A QUALITATIVE ANALYSIS OF VOCABULARY RESPONSES IN EARLY AND LATE MATURITY _ THESIS FOR THE DEGREE or PH. D. MICHIGAN STATE UNIVERSITY- JOHN C. EHRMANN 1955 This is to certifg that the thesis entitled \ A QUALITATIVE ANALYSIS OF I VOCABULARY RESPONSES IN ‘ EARLY AND LATE MATU‘RITX presented In] J OHN C . EHRI-IANN has been accepted towards fulfillment of the requirements for 4111924,— degree in Jambalogy Major , wrolessor t Date J 2'- 0-169 A QUALITATIVE ANALYSIS OF VOCABULARY RESPONSES , IN EARLY AND LATE MATURITY by JOHN 0. mm A THESIS Submitted to the School of Advanced Graduate Studies of michifian State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree or DOCTOR OF PHILOSOPHY Department of Psychology 1955 ‘ IHESIS ABSTRACT An attempt was made to substantiate the general hypdhh- esis that vocabulary definitions of aged individuals, when analyzed on a qualitative basis, will demonstrate marked intellectual decline. More specifically, it was assumed that aged normal individuals would produce word definitions qualitatively inferior to those produced by young normal individuals. An experimental group of 100 aged white males (Mean age 75 years) and a control group of 100 young white males (Mean age 25 years) were matched on the basis of intell- igence (Wechsler-Bellevue Full Scale IQ) and score correct on the Wechsler-Bellevue Vocabulary Test. All definitions for twenty nouns selected from this vocabulary test were subjected to qualitative analysis consisting of five cate- gories: 1) Class or categorical, 2) Descriptive, 3) Example, A) Functional and, 5) Error. The reliability of this scale was tested by having an independent rater classify all definitions of every fourth protocol, fifty in all. Comparison of the ratings indicated 96 per cent agreement. The following five specific hypotheses were tested: 1) The young group will produce significantly more Class or categorical definitions 2) The aged group will produce significantly more Descriptive definitions 3) The aged group will produce significantly more Example definitions a) The aged group will produce significantly more Functional definitions 5) The two groups will not demonstrate significant differences in errors of word definitions. Obtained results substantiated four of the five hypoth- eses. The young group produced a significantly greater number of Class or categorical definitions. The older group produced significantly greater numbers of Descriptive definitions and Functional definitions. Mean differences between the two groups were significant beyond the 0.5 per cent level of confidence. No significant difference was obtained between the two groups in their total number of failures for the twenty words.' Contrary to one of the specific hypotheses, no significant difference was noted for the two groups in the frequency of Example types of word definitions. In the discussion of the results, the following three aspects were emphasized: l) The theory that vocabulary represents "old learning" and therefore is resistant to decline in mental impairment is challenged. 2) The validity of psychological tests using 8 iii quantitative measure of vocabulary as an index of the in- dividual's previous level of functioning is challenged. 3) The efficiency of a qualitative analysis of vocabulary word definitions as an indication of the presence and extent of mental deterioration is emphasised. Date JUIY ‘ ___._ , .".-_ iv ACKNOWLEDGMENTS The author's primary debts of gratitude are to Dr. Albert I. Rabin and Dr. Donald M. Johnson. As the initial chairman of the Guidance Committee, Dr. Rabin provided constructive criticism and patient guidance throughout the formulation of this study. A particular debt of gratitude is owed to Dr. Johnson, who assumed the reaponsibility of Guidance Committee Chair— man in the absence of Dr. Rabin. He gave freely of his time, encouragement and invaluable advice. Other committee members who contributed their sage advice and offered constructive criticisms were Dr. Alfred G. Dietze and Dr. Paul Bakan. The Medical and Surgical Staff of the Veterans Admin- istration hospital at Saginaw, Michigan gave unlimited cooperation to the author and greatly simplified the matter of gathering the data. This research is dedicated to my wife, Dolores, whose contributions in editing and typing the final manuscript were imeasurable. TABLE OF CONTENTS INTRODUCTION............................................. Intellectual Functioning in Later Maturity...................................... The problem of mental deterioration.............. The results of intelligence tests................ The "Higher Mental Abilities".................... The results of Rorschach tests................... The Vocabulary Test................................. The measurement of vocabulary.................... Factor analysis of the vocabulary response....... Vocabulary and general intelligence.............. Vocabulary and mental deterioration.............. Qualitative analysis of vocabulary............... III-IE PROBI‘EIVI.OC.UOOOOCOOIOCCIQOCOCOIII.I....IOIOOOOIOOC... Subjects............................................ Matching............................................ The Vocabulary Test................................. iuantitatch score............................... Qualitative analysis............................. Reliability of scoring........................... vi Page 1 3 6 8 10 ll 13 13 1h 15 16 18 26 29 30 3h 3h 36 38 HypotheSGSOOOIOOOCOOC.......I...OOOIIIIOIOOCOCII. Ie-ESULTSOI.0...‘I..‘..II...IOI.0I... Hypothesis ACCOOOOIOOIODCICDOOIOOIOOIOOCOOICOOCOOOI. IIypOtheSiS Bocooocooo-ou-oo OOOOCOOOIOOOOOOIOOIIII. Iiypothesis CO...IIOOIOOICOIOOOOIOOIOOI.OIOIIIOIICIOC Hypothesis D0.0CCOOOICDUOOOOOIOIOOOOOIOIOOIOICOOOOII Hypothesis E.I..OOOUOIOOIOOIOCI.......0.000COQDOSGOI DISCUSSION AND COIquSIONSOOIOIOOCCOIOCOIOUI....OOOOCIIOO Smlfl‘mYOOCCCCIOOCOICOCO.I0......O..C...I...0.000....0000I APPEBIDIXOOUOOOOOCOIOOOOOOOOIOI.0.0...OCIOIOOOOOCODOOOOOI. BIBLImRAH’IY..Q.........I..I...OOI 38 1.0 1+1 #2 b3 u. #5 1+8 55 58 69 TABLE II III. IV. VI. VII. FIGURE LIST OF TLBLES.AND FIGURES Age, Education, IQ, and Vocabulary Matching Data for the Young and Aged Groups........ Comparison of the Educational Levels of the Young and Aged Groups with the U. S. Population at the Sanle AgeOOOIIOOOOOOCOOIOCOOCOOIOOOOOOOOICODO Comparison of the Two Groups on the "Categorical" Type of Response..................... Comparison of the Two Groups on the "Descriptive" Type of Response..................... Comparison of the Two Groups on the "Example" Type of Response......................... Comparison of the Two Groups on the "Functional" Type of Response...................... Comparison of the Two Groups on the "Error" Type of Response........................... Per Cent of Responses of Both Groups Falling Into Each gualitative Vocabulary- ClassificationOOOIIOQOO0......-IDIOOOOOOOOOOOOOOI. viii Page 33 35 Al 43 AA A? ‘M .- INTRODUCTION Within recent years a highly significant amount of research has been centered upon the psychological problems encountered in later maturity. It is unfortunate that the realization of the tremendous needs for such research with the aged and aging has been so long in coming, and that these investigations have, as yet, done little more than touch the surface. In comparison with the amount of work that has been devoted to other age groups, the field of gerontology is such a relatively new and unexplored area that the research worker is still in the process of finding and perfecting techniques and methodology applicable to the problems at hand. At the present time psychological explor- ations of this age range are devoted primarily to a descrip- tion of the nature of psychological changes in the aging process from the time of maturity up to old age. Well controlled psychological studies of aged people seem to have been restricted to investigations of abnormal groups including neurotics, psychotics, and various types of physically disabled individuals. Important as these studies are and will continue to be, in consideration of current population trends, it is certainly of the utmost importance for psychology to attempt to thoroughly investi- 1 “aw—7r 2 gate the normal aging process. Due to the nation's declining birth and death rates, fewer children are being born and more and more people are living into the older age range. Con- sequently, from decade to decade, a greater preportion of the population of this country falls into the age classification of later maturity. The statistics furnished by the United States Bureau of the Census and the United States Bureau of Labor Statis- tics provide a dramatic realization of current pepulation trends (82)(83)(84). For example, in 1900, the number of people over sixty-five years of age in this country totaled a little over three million, a ratio of one person of sixty- five or.more years to every 25 individuals; by 1950, the Census reported a total of 13.h million pe0ple aged sixty- five or older, a ratio of one oldster to every twelve in- dividuals. Thus the increase in life eXpectancy over a fifty year period was approximately twenty years, and the pOpulation aged sixty-five years or older increased by 37 per cent, whereas the number at ages under sixty-five in- creased only 13 per cent. Moreover, the Bureau of the Gen- sus anticipates still further pepulation gains for older peeple and estimates that by 1975, close to 21 million peeple in the United States will be sixty-five years or older, and that by the turn of the century at least 13 per cent of the country's pepulation will have reached the age of sixty-five. Realizing the significance of the above figures, community agencies throughout the country have insisted that the value to be derived from research de- voted to gerontological issues can no longer be denied. Intellectual Functioning in Later Maturity Psychology has, from a historical point of view, long been concerned with problems relating to the nature and measurement of intelligence. At the present time, the field of mental measurement has progressed into an advanced and rigorous science, but many problems are still unsolved. Although many valuable instruments have been perfected, it is now widely recognized that it is possible, through the use of these instruments, to obtain only an indication or estimate of actual native intelligence. We are unable to separate any measurement of an individual's functioning from such factors as the motivation to learn and to perform, the general condition and plasticity of the nervous system, and environmental opportunities. It is generally understood that no matter how carefully w and wisely a test of intelligence is constructed, any dis- cussion of its results must be based upon and interpreted in light of the particular manner in which the concept of intelligence is defined and the particular methodology or means utilized in its measurement. In consideration of these facts, the psychologist attempting a comprehensive statement of intellectual functioning during later maturity . W:f"u'.'e:‘~., l | ; l: . “r J, “—7. is confronted with additional difficulties peculiar to the aged individual. 0n the one hand, if an overall or general view of in- telligence is adepted, and intelligence is then measured by an overall test score, the gerontologist must be cognizant of the fact that any test or instrument is liable to sample different types of mental abilities from one age level to another. In one study, conducted by Balinski (5), a factor analysis of the Wechsler-Bellevue Scale demonstrated quite different factor loadings from one age range to another. Thus, a test may, perhaps, measure the same mental ability but with varying degrees of reliability and validity far different age groups. Even if the research worker is provided with some de- gree of certainty that the same ability can be effectively tapped at various age levels, he is still confronted with serious difficulties when he attempts to interpret his re- sults. Carefully controlled investigations during this period are very few in number and there is an almost com- plete lack of standardization and normative data to cover the upper age brackets. For example, the 1937 Revision of the Stanford-Binet, though often used to estimate the Intel- ligence Quotient of adult individuals, reports standardiza- tion data only up to eighteen years of age. Even such a widely adapted instrument for the measurement of adult in- telligence as the Wechsler-Bellevue Intelligence Scale demonstrates extremely inadequate samples from the upper age levels. The normative data provided in Wechsler's tables terminates at the age of 59 years and includes a total of less than 200 subjects beyond the age of 50 (86). Since our intelligence tests have primarily been standardized on young people and specifically designed for younger subjects, their use with older individuals is Open to the serious chal- lenge that they are not wholly valid in later maturity. “"’” 0n the other hand, if a more specific definition of intelligence is accepted and measurement is obtained through observation and detailed analysis of specific mental func- tions, the gerontologist is again faced with problems pecu- liar to the aged. Judgment, critical thinking, reasoning, and other so-called "higher mental abilities" do not readily lend themselves to quantitative measurement and, thus, are very difficult to compare from one age range to another. In addition, such abilities as reasoning and judgment, when .measured on a very narrow basis, tend to be highly correlated with the amount of previous education and training. Thus, in some studies devoted to an analysis of an individual's current functioning level on certain tasks, it is difficult, if not impossible, to estimate the native potential of the aged person because of the many Opportunities he has had for continued practice in comparison with a younger subject. If an attempt is made to adequately match groups of young and old subjects on such variables as education and training, it is necessary to consider the likelihood that the nature of the training of the older individual, which , Iv T9.m—hf “9:75le‘wm IF m, 1' 17‘17‘“ ‘3 . ..W.fl'!7" has usually occurred at a much earlier period, even though similar on a quantitative basis, was probably quite different qualitatively from the more recent training experiences of the younger subject. The Operation of the factor of motivation can readily be seen in the behavior of many aged people. Occasionally, a certain amount of intellectual decline may be directly attributed to the absence of any motivational factors. Since the learning process, with aged individuals, takes a longer amount of time, requires a greater effort, and produces re- sults which are not always as satisfactory as desired, some oldsters cannot or will not readily exert themselves. Thus, general interest wanes and a decrease of intellectual activity and the use of intellectual abilities is apparent. Since mo- tivation does seem to decrease in many aged persons, it is a factor that must always be considered in an investigation of intellectual functioning during later maturity. Because it cannot very easily be quantified, many investigators have chosen to ignore it. Thejproblem of mental deterioration. It is impor- tant to understand exactly what is meant by the concept of "mental deterioration" if one is to objectively study the nature of changes in intellectual functioning that may take place during the aging process. For purposes of diagnostic convenience, mental deterioration is often considered to be one of two rather different types--that which occurs during later maturity along with the natural increase of age, and that which occurs as a consequence of prolonged mental dis- ease or some brain lesion, and may occur at any age. However, from a psychological point of view there appears to be little difference between these two types. It has frequently been pointed out in the literature that the type of deterioration, evidenced on psychological test performance, encountered in normal old peOple is highly similar to that encountered in ,most organic brain diseases and mental disorders (32)(39)(h2) (h9)(86)- Therefore, at the present time, many peOple investigat- ing mental deterioration consider this concept as a very gen- eral one. For example,‘Wechsler states, "A person will be considered as giving evidence of mental deterioration when he is no longer able to carry on his intellectual tasks with the speed, accuracy, or efficiency previously characteristic of his functioning level"(86, p.5h). A definition such as this one accounts for the fact that mental decline.may and does ' take place independently of any specific organic or mental disease. The decline of mental ability with age should be considered a normal and eXpected part of the general aging process of the organism as a whole, according to Wechsler and others. Wechsler feels, "Every human capacity after attaining a maximum begins an immediate decline"(86, p.55). He points to the well established information on changes in the human brain that accompany normal old age—-a general loss in mean 8 brain weight and a gradual shrinking in size--and shows that corresponding changes in intellectual ability closely follow this pattern of deterioration. The actual task of investigating mental deterioration can be divided into three separate problems: 1. A reliable.measure of the individual's actual or present functioning level. 2. The evaluation of the individual's previous functioning level. 3. The eXpression of the difference between these two measurements in meaningful quantitative terms. At the present time a fully satisfactory solution has not been found for any of the above problems. The field of mental measurement has, however, established several instru- .ments that purport to reliably estimate the nature of the individual's present functioning. The major difficulties are encountered in the attempt to estimate the previous func- tioning level of the individual, and thus make a meaningful and quantitative analysis of any possible differences. The results of intelligence tests. The tremendous growth and expansion of the mental measurement movement dur- ing the period of World War I led into psychological surveys of extended age ranges. In one of the earliest studies, Jones and Conrad (#3), found a steady intellectual decline after a peak mental level of about 20 years. Their results indica- ted a decline to a fourteen year level at the age of fifty- ' "S‘s-D.- W‘Ffi'! '~’-..'" 2" r» ' j"- " ,3 W- b 9 five. Some of the most extensive research in this area was the series of Stanford University Studies on Later Maturity under the supervision of W.R. Miles (60)(6l)(62)(63)(6h). The results again showed a gradual decline during later years, but found wide individual differences. Beeson (6), found the average mental level of a group of aged individuals to fall within the "dull normal" level. Other studies by Rabin (68), Madonick (57), and Lorge (56), have shown.marked intellectual decline in old age. In general the results of intelligence tests given to many individuals at various periods throughout the life epan agree fairly well in showing that after reaching a peak of performance somewhere in the teens or twenties, there is a gradual decline in test scores from early to late maturity. The drOp in score between the ages of eighteen and eighty-five years has been found to be as high as 80 per cent of the orig- inal score. Weohsler has insisted that all mental abilities reach a maximum usually in the twenties and then begin to decline. He feels that this decline is at first very slow but soon increases perceptibly, and that, once started, it prOgresses continually (86). All of these investigations have shown that the amount of intellectual decline varies greatly with the ability being tested. The greatest amount of loss has been found on timed tests of any sort, and problems involving the flexibility of thinking, and the least loss on such tests as vocabulary and general information. There is a marked decline in the speed .. .--. -—- us" y'if“?~i~‘ .. :- " 9" v 5;: 10 of comprehension of a problem and much difficulty in compre- hending material that is entirely new or out of the indivi- dual's range of knowledge. The "Higher Mental Abilities". Scientific studies of the higher mental functions have demonstrated several interesting results. Reasoning and judgment, for example, are among the latest of all mental abilities to reach their peak, and likewise, are among the latest to decline (26)(27) (30)(32)(h9)(50). However, when speed and flexibility are included in the measurement of reasoning or judgment, decline is much more marked. The ability to form concepts has been shown to decline greatly in later maturity. The ability to learn new material and retain items al- ready committed to memory are often listed as common com- plaints of the aged individual. Investigations have also shown loss of this ability in later maturity (25)(26). One study of rote memory by Kubo (A9), found no serious decre- ment until the age of eighty-two, but several other invest- igators have shown a marked loss of both verbal and non- verbal memory and learning ability to old age (26)(32)(68) (75). These studies all point out that learning is labor- ious in the aged person, since he suffers from decreased flexibility of thinking. There are many examples in the literature of significant creative achievements during later maturity and old age, one indication of high mental level during this period. Dorland 11 has made several studies of the average age of men in various professions at the time of their most significant creative production, and has found this age to be a good deal older than commonly supposed (l8)(l9). W.R. Miles, tested imagin- ative responses to ambiguous stimuli and found no correlation between age in decades and creative imagination (60). Lehman and his associates have reported a series of investigations that showed significant contributions in various fields of science by individuals beyond1lm age of sixty (52)(53)(5h). Although some of the problems of gerontological research mentioned above are avoided when simply considering the-re- lationship between age and intelligence from the basis of the time of occurrence of outstanding examples of the indi- ,vidual's level of intellectual functioning, it is necessary to consider the possibility and even likelihood that the sig- nificant achievement of any individual has been maturing and developing over a considerable number of years. The results of Rorschach tests. Almost all testing of aged individuals with the Rorschach has indicated a decline in the speed of perceiving the ink blot stimuli as meaning- ful concepts. The amount of material encompassed in any single perceptive response tends to show a steady decline from decade to decade. Perceptional organization, as measured by the Rorschach test, shows important changes with older individuals. Old people tend to display marked tendencies to make false iden— , .. _ I . V , _ 7‘. .7 - T .w . _. . -- » . .. "7' ‘ 1 -,, . \ ‘4 ~ .ul 5- I ‘ ‘ k r . ‘u '1‘ Q‘— W W ‘ J . . ' . . .~ ’ .v '1- m‘. ‘f m “a. , W r. n x. 12 tification of objects and show a highly restricted child- like thought content demonstrated by a tendency to use ster- 'eotypes (66)(7h). Klopfer and others have stated that the Rorschach protocol of the aged individual is much like that of a child and that he tends to define things in terms of the mere utility of the object rather than in terms of its compo- sition or construction (h0)(h8). In a recent doctoral study, Rochwarg (72) found perceptual development undergoes a decline at senescence. The responses of the aged group seemed to lack appropriate organization and were characteristic of develop- mentally earlier levels. The.most complete Rorschach analysis of a normal aged group was carried out by Ames gt gl.(l). These investigators found the aged to be generally slower, less productive, and less efficient than younger persons. They feel that the aged are still able, for the.most part, to deal adequately with the practical problems of everyday life, but have a great deal of difficulty in organizing their experiences for future reference. They add that the aged person does not seem to be interested in abstract and theoretical things, or the formulation of original ideas, and indicate that the aged individuals produced many primitive intellectual mani- festations similar to those of children. They state, "As the subject deteriorates, his response becomes increasingly like thatcm a younger and younger person--both qualitatively and quantitatively". On the basis of a qualitative type of analysis of Rorschach factors, they grouped their subjects due imp: are' are 13 into the three classifications of "normal", "pre-senile", and "senile" (l). The Vocabulary Test The importance of vocabulary in determining how one person judges another has been heavily stressed recently in popular literature. Within the past few years there has been a steady stream of books dedicated to the aim of teaching an individual to improve his vocabulary. This surge of interest seems to be quite justified in considera- tion of the fact that a man's vocabulary can yield a great deal of information about himself. The skilled observer picks up such clues about the individual as his educational status, social status, cultural level, and approximate geographical area, all from the manner in which he speaks. The foremost position of vocabulary among all human char- acteristics by which a person is judged has been pointed out many times in scientific studies (l7)(55)(76). The measurement of vocabulary. It is never possible, of course, to obtain an exact quantification of an indivi- dual's vocabulary unless the individual is subjected to the impossible task of defining every known word. As in other areas of mental measurement, general sampling techniques are employed. Usually a given number of words are selected 11+ at. random from a standard dictionary with the assumption that. the individual's knowledge of these words will provide an index to his total vocabulary. In this fashion such major Vocabulary tests as those found. in the Stanford-Binet (81) and the Wechsler-Bellevue (86) were devised. Tests of vocabulary, for convenience, are usually class- 11" 16d in terms of the type of vocabulary they are designed ‘tica .measure, and in terms of the method of administration. The type of vocabulary tested may be either a general one Such as that found in the Stanford-Binet (78), the Wechsler- JESesllevue (86), the Columbia Vocabulary Test (12), and the wide Range Vocabulary Test (3), or specific tests designed ":<3 measure a person's vocabulary in specific fields, such 1lity to successfully organize one's thoughts into major coIleapts, has continually been shown to decline in later ma"Hilarity. Wechsler has suggested a system or (formula to indicate the amount of mental deterioration based upon a comparison 18 of scores on the sub-tests of the Wechsler-Bellevue Intell- igence Scale (86). He has divided his sub-tests into two groups--those which are expected to hold up with age, and The vocabulary sub-test, of course, The those which are not. is considered to be one of the best "Hold" scores. formula devised by Wechsler has been utilized by Rabin (68), Chesrow, Wosika and Reinitz (1h), Hunt (bl) Magaret and Simpson (58), and many others with quite conflicting results. Another popular test of mental deterioration, the Hunt- Minnesota Test for Organic Brain Damage, was devised utilizing the 1937 Stanford-Binet Vocabulary Test as a measure or the patient's original mental level and canparos this score with the score on tests of the formation and retention of new associations (#2). Studies utilizing this instrument have reported contradictory findings (12). Qualitative analysis of vocabulary. The scoring systems or most vocabulary tests assume that all correct answers are 01? approximately equal value and do not attempt to differen- tdiate between an excellently verbalized correct answer and a barely acceptable correct answer. The value of accounting for possible differences in the quality of word definitions has been emphasized repeatedly in the literature. Wechsler has stated: "Apart from its value as a measure of intelligence, the vocabulary is an especially desirable test to have on any scale because of its qualitative possibilities. In defin- ing a word a subject gives us more than its mere meaning. In many instances he tells us a good deal about himself, 19 or at least about the quality and character of his thought processes. These facts may not be immediately apparent, but can be brought to light frequently by analyzing the formal and contextual aspects of the subject's response" (86, p. 99). As early as 1901., Chambers (13) stressed his Opinion that in order to study the development of the use of language in young children, an analysis of the quality and growth of vocabulary was necessary. He tested the vocabulary of a large number of school children and found that he could classify all of the responses into four groups: (a) no answer; (b) wholly wrong answer; (0) vaguely right answer; rand (d) correct answer. His results indicated that word definitions of very young children tended to be vague and "a mere outline". Gray and holmes (31.) also noticed that the quality of word definitions changed as children grew <>JLder and suggested a qualitative analysis of vocabulary responses. Binet and Simon (8) reported on the results of their studies of intelligence testing with children. They found that children below the age of nine are prone to define words in terms of their function or usage, and Terman (81) has indicated identical findings. Kirkpatrick (1.7) discov- ered, in his work with children and adolescents, that defining a word by categorizing it or giving a synonym for it seemed to represent the highest level of develOpment. He found, 1t<>C>, that "descriptive" types of word definitions did not take place until a late developmental period. Watts (85) has suggested qualitatively grading vocabulary definitions 20 on a concrete--abstract continuum. Marx conducted a qualitative analysis of the first fifty words of the 1916 Stanford-Binet vocabulary. He correlated his findings with chronological age and discovered that the highest level of definition was one that categorized the word or gave a synomym for it. Descriptive types of respon- ses were also found to be of high quality, while use, example, and repetition types of responses were considered as inferior. Green (35) reported the results of her study in which 'she analyzed the vocabulary responses of large numbers of children and adults on a qualitative basis. She adopted a system of using weighted scores in accordance with the quality of the response and the deve10pmental level of the subject, and found that she could classify all of her data into the following five different levels of quality: (a) synonym (b) use and description (0) explanation (d) illustration, demonstration, and repetition (a) error By grading the responses according to this classification and assigning a differential score for each value of response, a higher correlation between this score and the total test IQ. could be obtained than was possible by using just the quantitative vocabulary score. Feifel and Lorge (23)(24) utilized the same system of 21 five-fold qualitative analysis and found highly similar re- sults. In their study older children significantly more often produced synonym types of responses, while younger children significantly more often gave inferior types of definitions. Yacorzynski took note of the qualitative differences possible in a vocabulary response and advanced the hypoth- esis that the often noted stability of vocabulary test scores in mental deterioration may be more apparent than real; this as a consequence of the fact that the usual score on a vocab- ulary test is calculated, independently of possible qualitative differences, on an "all-or-none" basis (88)(89). He argued that when responding to a vocabulary test a subject has a wide variety of acceptable responses available to him whereas in many other mental tests there is the possibility of only one correct answer. Certain physiological studies had indi- cated that many types of brain damage failed to destroy an ind :lvidual's ability to solve a problem but nevertheless Profoundly influenced the method by which the problem was SOlved, and Yacorzynski felt that it was most reasonable to assume that the mentally deteriorated individual would pro- duce vocabulary responses that were qualitatively much infer— 101‘ to those of the normal subject, even though both deterior- ated and normal. individual might receive the same "all-or- none" quantitative test score. Feifel (22) adopted the Yacorzynski hypothesis and attempted to verify it by testing a total of 370 normals and abnormals (consisting of schizophrenics, manic depres- 22 sives, and organics), aged 15 to 80 years, matched on the basis of age, education and quantitative vocabulary score. He utilized the five-fold classification schema of Green, which was mentioned earlier, and found that the abnormal subjects at all age ranges selected the inferior types of responses significantly more often than did the normal subjects, these differences being most marked in the respon- ses of older abnormals. Although his samples contained only a small number of cases within the upper age brackets, he :found a tendency for older normal individuals to use more inferior responses than young normals, but this difference was not statistically significant. Moran, Moran and Blake compared the vocabulary responses of normals and schizophren- ics with this same five-fold schema, utilizing a paper and pencil, multiple choice method. They did not find signi- ficant results between the two groups and felt that the differences obtained by Feifel may hold true for organics, but not for schiZOphrenics (65). Rabin, King, and Ehrmann (69) conducted a qualitative axItalysis of responses to the Ammon's Picture Vocabulary Test by groups of normals and schizaphrenics matched on the basis of age, sex and education. importance of considering the fact that schizophrenia is not These investigators stressed the a Strictly homogeneous group and that individual schizophren- ics may vary widely on such a major symptom as "contact with Accordingly, they utilized two groups of schizo- In reality". Db~3E‘enics-—a long term group and a short term group. 23 order to account for some of the bizarre verbalizations ob- t;ained from their schizophrenic groups, they devised the following qualitative scale: (a) Synonym or superior definition (b) Functional or descriptive (c) Concretistic (d) Correct, with irrelevant verbalization (e) Completely irrelevant Results using this scale showed significant differences taeetween normals and long-term schizophrenics, but no signif- icant differences between normals and short-term schizophren— ics. The authors conclude that with certain types of short- 1brmation in children by Reichard, Schneider, and Rappaport ('71). These investigators discovered that a child utilizes (3116 of three major stages in forming a concept: (1) the con- chretistic; (2) the functional; and (3) the conceptual. Ger- 81seim.pointed out that the same three stages could be utiliz- at: to classify definitional levels produced by verbal stimuli arui suggested using this as a three-fold system of qualitative analysis for vocabulary reaponses. She defined this qualita- tive system as follows: (a) Concretistic or descriptive--the most primitive 1i-Vpe of verbal definition implying memory of an object at a concrete sensory level. V (b) Functional or usage--a more complex method of 21+ verbalizing a definition--the subject recalls the use to which the object was put in the past. (c) Categorical or conceptual--the most abstract type of definition. Gerstein emphasized the relationship between the use of lan- guage as manifested by a vocabulary definition and the under- lying cognitive processes that may be revealed through a qual- itative analysis of these word definitions. Harrington and Ehrmann (38) utilized both a paper and pencil multiple choice vocabulary test and a verbally pre- sented vocabulary test analyzed according to the three-fold schema suggested by Gerstein and found significant results between normals and schiZOphrenics on the verbally presented vocabulary scale, but not on the multiple choice, paper and pencil test. These investigators suggest that the complexity of the type of vocabulary response required (simple indication or the correct item from a group of multiple choice answers as compared to producing a verbal definition to an orally Presented stimulus word) is important in the funtioning of schizophrenics. Stacey and Portnoy (80) conducted two studies ufiring Gerstein's suggested method of analysis. They tested groups of mentally retarded children and adults and compared them with normals, but found conflicting results. They agree that a qualitative analysis of vocabulary responses is im- Portant in attempting to differentiate between varying levels or intelligence but felt that their results indicated that "descriptive" types of word definitions were of a higher 25 level than were "functional" definitions. Chordorkoff'and Mussen (15) reported on the results of their study in which they utilized a slightly different type of qualitative vocabulary analysis. They felt that it might be profitable to separate the "descriptive" responses from the "functional" responses instead of classifying them at one level as most of the other workers in the area had done. Accordingly they utilized a four-fold scheme as follows: (a) Class or categorical (b) Descriptive (0) Example (d) Functional Utilizing a multiple choice version of this scale, they found significant differences between normals and schizophren- icsnthe normals choosing a greater number of categorical definitions, while the schiz0phrenics chose a greater number 01" example and functional reaponses. No significant differ- eflees were noted between the two groups in the frequency with which they chose descriptive definitions, but the normals c3"3-Ose descriptive definitims of high level while the schizo- Phrenics tended to choose descriptive definitions of a low 0 On ceptual level . THE PROBLEM The performance of deteriorated individuals on vocabulary tests has long been a controversial subject in psychological literature. Many investigators have shown that quantitative vocabulary score declines very little with the advent of old age , mental disorder, or organic pathology, while many other measures of mental functioning show marked impairment. It has been contended by some that the relative lack of impair- men-t; on this test provides the clinician with a long needed "index of previous functioning level" from which to judge the degree of mental impairment in other areas. Other inves- tigators have chosen to use the characteristic dispersion of 0t-her test scores around the allegedly stable vocabulary score as a device for the differential diagnosis of psycho- pathology. Several investigators have theorized that this stability of vocabulary test scores is due to the fact that Word meanings represent an example of "old learning" and as SUCh is one of the last types of mental functioning to be 1‘~'L‘1paired in mental deterioration. Other investigators have challenged the stable nature or vocabulary scores and contend that this alleged stability is simply an artifact of the instruments commonly used to measure word meaning. They contend that the reason the vo- 26 27 cabulary scores of deteriorated patients show little change is that there are many possible definitions of various levels of quality and, thus, the deteriorated individual is able to earn the same quantitative score as a normal individual, even though his responses are qualitatively much inferior. Qualitative analysis of vocabulary scores have repeatedly shown significant differences between the performances of nor- male and deteriorated individuals such as schizophrenics and organics. Some of the studies which have found discrepant results may be criticized on the basis of one or more of the fell owing: (a) Faulty criteria for selection of experimental 81‘°\-11:~s. (b) The total number of experimental subjects utilized was inadequate. (c) Lack of adequate matching of control and experi- mental groups on such important factors as age, education, and quantitative vocabulary performance. (d) The techniques utilized to present the vocabulary test to their subjects were not comparable to other studies. (6) The methods utilized for classification of the vonebulary reSponses were not comparable to the methods util- ized by others. Studies, mainly utilizing the Rorschach test, have found that perceptual organization undergoes a decline at later mel‘l'aurity. These studies show that the aged individual does not, seem to be interested in abstract reasoning and indicate 28 that his perceptual responses are qualitatively much inferior to those of a normal adult, but very similar to those of child- ren - The conceptual organization of the thought processes of the aged seems to be characteristic of a much earlier developmental level. Many investigators have pointed out that the mental dBterioration that occurs as a normal consequence of old age is not different from that which occurs as a consequence of Prolonged mental disease or some brain lesion. Consequently, they advocate defining a deteriorated individual as one who can no longer carry on his intellectual tasks with the speed, accuracy, or efficiency previously characteristic of his 1‘-“—1J£'.1<:1'.ioning level. Gerontologists have established the fact that among the mental functions that decline in later maturity 31‘6 such abilities as concept formation, discriminative judg- ment, flexibility of mental functioning, and verbal memory. Fa<‘:‘tor analytic studies of vocabulary tests have shown that it is precisely these same factors that make up the qualita- tive aspects of vocabulary responses. In consideration of the above findings it would seem most reasonable to make the assumption that a carefully coritrolled investigation of the qualitative aspects of vo- cabulary responses would demonstrate significant differences between groups of young normal and aged normal individuals. The present investigation proposes an intensive study of this 38neral assumption, utilizing a list of words selected from The Wechsler-Bells vue vocabulary scale. Based on the pre- 29 vious discussion and the review of the pertinent literature, the following general hypotheses are offered: Hypothesis I: There will be significant qualitative differences between the vocabulary performances of young and aged normal individuals. Hypothesis II: Aged normal individuals will utilize qualitatively inferior vocabulary responses significantly more Often than will younger normal individuals. Subjects The total pepulation of subjects selected for this Study consisted of an experimental group of 100 aged white males, aged 70 to 85, and a control group of 100 young white males, aged 20 to 35. All subjects were patients in the same Veterans Administration general-medical-surgical (GMBnS) hOSpital. All were classified as normal on the basis of haWing no history of emotional or behavior disorder as far as could be determined, and who in the estimation of the medical and psychological staff did not show any evidence or such a condition while receiving treatment at the hospital. In addition, any patients who presented a history or evidence or degenerative diseases or diseases seriously affecting the central nervous system were not included. The patients were S*elected from those with ailments that would be least likely t'0 affect performance on any general intellegence test battery. ‘v 30 After these initial criteria for selection were met, all of the patients were individually interviewed by the author. The general nature of the study and its purposes were briefly outlined, and a request was made for their Participation and cooperation. At this time all patients who did not express willingness to cooPerate and displayed no interest in the project were excluded. At this time, too a all patients with a bi-lingual background, or any indication of language handicap were excluded in order to rule out any affects such a background might have on their VOO ab ulary performances . Matching At the onset of this study it was hoped that the exper- jJuanital and control groups could be successfully matched on the basis of sex, intelligence, quantitative vocabulary 8901:6, and educational level. Sex was, of course, directly controlled by the selection of the subjects prior to the t'esting procedure . The level of intelligence was derived from the Wechsler- Bellevue Adult Intelligence Scale, Form I. This scale in its entirety was administered individually by the author to 3Sch of the subjects in the study. The intellectual level “as calculated on the basis of the Full Scale IQ, with the DIEnission of the Vocabulary sub-test, which is normally con- ‘U'_.L -‘vu‘e _ i___ 31 sidered as an alternative scale. In order to facilitate the final matching of the two groups, and to keep all subjects within practical variations of intelligence, it was decided to utilize only those patients who earned a Full-Scale Wech- sler—Bellevue IQ falling into Wechsler's normal classifications (90 to 110). Accordingly, the Wechsler-Bellevue Scale was administered as follows--the Verbal Scale subtests in their usual order of: Information, Comprehension, Digit Span, Ari-tohrnetic, and Similarities; this was followed by the Part ormance Scale subtests in their usual order of: Picture Arrangement, Picture Completion, Block Design, Object Assembly, and Digit Symbol. At this time a Full-Scale IQ, was calculated and if this IQ, did not fall into the range of 90 to 110, the sub.Ziect was not included in the study. If his score did fall 1[“50 the normal ranges of the IQ, the vocabulary sub-test was thSn administered in the usual way, except that each subject was presented with the entire list of words making up the VOGubulary scale, whereas the usual procedure of administra- tion is to continue down the list until 5 successive words have been failed, and then to count any remaining words as nutometic failures. All responses to the vocabulary test were copied in veI‘batim, and scored according to Wechsler's criteria. The raw scores were converted into Wechsler's weighted scores, in order to simplify matching, of the subjects-on this var- 181316. In order to determine the educational level of the 32 subjects , this study followed earlier investigations and credited either completion of or some attendance at a grade in establishing the number of years of schooling. Since the subjects for this study were not matched by pairs , mean scores for each of the variables were calculated fro(llil-ently and selection of succeeding patients was made accordingly. Table 1 presents all of the relevant information con- cerning the background of the two groups. It indicates that; the mean age of the young group is 25.16 years, the mean 383 Of the aged group is 75.9h years and that the two groups are Very well matched on the basis of intelligence, as meas- ured by the Wechsler-Bellevue IQ, and quantitative vocabulary score. Table 1 also indicates that the two groups did differ Significantly on the basis of their educational level. mateI’ling the two groups on this variable simply did not seem to be feasible, for almost invariably the normal aged indi- vidual received much less academic training than the young adult of today. When attempts were made to keep the mean eduea‘bional level of the two groups comparable, it was soon discovered that, by and large, only aged individuals of sup- erior intelligence completed the required amount of schooling. It Was decided that a close agreement between the two groups on intelligence test score might, therefore, assure better matching than would close agreement of their academic level. For the purpose of evaluating the "normalcy" of the .a" f I? A Fifi 4'3 TABLE I .AGE, EDUCATION, IQ, AND VOCABULARY MATCHING DATA FOR THE YOUNG AND AGED GROUPS Item (E13333 ($5330) figs Range 20-35 70-85 Mean 25.16 75.91. a“ 3.61. 4.21. 19, (W - B Full Scale) Range 90-110 90-110 mean 102.20 103.36 0’ 5.78 5.75 Education (In Years) Range 8-16 0-13 Mean 12.1.1; 8.12- a” 1.73 3.1.5 Vocabulary (W - B Weighted Score) Range 7-11; 7-11» Mean 10.80 10.86 4’ 2.21. 2.29 o ' I . I I. '- ,— ' ‘ . T .' ; k a." ‘ ‘ 5, q . w n a 3h educational level of the two samples utilized in this study, Table II presents a comparison of the mean academic level of the young and aged samples with the median educational achieve- ment of the U. S. pepulation of the same age. This table in- dicate s that the educational level of the two samples in this study is very similar to the educational level of the country at large, showing that the present samples were well selected on the basis of "normalcy". The Vocabulary Test Quantitative; score. The usual quantitative vocabulary Se Qre was computed from the responses given 130 the 1+2 items ma«k'iing up the vocabulary sub-test of the Wechsler—Bellevue In“telligence Scale. The vocabulary test from the Wechsler- B93.3.evue was selected for several reasons. The Wechsler SClale had been utilized in order to determine the intellec- tual level of the subjects, and the administration of its Ellternative sub-test, vocabulary, was a simple procedure Since this sub-test is presented on the usual test blank. Then, too, the reliability and validity of the Wechsler Vo- cabulary Test as a measure of the subject's "size" of vocab- u:Lary or "fund of word knowledge" has been repeatedly demon- 8trated in the literature. The test has been widely accepted and is, without a doubt, the most frequently utilized of all adult vocabulary scales. ”"" ‘~‘-' tsflaw' v- TABLE II COMPARISON OF THE EDUCATIONAL LEVELS OF THE YOUNG AND AGED GROUPS WITH THE U.S. POPULATION OF THE SAME AGE Educational Level Mean of present sample, age 20-35 12.1.1. Median of U.S. population, age 25-291 12.10 lMean of present sample, age 70-85 8.12 Median of U.S. population, age 75—792 8.20 Educational level of the nations population; School Life, 1953, No. 9, 138-139 36 Qualitative analysis. For purposes of the qualitative analys :L 8 itself, twenty words were finally selected. This selection was made from a larger number of words which had originally been compiled from the Wechsler vocabulary scale for a Dre-testing study. This larger list, thirty-five worélsn s VNas administered to approximately 100 individuals during the pro-testing stages. It was soon discovered, however, that many non-nouns did not seem to elicit a wide variety of qualitatively different responses, and so in the fiIl£353. list only nouns were used. In addition, several of the more difficult words toward the end of the list were failed 111th too frequently and also did not seem to permit a wide Var lety of responses and were thus eliminated. Some of the WC>3=Hiis originally selected presented serious scoring dif- fieulties, while others seemed quite ambiguous to many sub- jet>113 and all of these were likewise eliminated. Finally, tan-Eanw in this study only the following twenty words were s‘llbjected to a qualitative analysis: 1. apple 11. guillotine 2. donkey 12. nitroglycerine 3. diamond 13. stanza h. nuisance 1h. microscope 5. fur 15. belfry 6. cushion 16. affliction 7. bacon 17. pewter 8. nail 18. ballast 9. armory l9. catacomb 10. brim 20. Spangle A detailed analysis of vocabulary responses produced during the pre-testing period indicated that all of the responses could be classified according to the four-fold 37 qualitat, ive schema advanced by Chodorkoff and Mussen. This 801161118 seemed preferable to many others, for use in this StUdY. since it did not contain categories pertinent pri- marily to the responses of children and severly disturbed 5°h195c>i£>t1renics and not frequently encountered in normal adults. This Schema also separately categorizes "Descriptive" and "Functional" types of responses which have been shown to repre‘Sent wide differences in quality. «fixlthough one of the factors on which both the control and encperimental groups were matched was the score correct rece ived on the 1.2 item Wechsler-Bellevue vocabulary scale, it is evident, since’ only 20 words selected from this scale were qualitatively analyzed, that the groups could differ “id-61y in terms of the relative number of failures on these 8pecsific words. Accordingly, an error category was also utilized in order to compare the relative number of fail- ures for each group. The qualitative schema then, which ‘NElEB used in this study, was five-fold and consisted of the following categories:1 (a) Class or categorical responses-~the word was defined as a concept and placed into the class or category to which it belonged, or an adequate synonym was produced for it. (b) Descriptive responses-athe word was adequately dfiscribed in terms of its major properties. (c) Example responses--an adequate example of the 1Sample responses from both groups for all of the words qualitatively analyzed are offered in the appendix. 38 word Was offered. (d) Emotional responses--the word was defined in terms of its function or usage. (6) Error responses—-the word was incorrectly de— fined or no response was produced. Reliability of scoring. In order to obtain an es- timate of the reliability of scoring, every fourth protocol was scored by an independent rater. This independent rater was provided with a complete description of each of the 8col‘ing categories and instructions for scoring sample re- SDQrises. No other instructions or suggestions were given. All identification had previously been removed from the p:P'D‘lmcols. The scores given by the first examiner were, of course, unknown to the second. The scores for each protocol were then compared, item for item, to determine the degree of agreement. This agree- ment was initially computed at 97 per cent. However, a few mirlor errors in tabulation were encountered. The agreement “as again computed, this time at 96 per cent. This high Percentage of agreement between two independent scorers 1Iltlicates that the tests were scored in a reliable manner. Hypotheses. With reference to the outlined pro- cedure and the general hypotheses stated earlier, a reform- ulation of these hypotheses in specific operational terms 39 can now be offered. A - The young group will produce "Class or categorical" definitions significantly more often than the aged group. B - The aged group will produce "Descriptive" defini- tions significantly more often than the young group. C . The aged group will produce ”Example" definitions Significantly more often than the young group. D. The aged group will produce ”Functional" defini- tiona significantly more often than the young group. E. The two groups will not demonstrate significant differences in the frequency of their "Error" responses. RESULTS The tenability of five specific hypotheses concerning the qualitative nature of word definitions in an experimental group of aged normal males and a control group of younger normal males was examined in this study. Each of these hypo‘theses was examined by a t test of the significance of the difference between the means of the two groups. The t test furnished a measure of the validity of the null hypoth- 6318; i.e., a measure of the probability that an obtained 1'e’Sudt may be attributed to chance. A t level of 2.58 and hi gher seriously questions the validity of the null hypoth- 931 s, and the finding is regarded as highly significant. In each case, since the two groups were not matched pair by pair, the t formula for the difference between uncorrel- ated means was employed. A summary of the results obtained is provided in Tables III , IV, V, VI, and VII. In general, all but one of the hypotheses were confirmed by the data. This study indicates that older individuals do manifest qualitative deterioration in conceptualization, as measured by word definitions. The results are examined in detail, in reference to each of the 1‘ 1ve specific hypotheses . 1+0 1+1 Hypothesis A. The young group will produce "Class or categorical" definitions significantly more often than the aged group. Table III presents a comparison between the two groups in terms of the relative number of word definitions which were classified as "Class or categorical". This table in- dicates that the mean number of "Class or categorical" def- initions produced by the young group is 11.30 whereas, the mean for the aged group was only 9.56. The mean difference Of 1 -7h.was significant beyond the 0.5 percent level of confidence and may be interpreted as supporting hypothesis A. TABLE III COMPARISON OF THE TWO GROUPS ON THE "CATEGORICAL" TYPE OF RESPONSE C:11‘oup M / oJ-Diff t Significance Young 11.30 3.05 ‘ .l.15 t. .26 Beyond 0.5% Level Aged 9.56 2.68 1+2 The aged group will produce "Descrip- Hypothesis B. tive" definitions significantly more often than the young group. Table IV presents a comparison between the mean number or. definitions, produced by each group, which were classi- fied as "Descriptive". This table shows that the mean number of "Descriptive" responses for the aged group is 3 -71 and the mean for the young group is 2.95. This mean difference of 0.76 was significant beyond the 0.5 per cent level of confidence and may be interpreted as supporting hypothesis B. TABLE IV COMPARISON OF THE TWO GROUPS ON THE "DESCRIPTIVE" TYPE OF RESPONSE urn—— Group M 0/ (f-Diff t Significance Young 2.95 1.75 .287 3.3h Beyond 0.5% Level Aged 3.71 2.25 1+3 Hypothesis 0. The aged group will produce "Example" definitions significantly more often than the young group. Table V presents the relative number of word definitions Produced by each group which were classified as "Example". This table indicates that the mean number of "Example" definitions produced by the aged group was 1.19, while the mean number produced by the young group was 1.11.. tiii-is mean difference of 0.05 was insignificant. Thus the data of this study, since they demonstrate no significant difference between the two groups in terms of the relative Illlunber of "Example" definitions produced by each, do not 8 upport hypothesis 0 . TABLE v COMPARISON OF THE r110 GROUPS ON THE ’"EXAMPLE" TYPE OF RESPONSE Group M 0/ OJ-Diff t Significance Young 1 . 19 l . 33 \— . 191 . 26 Insignificant Aged l . 11+ . l . 37 1.1. Hypothesis D. The aged group will produce "Func- tional" responses significantly more often than the young group. Table VI indicates the comparative number of "Functional" responses produced by each group. This table shows that the mean number of "Functional" re5ponses produced by the aged group was 2.1.1 whereas the mean of the young group was only 1- - 1+4. This mean difference of 0.97 was significant beyond the 0.5 Per cent level of confidence and may be interpreted as supporting hypothesis D. TABLE VI COMPARISON OF THE TWO GROUPS ON THE "FUNCTIONAL" TYPE OF RESPONSE Group M 0’ JéDiff t Significance Young 1.1.1. 1.1.2 ‘— .278 3 .61. Beyond 0.5% Level Aged 2.41 2.21 \ 1+5 Hypothesis E. The two groups will not demonstrate Significant differences in the frequency with which they Produce "Error" responses. Table VII presents a comparison of the two groups on the frequency of "Error" responses. This table indicates that. the mean number of "Error" responses produced by the YOU-11g group was 3.11 while the mean of the aged group was 3- 18. The mean difference of 0.07 was insignificant. Thus, Since the data do not indicate any statistically signifi- cant. differences between the two groups, the results may be interpreted as supporting hypothesis E. TABLE VII COMPARISON OF THE TWO GROUPS ON THE ”ERROR" TYPE OF RESPONSE Group M 0’ {-Diff t Significance Young 3 . ll 2 . 06 ‘ . 215 .22 Insignificant Aged 3.18 2.51 A6 131 order to present a more graphic summary of the relative number of word definitions classified under each of 1311s: qualitative categories, a column graph was drawn show- ing the results for both the young and aged groups. This is Presented in Figure 1. This figure indicates that for the twenty word vocab- ‘¥1€i:?y'list the young group produced the following: 56.50 per cent "Categorical" definitions 1h.75 per cent "Descriptive" definitions 5.95 per cent "Example" definitions 7.20 per cent "Functional" definitions 15.60 per cent "Error" definitions This figure also indicates that for the twenty word ‘rc><=abulary list the aged group produced the following: h7.80 per cent "Categorical" definitions 18.55 per cent "Descriptive" definitions 5.70 per cent "Example" definitions 12.05 per cent *Functional" definitions 15.90 per cent "Error" definitions . I 1' “(w I: m, w. J‘YWEW.‘ . .HOHHH Sosa OBZH UZHHAfih mmbomo mfiom ho mHmZOmmHm mo BZMO mam .H .UHh ZOHBH9lfess were computed, of course, on actual test performance With no additional credit or correction being applied for age. I11 Eiddition, since both groups produced the same mean number or errors in defining the list of twenty specific words util- i1=5€36,for qualitative analysis, it seems certain that the c3-.‘1antitative vocabulary level of the two groups was the same. The findings of this study seriously challenge the basic 1:Jileories underlying current psychological tests of mental c1eteriorationnnamely, that vocabulary scores do not decline 5-r1 later maturity because vocabulary is an index or measure (>1? "old habits" and "early learning", in contrast to other Eit>ilities which show marked decline in later maturity and Etre theoretically assumed to involve more recent association. UDhe findings, rather, indicate that the stability of vocabu- lLary test scores in mental deterioration is an artifact of the usual scoring procedure. They support the Yacorzynski laypothesis that the aged individual can give a correct but inferior and "easier" response and thus not demonstrate his lowered efficiency; whereas a younger and non-deteriorated individual, with superior mental efficiency, will produce an equally correct, but more difficult and highly complex response. Further interpretation of the results of this study 51 leads to the conclusion that a vocabulary test score gives an eXtremely imperfect indication of an individual's early or Dre-deterioration level of functioning and that the test inatl‘mnents commonly utilized to measure mental deterioration PosSess extremely doubtful validity. It would appear then, that there is little Justification for future investigations 01' mental deterioration in which a quantitative vocabulary test. score is interpreted as an index of an earlier level or mental efficiency. It would seem much more Justifiable, in light of current knowledge, to utilize a vocabulary test, qualitatively analyzed, as providing some indication of the a(Btual loss of mental efficiency experienced by an individual, Slace his patterns of response in defining words have been Shown to be related to the impairment of mental functioning. Certainly the manner of approach and the method of pro- cedure used by an individual in defining a word are important in understanding his mental processes. Although a study of the qualitative level of his word definitions will indicate the presence and, to some extent, the degree of mental im— pairment, this indication will be very tentative and approx- imate unless a highly refined qualitative analysis is con- structed. The five-fold schema utilized in this study can be considered roughly, to represent five different levels of mental ability ranging. from superior to inferior. However, much caution must be applied in so interpreting the qualitative scale with results obtained from any single individual. Within each of the four correct categories a wide variation of qual- 52 it)? Of response is still quite possible. A class or categor- ical type of definition, which in general is quite superior to other types of definitions, can range from a simple syn- onym for the stimulus word to a highly complex genus type of definition in which the stimulus word is modified and pre- °iaely specified. A descriptive type of definition can range from an exact specification of the stimulus word in terms of its major properties, to a much inferior and limited descrip- ticn of the stimulus word in terms of certain of its very ccDrierete and relatively minor properties. It is necessary to consider that the level of quality of a word definition might possibly overlap somewhat for some of these qualitative categories. Actually, under special circumstances and in limited conditions, it is quite conceivable that a descrip- tive or functional type of definition would be superior to the more abstract and complex class type of definition. However, when considering large groups of individuals who are asked to verbally define a word, the qualitative cate- gories of 1) Class or categorical, 2) Descriptive, 3) Example, 1;) Function and 5) Error, may be considered to represent a scale of decreasing value of definition. The specific cat- egories, however, do not represent equal intervals of value along the scale. I In normal usage certain words tend to call forth spec- ific types of definitions. _Nothing is known, at present, about the exact frequency of the various qualitative types of word definitions the normal individual will produce. 53 Most frequently, of course, for the normal individual, are the superior definitions such as Class or categorical defin- itions. However, the more inferior types of word definitions will vary in terms of the frequency with which they are pro- duced, depending upon the Specific composition of the vocab- ulary list utilized. Since the vocabulary list utilized in this study consisted entirely of nouns that had been pre- v1Ously found to elicit a variety of different types of qualitative responses, no real conclusions pertinent to this question can be drawn from the data. Nevertheless, the relative percentages with which each category was elicited are of some interest. In this study both groups produced m“Binly superior definitions of the class or categorical type, with the younger group having a markedly greater percentage or these responses. Bothgroups produced a much smaller pro- Portion of inferior definitions of the descriptive and func- ‘t-ional types, with the aged group eliciting significantly high— er percentages of these. The percentage of functional defini- tions for the younger group was quite small. Both groups pro— C1L:.c:ed the same percentage of example types of definitions, but these definitions were very‘infrequent and pr0portionately I‘epresented ,the smallest of all categories for both groups. The relative proportions of descriptive and functional types of resPonses elicited by the aged group in this study My possibly have some relation to the Rorschach investigations Which have indicated that aged individuals tend to perceive things in a very practical and concrete light, and conceptu- \rm' 9"- firm-k hairy. 51+ alize stimuli in terms of their usage. However, in light or tdle current knowledge pertaining to the nature of decline in mental efficiency during later maturity, this tendency tC“NE1xwis descriptive and functional types of responses should be 1J1terpreted as signifying a decline in the ability to Su~f~3c.‘.eissfully cope with abstract and complex conceptualiza- tj-CDIis rather than indicating a specific kind of orientation or mental set of the aged individual toward his environment. A final word of caution toward interpretation of the findings of this study is offered. The aged individuals t’Ssmed in this investigation developed and matured in the cultural milieu existent approximately fifty year ago. Just how much this cultural environment differed from that of the Younger group is impossible to estimate. The matching data 8allowed, as one example, that educational levels for the two groups were vastly different. There is a possibility that c‘iifferences in the qualitative vocabulary definitions of the -liwo groups are, to some extent, a reflection of the many <3ultural changes that took place during this fifty year Period . In addition, the aged people have, for all practical purposes, been away from the formal academic environment for a considerable number of years, while a large propor- tion of the younger group have only recently completed their schooling. It is certainly possible, therefore, that the younger group in this study would be somewhat more prone or accustomed to offering verbal definitions ofs higher concep- tunl conceptual level. SUMMARY An attempt was made to substantiate the general hypoth- 6313 that vocabulary definitions of aged individuals, when aIlalyzed on a qualitative basis, will demonstrate marked intellectual decline. More specifically, it was assumed that aged normal individuals would produce word definitions clu-alitatively inferior to those produced by young normal individuals. An experimental group of 100 aged white males (Mean age 75 years) and a control group of 100 young white males (Mean age 25 years) were matched on the basis of intell- igence (Wechsler-Bellevue Full Scale IQ) and score correct on the Wechsler-Bellevue Vocabulary Test. All definitions for twenty nouns selected from this vocabulary test were ‘Subjected to a qualitative analysis consisting of five Categories: 1) Class or categorical, 2) Descriptive, 3) Example, 1.) Functional and, 5) Error. The reliability of this scale was tested by having an independent rater classify all definitions of every fourth protocol, fifty in all. Comparison of the ratings indicated 96 per cent agreement. The following five specific hypotheses were tested: 1) The young group will produce significantly 55 56 more Class or categorical definitions 2) The aged group will produce significantly more Descr 1ptive definitions 3) The aged group will produce significantly more Example definitions A) The aged group will produce significantly more Fume tional definitions 5) The two groups will not demonstrate significant differences in errors of word definitions. Obtained results substantiated four of the five hypoth- E3EBEBS. The young group produced a significantly greater I1‘JInber of Class or categorical definitions. The older gilfoup produced significantly greater numbers of Descriptive €1<3finitions and Functional definitions. Mean differences 1>