. I. ziwtulu .. . f bmflfi. $2.}... . 3...», 1...... .. J 6. .. g, be. 5‘ .203 . ”Ar: 1pm.} H.115“? L n . $8. H. . . ”mu .1. . .1 O IQ. .15” VT! 2. p A o .A . t. ‘3‘ 1.. .. {Vt \ .. , . . . . . , . a! 1 _, , .. . v v _ . .. . a . w .. . . ‘. .. . . . o. u ~ . I$_\_ . . o .. . all: h 11.2.... . 1 ¢........AVNI: .rq H. S... .892.” 1.5.... , . . _ k... 923%.... 11.....sz . . u u. . , ; ., .. . . .. 3’. £1 . .0. a... .. . \LMNSP 3 r3 .0 . v.1 . 9| . r s. :.¥¥xxarru1...nammm .hkfl . WA ‘HR .3)... 13.22.25, 1 am... .11.}: 3?... . u... ‘L I 3 r n . ~ .1 . U... many...) Lear...” ”.ufig .kn..iz.t¥.umwfi.i . .. Hung. . I ‘ . . v‘aL I It.- n l. “a . s)... .3. .v!)xsll?l.rtna.i.uv .. . . IIIhVnV|Q1pd8q .2... - 1... {ti 0 t {Hl't‘lxe g» .....!.....I.. (.1. huh“. . . . I. ...b.¢lI. I I A}... ..I!v..l..»0¢"hlo\, -. II. : ”z’fl’t’ I I . ‘ . .. xv . .C. . 117‘ I. ‘ la‘vwfii .a .. v A l x 11‘qu u. ti . pl .1 .V 12‘ 5. . .1! 3. L. 3:... Pvuv. nunxu.4..;.ziw #3:. .3... _ Ed \1 . , . Mr: I. I. 1!... ~ . . . . a ,. . I! , y t l (I. c . . .13.. 3 . ._ 1 I x .H‘ I I J t 5;le ”ANNE .. . . . . (Lyn 1 «8 rl. V II...» 1. 1.1.1111. .9! I.-- ‘ . . . .\ .\. riding“! n'tfipfi u .3“: shy-Ht“. ' $1.01“. a . .. . . t6 . iyklvtrrIIQ . A. i :Hnflzfiafizz ad). I - I . .nlavhrtllg;f.h. I! $3.1! i 11:”?! 5 i. 3.0.". :1. £A$!-Au f It‘ll" . I 1 $141... .1}; h: h. v. )1! I2! .1 n. p . Ifff .zil}. \! 1-1331,). Qifih. . 7...! _ 3.4!. . 0... Hr v... 15 3 ‘fi‘o‘l 1‘! 1...}- -22... 3.3.3 {$75.13; .33. an“ 3.. Rug... 3. (91...... .. . :7.) . .. .. futhuni, . \ v -. liner in... . a.) In»! I: I .. . Dawvri} . .. imfifi... an... an»... . .. . . . . . .. . 41.5.. 311 I} u. .. . .ll'fa. . i.p._x«o . . ..: Fill v E: . 3. r .r-......:....:. ., . . - A . .. , .. 412... l I .v I. 4 . . vnl... ' .. . in k..- \- .. 4. . . .p! -u- .. 2 . ., , haul...“ . .. . . nhjlgifiraifld nl‘vl . h... .,, UV“.M§. . a”. “0:93.... {\vaflnhuilnnr -.. H .1 .3 , . .I it. 3.. .. 0 J“ r . 3 C.-, 5% 333; unmawiu {flaw . S .. . 1 .. z .1 . .o 33.3.}..ihfinvfiié .. .t.-ivhuroffr...z “TWA “on” . . $3-...» 5121...“! 1-1.... It. BWMJB... and... .3 sh .fiohfifi!1\.?.1fl1cf s . l. t... :1}? . . . ‘9‘»)..‘58 . . hf], .13.... \a . It’lrl hnhL‘t ’Eiflhh! . . \ . s . 4. ..frr(l?(1.l. , . till \. at: .1. v...&-..lz.l ‘11.... , ‘.>O-'.?..’.KIIH . 3.7.4.5. , x . . Ifiuu 0...»: I :I'y..r.>.E.£ .. , .f {inuflg A. . .5. . Lui i t. . 7. 4.1.-. y}. (I. I... . avast-fungi; . v . ‘ :91}... .44 v. . I9. :13.“ §\.\4..J(>l)tltll4l.).l\ .. 11.32.113.97”: . , 1 3.1.1.3“) .. It , . . to)... 3‘}. . l: I .5. . I . 1 3.1!. vrii . . villi"; - . I»... . .V...Ll.ls.‘. . » .1... 'lr 01.2.}... - - :V..I|| . .. ’10.; I).I..\..!u. :1. m. i It; 1 fig“! R $3.171 i5!§f.rl($ its. If; t}..- a. “UNI! . . su-1‘.lll‘51.‘..n‘~.. . 3 [GI-3.. .7. 1 I .VTV. ii. .{i‘t lel. 5; +.P.:‘5L.X.. . . .. 1.31.... 'II l2.r\.ll,v.. ~|5 .. :2! 1.1.3» a gun“ Fhfirzhn itivhuyvrtltcbrto‘ {£3.49 . . ¢o$E§i¢tll . L 719.1; to. 3.. 1 v [.11. A... s. chit... .. x .1xf1..& .v ‘Axt Ir ~|!\,!rl‘l7v.rn!|. .vxc’v. . I by.’ .r kl‘l‘l I, sp’ll’l’ng .lx‘ltll’fluuf‘ 93.41.01.451 P2P. If... .P ._ f . r6. :lgi! . d1... .. 151'! ll l’ut‘ifiila . i.-- £.s.z..'fl\.n..fll;tti .. u trill. ll ’1.5~.~5DI (Ls? k» , .V f (it .; hv tun. ‘ Fifi... . R )9. .v .011\ '2’“; . LLIMXLJ 4:1“..“3. . .. .41....xi11. wfihfifvgne... 7..-: “Nu . p . urbnrwuixxlc-.akt.fllni€ ilki...§.ln1.}. L...+£.xar.él.v$.i .2... ... . nilG...i.ut§v . . idl.‘.c£;v.rl.?§. MI... -519... I». )3... .v. I r {i . a. . ... .1 (o. J. .. 1. .2111...» . r . .. 3.! -{Ils g... ... . _ ‘4... » .3...va I... [22... . .5, I. x. . A (I: L \~ It 5“ ‘1: ..L\. . . f. . I ion‘lliti? ‘.\Y:IR10{60«0|§.|>.1 i4-.. . . . .113! ......L;7.‘--.l.-l..< it!!! 2.1.!!! . . 5.1381? 1...)...Ifikin‘w . . . Sivnwl‘ttilflal. 9|... . . V . i . :1... . .1 RI}. . . . ~ . w ... . . o . . . . [(23, .. . : . . . . VIC.“ MICHIGAN STA HHHHH HHH 31310 HHHHHHHH HHH 928 HHH 0):; HH Inmmns This is to certify that the thesis entitled A COMPARISON OF WRITTEN DISCOURSE WITHIN FOUR ELEMENTARY SOCIAL STUDIES SERIES WITH REGARD TO SYNTACTIC COMPLEXITY, CONCEPTUAL DENSITY, CONCEPTUAL ABSTRACTNESS, AND VOCABULARY DIFFICULTY present tedb y Lynne Anne Harned has been accepted towards fulfillment of the requirements for PILD. deg-cc 11] Elementary and Special Education Major pro ssor MWAZM 07639 I - I“ 1c. 1 \ C 2-3,, m “.1 "H "-33,”! "4' v t ,‘fu- [Ivy 8 NWT '01! Michigan State LIBRAR Y University OVERDUE FINES: 25¢ per day per item RETURNING LIBRARY MATERIALS: Place in book return to remove charge from circulation records A COMPARISON OF WRITTEN DISCOURSE WITHIN FOUR ELEMENTARY SOCIAL STUDIES SERIES WITH REGARD TO SYNTACTIC COMPLEXITY, CONCEPTUAL DENSITY, CONCEPTUAL ABSTRACTNESS, AND VOCABULARY DIFFICULTY By Lynne Anne Harned A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Elementary and Special Education 1980 6/!‘1/‘772. C) Copyright by LYNNE ANNE HARNED I980 ABSTRACT A COMPARISON OF WRITTEN DISCOURSE WITHIN FOUR ELEMENTARY SOCIAL STUDIES SERIES WITH REGARD TO SYNTACTIC COMPLEXITY, CONCEPTUAL DENSITY, CONCEPTUAL ABSTRACTNESS, AND VOCABULARY DIFFICULTY BY Lynne Anne Harned Purpose The purpose of this study was to examine the written discourse within four elementary social studies series with regard to variables which may affect reading difficulty. More specifically, the purpose was to compare four aspects of language: syntactic complexity, conceptual density, conceptual abstractness, and vocabulary difficulty. Sample The series selected for inspection were Windows on Our World, Houghton Mifflin Company; The Holt Databank System, Holt, Rinehart and Winston; Concepts and Values, Harcourt, Brace, and Jovanovich; and Concepts and Inguiry, Allyn and and Bacon. From each of the four series, two levels of ma- terials, grade six and grade three, were examined. Lynne Anne Harned Procedure To analyze materials for syntactic complexity, concept density, and concept abstractness, five samples of ten con- secutive sentences were randomly selected from each text for each variable. Two variables which may affect syntactic complexity were tabulated for each sample: (I) prepositional phrases used as modifiers of nouns or verbs, and (2) complex sen- tences containing relative clauses which interrupt the Y subject—verb—object sequence of the independent clause. To determine concept density, the total number of concepts re- presented by words naming a person, place, animal, thing, quality, or idea was tabulated. To measure concept ab— stractness, the number of concepts represented by words naming a quality or idea which cannot be perceived by the senses was tabulated for each sample. To determine voca— bular difficulty, five samples of one hundred running words were selected from each text at each level. Vocabulary difficulty was measured by using The High Frequency Word List for Grades Three Through Nine and The American Heri- tage Word Frequency Book. Analysis of variance was used to test for differences among upper and lower texts and between upper and lower elementary levels. Post hoc comparisons permitted close inspections of the means. Lynne Anne Harned Findings Significant differences were found in: l. Syntactic complexity among the four series at each level and between levels within specific series as deter- mined by the number of prepositional phrase modifiers. 2. Concept density among the four series at each level and between levels within specific series. 3. Concept abstractness among the four series at the upper elementary level and between levels within specific series. 4. Vocabulary difficulty among the four series at the lower elementary level and between levels within specific series as determined by a word frequency list. 5. Vocabulary difficulty between levels.within spe- cific series as determined by a word frequency table. Conclusions This study demonstrates the feasibility of a procedure for examining social studies materials for specific syntac- tic, conceptual, and vocabulary factors. The study also indicates the need for such an analysis, since the instruc- tional materials which were examined did vary significantly in important factors which may affect readability. Lynne Anne Harned Recommendations 1. Textbook writers and publishers need to go beyond the use of readability formulas in analyzing social studies materials for factors which may contribute to reading diffi- culty. 2. Teachers need to learn how to examine instruction- al materials for syntactic, conceptual, and vocabulary fac- tors, and to adjust their instruction accordingly. 3. More research is needed to identify other features of written discourse which may contribute to reading diffi— culty. DEDICATION To Bud and Jill, who have brought me a lifetime of happiness ii ACKNOWLEDGEMENTS The writer wishes to express her most fond and sincere thanks to her chairperson and academic advisor, Dr. Lois A. Bader, whose outstanding professionalism, expertise, and trust greatly contributed to the successful completion of this doctoral program. The author is further indebted to Dr. Louise Sause, Dr. Ben Bohnhorst, and Dr. Eugene Pernell, members of the doctoral committee, for their guidance and support throughout this academic endeavor. A special thanks is extended to Dr. John Chapman for his interest and supportive friendship. The writer also expresses her deep appreciation to her parents, Mary and Jerome Piwarski, for their confidence, emotional support, and life~long guidance. Finally, the writer wishes to acknowledge with sincere and loving appreciation her husband, Bud, and her daughter, Jill, whose infinite patience, understanding, and love dur— ing the past five years will be treasured more than words can say. TABLE OF CONTENTS List of Tables CHAPTER I: THE PROBLEM Background . . Importance of the StUdy . Statement of Purpose Research Questions Delimitations Generalizability Definition of Terms Organization of Chapters CHAPTER II: RELATED LITERATURE AND RESEARCH Relationships Between Syntactic Structures and Comprehension . The Relationship Between COcnept Presen— tation and Comprehension in Social Studies Materials . The Relationship Between Word Frequency and Readability . . . Chapter Summary CHAPTER III: METHODOLOGY OF THE STUDY Introduction Population Selection of the Sample Procedure . Hypotheses . Method of Analyzing Data Summary . . CHAPTER IV: PRESENTATION AND ANALYSIS OF DATA Hypotheses and Statistical Tests The First Hypothesis . . The Second Hypothesis The Third Hypothesis The Fourth Hypothesis The Fifth Hypothesis The Sixth Hypothesis < H‘ NKOKOmU'IUI-L-‘H H H 14 21 29 37 .d The CHAPTER Seventh Hypothesis Eighth Hypothesis Ninth Hypothesis Tenth Hypothesis Eleventh Hypothesis Twelfth Hypothesis Thirteenth Hypothesis Fourteenth Hypothesis Fifteenth Hypothesis Sixteenth Hypothesis Seventeenth Hypothesis Eighteenth Hypothesis V: CONCLUSIONS AND RECOMMENDATIONS Introduction . . . . . . . Major Results and Discussions Implications ,. . . . . . Recommendations APPENDIX: Samples of Data Collection Sheets for Syn- tactic Complexity, Conceptual Density, Conceptual Abstractness, and Vocabulary DenSIty BIBLIOGRAPHY 96 10. LIST OF TABLES Analysis of Variance of Prepositional Phrase Modifiers at the Upper Elementary Level Homogeneous Subsets of Means of Pre- positional Phrase Modifiers at the Upper Elementary Level . . . Analysis of Variance of Prepositional Phrase Modifiers at the Lower Elementary Level Homogeneous Subsets of Means of Pre— positional Phrase Modifiers at the Lower Elementary Level . . . Analysis of Variance of Complex Sentences Containing Relative Clauses which In- terrupt the Subject-Verb—Object Se— quence of the Independent Clause at the Upper Elementary Level . Homogeneous Subsets of Means of Complex Sentences with Relative Clauses at the Upper Elementary Level Analysis of Variance of Complex Sentences Containing Relative Clauses which In- terrupt the Subject—Verb-Object Se- quence of the Independent Clause at the Lower Elementary Level . Homogeneous Subsets of Means of Complex Sentences with Relative Clauses at the Lower Elementary Level Analysis of Variance of Concept Density among the Four Series at the Upper Elementary Level . Homogeneous Subsets of Means of Concept Density at the Upper Elementary Level vi 49 50 50 51 52 52 53 54 55 55 .fl; ll. 12. l3. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. Analysis of Variance of Concept Density among the Four Series at the Lower Elementary Level . Homogeneous Subsets of Means of Concept Density at the Lower Elementary Level Analysis of Variance of Concept Abstract- ness at the Upper Elementary Level . Homogeneous Subsets of Means of Concept Abstractness at the Upper Elementary Level . . . . . . . . . . . . . . . Analysis of Variance of Concept Abstract- ness at the Lower Elementary Level Homogeneous Subsets of Means of Concept Abstractness at the Lower Elementary Level Analysis of Variance of Vocabulary Diffi- culty at the Upper Elementary Level as Determined by a Word Frequency List Homogeneous Subsets of Means of Vocabulary Difficulty at the Upper Elementary Level as Determined by a Word Fre- quency List . . . . . . . . Analysis of Variance of Vocabulary Diffi- culty at the Lower Elementary Level as Determined by a Word Frequency List Homogeneous Subsets of Means of Vocabulary Difficulty at the Lower Elementary Level as Determined by a Word Frequency List Analysis of Variance of Vocabulary Diffi- culty at the Upper Elementary Level as Determined by a Word Frequency Table Homogeneous Subsets of Means of Vocabulary Difficulty at the Upper Elementary Level as Determined by a Word Frequency Table Analysis of Variance of Vocabulary Diffi- culty at the Lower Elementary Level as Determined by a Word Frequency Table . Homogeneous Subsets of Means of Vocabulary Difficulty at the Lower Elementary Level as Determined by a Word Frequency Table vii 56 57 58 58 59 59 6O 61 61 62 63 63 64 64 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. Analysis of Variance Table for Significant Differences of Syntactic Complexity Between Levels as Determined by the Number of Prepositional Phrase Modi- fiers . . . . . . . . . . . . A Mean Summary Table of Syntactic Complex- ity Between Levels as Determined by the Number of Prepositional Phrase Modifiers . . . Analysis of Variance Table for Signifi— cant Differences of Syntactic Complex- ity Between Levels as Determined by the Number of Complex Sentences Con- taining Relative Clauses which Inter- rupt the Subject- -Verb- -0bject Sequence of the Independent Clause A Mean Summary Table of Syntactic Complex- ity Between Levels as Determined by the Number of Complex Sentences Containing Relative Clauses which Interrupt the Subject— —Verb- -Object Sequence of the Independent Clause . Analysis of Variance Table for Significant Differences in Concept Density Between Upper and Lower Elementary Levles A Mean Summary Table of Concept Density Be- tween Upper and Lower Elementary Levels Analysis of Variance Table for Significant Differences in Concept Abstractness Be— tween Upper and Lower Elementary Levels A Mean Summary Table of Concept Abstractness Differences Between Upper and Lower Ele— mentary Levels . Analysis of Variance Table for Significant Differences in Vocabulary Difficulty Be— tween Upper and Lower Elementary Levels as Determined by a Word Frequency List A Mean Summary Table of Vocabulary Difficulty Between Upper and Lower Elementary Levels as Determined by a Word Frequency List An Analysis of Variance Table of Vocabulary Difficulty Between the Upper and Lower Elementary Levels as Determined by. a Word Frequency Table . Vii]. 66 67 69 7O 71 72 73 75 76 77 79 36. A Mean Summary Table of Vocabulary Diffi- culty Differences Between Upper and Lower Elementary Levels as Determined by a Word Frequency Table . . . . . . . . 8O ix CHAPTER I THE PROBLEM Background Many reading authorities agree that students should be provided with instructional materials which match their reading ability levels. Generally, teachers are confronted with a number of obstacles in their attempts toward achiev- ing this goal. Teachers not only need to accurately deter- mine the students' levels of reading ability, but they also need to be aware of potential reading difficulties inherent in the texts and materials which will be utilized by stu— dents during instruction. The reading difficulty levels of instructional materials are of particular concern in the area of elementary social studies. In past years, there has been a growing amount of evi- dence indicating that elementary social studies texts may be too difficult for levels recommended by textbook compa— nies. Smith and Dechant (1961), upon summarizing a number of readability findings, reported that readability levels of one or two grades above designated grade levels were characteristic of elementary content area texts. Arnsdorf (1963), analyzing four basal social studies series, discovered that the range of difficulty within social studies texts was often greater than the range be- tween texts assigned to different grade levels. In a more recent study by Johnson and Vardian (1973), four readability formulas were utilized in the assessment of sixty-eight social studies texts from grades one through Six. Their findings indicated that the majority of the textbooks were appropriate for above average readers only. Similar results were obtained by Bader and Harned (1978). From their examination of four sets of elementary social studies materials, it was revealed that readability levels were generally higher than expected for most textsexamined. In reviewing the results of such studies, however, one might note that data have been compiled mainly through the use of readability formulas. While readability formulas may be useful in providing indices of difficulty, they usu- ally utilize basically one or two easily quantified vari- ables--generally word length and/or sentence length-—in es- timating the difficulty levels of the materials. This prac- tice has provided educators with no precise information re- garding the status of other variables which may affectread- ing difficulty levels of instructional materials. Dale and Chall (1949) have defined readability as the total of fill elements within a specific piece of printed material which affects the success of a group of readers may have with it. Chall (1956) cautioned that readability formulas should not be accepted as precise measures of reading difficulty, but rather as approximations, since they consider only limited aspects of difficulty. Harris (1969) reported in the Encyclopedia of Educa- tional Research that readability formulas are concerned mainly with components which are easily quantified but that investigators have noted that other more intangible factors also affect readability. It is suggested that factors such as conceptual difficulty and organizational structure of the materials be considered in assessing readability, .particularly of those materials dealing with highly abstract content, yet written in a deceptively simple style (p. 1073) Likewise, Lunstrum and Taylor (1977) note that a vari— ety of factors may influence the readability of a set of instructional materials, and that such factors ...include not only linguistic elements but also interest and purpose of the reader, format of materials, and so on (p. Because some of these factors are difficult to incorporate into formulas, Lunstrum and Taylor (1977) conclude: .there has developed a practice of utiliz- ing basically two simple variables-—word length and sentence 1ength--to derive estimates of the level of difficulty of printed materials (p. 15). Estes (1972) suggests that readability formulas do not provide any help in determining concept loading or inasses— sing the complexity and ambiguity of conceptual patterns, both of which may prove troublesome for students in compre— hending social studies materials. h-:A-__.xi‘\-—m—-Lz—" L -‘._.— Cohen (1975), using the cloze procedure, found some content area passages easier for junior high schoolstudents to comprehend than others, regardless of comparable reada- bility ratings. She notes the need for further research to: ...analyze the variety of linguistic struc— tures used in content textbooks, as the inter- action between "language" and ”information" poses distinctive problems for the reader (p. 250). In summary, the literature and research indicate that information about the reading difficulty levelcfifelementary social studies materials has been compiled mainly through the use of readability formulas, which generally utilizecnua or two easily quantified variables--word lenth and/or sen- tence length—-to estimate difficulty. This implies a need for further research which will elicit much more precise in- formation about other factors which may have a considerable effect on the reading difficulty levels of elementary so- cial studies materials. Importance of the Study The difficult reading levels of texts and instruction- al materials in the area of elementary social studies con- stitute a major problem for both teachers and students. While readability formulas have yielded information about some variables which affect readability—~namely word length and sentence length-—there have been few attempts to elicit precise data about the status of other influential varia- bles. It is anticipated that this study will provide evidence that precise data about many different variables which affect readability can be extracted from.elementary social studies materials. The author hopes that this in- formation will constitute a step toward the goal of produc- ing texts and other instructional materials which are more suitable for students in terms of reading difficulty levels Of equal importance is the possibility that the results from this study can be used to aid teachers in textbook se- lection and lesson preparation by creating an awareness of reading problems inherent in texts and instructional mater- ials. Statement of Purpose The purpose of this study was to demonstrate the feas- ibility of examining the written discourse within four so- cial studies series published for elementary school chil- dren with regard to the status of variables, other than word length and sentence length, which may affect the read- ing difficulty levels of instructional materials. More specifically, the purpose was to compare four aspects of the.language: syntactic complexity, conceptual density, conceptual abstractness, and vocabulary difficulty. From each of the four series, two levels of materials, upper elementary and lower elementary, were examined. Research QpestiOns The purpose of this study was to demonstrate the feas- ibility of comparing the written discourse within four elementary social studies series with regard to syntactic complexity, conceptual density, conceptual abstractness, and vocabulary difficulty. The major research questions are as follows: 1. Will the four series differ significantly in syntactic complexity at the upper ele- mentary level as determined by the number of prepositional phrase modifiers? 2. Will the four series differ significantly in syntactic complexity at the lower ele- mentary level as determined by the number of prepositional phrase modifiers? 3. Will the four series differ significantly in syntactic complexity at the upper elementary level as determined by the number of complex sentences containing relative clauses which interrupt the subject—verb—object sequence of the independent clause? 4. Will the four series differ significantly in syntactic complexity at the lower elemen- tary level as determined by the number of complex sentences containing relative clauses which interrupt the subject-verb-object se- quence of the independent clause? 5. Will the four series differ significantly in concept density at the upper elementary level? 10. ll. 12. 13. Will the four series differ significantly in concept density at the lower elementary level? Will the four series differ significantly in concept abstractness at the upper elemen- tary level? Will the four series differ significantly in concept abstractness at the lower elemen- tary level? Will the four in vocabulary mentary level quency list? Will the four in vocabulary mentary level quency list? Will the four in vocabulary mentary level quency table? Will the four in vocabulary mentary level quency table? series differ difficulty at as determined series differ difficulty at as determined series differ difficulty at as determined series differ difficulty at as determined significantly the upper ele- by a word fre- significantly the lower ele- by a word fre- significantly the upper ele- by a word fre— significantly' the lower ele- by a word fre- Will the upper and lower levels differ sig- nificantly in syntactic complexity as determined by the number of prepositional phrase modifiers? 14. Will the upper and lower levels differ sig— nificantly in syntactic complexity as deter- mined by the number of complex sentences con— taining relative clauses which interrupt the subject—verb—object sequence of the indepen- dent clause? 15. Will the upper and lower levels differ sig- nificantly in concept density? 16. Will the upper and lower levels differ sig- nificantly in concept abstractness? 17. Will the upper and lower levels differ sig- nificantly in vocabulary difficulty as deter— mined by a word frequency list? 18. Will the upper and lower levels differ sig— nificantly in vocabulary difficulty as deter- mined by a word frequency table? The above research questions were restated in null hy- pothesis form for statistical testing. These are presented in Chapter III. Delimitations 1. This study was limited to the examination of four social studies series which are widely used in Michigan schools. 2. The grade levels of the texts examined were limited to grade six, designated as an upper elementary level text; and, grade three, de— signated as a lower elementary text. 3. The syntactic factors for which each series was examined were limited to two: (a) prea positional phrase modifiers, and (b) complex sentences containing relative clauses which interrupt the subject-verb-object sequence of the independent clause. 4. The instruments used to determine vocabulary difficulty for each of the series were lim- ited to a word frequency list and a word fre- quency table. Generalizability The data from this study were drawn from only four ele— mentary social studies series. Further, only two levels from each series, level six and level three, were examined. Therefore, the findings from this study cannot be general- ized to all elementary social studies materials published for grades one through six. However, the findings may be suggestive of what may be found in samples of similar ma- terials. DefinitiOn of Terms In order that the reader may better understand this study, an explanation of relevant terms is provided. lO Syptax Syntax refers to the manner in which words are put to- gether to form the phrases, clauses, or sentences in a language. Syntactic Complexity Syntactic complexity refers to the difficulty of writ- ten materials with regard to specific syntactic variables which may negatively affect the readability of that mater- ial. Materials in this study were examined for two varia- bles, selected because they were found by Marcus (1971) to be among the most difficult for students to understand. 1. Prepositional phrases used as noun or verb modifiers. Example: Jillggave the letter on the table to Her mOther. 2. Complex sentences in which a relative clause interrupts the subject-verb-object sequence of the independent clause. Example: The boy who won the most prizes ran quidkly home. Concepts Carroll (1964) defines concepts as: ...the abstracted and often cognitively- structured classes of "mental" experiences learned by organisms in the course of their life histories (p. 80). This study is concerned with only those concepts which are represented by a word which names a person, place, animal, thing, quality, or idea. For the purposes of this study, 11 the investigator further distinguishes between the follow- ing types of concepts: 1. Concrete concepts are those which are repre- sented by a word which names a person, place, animal, thing, or class of things, having a physical existence. Concrete concepts are those underlined in the following sentences: Jim had never been on a train before. He redIIzed that the ways of a pullman dif- fered immensely from those of a farmhand. He knew, as he walked up the latform toward the engine, that he would maEe many blunders, but he counted on his speed of perception to see him through. On the latform stood an intimidating fellow, clothed in dirty jeans, a rain slicker, and leather boots. Jim thought the map to be of the lowest rank on board. But to his surprise, it was Charles Bennet, envineer of the diesel which would be headed for New Orleans. [‘0 Abstract concepts are those concepts which are represented by words naming a quality, abstraction, or idea; something which cannot be perceived by the senses; something having no physical existence. The abstract concepts are underlined in the following sentences: Jane had a re utation for absolute hon- estv. The thought that she would even aE:_ tempt to tell a lip seemed simply unbeliev- able. So, for her welfare, I decided to look into this accusation against Jane and abolish it. Conceptualize my surprise upon my g5— rival at her house, to find my friend in a mood of deceptiveness and evasion. Concept Density Concept density refers to the number of concrete and abstract concepts found within a designated sample of writ— ten material. 12 Vocabulary Difficulpy Vocabulary difficulty refers to the number of diffi- cult words found in a selected sample of written material as determined by a word frequency table or word frequency list. Word Frequencnyable The word frequency table used in this study is IRE American Heritage Word Frequency Book (1971) which was com- piled by John Carroll, Peter Davies, and Barry Richman. This is a detailed report of a word frequency study involv- ing over five million running words. In compiling the word frequency table, more thant ten thousand samples of five hundred words were selected from more than a thousand pub- lications that American school students are expected to read by assignment or voluntarily in grades three throUgh nine. Word Frequency List The word frequency list used in this study is a list of the one thousand words of the highest frequency in The American Heritage Word Frequency Book (1971). Organization of Chapters The content of Chapter I included a background of the problem, the importance of the study, the purpose of the study, research questions, the limitations of the study, the generalizability of the study, the definition of terms l3 pertinent to the study, and a presentation of the organiza- tion of subsequent chapters. In Chapter II, a review of the literature related to the study is presented. It includes sections on relation- ships between syntactic structures and comprehension, the relationship between concept presentation and comprehension in social studies materials, and the relationship between word frequency and readability. 'Chapter III describes the methodology used in this study. Chapter IV organizes, analyZes, and presents the data and findings of the study. Chapter V presents the conclusions, implications, and recommendations of the study as based on the findings. CHAPTER II RELATED LITERATURE AND RESEARCH The review of related literature and research in this chapter is organiZed under three major headings: (l) rela- tionships between syntactic structures and comprehension, A‘U . (2) relationship between concept presentation and compre— hension in social studies materials, and (3) relationship between word frequency and readability. Relationships Between Syntactic Structures and Comprehension In the past two decades, interest in syntactic struc- tures as they relate to reading comprehension has notice- ably increased. It appears that this occurrence is due primarily to the influence of transformational—generative grammar. Chomsky (1957, 1965), in his theory of transfor- mational-generative grammar, proposed that every sentence could be represented on two levels-—a surface structure level and a deep structure level. The surface structure level is the spoken or written form of a sentence; the deep structure level is the meaning representation ofzasentence. Syntax, or the arrangement of the smallest units of mean- ing, functions as an intermediary between sound and mean- ing. According to the theory of transformational-genera— tive grammar, a set of transformational rules provides for 14 15 the manipulation of syntax, and the complexity ofaasentence is.ascertained by the number of transformations required to reach the deep structure or meaning level. A number of studies has affirmed the idea that a sen- tence becomes more complex as more transformations are added. In a study by Miller (1962), sixty subjects were required to match sentences identical in meaning but varied in syntactic complexity. Findings indicated a relationship between the number of transformations and the reaction time involved in processing a sentence. Similar findings were obtained by Miller and.McKean (1964) where the timerequired to convert one type of sentence into another was measured. Sentences which were more complex syntactically required greater conversion time. In two studies by Gough (1965, l966),it was concluded that the time involved in the verification of a statement was affected by the number of transformations. Also, Fodor and Garret (1967) noted that sentences with fewer transfor- mations were more quickly and accurately paraphrased than those of greater complexity. Mehler, Bever, and Carey (1967) conducted a study in- volving the effects of syntactic structures on the eye movements of'a reader. AS subjects read sentences of vary- ing syntactic complexity, their visual scanning patterns were recorded. An examination of the readers‘ eye movement patterns revealed a relationship between grammatical struc- tures and duration of fixations. Also, Klein and Kurkowski 16 (1974), in a similar study involving the use of an eye move- 'ment cammera, noted that syntactic structures affected the number of eye movements produced by readers. As indicated by the aforementioned studies researchers have upheld the idea that a sentence becomes more complex as more transformations are added. In addition, there have been many investigations conducted to determine the surface structure variables which contribute to difficulties in reading comprehension. In a study with older subjects, Coleman (1962), using a cloze test, observed that technical passages divided into short sentences were significantly more comprehensible than those composed of longer sentences. Coleman and Blumenfeld (1963), also using older subjects and a cloze test measured (1) the comprehension of materials containing a high per- centage of abstract nouns nominalized from verbs, and (2) sentences where nominalizations were transformed to active verbs. Statistical data indicated sentences using active verbs were less difficult to comprehend than those using nominalizations of active verbs. Ruddell (1965) investigated the effect of the similar- ity of oral and written patterns of language structure on the reading comprehension of 131 fourth grade students. He found that reading comprehension secres on cloze tests which utilized high frequency patterns of oral language structure were significantly greater than scores over l7 materials that utilized low frequency patterns of oral language structure. Fagan (1971) conducted a study t0.05. No evidence of significant differences was found among the four series in the number of prepositional phrase modi- fiers at the upper elementary level (F==2.64l9,df==3,p>.05) The null hypothesis could not be rejected. To compare dif- ferences amongmeans, the Duncan post hoc analysis was run. An examination of Table 2 would indicate that the Allyn and Bacon series is notably different in the number of pre- positional phrases per sentence from the other three series. 50 Table 2. Homogeneous Subsets of Means of Prepositional Phrase Modifiers at the Upper Elementary Level (Duncan Procedure) Subset ‘Series Mean A Allyn and Bacon _ .8000 B Houghton Mifflin 1.2400 Holt, Rinehart, Winston 1.3137 Harcourt, Brace, Javanovich 1.3469 The Second Hypothesis Ho 2: There is no significant difference in syn- tactic complexity among the four series at the lower elementary level as determined by the number of prepositional phrase mod- ifiers. This hypothesis was analyzed by a one-way analysis of variance and Duncan's Multiple Range Test (see Tables 3 and 4). Table 3. Analysis of Variance of Prepositional Phrase Modifiers at the Lower Elementary Level Prepositional Phrase Modifiers at the Low- Mean F er Elementary Level quare df Ratio 2.7482 3 2.7218 p1<.05 There was a significant difference in syntactic com- plexity among the four series at the lower elementary level as determined by the number of prepositional phrasenwdifiers 51 (F = 2.7218, df = 3, p < .05). The null hypothesis could not be accepted. In order to clarify the nature of the differ— ences among means, a post hoc analysis was run. Table 4. Homogeneous Subsets of Means of Prepositional Phrase Modifiers at the Lower Elementary Level (Duncan Procedure) Subset Series Mean A Houghton Mifflin .5000 Harcourt, Brace, Jovanovich .7347 Holt, Rinehart, Winston .8600 B Harcourt, Brace,' Jovanovich .7347 Holt, Rinehart, Winston .8600 Allyn and Bacon 1.0600 Examination of Table 4 would indicate that the Houghton Mifflin series and the Allyn and Bacon series varied more from one another with regard to the average number of prepo- sitional phrase modifiers per sentence than from the other two series. The Third Hypothesis Ho 3: There is no significant difference in syn- tactic complexity among the four series at the upper elementary level as determined by the number of complex sentences contain- ing relative clauses which interrupt the subject-verb-object sequence of the inde- pendent clause. To evaluate this hypothesis, an analysis of variance was used. In addition, the Duncan post hoc analysis was run to allow for comparison among the means (see Tables 5 and 6) 52 Table 5. Analysis of Variance of Complex Sentences Containing Relative Clauses which Interrupt the Subject-Verb-Object Sequence of the Independent Clause at the Upper Elementary Level Complex Sentences with Relative Clauses at Upper Elementary Level Square d: Ratio .0400' 3 2.0632 p>.05 No evidence of significant differences was found in syntactic complexity among the four series at the upper ele- mentary level as determined by the number of complex sen— tences containing relative clauses which interrupt the sub— ject-verb-object sequence of the independent clause (F = 2.0632, df= 3, p) .05). The null hypotheses could not be re- jected. The Duncan post hoc analysis was run so that dif- ferences among the means might be observed. Table 6. Homogeneous Subsets of Means of Complex Sentences with Relative Clauses at the Upper Elementary Level (Duncan Procedure) Subset Sggies Mgpp A Harcourt, Brace, Jovanovich 0 Allyn and Bacon 0 Holt, Rinehart, Winston .0200 B Holt, Rinehart, Winston .0200 Houghton Mifflin .0600 53 The data in Table 6 suggests that at the upper elemen- tary level, the Houghton Mifflin series is more similar to the Holt, Rinehart, Winston series than to the remaining series, when considering the number of complex sentences with relative clauses Which interrupt the subject-verb- object pattern of the independent clause. The Fourth Hypothesis Ho 4: There is no significant difference in syntac- tic complexity among the four series at the lower elementary level as determined by the number of complex sentences containing re- lative clauses which interrupt the subject— verb-object sequence Of the independent clause. This hypothesis was statistically analyzed by an analy- sis of variance and Duncan's Multiple Range Test to show dif- ferences among the means (see Tables 7 and 8). Table 7. Analysis of Variance of Complex Sentences Containing Relative Clauses which Interrupt the Subject-Verb-Object Sequence of the Independent Clause at the Lower Elementary Level Complex Sentences with Relative Clauses at Mean F Upper Elementary Level Sguare df Ratio .0983 3 2.3221 PI>.05 No evidence of significant differences was found among the four series at the lower elementary level in the number of complex sentences containing relative clauses which 54 interrupt the subject-verb-object pattern of the independent clause (F= 2.3221, df= 3, p> 05). The null hypothesis could not be rejected. The Duncan test was used to allow for a closer inspection of the means. Table 8. Homogeneous Subsets of Means of Complex Sentences with Relative Clauses at the Lower Elementary Level (Duncan Procedure) Subset Series Mean A Allyn and Bacon 0 Holt, Rinehart, Winston .0200 Harcourt, Brace, Jovanovich .0600 B Holt, Rinehart, Winston .0200 Harcourt, Brace, ' Javanovich .0600 Houghton Mifflin .1000 Table 8 impliestflun:the Houghton Mifflin and the Allyn and Bacon series varied more from each other than from the other two series, in the number of complex sentences con- taining relative clauses which interrupt the subject-verb- object sequence of the independent clause. The Fifth Hypothesis Ho 5: There is no significant difference in concept density among the four series at the upper elementary level. An analysis of variance was used to statistically eval— uate this hypothesis. The Duncan post hoc analysis was run 55 to enable observation of differences among means (see Tables 9 and 10). Table 9. Analysis of Variance of Concept Density among the Four Series at the Upper Elementary Level Concept Density at the Upper Mean F Elementary Level Square df Ratio 8.6183 3 4.1311 p (.05 There was a significant difference in concept density among the four series at the upper elementary level (F = 4.1311, df= 3, p < .05). The null hypothesis couldnot be re- jected. In order to clarify these differences, Duncan's post hoc analysis was used; results are shown in Table 10. Table 10. Homogeneous Subsets of Means of Concept Density at the Uppe Elementary Level ' (Duncan Procedure) Subset Series Mean A I Allyn and Bacon 2.5400 Houghton Mifflin 2.6800 B Holt, Rinehart, Winston 3.2800 Harcourt, Brace, Jovanovich 3.3600 The data presented in Table 10 shows that the Allyn and Bacon series and the Houghton Mifflin series are similar 56 to one another in concept density. Likewise, the Holt, Rinehart, and Winston series is very similar totfluaHarcourt, Brace, and Jovanovich series with regard to concept density. The Sixth Hypothesis Ho 6: There is no significant difference in con- cept density among the four series at the lower elementary level. In analyzingthis hypothesis, an analysis of variance was used. In addition, Duncan's Multiple Range Test was used in studying differences among means (see Tables 11 and 12). Table 11. Analysis of Variance of Concept Density among the Four Series at the Lower Elementary Level Concept Density at the Upper Mean F Elementary Level Square df Ratio 11.0200 3 5.6184 p'<.05 There was a significant difference in concept density among the four series at the lower elementary level (F = 5.6184, df = 3, p <.05). The null hypothesis could not be accepted. In order to clarify the nature of the difference, Duncan's post hoc analysis was used. An examination of Table 12 indicates that the greatest difference in concept density was found between the Houghton Mifflin series and the Holt, Rinehart and Winston series. 57 Table 12. Homogeneous Subsets of Means of Concept Density at the Lower Elementary Level (Duncan Procedure) Subset Series Mean A Houghton Mifflin 1.9000 Harcourt, Brace, Jovanovich 2,3600 B Harcourt, Brace, Jovanovich 2,3600 Allyn and Bacon 2,5000 C Allyn and Bacon 2.5000 Holt, Rinehart, Winston 3.0400 An examination of Table 12 indicates that the greatest difference in concept density was found between the Houghton Mifflin series and the Holt, Rinehart and Winston series. The Harcourt, Brace and Jovanovich series and the Allyn and Bacon series are very similar in concept density at the low- er elementary level. The Seventh Hypothesis Ho 7: There is no significant difference in con— cept abstractness among the four series at the upper elementary level. This hypothesis was statistically treated using an ana- lysis of variance and Duncan's Multiple Range Test to allow for a comparison of differences among means (see Tables 13 and 14). There was a significant difference in concept abstract- ness among the four series in the upper elementary level tests (F = 3.8726, df=3, p .05). The null hypothesis could 58 Table 13. Analysis of Variance of Concept Abstractness at the Upper Elementary Level Concept Abstractness at the Upper Mean F Elementary Level Square df Ratio 6.1400 3 3.8726 p (.05 not be accepted. In order to note the nature of these dif— ferences, ZDuncan't post hoc analysis was run, the results of which are shown in Table 14. Table 14. Homogeneous Subsets of Means of Concept Abstractness at the Upper Elementary Level (Duncan Procedure) Subset Series ' Mean A Houghton Mifflin 1.0600 Holt, Rinehart, Winston 1.0600 Allyn and Bacon 1.2400 B Harcourt, Brace, Jovanovich 1.8000 An inspection of Table 14 suggests that the Harcourt, Brace, and Jovanovich series is notably different from the other three series in concept abstractness at the upper ele- mentary level. The Eighth Hypothesis Ho 8: There is no significant difference in con— cept abstractness among the four series at the lower elementary level. 59 This hypothesis was studied by using a one-way analysis of variance. To allow for a comparison of means, the Duncan procedure was used (see Tables 15 and 16). Table 15. Analysis of Variance of Concept Abstractness at the Lower Elementary Level Concept Abstractness at the Lower Mean F Elementary Level Square dfi Ratio 1.0983 3 1.0111 p >.05 No evidence of significant differences was found in concept abstractness among the four series at the lower ele- mentary level (F = 1.0111, df= 3, p 3 .05). The null hypothe- sis could not be rejected. Duncan's MUltiple Range Test was used to allow for an inspection of the means. Table 16. Homogeneous Subsets of Means of Concept Abstractness at the Lower Elementary Level (Duncan Procedure) Subset Series Mean A Holt, Rinehart, Winston .6600 Allyn and Bacon .8000 Harcourt, Brace, Jovanovich .9200 Houghton Mifflin 1.0000 60 The data in Table 16 suggest that all four series are very similar in concept abstractness in the lower elementary level texts. The Ninth Hypothesis Ho 9: There is no significant difference in vo- cabulary difficulty among the four series at the upper elementary level as deter- mined byja word frequency list. A one-way analysis of variance was used to test this hypothesis, and Duncan's post hoc analysis provided for an examination of the means (see Tables 17 and 18). Table 17. Analysis of Variance of Vocabulary Difficulty at the Upper Elementary Level as Determined by a Word Frequency List Vocabulary Difficulty/ Word Frequency List Mean F Upper Elementary Level Square d: Ratio .0744 3 .4545 p‘>.05 No evidence of significant differences was found among the four series in vocabulary difficult as determined by a word frequency list at the upper elementary level (F = .4545, df= 3, p >.05). The null hypothesis could not be re- jected. In order to observe any differences which might be present among means, the Duncan procedure for post hoc analysis was used. 60 The data in Table 16 suggest that all four series are very similar in concept abstractness in the lower elementary level texts. The Ninth Hypothesis Ho 9: There is no significant difference in vo- cabulary difficulty among the four series at the upper elementary level as deter- mined by,a word frequency list. A one-way analysis of variance was used to test this hypothesis, and Duncan's post hoc analysis provided for an examination of the means (see Tables 17 and 18). Table 17. Analysis of Variance of Vocabulary Difficulty at the Upper Elementary Level as Determined by a Word Frequency List Vocabulary Difficulty/ Word Frequency List Mean F Upper Elementary Level Square d: Ratio .0744 3 .4545 p'>.05 No evidence of significant differences was found among the four series in vocabulary difficult as determined by a word frequency list at the upper elementary level (F = .4545, df==3,13>.051 The null hypothesis could not be re- jected. In order to observe any differences which might be present among means, the Duncan procedure for post hoc analysis was used. 61 Table 18. Homogeneous Subsets of Means of Vocabulary Difficulty at the Upper Elementary Level as Determined by a Word Frequency List (Duncan Procedure) Subset Series Mean A Allyn and Bacon .7840 Holt, Rinehart, Winston .7877 Harcourt, Brace, Jovanovich .8020 Houghton Mifflin .8101 An examination of Table 18 shows that all series are very similar in vocabulary difficulty in upper elementary texts as determined by a word frequency list. The Tenth Hypothesis Ho 10: There is no significant difference in vo- cabulary difficulty among the four series at the lower elementary level as determined by a word frequency list. This hypothesis was statistically treated using an ana- lysis of variance and Duncan's Multiple Range Test (see Tables 19 and 20). Table 19. Analysis of Variance of Vocabulary Difficulty at the Lower Elementary Level as Determined by a Word Frequency List Vocabulary Difficulty/ Word Frequency List Mean F Upper Elementary Level Square df Ratio 1.2598 3 8.9750 p<.05 62 There was a Significant difference in vocabulary diffi- culty at the lower elementary level as determined by a word frequency list (F = 8.9750, df = 3, p <.05). The null hypo- thesis could not be accepted. To clarify the nature of the differences, Duncan't post hoc analysis was run. The re- sults are reported in Table 20. Table 20. Homogeneous Subsets of Means of Vocabulary Difficulty at the Lower Elementary Level as Determined by a Word Frequency List (Duncan Procedure) Subset Series _ Mean A Harcourt, Brace, Jovanovich .7900 Holt, Rinehart, Winston .8000 Allyn and Bacon .8240 B Houghton Mifflin .9012 Examination of Table 20 indicates that the Houghton Mifflin series is notably different from the other three series in vocabulary' difficulty at the lower elementary level..as determined by a word frequency list. The Eleventh Hypothesis Ho 11: There is no significant difference in vo- cabulary difficulty among the four series at the upper elementary level as determined by a word frequency table. In analyzing this hypothesis, a one-way analysis of variance was used. Also, Duncan's post hoc analysis was used to allow for an inspection of the means (see Tables 21 and 22). 63 Table 21. Analysis of Variance of Vocabulary Difficulty at the Upper Elementary Level as Determined by a Word Frequency Table Vocabulary Difficulty/ Word Frequency Table Mean F Upper Elementary Level Square 9: Ratio 322670574.8226 3 1.8665 p >.05 No evidence of significant differences was found among the four series in vocabulary difficulty at the upper ele- mentary level as determined by a word frequency table (F = 1.8655, df= 3, p) .05). The null hypothesis could not be re- jected. The Duncan post hoc analysis was used to note any differences among the means. Table 22. Homogeneous Subsets of Means of Vocabulary Difficulty at the Upper Elementary Level as Determined by a Word Frequency Table (Duncan Procedure) Subset Series Mean A Harcourt, Brace, Jovanovich 5639.1412 Houghton Mifflin 6157.0673 Allyn and Bacon 7035.3747 Holt, Rinehart, Winston 7388.6660 The data shown in Table 22 indicates that the four ser- ies are very Similar in vocabulary difficult at the upper elementary level as determined by a word frequency table. 64 'The Twelfth Hypothesis Ho 12: There is no significant difference in vo- cabulary difficulty among the four series at the lower elementary level as deter- mined by a word frequency table. 1‘An analysis of variance was used in evaluating this hypothesis. In addition, Duncan's Multiple Range Test per- mitted a close inspection of differences among the means (see Tables 23 and 24). Table 23. Analysis of Variance of Vocabulary , Difficulty at the Lower Elementary Level as Determined by a Word Frequency Table Vocabulary Difficulty/ Word Frequency Table Mean F Upper Elementary Level Square dfi Ratio 330057239.8044 3 1.1965 p>.05 No evidence of Significant differences was found among the four series in vocabulary difficulty at the lower ele- mentary level as determined by a word frequency table (F = 1.1965, df==3,]p>.051 The null hypothesis could not be re- jected. The Duncan Multiple Range Test was run to allow for an inspection of the means. An igspection of Table 24 reveals that all the series are very similareto one another in vocabulary difficulty at the lower elementary level as determined by a word frequency table. 65 The Thirteenth Hypothesis HO 13.0: Ho Ho Ho Ho 13. 13. 13. 13. l: 2: 3: 4: There is no significant difference in syntactic complexity between the upper and lower elementary level as determined by the number of prepositional phrase modifiers. There is no significant difference in syntactic complexity between the upper and lower elementary level of the Hough- ton Mifflin series as determined by the number of prepositional phrase modifiers. There is no significant difference in syntactic complexity between the upper and lower elementary level of the Holt, Rinehart, and Winston series as deter- mined by the number of prepositional phrase modifiers. There is no significant difference in syntactic complexity between the upper and lower elementary level of the Har— court, Brace, and Jovanovich series as determined by the number of preposition- al phrase modifiers. There is no significant difference in syntactic complexity between the upper and lower elementary level of the Allyn Bacon series as determined by the num- ,ber of prepositional phrase modifiers. The hypotheses were statistically treated using an ana— lysis of variance (see Table 25). There was a significant difference in syntactic com- plexity between the upper and lower elementary levels as de- termined by the number of prepositional phrase modifiers (F = 6.5000, df = l, p <.05). The null hypothesis was re- jected. There was a significant difference in syntactic complexity between the two levels of the HoughtOn Mifflin series as determined by the number of prepositional phrase 66 Table 25. Analysis of Variance Table for Significant Differences of Syntactic Complexity Between Levels as . Determined by the Number of Prepositional Phrase Modifiers Mean F Series Square df Ratio Entire population 7.4334 1 6.5000* Houghton Mifflin 13.6900 1 15.3118* Holt, Rinehart, Winston 5.1976 1 3.5983 HarcOurt, Brace, Jovanovich 4.5927 1 4.6580* Allyn and Bacon 1.6900 1 1.4945 7fi>(.05 modifiers (F==15.3118, df= 1,1>(.051 The null hypothesis could not be accepted. No evidence of significant differ- ences was found in syntactic complexity between the two le- vels of the Holt, Rinehart and Winston series as determined by the number of prepositional phrase modifiers (F==3.5983, df= l, p>.05). The null hypothesis could not be rejected. There was a significant difference in syntactic complexity between the two levels of the Harcourt, BraceanuiJovanovich series as determined by the number of prepositional phrase modifiers (F==4.6580, df==1,1p<.05) The null hypothesis could not be accepted. No evidences of significant differ- ences was found in syntactic complexity between the two le- vels of the Allyn and Bacon series as determined by the 67 number of prepositional phrase mOdifiers (F==1.4945, df==1, p.>.05). The null hypothesis could not be rejected. To allow for the observation of specific differences between means, a mean summary table is provided (Table 26). Table 26. A Mean Summary Table of Syntactic . Complexity Between Levels as Determined by the Number of Prepositional Phrase Modifiers Series Level Sum, Mean Entire population Upper 235.0000 .1750 Lower 157.0000 .7889 Houghton Mifflin Upper 62.0000 1.2400 Lower 25.0000 .5000 Holt, Rinehart, Winston Upper 67.0000 1.3137 Lower 43.0000 .8600 Harcourt, Brace, Jovanovich Upper 66.0000 1.3469 Lower 36.0000 .7347 Allyn and Bacon Upper 40.0000 .8000 ' Lower 53.0000 1.0600 The Fourteenth Hypothesis HO 14.0: HO 14.1: There is no significant difference in syntactic complexity between the upper and lower elementary levels as deter- mined by the number of complex sentences containing relative clauses which inter- rupt the subject-verb-object sequence of the independent clause. There is no significant difference in syntactic complexity between the upper and lower elementary levels of the Houghton Mifflin series as determined by the number of complex sentences containing relative clauses which inter- rupt the subject-verb-object sequence of the independent clause. 68 Ho 14.2: There is no significant difference in syntactic complexity between the upper and lower elementary levels of the Holt, Rinehart, and Winston series as deter- mined by the number of complex sentences containing relative clauses which inter- rupt the subject-verb-object sequence of the independent clause. Ho 14.3: There is no significant difference in syntactic complexity between the upper and lower elementary levels of the Har- court, Brace, and Jovanovich series as determined by the number of complex sen- tences containing relative clauses which interrupt the subject-verb-object se- quence of the independent clause. Ho 14.4: There is no significant difference in syntactic complexity between the upper and lower elementary levels of the Allyn and Bacon series as determined by the number of complex sentences containing relative clauses which interrupt the subject-verb-object sequence of the in- dependent clause. To analyze these hypotheses, an analysis of variance was used (see Table 27). No evidence of Significant differences was found in syntactic complexity between the upper and lower elementary levels as determined by the number of complex sentences con- taining relative clauses which interrupt the sUbject-verb— object sequence of the independent clause (F==1.9875, df==1, 1)).05) The null hypothesis could not be rejected. No evi- dence of significant difference was found in syntactic com- plexity between the upper and lower elementary levels of the Houghton Mifflin series as determined by the number of com- plex sentences containing relative clauses which interrupt the subject-verb-object pattern of the independent clause 69 Table 27. Analysis of Variance Table for Significant Differences of Syntactic Complexity Between Levels as Determined by the Number of Complex Sentences Containing Relative Clauses which Interrupt the Subject-Verb-Object Sequence of the Independent Clause Mean F Series Square df Ratio Entire population .0625 1 1.9875 Houghton Mifflin .0400 1 .5355 Holt, Rinehart, Winston .0000 l .0000 Harcourt, Brace, Jovanovich .0900 1 3.1277 Allyn and Bacon .0000' 0 .0000 p.>.05 (F= .5355, df= 1, p>.05). The null hypothesis could not be rejected. No evidence of significant differences was found in syntactic complexity between the upper and lower elemen- tary levels of the Holt, Rinehart and Winston series as de- termined by the number of complex sentences containing rela— tive clauses which interrupt the subjectiverb-ODjectpattern of the independent clause (F=O, df=0, p>.05). The null hypothesis could not be rejected. No evidence of Signifi- cant differences was found in syntactic complexity between the upper and lower elementary levels of the Harcourt,Brace and Jovanovich series as determined by the numberwxfcomplex sentences containing relative clauses which interrupt the subject-verb-object sequences of the independent clause G?= 70 3.1277, df= l, p ) .05). The null hypothesis could not be re- jected. No evidence of significant differences was found in syntactic complexity between the upper and lower elemen- tary levels of the Allyn and Bacon series as determined by the number of complex sentences containing relative clauses which interrupt the subject-verb-object sequence of the in- dependent clauses (F= .0000, df= 0, p) .05). The null hypo- thesis could not be rejected. 30 that specific differences between means may be no- ted, a mean summary table is presented in Table 28. Table 28. A Mean Summary Table of Syntactic Complexity Between Levels as Determined by the Number of Complex Sentences Containing Relative Clauses which Interrupt the Subject-Verb-Object Sequence of the Independent Clause Series Level Sum Mean Entire population Upper 4.0000 .0200 Lower 9.0000 .0450 Houghton Mifflin Upper 3.0000 .0600 Lower 5.0000 .1000 Holt, Rinehart, Winston Upper 1.0000 .0200 Lower 1.0000 .0200 Harcourt, Brace, Upper .0000 .0000 Jovanovich Lower 3.0000 .0600 Allyn and Bacon Upper .0000 .0000 Lower .0000 .0000 The Fifteenth Hypothesis Ho 15.0: There is no significant difference in concept density between the upper and lower elementary levels. 71 Ho 15.1: There is no significant difference in concept density between the upper and lower elementary levels of the Hough- ton Mifflin series. Ho 15.2: There is no significant difference in concept density between the upper and lower elementary levels of the Holt, Rinehart, and Winston series. Ho 15.3: There is no significant difference in concept density between the upper and lower elementary levels of the Har- court, Brace, and Jovanovich series. Ho 15.4: There is no significant difference in concept density between the upper and lower elementary levels of the Allyn and Bacon series. These hypotheses were evaluated statistically by using an analysis of variance (see Table 29). Table 29. Analysis of Variance Table for Significant Differences in Concept Density Between Upper and Lower Elementary Levels Mean F 'Series ' Square df Ratio Entire population 26.5225 1 12.3859* Houghton Mifflin 15.2100 1 8.3096* Holt, Rinehart, Winston 1.4400 1 .6031 Harcourt, Brace, Jovanovich 25.0000 1 12.4340“ Allyn and Bacon .0400 l .0214 *p:<.05 There was a significant difference in concept density between the upper and lower elementary levels (F = 12.3859, 72 df= 1, p< .05). The null hypothesis could not be accepted. There was a significant difference in cOncept density bee tween upper and lower elementary levels of the Houghton Mif- flin series (F= 8.3096, df= l, p <.05). The null hypothesis could not be accepted. No evidence of significant difference was found in concept density between upper and lower elemen- tary levels of the Holt, Rinehart and Winston series (F = .6031, df==l,15>.051 The null hypothesis could not be re- jected. There was a significant difference in concept den- sity between upper and lower elementary levels of the Har- court, Brace and Jovanovich series (F==12.4340, df==1, p< .05). The null hypothesis could not be accepted. No evi- dence of significant differences was found in concept den- sity between upper and lower elementary levels of the Allyn and Bacon series (F= .0214, df= l, p) .05). The null hypo— thesis could not be rejected. In order to permit an examination of the means, a mean summary table is included (Table 30). Table 30. A Mean Summary Table of Concept Density Between Upper and Lower Elementary Levels Series Level Sum I Mean Entire population Upper 593.0000 2.9650 Lower 490.0000 2.4500 Houghton Mifflin Upper 134.0000 2.6800 Lower 95.0000 1.9000 Holt, Rinehart, Winston Upper 164.0000 3.2800 Lower 152.0000 3.0400 Harcourt, Brace, Upper 168.0000 3.3600 Jovanovich Lower 118.0000 2.3600 Allyn and Bacon Upper 127.0000 2.5400 Lower 125.0000 2.5000 73 The Sixteenth Hypothesis HO 16.0: HO 16.1: HO 16.2: HO 16.3: HO 16.4: There is no Significant difference in concept abstractness beteween the upper and lower elementary levels.. There is no significant difference in concept abstractness between the upper and lower levels of the Houghton Mifflin series. There is no significant difference in concept abstractness between the upper and lower levels of the Holt, Rinehart, and Winston series. There is no significant difference in concept abstractness between the upper and lower levels of the Harcourt, Brace, and Jovanovich series. There is no Significant difference in concept abstractness between the upper and lower levels of the Allyn and Bacon series. An analysis of variance was used to evaluate these hy- potheses (see Table 31). Table 31 . Analysis of Variance Table for Significant Differences in Concept Abstractness Between Upper and Lower Elementary Levels Mean F Series Square df Ratio Entire population l9}8025 1 14.4513* Houghton Mifflin .0900 l .0618 Holt, Rinehart, Winston 4.0000 1 3.8416 Harcourt, Brace, Jovanovich 19.3600 1 10.2180* Allyn and Bacon 4.8400 1 5.0936* 7'~‘p<.05 74 There was a significant difference in concept abstract— ness between the upper and lower elementary levels (F = 14.4513, df= 1,1) .05) The null hypothesis could not be accepted. No evidence of significant differences was found in concept abstractness between the upper and lower elemen- tary levels of the Houghton Mifflin series (F==.0618, df==1, I) .05) The null hypothesis could not be rejected. No evi- dence of significant differences was found in concept ab- stractness between the upper and lower elementary levels of the Holt, Rinehart, and Winston series (F==3.8416, df==1, I) .051 The null hypothesis could not be rejected. There was a significant difference in concept abstractness between the upper and lower elementary levels of the Harcourt, Brace, and Jovanovich series (F==10.2180, df==1, p .05). The null hypothesis could not be accepted. There was a sig- nificant difference in concept density between the upper and lower elementary levels of the Allyn and Bacon series (F-5.0936, df= 1,1) .05) The null hypothesis could not be accepted. A summary table of means (Table 32) is presented in order that exact differences between means may be observed. Table 32. A.Mean Summary Table of Concept Abstractness Differences Between Upper and Lower Elementary Levels. Series Entire population Houghton Mifflin Holt, Rinehart, Winston Harcourt, Brace, Jovanovich Allyn and Bacon Level Sum Mean Upper 258.0000 1.2900 Lower 169.0000 .8450 Upper 53.0000 1.0600 Lower 50.0000 .0000 Upper 53.0000 1.0000 Lower 33.0000 .6600 Upper 90.0000 1.8000 Lower 46.0000 .9200 Upper 62.0000 1.2400 Lower 40.0000 .8000 The Seventeenth Hypothesis Ho 17.0: Ho Ho Ho Ho 17. 17. l7. 17. l: 2: 3: 4: There is no significant difference in vocabulary difficulty between the upper and lower elementary levels as deter- mined by a word frequency list. There is no significant difference in vocabulary difficulty between the upper and lower elementary levels of the Hough- ton Mifflin series as determined by a word frequency list. There is no significant difference in vocabulary difficulty between the upper and lower elementary levels of the Holt, Rinehart, and Winston series as deter- mined by a word frequency list. There is no significant difference in vocabulary difficulty between the upper and lower elementary levels of the Har- court, Brace, and Jovanovich series as determined by a word frequency list. There is no significant difference in vocabulary difficulty between the upper and lower elementary levels of the Allyn 76 and Bacon series as determined by a word frequency list. These hypotheses were studied statistically by using an analysis of variance (see Table 33). Table 33. Analysis of Variance Table for Significant Differences in Vocabulary Difficulty Between Upper and Lower Elementary Levels as Determined by a Word Frequency List Mean F Series Square df Ratio Entire population .5823 1 3.8142* Houghton Mifflin 1.05.94 1 8. 7264* Holt, Rinehart, Winston .0380 1 .2289 Harcourt, Brace, Jovanovich .0360 l .2213 Allyn and Bacon .4000 l .1113 7"p < .05 There was a significant difference in vocabulary diffi- culty between upper and lower elementary levels as deter- ndned by a word frequency list (F==3.8142, df==1, p<.05). The null hypothesis could not be accepted. There was a sig- nificant difference in vocabulary difficulty between upper and lower elementary levels of the Houghton Mifflin series as determined by a word frequency list (F==8.7264, df = 1, r><.051 The null hypothesis could not be accepted. No evi- dence of significant differences was found in vocabulary difficulty between upper and lower elementary levels of the 77 Holt, Rinehart, and Winston series (F==.2289, df==1, p .05). The null hypothesis could not be rejected. No evi— dence of significant differences was found in vocabulary difficulty between the upper and lower elementary levels of the Harcourt, Brace, and Jovanovich series as determined by a word frequency list (F= .2213, df=l, p .05). The null hypothesis could not be rejected. No evidence of signifi- cant differences was found in vocabulary difficulty between upper and lower elementary levels of the Allyn and Bacon series (F==.1113, df==1,]; .051 The null hypothesis could not be rejected. To provide for a closer inspection of the difference between means, a means summary table is presented (see Table 34). Table 34. A Mean Summary Table of Vocabulary Difficulty Between Upper and Lower Elementary Levels as Determined by a Word Frequency List Series Level §EE. Mean Entire population Upper 1593.0000 .7961 Lower 1654.0000 .8287 Houghton Mifflin Upper 401.0000 .8101 Lower 407.0000 .9012 Holt, Rinehart, Winston Upper 397.0000 .7877 Lower 400.0000 .8000 Harcourt, Brace, Upper 401.0000 .8020 Jovanovich Lower 395.0000 .7900 Allyn and Bacon Upper 392.0000 .7840 Lower 412.0000 .8240 78 The Eighteenth Hypothesis Ho 18.0: Ho Ho Ho Ho 18.1: 18.2: 18.3: 18.4: There is no significant difference in vocabulary difficulty between the upper and lower elementary levels as deter- mined by a word frequency table. There is no significant difference in vocabulary difficulty between the upper and lower elementary levels of the Hough— ton Mifflin series as determined by a frequency table. There is no significant difference in vocabulary difficulty between the upper and lower elementary levels of the Holt, Rinehart, and Winston series as deter- mined by a word frequency table. There is no significant difference in vocabulary difficulty between the upper and lower elementary levels of the Har- court, Brace, and Jovanovoch series as determined by a word frequency table. There is no significant difference in vocabulary difficulty between the upper and lower elementary levels of the Allyn and Bacon series as determined by a word frequency table. An analysis of variance was used in statistically treating these hypotheses (see Table 35). There was a significant difference in vocabulary diffi- culty between upper and lower elementary levelsansdetermined by a word frequency table (F==14.2124, df==l,19<.051 The null hypothesis could not be accepted. There was a signifi- cant difference in vocabulary difficulty between upper and lower elementary levels of the Houghton Mifflin series as determined by a work frequency table (F==7.8657, df==1, p< .05). The null hypothesis could not be accepted. No evi- dence of significant difference was found in vocabulary 79 Table 35. An Analysis of Variance Table of Vocabulary Difficulty Between the Upper and Lower Elementary Levels as Determined by a Word Frequency Table Mean F Series Square df Ratio Entire population .319E + 10 1 14.2124* Houghton Mifflin .188E + 10 1 7.8657* Holt, Rinehart, Winston .194E + 06 l .0010 Harcourt, Brace, Jovanovoch .3OOE + 10 1 14.0142* Allyn and Bacon .226E + 09 1 .9432 7'~‘p<§.05 difficulty between upper and lower levels of the Holt, Rine- hart, and Winston series as determined by a word frequency table (F= .0010, df= l, p) .05). The. null hypothesis could not be rejected. There was a significant difference in vo- cabulary difficulty between upper and lower levels of the Harcourt, Brace, and Jovanovich series as determined by a word frequency table (F==14.0142, df= 1,1)(.05) The null hypothesis could not be accepted. No evidence of signifi- cant differences was found in vocabulary difficulty between upper and lower elementary levels of the Allyn and Bacon series as determined by a word frequency table (F==.9432, df= 1, p) .05). The null hypothesis could not be rejected. The following mean summary table (Table 36) is pre- sented to permit the observation of specific differences between means. Table 36. A Mean Summary Table of Vocabulary Difficulty Differences Between Upper and Lower Elementary Levels as Determined by a Word Frequency Table, Series Entire pOpulation Houghton Mifflin Holt, Rinehart, Winston Harcourt, Brace, Jovanovich Allyn and Bacon Level Upper Lower Upper Lower Upper Lower Upper Lower Upper Lower Sum .13097E+08 .16677E+08 3016963 4458534 3694333 3680382 2875962 4545377 3510652 3993360 Mean 6522. 8338. 6157. 8917. 7388. 7360. 5639. 9090. 7035. 7986. 2311 8265 0673 0680 6660 7640 1412 7540 3747 7200 CHAPTER V CONCLUSIONS AND RECOMMENDATIONS Introduction The purpose of this study was to demonstrate that the written discourse within four social studies series pub- lished for elementary school children can be examined for variables, other than word length and sentence length, which may affect the reading difficulty levels of instructional materials. More specifically, the purpose was to compare four aspects of the language: syntactic complexity, concept density, concept abstractness, and vocabulary difficulty. From each of the four series, two levels of materials, upper elementary and lower elementary, were examined. The previous chapters presented a description of the problem, a discussion of the related literature and re— search, an explanation of the methodology, and a presenta- tion and analysis of the data generated by this study. The present chapter is organized as follows: (1) ma- jor results and discussions, (2) implications, and (3) re— commendations. Major Results and Discussions Ho 1: There is no significant difference in syn- tactic complexity among the four series at the upper elementary level as determined 81 82 by the number of prepositional phrase mod- ifiers. 1. There was no significant difference among the four series in the number of prepositional phrase modifiers at the upper elementary level. However, while the difference was not statistically significant, the Allyn and Bacon ser- ies did have notably fewer prepositional phrase modifiers per sentence than did the other three series. This particu- lar syntactic variable was selected for tablulation because it was found by Marcus (1971) to be one of the more diffi- cult structures for students to comprehend. Ho 2: There is no significant difference in syn- tactic complexity among the four series at the lower elementary level as determined by the number of prepositional phrase mod- ifiers. 2. There was a significant difference in syntactic complexity among the four series at the lower elementary level as determined by the number of prepositional phrase modifiers. The Allyn and Bacon series was found to have had the highest average number of prepositional phrase mod- ifiers per sentence (1.06), while the Houghton Mifflin ser- ies was found to have the lowest average number of preposi- tional phrase modifiers per sentence (.5). This syntactic structure was selected for tabulation because it was found by Marcus (1971) to be the most difficult for students to comprehend. Ho 3: There is no significant difference in syn— tactic complexity among the four series at the upper elementary level as determined by the number of complex sentences 83 containing relative clauses which interrupt the subject-verb-object sequence of the in- dependent clause. 3. There was no significant difference in syntactic complexity among the four series as determined by the number of complex sentences containing relative clauses which in- terrupt the subject-verb-object pattern of the independent clause. The occurrence of this syntactic structure was min- imal in the samples extracted from the four series.‘ The Harcourt, Brace, and Jovanovich series and the Allyn and Bacon series had no sentences which contained this structure. This structUre was selected for tabulation because it was found by Marcus to be one of the more difficult structures for students to comprehend. Marcus found that sentences whose basic components were not separated were more easily understood by students. Ho 4: There is no significant difference in syn- tactic complexity among the four series at the lower elementary level as determined by the number of complex sentences contain— ing relative clauses which interrupt the subject-verb-objective sequence of the independent clause. 4. There was no significant difference in syntactic complexity among the four series at the lower elementary level as determined by the number of complex sentences with relative clauses which interrupt the subject-verb-object pattern of the independent clause. As in the upper elemen- tary levels, this syntactic structure seldom occurred in sample sentences.‘ In the Allyn and Bacon series, no such structures were found. Marcus found this to be one of the 84 more difficult structures for students to grasp. Some stu- dents automatically assumed a noun-verb-noun pattern to be a subject-verb-object sequence. Ho 5: There is no significant difference in con- cept density among the four series at the upper elementary level. 5. There was a significant difference in concept den- sity among the four series at the upper elementary level. The Allyn and Bacon series contained the least number of concepts per sentence (2.5), Eddie the Holt, Rinehart, and Winston series presented the greatest number of concepts per sentence to the reader (3.4). Some researchers, including Taylor and Lunstrum (1977), have expressed concern that large numbers of concepts are likely to present difficulties for young readers. Ho 6: There is no Significant difference in con- cept density among the four series at the lower elementary level. 6. There was a significant difference in concept den- sity among the four series at the lower elementary level. The Houghton Mifflin series contained only 1.9 concepts per sentence. The Holt, Rinehart, and Winston series contained, an average of 3.0 concepts per sentence, which was the high- est average among the four series. Ho 7: There is no significant difference in con- cept abstractness among the four series at the upper elementary level. 7. There was a significant difference in concept ab- stractness among the four series at the upper elementary level. The Houghton Mifflin series presented an average of 1—S-mfi , _ 1 85 1.0 abstract concepts per sentence, which was the lowest average. The highest mean, 1.8, was recorded for the Har- court, Brace, and Jovanovich series. Several researchers have reported that abstract terms tend to be more difficult for young readers to comprehend. Garner and Sheldon (1954), among others, report that in the social science field, " the social studies in particular appear to contain an abun- dance of terms which serve to hinder comprehension" (p. 228) Concepts noted by Garner and Sheldon (1954) to be most dif— ficult to grasp were those which were of‘a more abstract na- ture. Ho 8: There is no significant difference in con- cept abstractness among the four series at the lower elementary level. 8. There was no significant difference in concept ab- stractness among the four series at the lower elementary level. The Holt, Rinehart, and Winston series had the low- est average number of abstract concepts per sentence (.67), while the Houghton Mifflin series had the highest average number (1 0) per sentence. Ho 9: There is no significant difference in vo- cabulary difficulty among the four series at the upper elementary level as deter- mined by a word frequency list. 9. There was no significant difference in vocabulary difficulty among the four series at the upper elementary level as determined by a word frequency list. The assump- tion was made that a series using large numbers of words which did not appear on the high frequency word list would 86 be more difficult to read than those using large numbers of words which appeared on the list. The Allyn and Bacon ser- ies had the most difficult vocabulary with seventy-eight percent of the words from the sample passages found on the word frequency list. The series having the least difficult vocabulary was the Houghton Mifflin series. Eighty-one percent of the words in the sample passages were found on the word frequency list. Ho 10: There is no significant difference in vo- cabulary difficulty among the four series at the lower elementary level as deter- mined by a word frequency list. 10. There was a significant difference in vocabulary difficulty at the lower elementary level as determined by a word frequency list. The Harcourt, Brace, and Jovanovich series had the most difficult vocabulary, with seventy-nine percent of its vocabulary words included on the word fre- quency list. The Houghton Mifflin series had the least dif- ficult vocabulary, with ninety percent of its vocabulary words included on the word frequency list. HO 11: There is no significant difference in vo- cabulary difficulty among the four series at the upper elementary level as deter- mined by a word frequency table. 11. There was no significant difference in vocabulary difficulty among the four series at the upper elementary level as determined by a word frequency table. The Holt, Rinehart, and Winston series had the vocabulary with the highest word frequency, while the Harcourt, Brace, and Jo— vanovich series had the lowest word frequency. 87 Ho 12: There is no significant difference in vo— cabulary difficulty among the four series at the lower elementary level as deter- mined by a word frequency table. 12. There was no significant difference in vocabulary difficulty as determined by a word frequency table among the four series at the lower elementary level. The Harcourt, Brace, and Jovanovich series utilized the vocabulary with the highest word frequency, while the Holt, Rinehart, and Winston series utilized a vocabulary with the lowest word frequency. Ho 13: There is no significant difference in syn- tactic complexity between the upper and lower elementary levels as determined by the number of prepositional phrase modi- fiers. 13. There was a significant difference in syntactic complexity between the upper and lower elementary levels as determined by the number of prepositional phrase modifiers. In examining data for the entire upper and lower level pOp- ulations, the researcher found significantly more preposi- tional phrase modifiers used in the upper level texts. This was also the case for the Houghton Mifflin series and the Harcourt, Brace, Jovanovich series. However, there was no significant difference in the number of prepositionalphrase modifiers between the upper and lower levels of the Holt, Rinehart, and Winston and the Allyn and Bacon series. The researcher was surprised to find that within the Allyn and Bacon series, there were more prepositional phrase modifi— ers used in the lower elementary level text than in the 88 upper elementary text. Marcus (1971), in his study using intermediate grade students, found this syntactic structure to be one of the more difficult for students to comprehend. Ho 14: There is no significant difference in syn- tactic complexity between the upper and lower elementary levels as determined by the number of complex sentences containing relative clauses which interrupt the subject-verb-object sequence of the in- dependent clause. 14. There was no significant difference in syntactic complexity between the upper and lower elementary levels as determined by the number of complex sentences containing relative clauses which interrupt the subject-verb-object sequence of the independent clause. This structure appeared rarely in sample sentences. In the entire population, this syntactic structure appeared only thirteen times and was found more frequently within the lower level texts. Within both levels of the Allyn and Bacon series, and within the upper level of the Harcourt, Brace, and Jovanovich series, there were no sentences containing this structure; Marcus found this to be one of the more difficult structures for students to comprehend. Ho 15: There is no significant difference in concept density between the upper and lower elementary levels. 15. There was a significant difference in concept density between the upper and lower elementary levels for the entire population. There was also a significant dif— ference in concept density between levels of the Houghton Mifflin series and the Harcourt, Brace, and Jovanovich 89 series. For all texts, the number of concepts found within upper elementary passages exceeded the number found within lower level texts. However, the difference was not signifi- cant within the Holt, Rinehart, and Winston and Allyn and Bacon series. While collecting data involving concept pre- sentation, the researcher noticed that meanings of concepts sometimes changed, widxflrmight tend to be confusing for young readers. For example, the concept "country" might be mentioned in a context whiCh contrasts it with "city." In the same text, ”country" may be used to name a particular place, as in a description of the United States as a great "country." Ho 16: There is no significant difference in con— cept abstractness between the upper and lower elementary levels. 16. There was a significant difference in concept ab— stractness between the upper and lower elementary levels for the entire population. There were also significant dif- ferences between levels for the Harcourt, Brace, and Jovano- vich series and Allyn and Bacon series. This significant difference did not occur within the Houghton Mifflin and Holt, Rinehart, and Winston series. In every case, concept abstractness was greater in upper level texts than in lower level texts. Ho 17: There is no significant difference in vo— cabulary difficulty between the upper and lower levels as determined by a word fre- quency list. 90 17. There was a significant difference in vocabulary difficulty between the upper and lower elementary levels as determined by a word frequency list for the entire popula- tion. When the data for individual series were examined, a significant difference was found only within the Houghton Mifflin texts. For three of the four series, the greatest number of high frequency words was found in the lower ele- mentary texts. However, it was surprising to find that the Harcourt, Brace, and Jovanovich texts contained the greatest number of high frequency words in the upper level text. Ho 18: There is no signifiCant difference in vo- cabulary difficulty between the upper and lower elementary levels as determined by a word frequency table. 18. There was a significant difference in vocabulary difficulty between the upper and lower elementary texts as determined by a word frequency table when the entire popu- lation was considered. This difference was also found with- in the Houghton Mifflin series and the Harcourt, Brace, and Jovanovich series. The investigator expected to find the vocabulary with the highest word frequency at the lower elementary levels. This occurred within all series except that published by Holt, Rinehart, and Winston. An examina- tion of the means for this series showed the vocabulary with the highest word frequency to be used in the upper le- vel text. Implications The presentation of data collected for this study i1- lustrates that the written discourse of social studies 91 materials can be examined for various factors which may af— fect the reading difficulty levels of these materials. Re- searchers, including Chall (1950), Estes (1972), and others, have noted that readability formulas leave many questions concerning the reading difficulty levels of materials un- answered, since these formulas generally involve the study of only one or two variables. The evidence presented here indicates that it is possible to inspect the discourse of social studies materials for precise data about many differ- ent variables which may affect readability. It has been further demonstrated that it is possible to make comparisons both between different series and be- tween levels within the same series with regard to certain factors which may affect the readability levels of social studies materials. The data collected in this study concerning syntactic structures indicate, in the opinion of the researcher, that prepositional phrase modifiers occur quite frequently with- in the written discourse of social studies materials. Mar- cus (1971) found this syntactic structure to be one of the most difficult for students to comprehend. For some series, there was very little difference between the upper and lower level texts in the number of prepositional phrase modifiers used per sentence. In the case of the Allyn and Bacon series, it was surprising to find that more preposi- tional phrase modifiers were used in the lower level text than in the upper level text. In the opinion of the 92 researcher, this syntactic structure, because of its high frequency, should be studied further for its possible ef- fects on the reading difficulty levels of social studies materials. According to the data presented in this investigation, syntactic complexity as determined by the number of complex sentences containing relative clauses which interrupt the subject-verb-object sequence of the independent clause, did not appear to be a prominent factor in any of the four ser- ies. This syntactic structure rarely occurred in sample sentences and, in several of the texts, did not occur at all. In this researcher's opinion, further study of this particular syntactic structure and its effect on the read- ability of social studies materials would be of lesser im- portance, since this structure appears so infrequently. The data compiled in this study pertaining to concept density appears to indicate that this is a factor which war- rants further study. There was a considerable discrepancy in the average number of concepts found in sample sentences from each series. In addition, in the Holt, Rinehart, and Winston series and the Allyn and Bacon series, there was almost no difference in the number of concepts mentioned in the upper level text and the number of concepts found in the sample sentences of lower level texts. This implies a possible assumption on the part of’these publishing compan- ies that sixth and third grade students can comprehend 93 equally well an equal number of concepts in an equal number ofhsentences. The data pertinent to concept abstractness indicate, in the opinion of this researcher, that a large number of concepts with which students are confronted in social stud- ies materials are of an abstract nature at both the upper and lower elementary levels. Carroll (1964) notes that ab- stract concepts are the most difficult for students to com— prehend, since they cannot be defined in terms of sensory qualities. The data also suggest that some publishing com- panies assume that third grade students can effectively as- similate approximately the same number of abstract concepts per sentence as sixth graders. Some of the research in this area, such as the studies by Gill (1962) and Arnsdorf (1963), suggests that the ability to understand selected abstract concepts increases from grade to grade. It is the view of this researcher that presentation of large numbers of concepts, and the abstractness of many of these concepts, should be studied further for effects on the comprehension of social studies materials by young readers. In addition, teachers' awareness of the presence of large numbers of concepts, many of which are abstract, may help them to better prepare students for the reading of social studies materials. The data collected in this investigation revealed that vocabulary difficulty as determined by a word frequency list or a word frequency table does not necessarily differ 94 greatly from the lower to the upper level texts. The re- searcher expected to find a greater numbercflfhigh frequency words utilized in the lower level texts, but in the case of the Harcourt, Brace, and Jovanovich series and the Allyn and Bacon series, the opposite proved to be true. While investigators have found that a high frequency vocabulary affects comprehension in a positive way, more research ap- pears to be needed in the area of social studies specifi- cally. In addition, the use of a word frequency list and a word frequency table clarified which series presented the most frequent vocabulary. However, in the opinion of this investigator, more research is necessary to determine the point at which an unfamiliar vocabulary begins to seriously hamper comprehensionxxfsocial studies materials. The re- searcher also found that while a word frequency table may be more precise in measuring word frequency, the word fre- quency list is more practical as far as time and effort is concerned. In the opinion of this researcher, the data pertinent to this investigation suggests two important implications for practice: (1) consumers of social studies materials need to be aware that there are many factors within these materials which may affect readability, and (2) textbook publishers should also be aware of these factors and their possible effects on readability as they produce social, studies materials for use by elementary school students. 95 Recommendations Some recommendations for further study which were generated from this study are: 1. This study should be extended using other varia— bles Which were not used in this study, but which may also negatively affect the reading difficulty levels of social studies materials. 2. A study should be conducted to investigate the ef- fects of prepositional phrase modifiers on the comprehen- sion of social studies materials by young readers. 3. More research is needed to determine the effects on comprehension of the presentation of large numbers of social studies concepts, especially those which are ab— stract in nature, and to determine whether or not the abil- ity to comprehend such concepts increases from grade to grade. 4. More research should be conducted to determine the point at which an unfamiliar vocabulary seriously begins to hamper the comprehension of social studies materials by elementary students. APPENDIX SAMPLES OF DATA COLLECTION SHEETS FOR SYNTACTIC COMPLEXITY, CONCEPTUAL DENSITY, CONCEPTUAL ABSTRACTNESS, AND VOCABULARY DIFFICULTY SAMPLES OF DATA COLLECTION SHEETS FOR SYNTACTIC COMPLEXITY, CONCEPTUAL DENSITY, CONCEPTUAL ABSTRACTNESS, AND VOCABULARY DIFFICULTY Syntactic Complexity Prepositional Phrase Modifiers Sample I: l. Suddenly, I turned my head. 2. I ran on the beach and I.swam in the surf. Prepositional Prepositional Phrase Phrase Series Level Modifiers Series Level Modifiers l 2 0 l 2 2 Sentence 1 Sentence 2 Key: Series 1 = Houghton Mifflin Level 2 = Lower Level 96 97 Sample 11: Complex Sentences with Relative Clauses which ' Interrupt the Subject-Verb-Object Sequence of the Independent Clause 1. One important group that you belong to is your community. If 2. What do the members of your community Share? Relative Relative Series Level Clause Series Level Clause 1 2 l l 2 O Sentence 1 Sentence 2 Key: Series 1 = Houghton Mifflin Level 2 = Lower Level Conceptual Density Sample 1: Concept Density 1. The man driving the car is lost. 2. So is the puppy. Series Level Concepts Series Levgi Concepts 1 2 2 1 2 l Sentence 1 Sentence 2 Key: Series 1 Houghton Mifflin Level 2 Lower Level 98 Conceptual Abstractness Sample I: 1. Concept Awareness bed and lay across it. The awful feeling did not go away, so I ran to my 2. I banged my fists down on my bed several times. Abstract - Abstract Series Level Concepts Series Level Concepts 1 2 l 1 2 l Sentence 1 Sentence 2 Key: Series 1 = Houghton Mifflin Level 2 = Lower Level_ Vocabulary Difficulty Sagple I: Vocabulary Difficulty as Determined by a Word Frequency List 1 2 3 1. Communities i3 g country need services. Word Word Word Series Level Value Series Level Value Series Level Value 1 2 0 l 2 l l 2 1 Word 1 Word 2 Word 3 Key: Series 1 = Houghton Mifflin Level 2 = Lower Level Word Value 1 Word Value 0 list word list Included in high frequency word Not included in high frequency 99 Sample 11: Vocabulary Difficulty as Determined by a Word Frequency Table l 2 3 1. All the planets are like spaceships traveling Word Word Word Fre- Fre- Fre- quen— quen- quen- Series Level gy Series Level gy 'Series Level qy l 2 3497 l 2 16745 1 2 59 Word 1 Word 2 Word 3 Key: Series 1 Level 2 Houghton Mifflin Lower Level BIBLIOGRAPHY BIBLIOGRAPHY Allen, R. Better reading through the recognition of gram- matical relations. The Reading Teacher, 12, December, 1964, 194-198. Arnsdorf, V. E. Readability of basal social studies mater- ials. The Reading Teacher, lg, January, 1963, 243-246. Bader, L., & Harned, L. In Michigan Social Studies Textbook Review. State of Michigan Department of Education, Lansing, Nfichigan, 1978. Baldwin, R. S. The effects of noncanonical sentences upon the reading comprehension of third grade children. Dissertation Abstracts, 1977, 37A, 4813. Bayne, S., & Cordova, J. P., Eds. The social sciences con- cepts and values. New York: Harcourt, Brace, and Jovanovich, 1975. Beattie, S. S., & Greco, D. Who are we? Boston: Houghton Mifflin Co., 1976. Bormuth, J. R., Carr, J., Manning, J., & Pearson, D. Chil- dren's comprehension of between and within syntactic structures. Journal of Educational Psychology, 61, 1970, 349—357. "I Branson, M. S. The way people live. Boston: Houghton Mifflin, 1976. Bruner, J. 8., & Postman, L. Emotional selectivity in per- ception and reaction. JoUrnal of Personality, lg, September, 1947, 69-77. Carner, R. L., & Sheldon, W. D. Problems in the develop- ment of concepts through reading. 'Elementary School Journal, 55, December, 1954, 226-229. Carroll, J. B. Words, meanings, and concepts. Harvard Ed- ucational Review, 34, 1964, 178—193. ‘ Carroll, J., Davies, P., & Richman, B. The American heri- tage word frequencybook. Boston: Houghton Mifflin Co., 1971. 100 101 Chall, J. This business of readability: a second look. EducatiOnal Research Bulletin, 35, 1956, 93. Chomsky, N. Aspects of a theory of syntax. Cambridge, Massachusetts: MIT Press, 1965. Chomsky, N. Syntactic structures. The Hague: MDuton, 1957. Cohen, J. H. The effect of content area material of cloze test performance. Journal of Reading, December, 1975, 247—250. Coleman, E. B. Developing a technique of written instruc- tion: some determiners of the complexity of prose. Symposium on Verbal Learning Research on the Technol- ogy of Written Instruction. Unpublished paper, Colum- bia University, 1966. Coleman, E. B. Improving comprehension by shortening sen— tences. Journal of Applied Psychology, 46, April, 1962, 131-134. ' Coleman, E. B., & Blumfield, J. P. Close scores of nomi- nalization and their grammatical transformations, us- ing active verbs. Psychological Reports, 13, 1963, 651-654. ”— Dale, E., & Chall, J. The concept of readability. Elemen- tary English, 26, January, 1949, 23. Estes, T. H. Reading in the social studies: a review of research since 1950. In J. Laffey (Ed.), Reading in the content areas. Newark: International Reading Association, 1972. Fagan, W. T. Transformations and comprehension. The Read- ing Teacher, 25, November, 1971, 169-172. Fielder, W. R., Ed. Inguiring about cities. New York: Holt, Rinehart, and Winston, 1976. Fodor, J. A., & Garrett, M. Some syntactic determinants of sentential complexity. Perception and Psychophysics, 11, July, 1967, 289-296. Francis, N. The Structure of American English. New York: The Ronald Press Company, 1958. Friedman, K. C. Time concepts of elementary school chil- dren. Elementary School Journal, 46, February, 1944, 337-342. '5 —“ 102 Gagne, R. M; The conditions of learning. New York: Holt, Rinehart, and Winston, 1965. Gill, C. C. Interpretations of indefinite expressions of time. Social Education, 26) December, 1962, 454-546. Gough, P. B. Grammatical transformations and speed of un- derstanding. Journal of Verbal Learning and Verbal Behavior, 1965,4107-111. Gough, P. B. The verification of sentences: the effects of delay of evidence and sentence length. Journal of Verbal Learning and Verbal Behavior, 5, 1969, 492-496. Harris, T. Reading. The Encyclopedia of Educational Re- search, R. L. Ebel, Ed. Toronto: MacMillan Co., 1969. Hash, R. J. The effects of a strategy of structured over- views, level guides, and vocabulary exercises on stu- dent achievement, reading comprehension, critical thinking, and attitudes of junior high school classes in social studies. Ph.D. dissertation, State Univer- sity of New York, Buffalo, 1974. Herman, W. L. Reading and other language arts in social studies instruction: persistent problems. A new look at reading in the social studies, R. C. Preston, Ed. Newark: International Reading Association, 1969. Howes, D. H., & Solomon, R. L. A note on McGinnies emotion~ ality and perceptual defense. Psychological Review, 51, July, 1950, 229-235. Howes, D. H., & Solomon, R. L. Visual duration threshold as a function of word probability. Journal of Ameri- can Psychology, 41, June, 1951, 401—410. Jarolimek, J., & Foster, C. D. Quantitative concepts in fifth grade social studies testbooks. Elementary School Journal, May, 1959, 437—442. Johnson, F. C. The effect of frequency, pronounceability, and word length on word recognition ability of elemen- tary school children. Dissertation abStracts, 54, 1974, 7585. Johnson, R. M. A critical analysis of the treatments given representative social science ideas in leading fifth and eighth grade American history textbooks. Disserta— tion Abstracts, 25, 1967, 1988. 103 Johnson, R. E., & Verdian, E. B. Reading, readability, and the social studies. The Readinngeacher, 26, February 1973, 483-488. ‘ . Johnson, R. C., Thompson, C. W., & Frincke, G. Word values, word frequency, and visual duration thresholds. Psy- chological Review, 62, September, 1960, 332-342. Kierzek, J. M., & Gibson, W. The MacMillan handbook of English. New York: The MacMillan Company, 1964. King-Ellison, P., & Jenkins, J. J. The durational threshold of visual recognition as a function of word frequency. American Journal of Psychology, 62, December, 1954, 700-703. Klare, G. R. The role of word frequency in readability. Elementary English, 45, January, 1978, 12-22. Klien, G. A., & Kurkowski, F. Effect of task demands on relationship between eye movements and sentence com- plexity. Perceptual and Motor Skills, 59, 1974, 463- 466. Lunstrum, J. P., & Taylor, R. L. Teaching reading in the social studies. Boulder, Colorado: Social Science Education Consortium, Inc., 1977. Lyda, W. J., & Robinson, V. M. Quantitative concepts in selected social studies textbooks for second grade. Elementary School Journal, 65, December, 1964, 159-162. Marcus, A. D. The development of a diagnostic test of syn- tactic meaning clues in reading. Ph D. dissertation, New York University, 1969. Marcus, A. D. The development of a diagnostic test of syn- tactic clues in reading. In R. E. Leibert (Ed.), D;- agnostic viewpoints in reading. Newark, Delaware: International Reading Association, 1971. Marcus, A. D. Reading as reasoning; reading as ambiguity: understanding structure. Paper presentedzn:the annual meeting of the National Council of Teachers of English, Las Vegas, November 25-27, Markle, S. M., & Tiemann, P. W. Conceptual learning and instructional design. In M. D. Merrill (Ed.), Instrue tional design readings. Englewood Cliffs, New Jersey? Prentice Hall, Inc., 1971. 104 Marks, C. B., Doctorow, M. J., & Wittrock, M. C. Word fre- quency and reading comprehension. Journal of Educa- tional Research, 62, February, 1974, 259-262. Martorella, P. Concept learning: des2gns for instruction. Scranton: International Textbook Company, 1972. Martorella, P. Concept learning in the social studies: models for structuring curriculum. Scranton: Inter- national Textbook Company, 1971. ’ McGinnies, E. Emotionality and perceptual defense. Psy- chological Review, 56, September, 1949, 244-251. McGinnies, E., Comer, P. B., & Lacey, O. L. Visual recog- nition thresholds as a function of word length and word frequency. Journal of Experimental Psychology, 44, July, 1952, 65-69. Mehler, J., Bever, T. C., & Carey, B. What we look at when we read. Perception and Psychophysics, 2, 1967, 213- 218. Michnay, A. M., & McCarthy, M. C. (Eds.) The age of western expansion. Boston: Allyn and Bacon, 1975. Michnay, A. M., & McCarthy, M. C. The interaction of cul— tures. Boston: Allyn and Bacon, 1975. Michnay, A. M., & McCarthy, M. C. Lands of Latin America. Boston: Allyn and Bacon, 1976. Michnay, A. M., & McCarthy, M. C. ‘The makingcflfour America. Boston: Allyn and Bacon, 1974. ‘ Michnay, A. M., & McCarthy, M. C. New world and Eurasion cultures. Boston: Allyn and Bacon, 1975. Miller, G. A. Some psychological studies of grammar. Amer- ican Psychologist, 22, November, 1962, 748-762. Miller, G. A., & McKean, K. A chronometric study of some relations between sentences. Quarterly Journal of Ex- perimental Psychology, 26, November, 1964, 297-308. Newbigging, P. S. The perceptual reintegration of words which differ in connotative meaning. Canadian Journal of Psychology, September, 1961, 133-142. Newbigging, P. L., & Hay, J. M. The practice effect in re- cognition threshold determinations as a function of word frequency and length. canadian Journal of Psy- chology, 26, 1962, 77-84. 105 Noble, C. E. An analysis of meaning. Psychological Review 52, November, 1952, 421-430. Noble, C. E. The effect of familiarization upon serial ver- bal learning. Journal of Experimental Psychology, 45, May, 1955, 333-338. Noble, C. E. The meaning-familiarity relationship. Psy- chological Review, 66, March, 1953, 89-98. Peters, C. W. The effects of systematic restructuring of material upon the comprehension process. Reading Re- searchgguarter2y, 22 (1975076), 87-111. Ratcliffe, R. H. A critical analysis of the treatments given representative social science ideas in leading eleventh grade American history textbooks. Disserta- tion Abstracts, 22, 1967, 2041. Richek, M. A. Effect of sentence complexity on the reading comprehension of syntactic structures. ‘Journal of Ed- ucational Psychology, 66, December, 1976, 800-806. Richek, M. A. Reading comprehension of anaphoric forms in varying linguistic contexts. ReadinggResearch Quar- terly, 22, 1976-77, 146-165. Roberts, P. Understanding grammar. New York: Harpen 1954. Ruddell, R. B. The effects of the similarity of oral and written patterns of language structure on reading com- prehension. Elementary English, 42, April, 1965, 403- 410. V _— Smith, C. W. The vocabulary of factual reading materials. The EduCational Forum, May, 1963, 443-447. Smith, H. P., & Dechant, E. U. Psychology in teaching read- ing. Englewood Cliffs, New Jersey: Prentice-Hall 1961. Solomon, R. L., & Postman, L. Frequency of usage as a de- terminant of recognition thresholds for words. Jour- nal of Experimental Psychology, 43, March, 1952, 19 - 201. ‘— Thorndike, E. L. Reading as reasoning: a study in mis- takes in paragraph reading. Journal of Educational Psychology, 6, 1917, 323-332. Thorndike, E. L. The psychology of thinking in the case of reading. Psychological Review, 24, 1917, 220-234. 106 Walker, C. M. High frequency word list for grades three through nine. The Reading Teacher, 52, April, 1979, 803-812. Whitely, D. (Ed.) The social sciences concepts and values. New York: Harcourt, Brace, and Jovanovich, 1975} Wilson, M. C. The effect of amplyfying material upon com- prehension. Journal of Experimental Education, 25, September, 1944, 5-8. "I11111111171111111111