column whore the mean frequency was greater.. 67













LIST OF AGURES


Figure Page

1 Mean frequency of scores on 76 dependent
variables during Condition A and Condition
B for Group I subjects.................... 40

2 Mean frequency of scores on 76 dependent
variables during Condition A and Condition
B for Group II subjects...................... 44

3 Mean frequency of scores on 76 dependent
variables for Group I and Group II subjects
during Condition A........................... 47

4 Mean frequency of scores on 76 dependent
variables for Group I and Group II subjects
during Condition B......................... 50








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

AN ANALYSIS OF THE CONTENT
OF ORAL LANGUAGE PATTERNS
OF CHILDREN

By

Shirley J. R. Pine

August, 1970


Chairman: Dr. G. Paul Moore
Co-Chairman: Dr. E. C. Hutchinson
Major Departments Speech


The language performance of a selected group of four-year-old

and seven-year-old children under two conditions was investigated

through a content analysis research design. Content is the denotative

meaning of objects, concepts and processes as symbolized by verbal

utterances. A study of the content of language contributes data

concerning the referential dimension of speech. This investigation

utilizes a content analysis category system derived from psychological

and sociological theory in a study of the referential dimension of

child language.

To investigate the content of oral language as a function of

setting and age, the following procedures were carried out. Fourteen

pairs of siblings were selected from a population of four-year-old

and seven-year-old children. The two experimental stimulus conditions

were (1) a structured clinic setting and (2) a nonstructured home

setting. Under the first condition, the stimuli presented the








subject were a series of items from a children's projective test

and five magazine illustrations. For the second stimulus condition,

each subject wore a wireless micro hone transmitter in his home.

The language performance of the subjects was recorded.

Speech samples consisting of 127 consecutive communication

units were entered on electronic data processing cards for analysis

through an automated system of content analysis. Objects, concepts

and processes represented in verbal expressions were reduced and

grouped into categories; the meaning of any word or group of words

was summarized by listing the category under which it occurred. The

dependent variables for this investigation were quantitative data

from the 76 categories of the Harvard III Psychosociological

Dictionary. Objects are grouped into three major areas social,

cultural and natural events. Behavioral dimensions or psychological

processes include emotions, perceptions, thought processes and

evaluations, and impersonal and social emotional actions. A third

set of categories is comprised of words used to modify or amplify

the meaning of nouns and verbs.

To assess how language performance changes as a function of

age and setting, each dependent variable was tested using the

Mann-Whitnev U statistic. There were content differences attribut-

able to setting.

The question of the relation between age and content was partially

resolved. Differences in the content of speech between subjects

four years of age and seven years of age were statistically


viii







significant for 19 of 76 dependent variables. Within the conditions

of this study, four-year-old children and seven-year-old children

have the same speech content in 57 categories.

For subjects, four years of age, 19 categories were found to

be statistically different, and for subjects, seven years of age,

30 categories were found to be statistically different between

settings. A more comprehensive sample of language was obtained

under a structured clinic setting with preselected stimuli. However,

this is not a representative example of typical verbal behavior.

Patterns of equivalent speech usage among varying conditions were

identified.

The Harvard III Psychosociolopical Dictionary, modified to

meet the expressions of children, can be utilized as a tool to

study language performance. The usefulness of an automated system

of content analysis in efficiently and reliably handling large

amounts of data was demonstrated in this investigation.

Content analysis is a relevant procedure for the study of the

referential dimension of speech in young children.













CHAPTER I

PURPOSE AND REVIEW OF THE LITERATURE


Purpose

An investigation into language performance is confronted

with a number of formidable issues. One of these is the

description and measurement of the referential dimension of oral

language. Research and clinical tools developed to evaluate

language performance have described phonology and grammar while

largely ignoring the referential area. Content analysis procedures

can be a valuable adjunct in an investigation of the referential

dimension. The content of an utterance is the denotative meaning

of objects, concepts and processes as symbolized by verbal utterances.

Content analysis is a technique for making inferences based on the

systematic and objective identification of these objects, concepts

and processes (Holsti, 1968). The development of electronic data

processing programs and techniques facilitates the assessment of

content analysis as a research instrument in an investigation of

the oral language of young children.

The purpose of this investigation is to employ the methodology

of content analysis in a description of the language performance of

children. Content analysis provides a systematic set of procedures








for measuring? the manifest meaning of verbal statements (Holsti,

1968). It is proposed to examine the content of speech samples as

a function of stimulus situation and as a function of age. To

accomplish this purpose, the following questions will be asked:

1. What are differences in content in the oral language of

four-year-old and seven-year-old children that can be

attributed to the age differences of children?

2. What are differences in content in the oral language of

four-year-old and seven-year-old children that can be

attributed to different stimulus situations?

3. What is the frequency distribution among the dependent

variables (categories used for content analysis) for

each age group and for each treatment?

An important consideration in this study is the generation

of hypotheses for future research. Normative data on content

performance can be obtained to use with speech pathology cases.

Thematic analyses and content analyses may differentiate pathological

language conditions among children. Content analysis may contribute

information relative to language differences between the child who

speaks standard American English and the child who speaks a

patois. With data compiled from this investigation, a new and

different perspective on the development of language will be

possible.


Review of the Literature

Literature related to this investigation will be discussed

under throe general areas (1) experimental settings for sampling








speech of young children; (2) traditional methods for evaluating

language performance; and (3) the rationale for content analysis of

speech.


Experimental Settings for Sampling Speech

Samples of the oral language of children have been collected

in a variety of settings, including the laboratory, the clinic, the

classroom and the home (McCarthy, 1954; Ervin-Tripp, 1966; Hutchinson,

1967). It has been assumed that children use different words and

different frames of reference with every change of environment.

The majority of speech samples have been collected following

a standardized protocol in a structured clinic setting using

experimenter-selected stimuli. Stimuli presented to the subject

included objects and pictures. The purpose of these studies has

been chiefly quantitative. Templin (1957) and others (Wellman,

et al., 1931; McCarthy, 1954) have assessed speech development in

children in terms of articulation errors, picture and object

naming, mean length of response, number of responses, complexity

of sentence structure and vocabulary. Normative studies were then

conducted assessing the relation of these variables to sex, age,

occupational status and socio-economic group.

Oral language records have been taken for the same or matched

subjects in two or more nonstructured settings by investigators

of early child development. Smith (1933, 1935) analyzed 305

conversational records of 220 children between the ages of 18

months and six years. She contrasted a nonstructured play condition

with a nonstructured adult-child interaction. It was observed that









in situations where the child was conversing with an adult,

rather than playing with peers, the child used longer and more

complex sentences and asked more questions.

Young (1941) obtained language samples (74 subjects, 30 to

65 months of age) under four conditions. The settings were

outdoor play, indoor activity, mealtime and responses to adult-

presented pictures. Frequency of words was greater outdoors at

play and in response to pictures than during indoor activities or

at mealtime. There was greater variety to parts of speech when

responses were elicited by a picture. Enumeration was pronounced

when pictures were presented by the adult to the child.

Williams and Mattson (1942) investigated the responses of

nursery school children (six subjects, 41 to 46 months of age)

in controlled social settings. The variables under investigation

were size of group, function of speech, parts of grammar and

length of sentence. The three settings were (1) one child and

observer, (2) two children and observer and (3) three children and

observer. Average number of words per sentence remained constant

for the three social situations. Speech was less egocentric and

more social with increase in group size from two individuals to

three individuals.

Hahn (1948) observed that first-grade children (116 subjects)

use more "nonsentences" and that mean length of sentence is shorter

when a child speaks before a peer group than with an adult. It was

observed also that length and completeness of sentence structure

depend upon the topic of the remarks.








McConnon (McCarthy, 1954) recorded two samples of conversa-

tion in six different situations for 28 nursery-school children.

The settings represented spontaneous speaking situations, a lunch

meal, morning outdoor play period, indoor free play situation,

table play, afternoon outdoor play and an outdoor play situation

at home. Twenty-five remarks were taken for each sample. This

was a methodological study where the two speech samples were tested

for internal consistency using quantitative, functional and parts

of speech analysis. Low coefficients of agreement were found.

When speech samples from the same subjects have been collected

in more than one setting, the independent variable has not always

been the setting. The independent variable in a study by Pringle

and Tanner (1958) was the home environment of nursery-school

children. The sample included 1oth spontaneous talk during play

and verbal remarks made during intelligence testing. Contrasts of

speech behavior between settings were not tested.

Pilot studies have been undertaken to develop techniques for

collecting and analyzing samples of spontaneous talking behavior in

a field environment without an investigator present (Soskin and John,

1963). Hutchinson (1967) used a transistorized wireless microphone

to gather data in a house setting on linguistic patterns from pre-

school children (20 subjects). The speech samples were described In

terms of vocabulary, grammatical forms, distribution of syllables by

length, distribution of remarks by length and functional categories.

Research on phonology and grammar acquisition has been

collected by direct observation and recording of the child in his

home environment (Miller and Ervin, 1964; Brown and Fraser, 1964).









The linguistic performance of the child has been examined with

emphasis on deductive inferences about the child's underlying

knowledge of grammatical structure and syntactic rules.

Integration of these investigations is difficult because

each study was independent as to method and purpose. Additional

inquiry into setting as an independent variable would contribute

to a clarification of the role of the stimulus condition. An

investigation of data collected from the same subjects under two

different settings and analyzed by the same procedure would

contribute direct information to the interaction between verbal

behavior and setting.


Methods for Evaluating Language Performance

Investigations into speech and language development can be

grouped into two general areas. One set of studies inquires into

the phonological and grammatical fields. Another group of investi-

gations describes the referential dimension of language chiefly

through vocabulary counts and classification of remarks by function.

Phonological and grammatical studies. The level of speech

concerned with the individual sound system of language is the

phonological. The grammar of language consists of its morphology

and syntax. Syntax refers to the grammatical rules that apply to

arrangement of morphemes within words (Ervin-Tripp, 1966).

The productive acquisition of speech sounds by children

under eight years of age has been comprehensively assessed (Wellman,

et al., 1931; McCarthy, 1954; Templin, 1957). Norms for development

of each sound are given; by eight years of age the production of

speech sounds matches the adult model.








By his fourth year, the child's passive phonological knowledge

approximates the adult model (Ervin and Miller, 1963). Michel

(1965) and Messer (1967) found that three-year-old children selected

more nonsense words using English phone patterns than nonsense

words using nonEnglish phone units.

The morphological level of grammar is more difficult to

analyze and little data are available. Early studies largely

enumerated verb forms and parts of speech (McCarthy, 1954). This

research was based on both oral language and written language.

Fisher (1934) studied the sylistic alternation of the participial

suffix -ing versus -in among children in reference to social class.

The -ed suffix is used in a generalized manner by four years of

age (Ervin-Tripp, 1966).

Studies in morphology treat the application of rules to

nonsense materials. Berko (Berko and Brown, 1960) and Anisfeld

(1967) have found that by four years of age children have mastered

certain fundamental morphological features of their language.

Skill in use of morphological rules continues to increase with age.

Syntax units are classes of functional equivalents where

members have similar privileges of occurrence in grammatical

constructions. The classes include subjects, nouns, verbs, adjec-

tives and so on. On word-association tests, children tend to

group terms of general functional equivalence together (Entwistle,

1966). In a preliminary analysis of the grammatical concept "verb,"

Porter (Berko and Brown, 1960) found that the position of the word

carried the most information as to class.









Templin (1957) has presented a comprehensive description of

grammar usage (60 subjects, three to eight years of age). Remarks

were classified into six categories in reference to grammatical

structure varying from incomplete to functionally and structurally

complete. Noun, adjectival and adverbial clauses were examined.

Parts of speech counted were noun, verb, adjective, adverb, pronoun,

conjunction, preposition, article and interjection.

Results of the Templin (1957) investigation and a group of

like investigations indicate a correlation between age and approxi-

mation of adult model in sentence structure. Changes were measured

in length of remark, use of complex sentences and use of inflected

verb forms and abstract nouns (McCarthy, 1930; Smith, 1935;

Williams, et al., 1937; Young, 1942a).

Menyuk (1963a, 1963b), Brannon (1968) and O'Donnell, et al. (1967)

and others (Hass, 1965) have described the development of grammar

using a transformational grammar model. A positive relation was

found between age and variety of transformations in children between

the ages of two years, ten months, and seven years, one month.

O'Donnell, et al. (1967) measured oral responses of kindergarten

children and oldtr elementary-school children. He found correlations

between age and word length and various phrase and clause constructions

using a transformational analysis developed for written language.

Loban (1963), in a comprehensive longitudinal study (338

subjects) of elementary-school children, examined structural

patterns within communication units. Two types of analysis were

applied to the data (I) amount of subordination; and (2) classifi-

cation by conventional grammatical usage. His analysis contrasted









a high achieving group of grade-school children against a low

achieving group rather than by age.

Investigations into phonology and grammar have been extensive

and comprehensive. When age is allowed to vary, the features

investigated are correlated with age.

Referential dimension. The reference system is that area

of language which studies the relations between verbal signs and

symbols and what they mean. Reference is a correlation between a

linguistic form and its nonlinguistic recurrences (Berko and Brown,

1960).

Studies of the referential area can be grouped under two

headings: (1) studies where reference is described in terms of

vocabulary measures; and (2) studies where language is classified

by functional or content categories.

Vocabulary measurement has raised a number of issues. There

has been no standard method to distinguish between root and deriva-

tive words, to distinguish between common and multiple usages and

to identify special terms such as proper nouns, technical, foreign,

obsolete and provincial terms (McCarthy, 1954). Different methods

of sampling have been used: lengthy samples of speech are recorded,

some tests require eliciting of words or stories, other studies

involve pointing to pictures and other methods request vocabulary

definitions (Irwin, 1960).

Studies of vocabulary definitions from vocabulary scales

have classified definitions as a function of age (Entwisle, 1966;

Sigel, 1966; Al-lssa, 1969). The purpose of procedures that

examine definitions is to discriminate among individuals taking the









test. Emphasis is on the comprehension of language rather than

the production of language. Children between five and eight years

of age use description, illustration and demonstration types of

definitions, while older children respond more often with synonyms

and explanatory responses.

Vocabulary can be scaled by parts of speech and by word

counts. When vocabularies of young children are counted by parts

of speech, the count yields a large proportion of nouns because

nouns are predominant in language. Naming is prominent in speech

of very young children, but drops to about ten per cent at the age

of 4-1/2 years (Smith, 1935). Naming accounts for six per cent

to eight per cent of speech among children between the ages of 5-1/2

and 9-1/2 years of age. It has been presumed that a young child

goes through a naming stage followed by a question stage (McCarthy,

1954).

When a lengthy sample of spontaneous speech is taken and

tabulations of word frequency are made, pronouns, articles and

linking verbs account for a high percentage of word forms (Ervin-

Tripp, 1966; Wepman and Hass, 1969; Hutchinson, 1967). Tabulations

of vocabulary do not stress content since many contentive words

having referential function are lower in per cent of usage and

appear far into the enumeration.

Investigation into the variability of vocabulary items are

rare. Uhrbrock (1936) recorded one child from six weeks of age

through her fourth year. Forty per cent of the total number of

different words occurred only once. In separate samples of 1000

words, there were an avorago of 290 different words. Only 52 new








words appeared in the 24th thousand. Hutchinson (1967) recorded

20 children between the ages of three years and five years in their

home environment. Using a 10,000 word sample, he found three-year-

old children used 1,055 different words, four-year-old children

used 1,076 different words and five-year-old children used 1,179

different words.

Jersild and Ritzman (1938) and Hutchinson (1967) have used

the ratio of number of different words to total words (tokens).

Jersild and Ritzman analyzed verbatim speech records of superior

preschool children. About one word in five was "new" to the children

between 42 and 47 months of age. Ratios of different words to

total words declined with age. Hutchinson found that approximately

one word in ten was different for his sample of preschool children.

Loban (1963) demonstrated greater diversity in vocabulary among

high-level achievers than low-level achievers by use of the type-

token ratio.

The problem of selecting a comprehensive set of categories

for classifying the content of children's speech and language is

not a simple one. A number of classification systems have been

used. They have not been concerned with word class, but have

emphasized either (1) the topic or theme of the remark, or (2)

the function of language, i.e., the expression of feelings and

desires.

One group of studies has investigated the function of speech

according to the theories of Jean Piaget. Piaget (McCarthy, 1930)

identified two types of speech: egocentric and socialized speech.

Egocentric speech, exemplified by echolalia, repetitions and monologues









is typical of children between the ages of three and five years.

Socialized speech is typical of children around seven and eight

years of age. Socialized speech includes adapted information,

criticisms, commands, requests, emotional responses, threats,

questions and answers.

McCarthy (1930), Fisher (1934) and others (McCarthy, 1954;

Fay, 1967) applied the Plaget categories to speech data. In general,

Plaget's classifications are applicable. A number of observations

pertinent to the proposed investigation have been made in these

studies. From 34 per cent to 41 per cent of the remarks of pre-

school children contains self-reference (McCarthy, 1954). As the

group becomes larger, speech becomes more social and less egocentric

(Williams and Mattson, 1942). The purposes and methods of collecting

speech samples were similar, but presentation of results is varied.

Further, the categories of egocentric and socialized speech were

not consistently defined from study to study (McCarthy, 1954; Irwin,

1960).

Other sets of functional categories have been devised to meet

the purposes of the investigator and the situation under which the

language sample is taken. Loban (1963) used seven categories

facts and unelaborated perceptions, interpretations, personal

associations, tentative statements or suppositions, generalizations,

irrelevancies, direct questions and figurative language. Other

experimenters have noted requests, threats, criticisms, commands

and questions.

Hutchinson (1967) used an 11-point scale for function of

speech which had been utilized in adult studies. These include








social manipulation, reasoning, nondirected discourse, criticism,

imparting Information, inquiry, argumentation, imaginative discourse,

incoherent verbalizing, expletive expressions and salutation. All

but eight per cent of 2700 remarkseby preschool children were

classified using these categories.

Shirley (1938) and Hahn (1948) have described by topic the

content of the verbal language of children in the age range proposed

for this study. Hahn described the spontaneous speech of first-

grade children during a "share and tell" period. She grouped

language content into six topical areas object shared, home

play, family activities, family outings, accounts of movies and

accounts of an animal. The speech was highly personal, centering

about the child and his life.

Shirley (1938) recorded verbatim records of preschool

children (336 subjects, ages two to five years) during a play

period. It was found that one-third of the common word concepts or

topics related to mother, home, father and siblings. There was a

high frequency of make-believe or fantasy concepts. Remarks

associated with the situation at hand increase with age. The 11

most frequently used word concepts seemed to carry an emotional

tone; concepts seemed to arise out of the common needs of the

children.

Metraux (1950) described the language content of children

18 to 54 months under the two categories: relation to an activity

and relation to others. Ames (1946, 1948) counted expressions

concerned with time and space in the speech of nursery-school

children. Smith (1932) observed slight but regular increases in

use of criticism between the ages of two years and five years.









Content analysis. The purpose of content analysis is to

measure the referential and semantic components of a test. It is

any research technique used to make inferences by systematically

and objectively identifying specified characteristics within a

sample of language (Stone, et al., 1966; Holsti, 1968). Content

analysis examines the manifest attributes of the text. It is

typically a thorough study of a small amount of data with frequency

counts and distribution patterns as chief sources of inference

(Pool, 1959; Holsti, 1968). Objects, concepts and processes are

reduced and formulated into classes; the meaning of any word or

group of words can be summarized by listing the class or category

under which it occurs. This lexical classification aspect of

content analysis is its most useful attribute (Stone, et al., 1966).

A content analysis program suitable for language investiga-

tion is the Harvard III Psychosociological Dictionary developed by

the Department of Social Relations, Harvard University, Cambridge,

Massachusetts (Stone, et al., 1966). A general dictionary, such as

the Harvard III Psychosociological Dictionary, provides an opportunity

to explore a wide range of variables. Each category within the

dictionary consists of a number of language signs, words, idioms

and phrases that together are a variable. The meaning of a word is

summarized by listing the category or categories under which it occurs.

Each category is a natural language unit. Natural language units

are denotative distinctions that are accepted across the language

community. Agreement among language users as to which words should

be entered into a category is the measure of denotative meaning

(Stone, et al., 1966).








The Harvard III Psychosociological Dictionary was developed as

part of a computer content analysis program called the General

Inquirer (Stone, et al., 1966; Miller, 1970). General Inquirer and

Inquirer II electronic data systems include dictionaries, data

preparation systems and data analysis programs to be used for

studies in speech, psychology, anthropology, sociology, education

and political science.


Summary

A review of the literature indicates that a number of studies

have assessed the speech and language performance of children.

However, data collection procedures and methods of analyses were

varied. It has been assumed that the situational characteristics

under which the data are gathered affect the findings concerning

vocabulary, grammatical usage and style. There has been little

effort to explore this assumption directly. An investigation of

the language performance of the same subjects to explore the

interaction between speech and setting would provide evidence in

support of or rejection of this relationship.

It has been generally assumed that language and speech

functions change with age. Normative scales have been developed

for the phonology and grammar of oral language. It has been

observed, using cross-sectional samples, that phonology and grammar

change in increments toward an adult model with age. A knowledge

of which word expressions are used at a particular age would offer

important insights into patterns of language.









The referential dimension of language has been investigated

through vocabulary counts and the identification of the function

of verbal expression. A number of issues can be raised. There is

a need to distinguish between root and derivative words and to

disregard high frequency nonlexical words. Rules for assignment

of lexical items into categories can account for root and deriva-

tive words, drop technical, obsolete and rare terms, assign words

on the basis of denotative meaning and omit high frequency function

words. A research design using two groups of subjects selected

according to age, using the same methods of data collection and

one method of analysis, would demonstrate whether the probability

Is increased that differences in speech content can be attributed

to age and to setting.

One difficulty in determining the content of the speech of

children four years and older has come from the use of hand-

tabulation methods on large amounts of data. There has been a

need for an instrument that is consistent in the tabulation of

speech data. A computer program designed to treat objects, concepts

and processes within a speech sample would provide greater rell-

ability, process large samples of speech and contribute additional

information (Alexander, 1967; Beler, et al., 1967; Borko, 1968;

Gerbner, et al., 1969).

A research design using content analysis has the purpose of

describing characteristics of a communication. A comparison of a

verbal text across time and across situation will (1) describe

trends in communication content and (2) relate known character-

istics of sources to the messages they produce (Holsti, 1968).






17


Observable speech behavior, especially the lexical-content features,

is processed and analyzed to provide measures for statistical

operations. The usefulness of thibdesign has not been explored

previously in the area of speech and language development.

Research providing for the examination of the content of

speech as a function of stimulus situation and as a function of

age using the methodology of content analysis is needed. Such an

investigation would assess the probabilities of age and setting as

independent variables. It would contribute information as to the

feasibility of a content analysis research design in the area of

speech and language.













CHAPTER II

PROCEDURE


To study the applicability of content analysis in a

description of the referential dimension of oral language, the

following procedures were carried out. Fourteen pairs of siblings

were selected from a population of normal four-year-old and seven-

year-old children. The primary data were recorded samples of

speech collected under two experimental conditions. One condition

was a clinic setting where the subject was presented experimenter-

selected stimuli. The second condition was a home setting where a

sample of spontaneous talking behavior was obtained.

Preliminary preparation of data included transcription of

speech samples, the segmentation of speech into communication units

and editing of transcriptions for electronic data processing.

For data analysis, a general content analysis dictionary,

the Harvard III Psychosociological Dictionary (Stone, et al., 1966)

was selected. The dependent variables are the 76 lexical categories

of the dictionary. The contentof the speech samples is assigned

on the basis of denotative meaning into the categories. The

Inquirer II (Arp, et al., 1968) is a computerized system which

implements the data processing.









The procedures used to investigate the content of language

in relation to the age of the subject and under two conditions are

described in this chapter.


Subjects

Fourteen pairs of siblings were selected; 14 subjects from a

population of four-year-old children and 14 subjects from a popula-

tion of seven-year-old children. The criteria for subject selection

were age of subject, sibling relationship, language background,

hearing acuity, intelligence and physical status.


Age

Group I. The age range of the 14 subjects in Group I was

3 years, 9 months to 4 years, 7 months, with a mean age of 4 years,

3 months.

Group II. The age range of the 14 subjects in Group II was

6 years, 9 months to 7 years, 7 months, with a mean age of 7 years,

2 months.

See Table 1 for the ages of the individual subjects in Group I

and Group II.


Sibling Relationship

Subjects in Group I were paired with a sibling subject in

Group II. There were eight pairs of brothers and one pair of

sisters. Five pairs had one child of each sex.


Location of Subjects

Group II subjects were attending first or second grade in the

elementary schools of Gainesville, Florida. Names of children with
9














Table 1. Ages and intelligence scores for individual subjects in
Group I and Group II. The intelligence test was Peabody
Picture Vocabulary Test.


Group I
Intelligence
Family Age Score


4-7

4-3

4-0

4-5

4-4

4-6

4-6

3-11

3-9

4-6

4-4

4-5

4-1

3-10


Family

1

2

3

4

5

6

7

8

9

10

11

12

13

14


Mean 4-3


Group II
Intelligence
Age Score

6-10 91

7-0 95

7-5 108

6-9 106

7-8 112

7-8 117

6-10 114

7-8 107

7-6 97

7-5 108

7-7 105

6-10 100

7-0 118

7-4 116


7-2 106.7


102.5









younger brothers and sisters, ages 3 years and 4 years, were

provided by the staff of the following elementary schools: Metcalfe,

Littlewood, Glen Springs, Idylwild and St. Patrick's Catholic

School.

A letter was sent to the principal of each school describing

the purpose of the investigation. Parents of potential subjects

were contacted by letter from the investigator. These letters

appear in Appendix A.


Language Background

Subjects used in this study spoke American English; no

foreign language was spoken by any subject.


Hearing

Each subject passed a hearing screening test in both ears at

20 dB at 500 Hz, 1000 Hz and 2000 Hz. The audiometer used was

Beltone 15C. It is assumed that children passing this screening

procedure have hearing within normal limits.


Intelligence

The subjects in this study were administered the Peabody

Picture Vocabulary Test, Form B (Dunn, 1959), by the investigator.

The resulting intelligence scores for the four-year-old subjects

ranged from 90 to 115, with a mean score of 102.5. The resulting

intelligence scores for the seven-year-old subjects ranged from

91 to 118, with a mean score of 106.7. The results of a t test

revealed no significant difference (t 0.4568; df 26) between

group mean scores.









See Table 1 for intelligence scores for individual subjects.


Physical Status

Children who demonstrated obvious neuro-muscular involvement

or who presented a history of these involvements were excluded.


Experimental Procedure


Data Collection Procedures

The primary data were 56 oral language samples (two from

each subject) drawn under two experimental conditions. Experimental

Condition A was a clinic setting where the child was presented

experimenter-selected stimuli. Experimental Condition B was a

home setting where a sample of spontaneous talking behavior was

obtained.

Experimental Stimulus Condition A. The subjects were tested

individually by the investigator in a clinic office provided by the

Speech Department, University of Florida. The order of presentation

of the tasks was: (1) hearing screening; (2) Peabody Picture

Vocabulary Test; and (3) picture stimuli.

Under Condition A, the stimuli presented to the child were a

series of items from the Children's Apperception Test, human form

(Bellak and Bellak, 1950), and five magazine illustrations of

children in school, home and play settings. This set of 15 pictures

was used to elicit stories and remarks.

The investigator sought to create an atmosphere where the

child felt relaxed. The subject was encouraged to talk freely

about the selected pictures. Instructions were as follows:








Do you know what a story is? We're going to tell stories
about pictures. I will show you a picture and you tell
me what is going on; what the children are doing. Be
sure you make up a story.

Remarks of encouragement werpmade by the investigator. These

were usually general in nature. Examples of remarks are:

What else? Is there more to your story? What will happen
in the future? Why do you think did ? Can you
think of anything more? Take your time. And what happened
after he And then what did he do?

Responses of the subjects were recorded magnetically on a

Wollensak 1520 tape recorder.

Experimental Stimulus Condition B. Under Experimental

Condition B, the subject wore an FM wireless microphone in his

home. Tape recordings were made of the speech of the child during

typical home activities. The parent and the child were counselled

concerning the desirability of the child following a normal

pattern of behavior. The child was free to engage in whatever

activity he chose. The minimum length of time recorded was 1-1/2

hours and the maximum time was 3 hours.


Equipment

Magnetic recordings of the oral production of the child under

Condition A were made using a Wollensak 1520 tape recorder. The

recording speed was 3-3/4 inches per second; Scotch magnetic tape

111 and 175 were used.

The transistorized FM wireless microphone which was worn by

the children under Condition B was a Model EC lavaller-type Vega-

Mike. It was placed in a cloth harness and worn about six inches

from the mouth of the child.









The signal generated by this microphone-transmitter was

received by a Vega-Mike FM receiver and fed into a Wollensak 1520

magnetic tape recorder. The recording speed was 3-3/4 inches per

second. A 0.5 mil, number 290, Scotch-brand magnetic tape was used

to provide up to three hours of recording.

The Wollensak 1520 tape recorder served as a transcription

unit also.


Data Preparation

Three procedures were involved in data preparation. These

were: (1) transcription of the speech of each subject; (2)

segmentation of the language into units; and (3) preparation of

data for electronic data processing.

Transcription of speech. A typed transcription of the speech

of the subject was made. Conventional American English spelling was

used for all words. Contractions were spelled as two words. Noises

were typed when they were an integral part of the sentence.

Utterances in which the words were incomprehensible or not understood

by the investigator were omitted from the transcriptions.

Segmentation of data. The purpose of segmentation was to

equalize sample size. A semantic and syntactical unit was selected

rather than a word or morpheme unit because it is the semantic

dimension of speech under observation. The basic unit for this

investigation is the communication unit. Each sample consists of

127 communication units.

A communication unit is a group of words which cannot be

further reduced without the loss of essential meaning (Loban, 1963;








Watts, 1948). The communication unit is comparable to remark units

used in speech pathology, such as expression units (Williams, 1937),

verbalizations and utterances (McCarthy, 1930; Templin, 1957; Winltz,

1959).

A communication unit must have essential meaning in both a

semantic and syntactic sense. A communication unit may be a

sentence, independent predication, noun phrase, verb phrase,

enumeration, answers to questions lacking repetition of the question,

or short utterances such as yes, no, ok. When it is more than one or

two words, the communication unit usually is a grammatically indepen-

dent clause with any of its modifiers between two silences. Excerpts

from transcriptions and clarification of procedures used in

segmenting the flow of oral language are placed in Appendix B.

Another unit employed was the maze. This is a cluster or

fragment of speech which does not have semantic or structural

unity (Loban, 1963). Other investigators have referred to this

type of verbal production as hesitations, repetitions, incompleted

words (Winitz, 1959), false starts or faltering (Miller, 1970),

or shifts in expression while talking. Semantic unity of a remark

is present when the maze is omitted. Mazes were excluded from the

language sample used in this investigation.

Recorder reliability. The reliability of the investigator

to identify communication units was tested. Sixteen speech samples

were randomly selected; four from each experimental condition and

each age group. Five-minute segments from the 16 different speech

samples were dubbed onto magnetic tape using two Wollensak 1520

tape recorders. Five seconds of silence were presented between









samples. Approximately 11 per cent of the communication units of

this investigation were represented in the reliability recording.

The reliability recording was presented to two graduate

students in speech pathology. Each judge listened to the tape-

recorded speech samples and indicated on a typed manuscript what

he considered to represent the communication unit. Appendix B (3)

was used to instruct the judges. The investigator assisted judges

in marking the first two speech samples as part of the training

procedure. The judges placed a slash mark at the completion of each

communication unit and drew a line through each maze.

The judgments made by the two judges were compared with those

of the investigator. Marks which differed were counted and a

percentage of agreement was obtained. The investigator identified

670 communication units which were used to determine the percentages.

The number of judgments at variance from the investigator for Judge A

was 34, and the percentage of agreement was 94.93 per cent. The

number of judgments at variance from the investigator for Judge B

was 31, and the percentage of agreement was 95.35 per cent. It was

predetermined that an agreement of 90 per cent or greater would

indicate satisfactory reliability in the identification of communi-

cation units. The percentage of agreements was greater than 90 per

cent.

Preparation of data for computer processing. The following

steps were taken to prepare the speech samples to be entered on

electronic data processing cards. A + was used for terminal

punctuation of communication units to avoid confusion with

standard English punctuation. Punctuation was omitted with the








exception of apostrophes. A J indicated a proper name. Periods

were omitted following abbreviations since the computer reads a

period as a terminal punctuation mark. Parentheses were used to

enclose words of clarification which allowed assignment of words

not in the Harvard III Psvchosociological Dictionary. Expressions

not entered in the dictionary and which appear frequently in the

text, such as references to animals, children's activities, time

and numerical quantities, are not tagged unless an appropriate

dictionary word is added in parentheses. Proper nouns referring to

days of the week and business firms are also clarified with a word

from the dictionary. Clarified words are printed on a leftover

list; they are also categorized for analysis.

Words were clarified in the following manner:

sleep (norm) numerical words (quantity) or (time)
play (norm) inside (space)
football (game) monkey (animal)
bread (food) hat (clothing)
flashlight (tool) spanking (discipline)
tugawar (game) hard (difficult)
cards (game) cookies (food)
tricycle (toy) tape (tool)
bicycle (toy) telescope (tool)
mom (mother) pajamas (clothing)
dad (father) Friday (day)
Howard Johnson (store) car (automobile)


Data Analysis

The dependent variables in this study are quantitative data

from the lexical categories of the Harvard III Psvchosociological

Dictionary. The verbal texts, entered on electronic data processing

cards, are computer sorted into 55 first-order and 21 second-order

word concept categories. The computer program provides for (1)

text and list of category assignments, (2) tag tally by raw scores,








(3) concept index score (the ratio of occurrence of tagged lexical

units to total assignments) and (4) leftover list of words not in

dictionary (Stone, et al., 1966; Miller, 1970).

The categories and concepts used in the dictionary are

derived from sociological and psychological theory. Table 2 gives

brief definitions of each category and examples of lexical items

used by subjects in this investigation which are entered under that

category. The first-order tags represent a set of discrete

denotative categories. Each entry word is tagged under only one

first-order category. The social realm, cultural realm and natural

realm typically have nouns or objects assigned. Social categories

refer to persons, social roles and activities. Cultural categories

refer to food, clothing, tools, buildings and certain cultural

patterns dealing with values and natural objects. The natural

realm includes body parts, natural objects and natural world

objects. Behavioral dimensions or psychological processes include

emotions, perceptions, thought processes and evaluations, and

impersonal and social-emotional actions. A third set of categories

lists words used to modify or amplify the meanings of nouns and

verbs. These are classified under the entry words of time, space,

quantity and quality.

Second-order categories or tags represent both denotative

and connotative levels of meaning. When a word has been assigned a

first-order tag, its meaning can then be enlarged with a second-

order tag. Second-order tags are not independent entities since the

meaning of any entry word can be tagged with more than one second-

order tag. For instance, the word "teacher" has a first-order tag









Table 2. Harvard TII PsychosocioloZical Dictionary content analysis
categories. Examples of entry words used by subjects in
this investigation are given within parentheses following
each category.




FIRST-ORDER TAGS


Social Realm

Persons

1. self all pronoun references to the personal self
(I, me, mine, myself)

2. selves all pronoun references to the inclusive self
(we, us, ours)

3. other all nonsex-specific pronouns for others
(you, yours, they, their)

Role

4. male role all roles with specific male references
(father, boy, king, himself, uncle)

5. female role all roles with specific female references
(mother, girl, aunt, herself, women)

6. neuter role all role names not connoting sex or occupation
(baby, anybody, child, friend, ghost, person, somebody)

7. ob role all roles with clear occupational reference;
open to both sexes
(lawyer, maid, teacher)

Collectivities

8. small group groups in which members are usually able to have
face-to-face interaction
(team, class, meeting, club)

9. large group collectivities (groups) usually too large for
face-to-face interaction
(people, party, company, church)








Table 2 Cont'd


Cultural Realm

Cultural Objects

10. food articles or types of food
(bread, butter, candy)

11. clothing articles or types of clothing
(purse, dress, shirt)

12. tools instrumental objects or artifacts of any kind;
broader than hand tools
(pencil, nail, rope, automobile, toy, boat, mirror, dishes,
chair)

Cultural Settings

13. social place buildings and building parts; political, social
and economic locations
(store, neighborhood, office, tent, school)

Cultural Patterns

14. ideal value --culturally defined virtues, goals, valued
conditions and activities
(kind, pretty, busy, fun, frank)

15. deviation culturally devalued goals, conditions and types
of activity
(crazy, dumb, stupid, lie)

16. action norm normative patterns of social behavior
(dinner, game, wedding, jobs)

17. message form names of communication media including art
objects, money and traditional units
(flag, movie, name, page, stories, newspaper)

18. thought form units and styles of reasoning
(why)

19. nonspecific objects abstract references to objects
(it, something, thing, stuff, what, other, that)


Natural Realm

20. body part parts of the body
(feet, hair, face)








Table 2 Cont'd

21. natural object objects not made by man
(pearl, lion, rock, worms, animal, tree, flower)

22. natural world geographicalgplaces, weather reference and
cosmic objects
(moon, rain, cave, forest, sea)


Qualifiers

23. sensory reference smell, colors, tastes
(cold, dirty, red, sound, sharp, quiet)

24. time reference references to measurement of time
(morning, now, once, again, last, before, clock)

25. quantity reference references to units and measures of
quantity
(hundred, inch, more, both, first, many, these)

26. space reference references to spatial dimensions
(about, out, outside, under, western, edge)


Psychological Processes

Emotions

27. arousal states of emotional excitement
(wake, curious, awake)

28. urge drive states
(want, need, wish, dream)

29. affection indicants of close interpersonal relationships
(love, goodby, smile, kissing)

30. pleasure states of gratification
(funny, laugh, rest, surprised)

31. distress states of despair, fear, guilt, shame, grief,
failure or indecision
(cry, difficult, scare, afraid, sick, trouble)

32. anger forms of aggressive expression
(angry, burn)

Thought

33. sense perceptions and awareness
(listen, look, smell, hear, see)








Table 2 Cont'd

34. think cognitive processes
(know, means, remember, puzzle, wonder, guess)

35. if conditional words
(might, maybe, almost, probably, else, except)

36. equal words denoting similarity
(same)

37. not words denoting negation
(cannot, none, no, nothing)

38. cause words denoting a cause-effect relationship
(because, cause)

39. defense mechanism words denoting rationalizations
(pretend)

Evaluation

40. good synonyms for good
(better, nice, right, clean)

41. bad synonyms for bad
(awful, bad)

42. ought words indicating a normal imperative
(must, would)


Behavioral Processes

Social-Emotional Actions

43. communicate processes of transmitting meaning
(talking, said, telephone, told, sings, whisper)

44. approach movement toward
(give, marry, visit, come)

45. guide assistance and positive direction
(lead, help, let, point, teach)

46. control limiting action
(shut)

47. follow submissive action
(please, promise)

48. attack destructive, hostile action
(fight, hit, bite, knock, hurt, punish)








Table 2 Cont'd

49. avoid movement away from
(lost, went, forget, hid, go)

Impersonal Actions 0

50. attempt goal-directed activity implying effort
(try, climb)

51. get obtaining, achieving action
(catch, steal, get, buy, take, win)

52. possess owning, consuming
(keep, ate, wear, lock, own)

53. expel ejecting
(scratch, threw, drop, push)

54. work task activity
(make, work, draw, wash, fix)

55. move activity involved in physical movement
(carry, fell, pull, sit, hurry, flew, swim, start)


SECOND-ORDER TAGS


Institutional Contexts

56. academic (school, read, page, class, teacher)

57. artistic (draw, picture)

58. community (neighborhood, people, visit)

59. economic (office, pearl, store, automobile, sells, spend)

60. family (married, birthday, kid, pan, mother, home)

61. legal (lawyer, promise, punish)

62. medical (sickness, hospital, ambulence)

63. military (attack, fight, fort, gun)

64. political (flag, king, castle)

65. recreational (swim, game, tent, toy, coach, club)

66. religious (Christmas, church, prayer)

67. technological (knife, make, nail, rope, fix, jobs)








Table 2 Cont'd


Status Connotations

68. higher status (aunt, king, lady, father)

69. peer status (friend, cousin)

70. lower status (baby, child, maid, son)


Psychological Themes

71. overstate (bad, easy, nothing, very, all, strong, especially)

72. understate (but, almost, probably, some)

73. sign strong (catch, fight, fort, lion, right)

74. sign weak (cry, fell, kid, sad, wrong, poor)

75. sign accept (friend, give, kind, ok, please, visit)

76. sign reect (hate, dirty, scratch, steal, secret, punish,
interrupt)









of job role and second-order tags of higher status and academic;

"swim" has a first-order tag of move and a second-order tag of

recreational. Hence, second-order tags transcend the initial basis

for classification and are used to supplement or fill out the

primary meaning of a word.

Concern is with the semantic or content aspect of language.

Therefore, certain high-frequency words, such as determiners,

linking verbs and functors, do not enter into frequency counts.

Examples of these null words are: a, an, the, is, to and will.

The Harvard ITT Psychosoclological Content Analysis Dictionary

was developed and refined through a number of stages. Words to be

Included in a category were defined as natural language units;

agreement among language users as to which word was entered into a

category was the measure of denotative meaning.

Assignment of words to categories was based on two methods.

First, words drawn from the Thorndlke-Lorge count list were put on

cards. Word assignments were made by a group of judges drawing upon

their experience as language users. They selected which meaning

was most common and which tag category should be chosen for the

best representation of that meaning. Second, a sample of words was

drawn from practical research situations. These included conversa-

tional material, personal documents, dreams, responses to survey

research questions, nondirective psychological test protocols,

literary sources, editorials, speeches and folktales. A key-word-

in-context (KWIC) strategy was employed to relate the words to

categories. Each occurrence of a word together with the context of

the key word is printed. This gives direct information as to





36


meaning and frequency of a particular word. The third revision of

the Harvard Psychosociolorical Dictionary has classified 3,564

language signs (Stone, et al., 1966).

To assess the amount of change due to age and due to treatment,

each dependent variable (the 76 tags or categories) will be tested

using the Mann-Whitney U statistic. For descriptive purposes,

frequency polygons among the dependent variables for each age

group and each treatment will be plotted,















CHAPTER III

RESULTS


This study is concerned with the content of speech samples as

a function of setting and of age. Speech data, obtained from 28

children (14 pairs of siblings), were recorded in a clinic setting

and in a home setting. These data were processed through a computer

content analysis research design using the Harvard III Psycho-

sociological Dictionary and the General Inquirer and Inquirer II

systems. The dependent variables were the categories of the

Harvard III Psvchosociological Dictionary. This chapter presents

the results obtained by statistical and descriptive analyses of the

content of the speech samples.

To assess the amount of change due to age and due to setting,

each dependent variable was tested using the Mann-Whitney U

statistic. The level of confidence chosen was 0.05. A U value of

55 or less indicates that the null hypothesis is rejected (Siegel,

1956). The null hypotheses were

1. There will be no content differences between speech samples

obtained under spontaneous play situations and samples

obtained when speech is elicited by preselected stimuli

that can be differentiated using the Harvard III Psycho-

soctological Dictionary.









2. There will be no developmental changes in use of verbal

language between four-year-old and seven-year-old children

that can be differentiated using the Harvard III Psycho-

sociological Dictionary.

For descriptive purposes, frequency polygons among the dependent

variables using mean scores for each age group and each treatment

were plotted. Concept Index Scores were computed for each dependent

variable for both experimental conditions and both age groups. Data

not processed by the dictionary routine are presented in a leftover

list of words.


Analysis of Group I Subjects Comparing
Condition A and Condition B

Table 3 presents the results of Mann-Whitney U tests comparing

the scores for the 76 dependent variables between Experimental

Condition A (clinic setting) and Experimental Condition B (home

setting) for Group I (four-year-old) subjects. A U value of 55 or

less indicates the rejection of the null hypothesis. Significant

values were obtained between experimental conditions for the following

dependent variables: self, male role, large group, tool, non-

specific object, social place, action norm, thought form, arousal,

ur sense, think, cause, good, attempt, move, peer status, lower

status and sign accept. For these four-year-old subjects, there

were nonsignificant values on 57 of the 76 dependent variables.

Figure 1 contrasts two frequency polygons of the mean scores

of Group I subjects for Experimental Condition A and Experimental

Condition B for the 76 dependent variables. A number of these

categories had low frequencies under both conditions. These









Table 3. Summary of statistical analysis between Condition A and
Condition B, Group I subjects. U 55; p 0.05.


U Signif- U Signif-
Dependent Variable Value leant r Dependent Variable Value leant


Social Realmi Persons
1. self 16.0 p<
2. selves 58.5
3. other 81.0
Social Realm: Role
4. male role 6.0 p<
5. female role 94.0
6. neuter role 67.0
7. job role 91.0
Social Realm: Collectivities
8. small group 86.5
9. large group 42.5 p<
Cultural Realm: Objects
10. food 91.5
11. clothing 94.5
12. tool 20.0 p<
Cultural Realm: Setting
13. social place 49.5 p<
Cultural Patterns
14. ideal value 72.0
15. deviation 97.5
16. action norm 32.5 p<
17. message form 81.0
18. thought form 50.5 p<
19. nonspec. obj. 29.5 p<
Natural Realm
20. body part 77.0
21. natural obj. 65.5
22. natural world 77.5
Qualifiers
23. sensory ref. 87.0
24. time ref. 73.5
25. quantity ref. 73.0
26. space ref. 73.0
Psych. Processes Emotion
27. arousal 49.0 p<
28. urge 24.0 p<
29. affection 64.0
30. pleasure 96.5
31. distress 73.5
32. anger 83.0
Psych. Processes Thought
33. sense 37.5 p<
34. think 44.0 p<
35. if 75.5
36. equal 91.0
37. not 92.5


.05



,05





.05



.05

.05



.05

.05
.05
.05


38. cause 46.5 p<,05
39. defense mech. 90.5 -
Psych. Processes: Evaluation
40. good 47.0 p<.05
41. bad 92.0 -
42. ought 94.0 -
Behavioral Processess
Social Emotional Actions
43. communicate 89.0 -
44. approach 60.0 -
45. guide 98.0 -
46. control 91.0 -
47. follow 56,0 -
48. attack 94.5 -
49. avoid 77.0 -
Behavioral Processes:
Impersonal Actions
50. attempt 47.5 p<.05
51. get 98.0 -
52. possess 87.5 -
53. expel 80.0 -
54. work 81.0 -
55. move 47.0 p<.05
Institutional Contexts
56. academic 89.5 -
57. artistic 76.5 -
58. community 64.0 -
59. economic 78.0 -
60. family 71.5 -
61. legal 77.0
62. medical 84.0
63. military 68.0 -
64. political 79.0
65. recreational 63.0
66. religious 80.0 -
67. technological 64.5 -
Status Connotations
68. higher status 76.5
69. peer status 49.0 p<.05
70. lower status 1.5 p<.05
Psych. Themes
71. overstate 92.0
72. understate 85.5
73. sign strong 79.0
74. sign weak 65.5
75. sign accept 46.5 p<.05
76. sign reject 89.0


.05
.05





.05
.05
*






40
















I-








019 -
-2





















~-'
Su _-





-0-T-r



-----4- 2

































"- --








categories provide little descriptive data for this study. The

following dependent variables had mean scores of less than 1.00

under both conditions lJob role, deviation, pleasure, anger, equal,

defense mechanism, bad control, artistic, glgal, medical, political

and religious.

Under Experimental Condition A, the categories of speech

content with a mean frequency of ten or greater include the following

dependent variables: self, other, male role, female role, non-

specific object, time reference, space reference, quantity

reference, social place, action norm, sense think, not, avoid, get,

move family, higher status, lower status, understate, sign strong,

sign weak, sign accept and sign reject. These categories provide

the most descriptive information concerning content areas. However,

having an average frequency of ten or greater does not indicate

that these categories are significant to either age or setting.

Under Experimental Condition B, those dependent variables

having a mean frequency of ten or greater includes self, other,

female role, nonspecific object, time reference, space reference,

quantity reference, not, avoid, get, sign strong, sign accept and

sign reject.

The configuration of the frequency polygon among the dependent

variables under Experimental Condition A approximates the configuration

among the dependent variables under Experimental Condition B. There

is a correspondence in the patterns of content distribution for

four-year-old children in a clinic and home setting. An inverse

pattern occurs in the following categories: male role, pleasure and

lower status.








Analysis of Group II Subjects Comparing
Condition A and Condition B

Table 4 presents the results of Mann-Whitney U tests comparing

the scores for the 76 dependent variables between Experimental

Condition A (clinic setting) and Experimental Condition B (home

setting) for Group II subjects. Significant values were obtained on

the following 30 dependent variables: self, selves, male role,

neuter role, smal group, food, clothing, natural object, non-

specific object, space reference, social place, action norm, arousal,

urge, distress, cause, good, guide, attack, avoid, attempt, get,

possess, move, family, recreational, lower status, understate, sign

weak and sign reject. For these seven-year-old subjects, 46 of the

dependent variables were nonsignificant.

Figure 2 presents frequency polygons of Group II subjects

comparing Experimental Condition A and Experimental Condition B.

The following dependent variables which had mean scores of less

than 1.00 provide few data: job role, small group, deviation,

arousal, measure, anger, equal, n, bad, follow, medical, military

and religious.

Categories with a mean frequency of ten or greater represent

content areas prevalent in the language of seven-year-old children.

There was a somewhat different group of categories with an average

of ten or greater with changes in setting. Under Experimental

Condition A, the dependent variables with a mean frequency of ten

or greater include the following categories: self, other, male role,

female role, nonspecific object, space reference, quantity reference,

social place, think, not, avoid, move, family, lower status, under-

state, sign weak sign accept and sign reject.








Table 4. Summary of statistical analysis between Condition A and
Condition B, Group II subjects. U 55; p 0.05.


U Signify.
De endent Variable Value ica t


Social Realm: Persons
1. self 48.0 p<,
2. selves 53.5 p<,
3. other 86.0
Social Realm: Role
4. male role 2.0 p<,
5. female role 60.0
6. neuter role 42.5 p<,
7. job role 83.5
Social Realm: Collectivities


8. small group 38.5
9. large group 67.0
Cultural Realms Objects
10. food 38.0
11. clothing 52.5
12. tool 77.0
Cultural Realms Setting
13. social place 33.5
Cultural Patterns
14. ideal value 91.5
15. deviation 61.0
16. action norm 3.5
17. message form 76.0
18. thought form 72,0
19. nonspec. obj. 15.0
Natural Realm
20. body part 63.0
21. natural obj. 17.5
22. natural world 64.0
Qualifiers
23. sensory ref. 88.0
24. time ref. 65.0
25. quantity ref. 97.0
26. space ref. 23.0
Ps rh Prmocsses REmtir


27. arousal
28. urge
29. affection
30. pleasure
31. distress
32. anger
Psych. Processes:
33. sense
34. think
35. if
36. equal
37. not


42.0
48.5
92.5
81.5
16.0
91.0
Thought
81.5
87.5
70.5
90.0
55.6


.05

05


p<.05


p<.05
p<.05


p<.05



p<.05


p<.05


p<. 05





p<.05

p<.05
p<.05


p<.05


U Signif-
Denendent Variable Value icant


38. cause 44.5 p<.05
39. defense mech. 91.0
Psych. Processes: Evaluation
40. good 52.0 p<.05
41. bad 86.5
42. ought 62.0
Behavioral Processest
Social Emotional Actions
43. communicate 93.0
44. approach 89.0 -
45. guide 37.5 p<.05
46. control 89.0 -
47. follow 85.0 -
48. attack 41.0 p<.05
49. avoid 41.0 p<.05
Behavioral Processess
Impersonal Actions
50. attempt 37.5 p<.05
51. get 51.5 p<.05
52. possess 28.5 p<.05
53. expel 85.0 -
54. work 81.0 -
55. move 46.5 p<.05
Institutional Contexts
56. academic 77.5
57. artistic 80.0
58. community 68.0
59. economic 75.0
60. family 33.0 p<.05
61.' legal 96.0 -
62. medical 98.0 -
63. military 57.0 -
64. political 75.0 -
65. recreational 32.5 p<.05
66. religious 83.0 -
67. technological 84.0
Status Connotations
68. higher status 60.5
69. peer status 56.5 -
70. lower status 5.0 p<.05
Psych. Themes
71. overstate 94.0 -
72. understate 44.0 p<.05
73. sign strong 76.0 -
74. sign weak 40.5 p<.05
75. sign accept 62.0 -
76. sign reject 10.5 p<.05


n


p -


- -- ---






44

c








--2

2






I
g2

I' I -3A


:1I1
--" 7





s 0I





















-l^ ^ ^ - --. -




___~~~~~~ -__ _ _ _ _ _-_-- ^-.-. .-.----..-- ---- 7E












___J

scs
---. -- --
frtTr e-1 1r1-1-TrTTTTTTFr1Til1TF
*~ 71 2cI* 7 772 77 71 7









Dependent variables having a mean frequency of ten or greater

under Experimental Condition B for Group II subjects include: self,

other, female rolo, nonspecific object, time reference, space

reference, quantity reference, not, family sign accept and sign

reject.

Similarities in content distribution between the clinic

setting and the home setting are noted in the configurations of the

frequency polygons. The pattern found under Condition A approximates

the configuration among the dependent variables under Condition B

for Group II subjects. An inverse pattern occurs in the following

three categories: male role, quantity reference and sign weak.


Analysis of Group I and Group II Subjects
Under Condition A

Table 5 presents the results of Mann-Whitney U tests comparing

the scores for the 76 variables between Group I and Group II

subject during Experimental Condition A. Significant U values of

55 or less were obtained on the following 12 variables male role,

other, time reference, space reference, action norm, distress, sense,

communicate, attack, Ret, recreational and sign reject. Sixty-four

dependent variables were nonsignificant between Group I and Group II

in a clinic setting with preselected stimuli.

Figure 3 presents the frequency polygons for Group I and

Group II subjects under Experimental Condition A. The configuration

of the frequency polygon among the mean scores of the dependent

variables for Group I approximates the configuration for subjects

in Group II. An inverse pattern occurs in the following dependent

variables larre froup, quantity reference, space reference and








Table 5. Summary of statistical analysis between Group I and Group II
subjects, Condition A. U 55; p 0.05.


U Signif-
Dependent Variable Value icnnt


Social Realm: Persons
1. self 95.5
2. selves 77.5
3. other 52.5 p<
Social Realm, Role
4. male role 40.5 p<
5. female role 71.0
6. neuter role 71.5
7. job role 75.0
Social Realm: Collectivities
8. small group 60.5
9. large group 71.5
Cultural Realm: Objects
10. food 62.0
11. clothing 57.0
12. tool 95.0
Cultural Realm! Setting
13. social place 57.5
Cultural Patterns
14. ideal value 88.5
15. deviation 97.5
16. action norm 34.5 p<
17. message form 66.0
18. thought form 97.0
19. nonspec. obj. 82.0
Natural Realm
20. body part 97.5
21. natural obj. 71.5
22. natural world 87.0
Qualifiers
23. sensory ref. 95.5
24. time ref. 8.0 p<
25. quantity ref. 57.0
26. space ref. 24.0 p<
Psych. Processes: Emotion
27. arousal 87.5
28. urge 83.5
29. affection 59.5
30. pleasure 94.0
31. distress 33.0 p<
32. anger 83.0
Psych. Processes Thought
33. sense 49.0 p<
34. think 84.0


if
equal
not


64.0
84.0
73.5


U Signif-
Dependent Voriablo Value clant


I


38. cause 93.5 -
S 39. defense mech. 98.0 -
- Psych. Processes Evaluation
.05 40. good 73.5 -
41. bad 86.5 -
.05 42. ought 77.0 -
- Behavioral Processesi
- Social Emotional Actions
- 43. communicate 52.5 p<.
44. approach 62.5 -
- 45. guide 88.0 -
S 46. control 84.0 -
47. follow 77.0 -
48. attack 36.0 p<.
S 49. avoid 60.5 -
- Behavioral Processess
Impersonal Actions
S 50. attempt 89.5 -
51. get 19.0 p<.(
-52. possess 75.0 -
53. expel 64.0 -
.05 54. work 60.5 -
- 55. move 81.5 -
Institutional Contexts
56. academic 71.0 -
57. artistic 81.0 -
58. community 75.0 -
S 59. economic 61.5 -
60. family 72.5 -
61. legal 70.0 -
S 62. medical 84.0 -
.05 63. military 83.0 -
- 64. political 72.5 -
.05 65. recreational 48.5 p<.(
66. religious 81.0 -
67. technological 91.5 -
Status Connotations
-68. higher status 56.0 -
- 69. peer status 69.0 -
.05 70. lower status 92.5
Psych. Themes
71. overstate 70.5 -
.05 72. understate 56.0 -
- 73. sign strong 60.5 -
74. sign weak 91.5 -
75. sign accept 59.5 -
76. sign reject 18.0 p<.


05


-- -- ----


05







05










35







05






47

















-- --------------~-







~ -








IR
Lj --
- .. --











I--
~-N
-- ------ -- -- --


































n I-rT--r--M,7-T^M ^-rT -i--TT-T-r[ 7n. II 1
__ __ __ -
I""









overstate. The configuration of the frequency polygons indicates

that there is a regularity to the language performance of the

subjects when a speech sample is obtained in the clinic with

preselected stimuli.


Analysis of Group I and Group II Subjects
Under Condition B

Table 6 presents the results of the Mann-Whitney U test comparing

the scores for the 76 dependent variables between Group I and Group II

under Experimental Condition B. Significant differences were

obtained on the following seven variables: other, natural object,

time reference, distress, guide, get and economic. Sixty-nine

dependent variables were nonsignificant between Group I and Group II

subjects when speech samples were taken in the home environment.

Figure 4 presents the frequency polygons for Group I and

Group II subjects under Condition B. The configuration of the

frequency polygon among the mean scores of the dependent variables

for Group I approximates the profile for subjects in Group II. An

inverse difference occurs in one dependent variables food. There

is a homogeneity to the pattern of content distribution among the

two groups of subjects in a home setting.


Concept Index Score

Table 7 presents the Concept Index Scores for each group of

subjects and each experimental condition. This score provides for

a comparison of speech samples of varying word lengths. The score

is a ratio score between the frequency assigned to a given category

and the total number of concepts assigned in the verbal sample under








Table 6. Summary of statistical analysis between Group I and Group II
subjects, Condition B. U 55; p 0.05.


U Signif-
Denendent Variable Value icant


Social Realm: Persons
1. self 75.5
2. selves 67.0
3. other 50.5 p<,i
Social Realms Role
4. male role 65.5
5. female role 91.5 -
6. neuter role 94.5
7. job role 83.5
Social Realm: Collectivities
8. small group 84.5
9. large group 78.5
Cultural Realm: Obiects
10. food 56.5
11. clothing 92,0
12. tool 63.5
Cultural Realm: Setting
13. social place 93.5
Cultural Patterns
14. ideal value 66.0
15. deviation 63.0
16. action norm 91.5
17. message form 740 -
18. thought form 64.0
19. nonspec. obj. 71.0
Natural Realm
20. body part 93.0
21. natural obj. 50.0 p<.(
22. natural world 75.0 -
Qualifiers
23. sensory ref. 87.5 -
24. time ref. 43.5 p<,.
25. quantity ref. 90.5
26. space ref. 62.5
Pqvch. Procsses:c Emotion


27. arousal
28. urge
29. affection
30. pleasure
31. distress
32. anger
Psych. Processest
33. sense
34. think
35. if
36. equal
37. not


98.0
65.5
73.5
75.5
47.5
91.5
Thought
89.5
63.0
58.0
83.5
85.0


p<.05


U Signif-
Dependent Variable Value ic t


38. cause 98.0 -
39. defense mech. 97.0 -
Psych. Processes: Evaluation
40. good 89.5 -
41. bad 92.0 -
42. ought 84.0 -
Behavioral Processes:
Social Emotional Actions
43. communicate 56.0 -
44. approach 92.5 -
45. guide 51.5 p<.05
46. control 83.0 -
47. follow 65.0 -
48. attack 78.0 -
49. avoid 59.5
Behavioral Processes:
Impersonal Actions
50. attempt 89.0
51. get 53.5 p<.05
52. possess 59.0 -
53. expel 96.5 -
54. work 70.5 -
55. move 97.5 -
Institutional Conte,,ts
56. academic 59,5 -
57. artistic 92.5 -
58. community 62.0 -
59. economic 40.5 p<.05
60. family 68.0 -
61. legal 93.0 -
62. medical 98.0
63. military 97.0
64. political 82.0
65. recreational 97.5
66. religious 890 -
67. technological 70.0
Status Connotations
68. higher status 68.5
69. power status 75.0
70. lower status 80.0
Psych. Themes
71. overstate 72.5 -
72. understate 910 -
73. sign strong 97.0
74. sign weak 73.5
75. sign accept 81,0
76. sign reject 79.5


05


I T, -- r ,


---- ~--~'-


- . . .


--













IC











ol-a.* -<
ji V























I :1: -;:
--2
r2





















-C
d2





0I -,
-






















A- -- = =--



































I ,'T- I I .2 r r_ ; i E ^ Ts --R"p "I--r'T- T i- T- -T- --' - '-
eII

C -




S S

-- a




:~ ae


'-. a









CF

II C









/ PP
&



_-/ 2-






f-J- 1 r-~771;--~l-l Tmj-f-fTflr-rrrrnryTrfl-f Tfl-T-l-TF-rFFF
2 rg p ~ C ; ~ 2 % r r








Table 7. Summary of Concept Index Scores for Experimental Condition A
and Experimental Condition B for Group I and Group II subjects.



M Group I Group II
Condi-' Condi- Condi- Condi-
Dependent Variables tion A tion B tion A tion B


I. Persons
self
selves
others
II. Roles
male role
female role
neuter role
job role
III. Collectivities
small group
large group
IV. Cultural Objects
food
clothing
tools
V. Cultural Settings
social place
VI. Cultural Patterns
nonspecific object
ideal value
deviation
action norm
message form
thought form
VII. Natural Realm
body part
natural object
natural world
VIII. Qualifiers
sensory reference
time reference
space reference
quantity reference
IX. Emotions
arousal
urge
affection
pleasure
distress
anger


5.02 9.10
0.41 0.89
3.46 3.95


7.00
2.83
1.38
0.30

0.12
0.64

0.52
0.29
1.99

2.32

6.84
0.29
0.08
1.45
0.59
0.20

0.76
1.39
0.65

0.94
2.11
3.70
4.76

0.15
0.27
0.62
0.14
0.36
0.09


2.02
2.52
0.83
0.02

0.05
0.08

0.92
0.34
0.83

1.44

11.04
0.09
0.09
0.70
0.41
0.61

0.53
1.23
0.47

1.30
3.02
3.24
6.74

0.00
1.59
0.87
0.16
0.75
0.02


3.11 7.42
0.52 1.32
3.69 5.23


8.30
2.59
1.32
0.08

0.26
0.19

0.69
0.41
1.47

2.50

5.65
0.24
0.04
2.23
0.60
0.09

0.50
1.35
0.39

0.60
5.04
4.94
3.95

0.18
0.28
0.80
0.11
0.74
0.01


1.48
2.30
0.84
0.06

0.01
0.12

0.19
0.27
1.54

1.49

11.23
0.27
0.17
0.55
0.58
0.23

0.36
0.51
0.32

1.23
4.46
3.65
5.17

0.00
0.77
1.06
0.06
0.17
0.00








Table 7 Cont'd



Group I Group II
Condi- Condi- Condi- Condi-
Dependent Variables tion A tion B tion A tion B

X. Thought
sense 1.18 1.98 1.62 1.64
think 3.08 1.31 1.71 2.00
if 0.80 0.55 1.50 0.84
equal 0.00 0.02 0.02 0.06
not 4.43 4.85 2.59 4.62
cause 0.92 0.31 0.64 0.33
defense mechanism 0.01 0.05 0.01 0.03
XI. Evaluation
good 0.53 1.02 0.38 0.91
bad 0.08 0.12 0.12 0.07
ought 0.39 0.37 0.42 0.27
XII. Social Emotional Actions
communicate 1.12 0.94 1.50 2.01
approach 0.44 0.84 0.57 0.91
guide 0.39 0.47 0.24 0.77
control 0.00 0.02 0.02 0.09
follow 0.00 0.12 0.05 0.03
attack 0.35 0.44 0.70 0.32
avoid 2.33 1.81 0.02 2.23
XIII. Impersonal Actions
attempt 0.42 0.16 0.05 0.14
get 1.73 1.75 2.83 2.88
possess 1.00 0.95 0.91 0.49
expel 0.23 0.22 0.08 0.20
work 0.50 0.83 0.57 1.17
move 3.14 1.95 2.07 1.81
XIV. Institutional Contexts
academic 0.35 0.52 0.45 1.01
artistic 0.32 0.30 0.15 0.07
community 0.68 0.30 0.26 0.71
economic 0.30 0.20 0.45 0.78
family 3.15 2.64 3.09 1.83
legal 0.00 0.14 0.07 0.06
medical 0.04 0.02 0.01 0.01
military 0.15 0.03 0.16 0.06
political 0.12 0.17 0.08 0.03
recreational 1.14 0.84 1.36 0.81
religious 0.14 0.13 0.06 0.03
technological 0.94 0.73 0.73 0.93
XV. Status Connotations
higher status 1.41 2.16 1.63 1.54
peer status 0.36 0.06 0.33 0,23
lower status 2.58 0.42 1.89 0.38








Tablo 7 Cont'd



Group I Group II
Cgndi- Condl- Condi- Condl-
Dependent Variables t'on A tlon B tion A tion B

XVI. Psychological Themes
overstate 1.56 1.38 1.45 1.87
understate 2.23 1.88 2.85 1.94
sign strong 1.71 2.09 1.77 2.06
sign weak 2.33 1.88 1.81 1.36
sign accept 2.20 3.47 2.13 3.49
sign reject 3.82 3.72 4.46 3.86








investigation. The quotient is multiplied by 100 to obtain the

Concept Index Score (Stone, et al., 1966).

The Concept Index Score provides information relative to the

model used in the development of the Harvard III Psvchosociological

Dictionary. The dictionary is organized to measure language

texts for 16 concept clusters on 76 dependent variables. The

dictionary identifies verbal lexical events that can be assigned a

content category. The verbal text in this investigation has

measurable examples of the concepts and variables included in the

Harvard III Psvchosoctological Dictionary.

Concept Index Scores tend to supplement and support statistical

analyses. The following observations are pertinent:

1. There are nine dependent variables which have higher

Concept Index Scores for Group I and Group II subjects

in a structured setting than in a home setting. These

include: male role, neuter role, social place, action

norm, move, recreational, family, lower status and

understate.

2. Five dependent variables have higher Concept Index Scores

for both groups of subjects in a home setting than in a

structured setting. These includes self, nonspecific

obLec, quantity reference, sign strong and sign accept.

3. Four-year-old subjects during a clinic interview had

higher Concept Index Scores for three categories: large

group, think and sign weak. In a home setting, they had

higher Concept Index Scores for urge and higher status.








4. Seven-year-old subjects during Condition A and Condition B

had higher Concept Index Scores for the categories time

and get. In the home setting, they had higher Concept

Index Scores for others, work and academic.

The Concept Index Scores increase the descriptive data

available to this investigation. The score serves as a relative

index of frequency in any one category.


Leftover Words Not Processed

Data, not handled by the dictionary routine of the Inquirer II

program, were expressions which were not entered in the Harvard III

Psychosociological Dictionary. These data were not assigned into

any of the 76 dependent variables. The major groupsof words which

were not assigned tags were proper names, brand names and titles

of books or television programs. In addition, the content words

listed in Table 8 were not in the dictionary and were used by the

subjects of this investigation. Words on the list appeared with a

frequency of three times or greater in the verbal output texts under

examination. Words with an asterisk (*) appeared with a frequency

of ten or more times in the speech of the subjects of this study.

Words with a plus sign (+) were edited by the investigator so that

the category was assigned; data for these words appear in the

statistical and descriptive analyses of the speech samples.

Five general content areas can be identified from this left-

over lists (1) clothing, (2) food, (3) animals, (4) toys and (5)

parents. In addition, a number of high-frequency verb items was

not included in the Harvard III Psychosociological Dictionary. These

include: had, has, have, hope, hate, play, sleep, use and watch.









Table 8. Leftover list of words not entered in the Harvard III Psycno-
sociological Dictionary spoken by subjects in this investiga-
tion.




Group I Group II
Condition A Condition B Condition A Condition B


+alligator
+asleep
+balloon
bathtub
bet
+bicycle
bike
+car
crib
crawl
+*daddy
diapers
everything
+football
grab
+grass
*had
*has
*have
*here
+inside
lamp
+mama
+*mommy
monster
+pajamas
+*play
+rooster
+*sleep
+sleepy
sneak
+spank
+three
+tricycle
truck
through
+*two
used
watch


+bee
butt
caps
+carrot
+cookies
+*daddy
dirt
done
+dummy
+eight
+*four
+five
*had
*has
*have
hope
*here
+leopard
+jacket
made
+*mama
+*mommy
part
+peanut
pill
+*play
quit
rake
+seven
+six
spill
spook
stove
+ton
+*three
+*two
*watch
undone


+asleep
+*balloon
+baseball
+bike
+car
cane
crib
crow
couch
+*daddy
+*dad
dirt
+frog
+football
four
grandaddy
grandmother
*had
*has
+hat ,
*have
*here
hump
+Inside
mad
made
+mama
measles
+*mom
+mommy
naughty
o'clock
+owl
+*play
ran
+*rooster
sent
+*sleep
+slept
smoke


+*spank
stew
+tricycle
+trike
+three
through
+*tugawar
+*two
TV
undressed
vampire
*watch
whatever


antenna
+bike
+*car
care
disqualified
+*eight
+eighteen
envelope
everything
everytime
+*four
+*five
*has
*have
*here
hi
hook
hope
made
+*mama
mess
+*mom
+*mommy
+Monday
+nine
+nineteen
+*play
+*seven
+*six
star
+tape
+thirty
+*three
through
tough
+twelve
+twenty
+*two
use
watch


Symbols: + edited
than ten.


by investigator; frequency of use greater


~ ---








Summary

The descriptive and statistical analyses revealed that a

number of dependent variables are significant to each experimental

condition and to each group of subjects. The patterns of distribu-

tion among the dependent variables as described by frequency polygons

vary in only a few categories between ages and between treatment

conditions. Concept Index Scores indicate that the speech content

of the subjects of this investigation is distributed among the 76

dependent variables. A list of expressions not processed by the

computer program was presented. The results of this investigation

offer data for a discussion of the function of age and setting in

the content of speech.















CHAPTER IV

DISCUSSION


The issue Involved in this investigation was whether a content

analysis procedure for the description of the referential dimension

of oral language could be utilized to analyze differences in speech

content between age groups and between two experimental conditions.

Two questions were asked:

1. What are the differences in content in the oral language

of four-year-old and seven year-old children that can be

attributed to different stimulus situations?

2. What are the differences in content in the oral language

of four-year-old and seven-year-old children that can be

attributed to the age differences of the subjects?

To answer these questions, the descriptive and statistical

results obtained in this investigation will be discussed with respect

to the following areas: (1) results of the analyses of the content

of oral language as a function of setting; (2) results of the

analyses of the content of oral language as a function of age; (3)

procedures for revising the Harvard III Psychosociological Dictionary

to be used more efficiently in child language studies; and (4)

proposals for research generated by this investigation.








Discussion of Results of Analyses
as a Function of Setting

It has been assumed that the referential dimension of speech

varies with change of setting. Thearesults of this investigation

contribute support for this view. There are, however, a number of

content categories that indicate no differences attributable to

setting. This implies that some areas of language performance are

not embedded in the situation or setting. Within the conditions of

this investigation, generalizations may be made between settings

for 35 content categories.

Statistically significant differences were found between

Experimental Condition A, the structured clinic setting where

responses were made to preselected stimuli, and Experimental

Condition B, the home environment where speech was spontaneous in

response to normal family activity. Eleven variables were significant

in distinguishing between the home setting and the clinic setting

for both groups of subjects. In addition, eight categories were

significant in distinguishing between settings for the four-year-old

subjects and 21 categories were significant in distinguishing between

settings for seven-year-old subjects. To clarify the discussion of

these results, three subdivisions and a table will be employed.

The subdivisions are: differences between settings for Group I and

Group II subjects, differences between settings for Group I subjects

and differences between settings for Group II subjects.


Differences Between Settings for Group I and Group II Subjects

Eleven dependent variables were significant in distinguishing

between the clinic setting and the homo setting for Group I and

Group II subjects.









Categories with higher mean frequencies in structured

setting. Variables which had a greater magnitude under Condition A

are listed below:

Categories Expressions

male role father, himself, king, boy
social place home, tent, school, store
action norm game, lunch, play, breakfast, job
arousal awake
cause because
attempt try, pull, climb
move sit, fell, rush, jump, start
lower status baby, child, boy

In a setting where a subject relates a story in response

to a picture, the likelihood of these categories occurring is

apparent. Examples from the subjects in this study demonstrate

this conclusion.

That is a daddy sitting down smoking his pipe.
It fell with her.
And eat their breakfast.
Then they start the camp fire.
She is telling the boy what to do.
But he sings like a lady.
I think he is going to climb a big tree.
I bet he might come home with his knee hurting.
They are inside that little cave or tent.
Because the wolf was trying to eat her up.

It seems evident that the use of pictures to stimulate

stories resulted in significantly different speech content

for the categories listed above. The category of male role

may reflect the fact that a majority of the subjects were male. The

other items are used in developing a story; a place to go, an action

or behavior takes place and a child is a participant. Those








categories related to the clinic setting seem to be correlated to

the activity of telling stories.

Categories with higher mean frequencies in the home setting. -

Three dependent variables, nonspecific object, urge and good, had a

higher frequency during Condition B for both groups of subjects.

Words from the category nonspecific object had a high frequency

of use under both conditions, but a higher frequency of use in the

home environment. Examples of expressions from this tag include:

it, that, anything, thing, other and stuff. Objects and persons

located near the child appear to be referred to in nonspecific terms.

This suggests that much of the communication between persons is

based on mutual knowledge or recognition of the topic, issue or

object being discussed.

Expressions of desire and need, such as "I want" or "I wish,"

occurred significantly more often in the home. This seems logical

since there are caretaking responsibilities assumed by adults in the

home. Children, no doubt, use expressions of need with these

adults. The subjects also interact with other children and may

express themselves negatively and say "I don't want to do it."

The significant results for the category urge probably are a

demonstration of this verbal behavior.

Personal evaluations are stated more often under Condition B.

Synonyms for good, such as nice and right, are expressed in the

home environment with greater frequency than in a clinic environment.

The subjects in this study were in a different interpersonal

relationship in the home setting. They are not responding to

preselected stimuli. The use of expressions such as good, nice and







right suggest they were conversing on a more equal status in the

home setting.

Concept Index Scores indicated that self, nonspecific object,

quantity reference, sign strong and sign accept had higher scores

under Condition B. Pronoun references to oneself are more common

in the home. The high scores on the quantity reference can be

attributed to the expressions entered in that category. Not only

obvious numerical units are included, but words such as big, great,

tiny, all, some, most, many, even and much are tagged into quantity

reference. Children use these qualifiers in many circumstances.

Concept Index Scores for sign strong and sign accept categories

are indications of the nouns and verbs children use during play.

These are connotative tags and, under sign strong, are words such

as brave, big, catch, right and help. For sign acceptsubjects

used expressions such as party, ask, follow, friend, gift and

ok. These words and analogous words are utilized in the home

at a greater ratio than in the clinic.

Content in these categories can be attributed to the home

setting. Speech with family and friends is embedded in the situation

and likely to be less varied and more repetitive. Categories

significant to Condition A are related to the activity of telling

stories. These 11 dependent variables which are statistically

significant for both Group I and Group II subjects have the most

explanatory power in describing the content of language as a function

of setting.








Differences Between Settings for Group I Subjects

In addition to the 11 dependent variables stated above, eight

dependent variables were significant in differentiating between

speech samples obtained in a clinic setting and those obtained in

the home setting for four-year-old subjects.

Categories with higher mean frequencies in the clinic setting.

- The dependent variables, large group (church, party), tool (nail,

rope, toy), think (wonder and know) and peer status (cousin, friend),

had greater frequency under Experimental Condition A.

Four-year-old children told stories that included content

areas relating to groups of people and that involved artifacts and

objects such as vehicles, machines and utensils. They said

"I don't know" in answer to questions from the adult; this contributes

to the significant difference found for the category think. Their

stories included references to cousins, sisters, brothers and

friends.

Categories with higher mean frequencies in the home setting. -

The dependent variables, self (I, my, me), thought form (why), sense

(look, see) and sign accept (give, ok, yes), were observed with

greater frequency under Experimental Condition B.

Expressions from these categories are among the most frequent

vocabulary words for children three and four years of age.

Hutchinson (1967) listed the same words among the 50 most common

vocabulary items for his subjects who were also recorded in the

home environment.

For subjects of Group I, 19 variables differentiated between

experimental conditions. Homogeneity of content between the home









and clinic is indicated by the 57 nonsignificant variables. Within

the conditions of this study, children, four years of age, employ

the same speech content under both experimental conditions for 57

categories.


Differences Between Sottinps for Group II Subjects

Group II subjects had significant differences for 30 dependent

variables. In addition to the 11 variables discussed under the first

unit of this section, 21 other variables were statistically signifi-

cant in differentiating the oral language sample obtained in the

clinic from the oral language sample recorded in the home setting.

Categories with higher mean frequencies in clinic setting. -

The 18 dependent variables discriminating Condition A from Condition

B for seven-year-old subjects are listed below with examples of

entry data. Higher mean scores were found under the clinic

condition for these variables.

Categories Expressions

self I, me, my, mine
selves we, us, ours, our
neuter role child, friend, ghost, person
small group class, picnic, club
food milk, honey, coffee
clothing dress, shirt, clothes
natural object cat, chicken, branches, water
space reference outside, about, across, edge
distress cry, scare, trouble, difficult
attack cut, shot, punish, kick, fight
avoid lost, went, hid, forget
get catch, take, gather, finish
possess keep, ate, own, swallow
family married, parents, cousin
recreational game, tent, toy, party
understate but, probably
sign weak fell, kid, wrong, poor
sign reject hate, dirty, scratch, kick








The dependent variable self has a greater frequency in the home

environment for the seven-year-old subjects, whereas this tag was

found significant for four-year-old subjects in the clinic environment.

Subjects in Group II tended to tell stories about persons or friends

rather than about themselves. The number of categories which are

different statistically suggests that presentation of pictures

provides an opportunity for variation in response that is not offered

in a home setting. This tendency is more pronounced for these seven-

year-old subjects. Language skills provided through schooling may

account for this trend.

Category with higher mean frequency In home setting. One

dependent variable, guide, had greater frequency of response under

Condition B. Entry words tagged guide include lead, help and let.

These Items deal with assistance and positive direction. This

suggests that seven-year-old children like to boss and give directions

interacting in a different manner in the home environment than In a

structured clinic setting.

The older subjects had greater Concept Index Scores for others,

work and academic. Use of pronouns referring to others increased with

age according to Young (1942). This is suggested by the score from

the category others in this investigation. A large collection of

expressions employed by children is placed in the category work,

Within the limits of this study, this cluster of words does not

discriminate statistically between conditions under which the sample

was obtained. Expressions entered into the category of work include:

found, wash, fix, sell, draw, make, spend and build. Academic terms

used by the older subjects include: book, teacher, grade, school,









class and read. The Concept Index Score suggests that seven-year-

old children who are in school talk about it somewhat more in the

home environment.

Table 9 summarizes content differences as a function of

setting for Group I and Group II subjects. Greater mean frequencies

according to experimental condition and age are indicated.

Smith (1935) observed that there was superior grammatical and

vocabulary usage by children in an adult-child interaction setting.

This investigation supports this view. The more inclusive description

of the content of speech for subjects in this study was obtained in

a formal adult-child situation. Presentation of pictorial stimuli

provided a more challenging experimental condition than a nonstructured

home setting. There were more differences for seven-year-old subjects

than for four-year-old subjects. The effect of learning in school

may explain the greater number of significant differences for older

subjects.

A speech sample obtained in a laboratory setting is not

representative of typical verbal behavior. It is not a sample of

how speech is employed in meeting the demands of day-to-day communica-

tion. If a representative picture of speech behavior is to be

obtained, comprehensive research into the area of speech and language

development should include data obtained under more than one setting.

Consideration should be given to a definition of the situation within

the setting. Setting is the locale of speech. Situation would define

the nonverbal behavior occurring, such as family meal, interaction

with mother, play with other children or solitary play. Patterns of

equivalent speech usage among the varying situations could be














Table 9. Summary table of statistially significant differences
between Condition A and Condition B. Dependent variables
are located in column where the mean frequency was greater.




Mean frequency
greater for Mean frequency Mean frequency
Group I and greater for greater for
Group II subjects Group I subjects Group II subjects


Condition A


male role
social place
action norm
arousal
cause
attempt
move
lower status


large group
tool
think
peer status


self
selves
neuter role
small group
food
clothing
natural object
space reference
distress
attack
avoid
get
possess
family
recreational
understate
sign weak
sign reject


nonspecific objects
urge
good


Condition B

self
thought form
sense
sign accept


guide








identified and interactions that are significant between situations

can be found.

These data demonstrate that characteristics of the content of

speech as a function of setting can be delineated using Harvard III

Psychosociological Dictionary. The dictionary measures language

performance by identifying categories unique to a setting. It

appears to be a useful tool for describing speech behavior in

different settings.


Discussion of Results of Analyses
as a Function of Age

This study inquired into content differences in speech samples

attributable to age. Differences in the content of speech as a

function of age were not found in the majority of the dependent

variables. It appears that children four years of age and seven

years of age converse with much the same speech content. The

following discussion is based on the statistical and descriptive

analyses obtained in this investigation. The failure to reject the

null hypotheses in 57 of 76 categories may be attributed to a number

of factors which will also be discussed.


Differences Between Age Groups During Conditions A and B

Under both Experimental Condition A and Experimental Condition

B, the older subjects used two dependent variables with greater

frequency than did the younger subjects. The categories were

pronouns referring to others, such as you, yours and they; and

content words from the variable get, such as catch, steal, win, buy

and take.









A conclusion drawn from those data sl that the use of pronouns

referring to others and content words from the dependent variable get

is indicative of maturation of speech skills as demonstrated in both

Condition A and Condition B. Young (1942) observed pronouns as

indices of speech development. She noted a decrease in possessive

self pronouns with age and an increase in pronouns related to others

with age. The data presented here support the observation,


Differences Between Age Groups During Condition A

Statistically significant age differences found in a clinic

setting with greater mean frequencies for seven-year-old subjects

were the variables of male role, time reference, space reference,

action norm, distress, sense, communicate, attack, recreational

and sin reject. This indicates that developmental differences can

be observed in the content of the following areas: (1) male

reference; (2) qualifying references of time and space; (3) normal

patterns of social behavior, such as job, game and lunch; (4)

states of fear, grief and indecision; (5) sense perceptions, such as

look, listen and see; (6) communicative words, such as said and told;

(7) hostile actions, such as fight; (8) recreational activities; and

(9) words such as dirty, scratch and steal.

The Concept Index Scores emphasize that the subjects selected

male heros and male persons with higher relative frequency for the

stories they told in the clinic setting. In the course of telling a

story, references that modify and amplify a tale are employed by the

seven-year-old subjects. Action and behaviors were described. The

pictures used as atimuli offered the opportunity to those older








subjects to employ a larger vocabulary than they might use in

another setting.

When stimuli and setting were structured by the investigator,

differences attributable to age were found in ten dependent variables.

Magnitude was related positively with age.


Differences Between Age Groups During Condition B

In a home setting, four-year-old children employed the

categories of natural object and distress with greater frequency

than seven-year-old children. The older subjects made more state-

ments utilizing time reference, guide and economic dependent variables.

The younger subjects spoke of animals and nature and used words

conveying states of fear and despair with greater frequency in the

home setting than did the older subjects.

The seven-year-old subjects made more references to measurement

of time, to the institutional context, economic (office, store and

automobile), and employed verbs conveying orders or positive direction,

such as let and help, in the home setting. Significant content

analysis differences under Experimental Condition B were found in

five variables. The magnitude of two of the differences was inversely

related to age; three were positively related.

These differences are not sufficient to make a secure statement

that maturational indices can be identified employing the Harvard III

Psvchosociological Dictionary. A cautious view must be maintained.

Oral speech skills under both experimental conditions, as demonstrated

by Group I subjects and Group II subjects, approximated each other on

the majority of the dependent variables. The frequency polygons and

Concept Index Scores indicate that the distribution of the dependent









variables corresponds in configuration under both experimental

conditions. Within the limits of this investigation, children,

four years of age and seven years qf age, used the same content

categories in 57 of 76 dependent variables.

Three explanations can be offered to account for these

results. The observation that children, four years of age and

seven years of age, have the same content categories would be in

accord with the view that children demonstrate fundamental phono-

logical and grammatical speech and language skills by four years

of age (Ervin-Tripp, 1966; Messer, 1967). Any changes in speech

content attributable to age for children four years of age and

older would be in the refinement and expansion of entry words

within each of the dependent variable categories.

A second explanation for the lack of differences is that a

general dictionary, such as the Harvard III Psychosociological

Dictionary, may measure indices that do not discriminate maturational

variables. An original dictionary developed with variables specific

to anticipated differences might be more effective in identifying

developmental indices.

The third explanation concerns the view that there may be a

wide dispersion in the content of oral language behavior at each

age level, which Is demonstrated in a failure to reject the null

hypotheses on the majority of variables. Williams, et al. (1937)

and Smith (1941) noted considerable overlap in vocabulary skills at

each age and grade level in children. Templin (1957) noted

increases in recognition vocabulary with age, but a less definite

trend in speech vocabulary. A future appraisal of content in the









speech of children should investigate the dispersion of dependent

variables within one age group.

In surveying the tables and figures provided by this study,

the underlying systematic unity of language performance can be

observed under both conditions and between ages. The configuration

of the frequency polygons indicates that a correspondence is

present between setting and between ages in each of the categories.

Frequencies of occurrence are available for a number of categories

which were not critical to the hypotheses of this study. Thirty-

six of the dependent variables demonstrated no significant

differences. The dependent variables describe the content of

speech and relate speech to a framework of social and psychological

theory.


Dictionary Revision

This investigation has considered the utility of the Harvard

III Psychosociological Dictionary for data analysis. The dictionary

is more inclusive than earlier studies in that those topics Shirley

(1938) observed in the speech of her subjects, the content that

Hahn (1948) listed for her first-grade children and the two general

areas introduced by Metraux (1950) are handled in the routines of

the dictionary. Hutchinson (1967) listed word classes for foods,

numerals, colors, body parts, parental identifications, television

references and taboo words. Only the name of television references

is not accounted for in the dictionary. The words selected by

Ames (1946, 1948) as conveying a sense of time and sense of space

appear in the dictionary. In effect, it is possible to process

the data from these six studies in one effort.








However, revision of the Harvard III PsvchosocioloRgcal

Dictionary would be useful for future investigations of child

language. Revision would make it possible to analyze a greater
*
number of items of content in a verbal text and would facilitate

data preparation. Two procedures are involved in dictionary

revision: (1) reduction in the dependent variables; and (2)

addition of new words into the dictionary.

The reduction of dependent variables would be based on a

consideration of those categories which had negligible frequency or

zero cells. For example, consideration might be given to consoli-

dation of the small group dependent variable and large group

dependent variable into one category entitled collectivities

since the two share a common aspect as well as low frequency of

occurrence. The dependent variables equal, defense mechanism and

thought form did not provide much information for this investiga-

tion and could be omitted from the dictionary. Other dependent

variables with low frequency appear to be useful for investigation

of child language. A decision to omit or to consolidate the

following categories should await further investigation:

First-Order Tags Second-Order Tags

job role legal
control political
follow military
ideal value religious
bad medical
expel economic
deviation artistic
arousal
pleasure
anger









The addition of words to the Harvard III Psychosociological

Dictionary would provide a children's version of the dictionary.

Three methods of adding words to the dictionary are proposed.

First, there is a group of words which can be categorized by the

primary denotative meaning of the word. These include terms that

can be tagged into food, clothing, tool, natural object, male role

and female role categories. Examples drawn from this Investigation

include: carrot, cereal, cookies, jacket, pants, balloon, tricycle,

dad, daddy, mommy, mama and mom. Second, there is a group of words

in which a key-word-in-context strategy can be used to relate words

to categories. Such words are more likely to be verbs. Examples

drawn from this investigation include: grab, play, sleep, sneak,

watch, done, hope, made and quit. There is a third group of words

which cannot be tagged to a dependent variable due to the difficulty

of assigning a denotative meaning. These words can be entered in

the dictionary using key-word-in-context analysis, word association

methods and disambiguation methods. Examples of words from this

investigation that could be entered in the dictionary following

these procedures are: had, has, have, use, kind, like, leave,

board, hard and bit.


Implications for Research

A number of areas for further study have been indicated by

this investigation. Tentative normative data and methodological

clues for additional investigations of speech and language have

been provided. Research generated by this study can be divided

into four general areas first, research related to refinement of

procedures to identify the significant content variables which aru








a function of age; second, studies related to measurement through

content analysis of naturalistic settings with speech situation

held constant; and, third, research related to the description of

pathologies of speech and language as demonstrated in content

analysis. Finally, consideration is given to the variable of

personality in speech and language research with nonpathological

subjects.


Content Analysis and Age

Research related to refinement of procedures which identify

content language variables as a function of age has been suggested

previously in this chapter. Two possible studies concern different

approaches to the verbal texts obtained for this study.

Two selected subgroups of subjects identified as high and

low groups on a language proficiency or fluency measure could be

tested to investigate the hypothesis of variability of speech

content within one age group. A hypothesis of no difference between

groups could possibly be rejected.

It may not be the categories but the entry words into the

categories which delineate the maturational dimension of speech

content. Loban (1963) observed that statements which expressed

tentativeness, supposition, hypotheses and conditionality occurred

more frequently in the speech of subjects skilled in language.

Use of subordinating connectives, such as although, because and

unless, developed with age. Ames (1946) observed that children

younger than 42 months used specific time references before acquiring

general time references. Smith (1926) noted an increase with ago









in use of indefinite nouns, such as somebody and anything, and

expressions of time, position and quantity. The increasing

versatility of expressions might be isolated by investigating

within the categories. Therefore, relevant variables in the

Harvard III Psychosociological Dictionary might be found in a

detailed examination of the entry words assigned to persons,

qualifiers, nonspecific objects, sense, think, if, not and cause

dependent variables.


Content Analysis and Setting

The design of an investigation to hold home environmental

situation constant is more difficult to achieve than one where home

environment is free to vary, but it is not impossible. The use of

a wireless microphone worn by the subject permits naturalistic

observations in the absence of an investigator. Counseling with

individuals in the home environment can assure the desired situation

or a variety of situations while gathering natural or field observa-

tions. The purpose would be to locate those categories which have

the most explanatory power in relation to setting. Speech between

parents and child during family mealtime may approximate speech

samples collected in a structured clinic setting. Having the

subject and the parent or model in verbal interaction introduces

data relative to acquisition of speech. This type of investigation

is larger in view than speech development per se; it would ultimately

overlap the fields of sociology and social psychology where such

studios have been initiated (Ervin-Tripp, 1964; Barker, 1963).









Content Analysis and Speech Pathology

Content analysis would contribute an objective and consistent

identification of deviating speech and language patterns. Questions

that might be asked are: Do content and thematic analyses of the

stories of cleft lip and cleft palate subjects indicate variations

in content that can be attributed to the physical deformity? Can

varying forms of aphasia or language disorders be discriminated by

content analysis? Can paucity of responses in some categories be

utilized in remedial programs? Identification of the content of

the speech of predefined classes of speakers, such as minority

groups, dialect groups and social classes, would contribute

information relative to the referential dimension of language for

these groups.


Personality in Speech Research

When gathering data for normative studies, the personality

of the child is largely ignored. Identification of subjects prior

to experimental treatment using a personality assessment procedure

would prevent personality variables from obscuring age and setting

differences. This investigation sought to obtain a homogeneous sample

of subjects in terms of age, genetic and environmental background,

hearing acuity and intelligence. The differences found may reflect

personality variables as well as those of age and setting. An

investigation of content of speech as a function of age, personality

and setting would contribute information relative to this issue.

This investigation has demonstrated that content analysis is

one strategy for the study of the oral language of children, four





78


years of age and seven years of age, under the conditions of a

structured clinic setting and a home setting. Dependent variables

which were statistically significant were discussed; the limitations

of the investigation were suggested. Procedures for modification of

the Harvard III Psychosociological Dictionary were proposed to

facilitate analysis of additional data from a verbal text.

Implications for research were stated.















CHAPTER V

SUMMARY


The language performance of a selected group of four-year-old

and seven-year-old children under two conditions was investigated

through a content analysis research design. Content is the denotative

meaning of objects, concepts and processes as symbolized by verbal

utterances. A study of the content of language contributes data

concerning the referential dimension of speech. This investigation

utilizes a content analysis category system derived from psychological

and sociological theory in a study of the referential dimension of

child language.

To investigate the content of oral language as a function of

setting and age, the following procedures were carried out. Fourteen

pairs of siblings were selected from a population of four-year-old

and seven-year-old children. The subjects were homogeneous in terms

of language background, hearing acuity, intelligence and physical

status.

The two experimental stimulus conditions were (1) a structured

clinic setting and (2) a nonstructured home setting. Under the

first condition (Condition A), the stimuli presented the subject

were a series of items from a children's projective test and five









magazine illustrations. For the second stimulus condition (Condition

B), each subject wore a wireless microphone transmitter in his home

where the subject pursued typical home activities. Responses of the

subjects in the laboratory and vocal output in the home setting

were recorded.

Speech samples consisting of 127 consecutive communication

.units were entered on electronic data processing cards for analysis

through an automated system of content analysis. Objects, concepts

and processes represented in verbal expressions were reduced and

grouped into categories; the meaning of any word or group of words

was summarized by listing the category under which it occurred. The

dependent variables for this investigation were quantitative data

from the 76 categories of the Harvard III Psychosociological

Dictionary. Objects are grouped into three major areas: social,

cultural and natural events. Behavioral dimensions or psychological

processes include emotions, perceptions, thought processes and

evaluations, and impersonal and social emotional actions, A third

set of categories is comprised of words used to modify or amplify

the meraing of nouns and verbs.

To assess language performance as a function of age and

setting, each dependent variable was tested using the Mann-Whitney

U statistic (U 55; p m 0.05). The null hypotheses were rejected

on a number of the dependent variables. There were content

differences between subjects who wore four years of age and seven

years of age and between speech samples obtained during activities

in a home setting and samples obtained when speech was elicited by

an adult with preselected stimuli in a clinic setting.








Differences in the content of speech between Group I subjects

and Group II subjects were statistically significant for 19 of 76

categories. The 19 categories were other, natural object, get,

male role, time reference, space reference, action norm, distress,

sense, communicate, guide, attack, recreation, economic and sign

reject.

Within the conditions of this study, children, four years of

age and seven years of age, had analogous content in 57 categories.

These nonsignificant variables increase the probability that the

speech content of four-year-old children is the same as the speech

content of seven-year-old children. Three alternative explanations

can be proffered. First, Group I and Group II subjects demonstrate

the same content in language performance because changes in speech

'content attributable to age are in the refinement and expansion of

vocabulary expressions within each of the categories. Second, the

categories of the Harvard III Psychosociological Dictionary do not

measure variables critical to maturation of speech skills. Third,

there is a wide dispersion of language skills within each age level

this variability prevents isolation of variables which are a func-

tion of age.

This study is not unique in finding descriptive differences

in speech samples between settings. However, the present data are

unique in two ways. First, the data were collected from the same

subjects under two different stimulus conditions following parallel

procedures. Second, the present data are unique because statistical

analyses were performed. This increases the probability that those

inferences made concerning the relation of speech sample and setting

are true differences.








For Group I subjects, 19 categories were found to be statis-

tically different, and for Group II subjects, 30 categories were

found to be statistically different between settings. Under a

structured clinic setting, a more comprehensive picture of language

performance is obtained. However, this is not a sample of how

speech is employed to meet the demands of day-to-day communication.

It is desirable that speech data be obtained in a wide variety of

settings and situations. Patterns of equivalent or nonequivalent

speech usage among the varying conditions can be identified;

interactions that are significant between topics, situations and

participants can be described.

The Harvard III Psychosociolon cal Dictionary is a tool that

can be utilized in studies of speech and language. A sample of

speech was reduced into object, concept and process categories.

These categories, drawn from psychological and sociological theory,

provided a model for the study of the speech content of children.

A number of expressions common to the speech of children were not

entered in the dictionary. It was proposed that these expressions

be placed in the dictionary to provide for future investigations.

The usefulness of the General Inquirer and Inquirer II automated

systems of content analysis in efficiently and reliably handling

large amounts of data was demonstrated in this investigation.

An assessment of language performance in young children is

incomplete without information concerning the referential dimension.

Content is a relevant area for research in language development

because the data contribute information concerning this dimension

of speech and language.















APPENDIX A


LETTERS SENT TO ADMINISTRATIVE STAFF OF SCHOOLS
AND TO PARENTS CONCERNING THE INVESTIGATION


1. The following is the letter that was sent to the principals

of five elementary schools.




UNIVERSITY OF FLORIDA
Gainesville, 32601

Reply To:

Speech and Hearing Clinics
436 Arts and Sciences Building


Dear

As part of the requirements for the doctorate, Mrs. Shirley
Pine is conducting a study in normal language development. Children
enrolled in first and second grades who have siblings in the three-
and four-year-old age groups will be invited to participate.

The purpose of the investigation is to use the methodology
of content analysis in the description of the oral language of
children. It is proposed to analyze the content of speech samples
as a function of stimulus situation and as a function of age. The
variable under investigation Is oral language; no measurements will
be made in reference to individual children.

As part of the research procedure, both language and hearing
screening tests will be given. In addition, a sample of the child's
oral language will be recorded. Stimuli presented the child by the
investigator will be pictoral.








It is our feeling that children from the public elementary
schools would be appropriate subjects, and we would like to secure
permission to invite some of the first- and second-grade children
to participate. Mr. Tomlinson has been consulted concerning this
investigation. It will not require the students giving up part of
their instructional day.

Sincerely yours,



E. C. Hutchinson, Ph.D.
Associate Professor




2. The following is the letter and form that was sent to

the parents of subjects.




UNIVERSITY OF FLORIDA
Gainesville, 32601

Reply To:

Speech and Hearing Clinics
436 Arts and Sciences Building


As part of the requirements for a graduate degree at the
University of Florida, I am conducting a study in normal language
development. The study is concerned with content of language of
children at two different ages and under two different conditions.
This study will help in the understanding of speech and language
of growing children. It will assist in planning educational
programs.

Two children from a family will be invited to participate.
One child should have a birthday between May 31, 1962, and May 31,
1963; the younger child should have a birthday between May 31,
1965, and May 31, 1966. Should your children have birthdates
within these periods, I would like you to participate in this
study.

The research will be conducted at the Speech Department,
University of Florida. As part of the research procedures, the
children will be given language and hearing screening tests.
Results of the screening tests will be available to the parents.
The study will require approximately one hour at the University.








Each child will also be asked to wear a wireless microphone while
playing after school one afternoon. Appointments will be scheduled
at your convenience. Transportation to the Speech Department can
be provided for the parent and children.

A two-dollar gift certificatP from a local toy store will be
given each child.

Since children within the appropriate age groups and within
the same family will be difficult to locate, I would appreciate
your participation. If you have any questions or wish further
information, please telephone me, 378-7650, or E, C. Hutchinson,
Associate Professor, Speech Department, 392-2041, I am enclosing
a form for you to return with an addressed envelope.

Thank you.


Sincerely yours,


Mrs. Shirley J. Pine, M.A.



E. C. Hutchinson, Ph.D.
Associate Professor


Enclosure


Yes, we have two children, the following ages, and will
be able to participate in the language study. Our
children and their birthdays are listed below.


Name

Name


Date


Date


__No, we will be unable to take part in the language
study.


Yes, we will need transportation to the Speech Depart-
ment.


Thank you very much.





86



Signed

Address

Phone


RETURN FORM TO: Mrs. Shirley J. Pine
Department of Speech
University of Florida
Gainesville, Florida
32601














APPENDIX B

EXCERPTS FROM TWO TRANSCRIPTIONS AND EXAMPLES
OF PROCEDURES FOR TYPING, SEGMENTING AND EDITING SPEECH SAMPLES


1. The following is an excerpt from one transcription of a four-

year-old girl in the clinic setting.




Subject <1341> people are pulling ropes+ they will pull it down to

the sea+ because he is almost down there+ they will float+ they will

swim+ let us see+ a man is sitting in a chair with a boy beside him+

watching television+ I think he is watching #red #skelton+ he has

a red face+ he says people+ #tiny #tim+ sometimes but not all the

time+ but he sings like a lady+ you know+ that is a lady carrying

a baby+ and somebody is riding a tricycle (toy) behind her+ I think

they are going to get some milk+ there is a milk carton+ they will

bring it back home+ they will drink it+ because she has got to hurry

back for somebody to drink it+ looks like somebody is setting a baby

bed by the bed+ it looks like+ talking+ probably about tomorrow+

going to school+ no+ I stay home with the maid+ in five (quantity)

days+ these are big people+ and a little one sitting out in cold

weather+ because there is no leaves (tree)+ because they camp out+

yes+ we camped out last summer in the woods down in Panamaa #city+








we used to go to Panamaa #city #gainosville Panamaa #city #gainesville

Panamaa #city+ we kept going back and forth every month+ because

mama (mother) said let us move back and forth every day+ my cousin

lives down there+ they have a great big woods+ and we used to go down

there you know and play (norm)+ and we used to camp out down there+

that is a giant trying to catch that little boy+ tie him up+ he will

I think eat him+ because he does not like him+ because he tried to

do a kind thing+ but he did not do it right+ there is a mother and

another mother telling the father+ and there is a mother telling the

#chinese boy something+ I think she is telling him that day after

tomorrow+ because he has got that kind of face+ about the day before

yesterday+ yes+ there is a baby in a bed+ and they doing something+

and she thinks somebody will come in and get it+ she will holler+

because she would not have her mother with her+ is that all the

picture we have+ there is a daddy (father) spanking (discipline) a

little boy+ I do not know+ I do not know+ he has a shirt on+ he will

play+ cowboys and indians+ I play with my brother+ I am an indian

girl+ there is three (quantity) little birds eating+ the three

(quantity) birds were eating+ and they were eating some worms+ and

they had some bowls+ and they had some napkins+ and the wind started

blowing+ and it blew everything over+ they did not have anything to

eat anymore for lunch+ kind of sad+ probably they flew around+ and

went down to get some more worms+ this is a whole picture+ yes+ it

is a real picture too+ that is some people looking at something+

I think+ I wonder what they are looking at+ I cannot even see+ I

cannot even soe+ I can see the class+ but I cannot seo anything

else+ they would be looking at a squirrel (animal)+ ok+ that is









another real picture+ that is a daddy (father) talking to a boy+

he is talking about day before #friday (day)+ at #friday (day)

you do things+ and then you do not have to go to school+ he is saying

hip hip hooray+ let that be all the pictures+ three (quantity) more+

that is somebody getting out when they just got through getting

married+ because there is her+ she is the one that got married+ and

she is a wedding flower girl+ and there is their car (automobile)+

see+ she got married+ with him+ and there is the car (automobile)

all fluffered with something+ it has all kinds of things on it+ it

has bathroom paper on and everything else+ yes+ this is a lot of

people out taking a picnic+ looks like she is going to have another

baby+ because there is her son right there+ and I bet she is going

to have another baby+ her food is over there+ yes+ she already had

a baby+ and it looks like+ I do not know anything else+ cars

(automobiles)+ and there is one just like ours except it is not

yellow+ because this back of it is like that+ look at that dog+

sure a fat dog+ I think it is a girl that is going to have puppies

(animal)+




2. The following is an excerpt from one transcription of a seven-

year-old boy in his home environment.




Subject <1472> now this is in a parade+ it has got to go slow+ no

a little faster+ faster+ #danny you cannot get right there+ and

#laurie you are not going to get there either+ you get right here+

come on+ I will show you how you get on+ here watch+ see like #roy+








you just put your hand down like here+ then you can go fast+ ok+

ok #laura will push the middle+ you can be the policeman though+ but

you have got to help us too+ #danny you stop+ you got to remember+

yes my name is #mr #john #wade+ yes I am the guy who is part of

the #kennedy family+ and I do drive a sport's car (automobile)+ I

just love cars (automobile)+ this is an old fashioned racing car

(automobile) that I am driving now+ it is real neat+ back up+ slow

as anything+ we are driving around hunting for old fashioned cars

(automobile) that have been wrecked+ ok look+ nineteen twenty eight

(date)+ remember nineteen twenty eight (date) nineteen twenty seven

(date)+ nineteen twenty six (date)+ nineteen twenty eight (date)+

look+ turn+ this is one of those covered wagons+ a big huge you

know what+ no+ I thought you said you were a person who had run out

of gas+ a covered car (automobile) I mean+ ready to carry some

bikes (toy)+ now this called an old fashioned coupe (automobile)+

no #ray+ no we do not+ we know how it runs+ just get off+ I told

you+ I got the microphone+ want to carry it on the covered wagon+

get off+ #greg carried a string where the parade is+ it is called

the oval coupe (automobile) chicken+ it loads on to the back+ see

what happens when you back up+ this can be the old fashioned motor+

on the side+ push her up+ do not you know anything+ like this+ so

we do not fall down+ got that #william+ two (quantity)+ now wait+

you mean the motor+ you mean that+ that is kind of like a motor on

the back+ you can see it+ they broke it out+ #terry wear my hat for

a few minutes+ what+ what do you mean by eighteen (quantity)+ ok+

I got what you mean+ why you going to stand there like a dumdum

(stupid)+ you are a dumdum (stupid)+ sea+ now this one will go









pretty slow+ this one of those kind+ you keep making sure the wheel

is going+ slow down+ no+ what+ you just leave it to me+ you do not

push+ got to wear it+ but it is someone else's+ that lady you

know+ not so fast+ see this will not go around+ it is the motor+ but

you are not supposed to push it by yourself+ see+ the wheel is

pushing it+ see you would have to push+ see they call it an old

pump+ one of these are pretty strong with cars (automobile) #danny+

#richard+ no+ not yet you+ #mike will you get it+ I know+ but look+

it is swings back and forth+ because of you+ you ruined it+ I know+

you got to pull it back+ now let it up some+ now just make sure

you do not touch anything+ now it is just falling apart+ get out+

I can slug (hit) you+ seven (time) to six (time) fight+ now you

would not like that do you+ I do now even think #mike would like a

seven (time) to six (time) fight+ he would just say help as soon

as he got near me to fight+ would you+ would you just wait a minute+

darn it+ I just think I will work on it in+ it is not+ there what+

yes I know+ I am sure glad I do not have a brother like that+ I

hope I am going to have another brother+ but I sure would not like

to have a brother like #jeff+




3. Examples of segmentation of speech samples.

A. Compound sentences which could be divided without

essential loss of meaning are edited as two communication

units.

Example: Mommy decided to go to the store to get
something+ and so a boy got on his tricycle+ and she
got the baby+ and when they came back mommy had all she
wanted in the basket+ and her boy was riding behind her+




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

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs