“HES’. III ml 11 1| m ll Ll l “M Ill 1141qu 3 1293 This is to certify that the thesis entitled EFFECTS OF QUANTITATIVE AND NON- QUANTITATIVE LITERACY ON THE KNOWLEDGE AND ADOPTION OF TECHNOLOGICAL INNOVATIONS . presented by Mwanika Ok'Ogule Mwanika has been accepted towards fulfillment of the requirements for ' 5' . Ph . D . Communication I degree in Major professor Date May 31, 1979 07639 ,2, *his {'14 I V'“ Afi--\\\\ t i OVERDUE FINES: 25¢ per day per item RETURNING LIBRARY MATERIALS: Place in book return to remove charge from circulation records .0... "IID .. Ili‘l'lllvlri 'lu'l‘f.ll COpyright by MWANIKA OK ' OGULE MWANIKA 1979 © EFFECTS OF QUANTITATIVE AND NON-QUANTITATIVE LITERACY ON THE KNOWLEDGE AND ADOPTION OF TECHNOLOGICAL INNOVATIONS BY Mwanika Ok'Ogule Mwanika A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 1979 Accepted by the faculty of the Department of Communication, College of Communication Arts and Sciences, Michigan State University, in partial fulfillment of the requirements for the Doctor of Phi1030phy degree. Gui dance Committee : ABSTRACT EFFECTS OF QUANTITATIVE AND NON-QUANTITATIVE LITERACY ON THE KNOWLEDGE AND ADOPTION OF TECHNOLOGICAL INNOVATIONS BY Mwanika Ok'Ogule Mwanika Literacy has been used extensively in the study of factors influencing knowledge and adoption of technological innovations. The results have been mixed but generally show positive correlations. This study's literature review revealed two omissions in prior explications of literacy which have prevented a more thorough test of the relative impact of literacy on knowledge and adoption of technological innovations, partic- ularly in Developing Countries. These are quantitative lit- eracy, which refers to the skill in the use of quantitative symbols and concepts, and literacy in English in those areas such as Africa where English is frequently not the dominant language. The importance of both quantitative and non-quantita- tive (English) literacy is discussed. The study notes that the change agents and their rural clients of innovations generally differ in their linguistic repertoires with respect to the "language" of diffusion. While the former are gener- ally college graduates who are generally educated and trained in English, the latter are not only generally illiterate in their native languages but, more importantly, most of them do not speak nor read and write in English--the language of Mwanika Ok'Ogule Mwanika technology; and generally they have an unelaborate numbering system in their native languages or dialects. The implica- tions for these problems are discussed. The study reconceptualizes literacy into four levels: Level I comprising illiterate individuals; Level II for lit- eracy in native language only; Level III being literacy in both a native language and in English (i.e., biliteracy), and Level IV which adds literacy in quantitative symbols and concepts to the Levellflflistandard. The study's basic purpose was to explore, through stepwise multiple regression analysis: the relative effects of quantitative and non-quantitative literacy in predicting awareness/knowledge and adoption of innovations; and the relative effects of non-quantitative literacy types and edu- cation measures on quantitative literacy. In addition, the study explores the correlations among literacy types, educa- tion, and awareness/knowledge and adoption of innovations; the relationship between education and propensity of adoption of complex technological innovations; and the relationship between the four constructed levels of literacy and knowl- edge and adoption of technological innovations. The study's data derived from two sample spaces; viz., Nigeria - Ilewo (N = 364), and the USA - Michigan (N = 230). The latter sample space was subdivided into the native English- speaking group (N = 169), and the native Spanish-speaking group (N = 61) for purposes of analysis. Mwanika Ok'Ogule Mwanika Overall, the results indicated generally strong posi- tive correlations among the measures of literacy and educa- tion. In predicting quantitative literacy, awareness/ knowledge and adoption of innovations, literacy in English tended to be a stronger predictor than non-English native language; and in predicting knowledge and adoption, quanti- tative literacy tended to be a stronger predictor than literacy in English. With the Spanish-speaking respondents, bilateracy was stronger than either Spanish or English literacy in pre- dicting quantitative literacy, knowledge and adoption of innovation. Relatively low percentage of the variance in adoption was explained in all of the tests. The Chi square tests indicated that level of educa- tion is generally positively related to the propensity to adopt those innovations which are more complex. Finally, the relationships between the four constructed levels of literacy and knowledge and adoption of innovations were not statistically significant, using one-way ANOVA, although the means were higher with higher levels of literacy. Applications of the findings and ideas for future research were suggested. My father, the late Benjamin Ogule who initiated and inculcated into me the norm of questioning and whose determination was always to propel me as far as humanly possible along the formal road to truth; and my mother, Kolobina Nakiria who continued this process undaunted despite greater social and economic problems, to them I dedicate this dissertation. ii ACKNOWLEDGMENTS No dissertation is a product of a single individual and this is especially true of the present one. From the in- ception to the completion of this study, the author incurred much indebtedness. He received a great deal of help from many others, in the form of cooperation, finance, sugges- tions, criticisms, and just plain instruction. I cannot be- gin to acknowledge here the aid of all those who have helped but to say that I acknowledge my debt to each and all of them. However, my conscience requires that I make special mention of a few. Certainly I am very indebted with grate- ful acknowledgment and appreciation to Dr. Lawrence E. Sarbaugh, my Academic Advisor, and Chairman of Doctoral Guidance Committee, for his helpful suggestions and criti- cisms in planning this study and preparing this manuscript, for his encouragement and his assistance in completing the data analyses, and through whom I was able to obtain from the Office of the Dean of the College of Communication Arts and Sciences at Michigan State University the secondary data from Nigeria - Ilewo. Throughout my study at MSU, Dr. Sarbaugh remained a true and exceptionally generous friend and mentor who never seemed to spare time for himself iii whenever I needed his help. I cannot express adequately my indebtedness to him. It is also a pleasure to acknowledge the great debt which this study owes to Dr. Joseph Woelfel, my Doctoral Guidance Committee member, who first got me interested in the concept of "quantity" during his Advanced Measurement Seminar at MSU, Department of Communication, and whose methodological guidance was invaluable. I also wish to recognize the contributions of my other Guidance Committee members--Dr. Richard V. Farace, Dr. Bradley Greenberg, and Dr. Thomas A. Muth. I wish to thank the School of Graduate Studies and the Department of Communication at MSU for awarding me a graduate assistantship which enabled me to pursue advanced graduate study at MSU. Thanks also to the Department of Communication at MSU for allowing extra credit to be given to Communication 100 students who participated in this study. I acknowledge my intellectural debt to the members of the Communication Department at MSU where I learned most of what I know about communication theory and research and other areas of specialization. The intellectual stamp of that Department is visible throughout this manuscript. But, of course, the Department is not responsible for what I did not learn nor is any one else responsible for any errors of commission and omission which this manuscript may contain. I take full responsibility for them. iv I wish to extend my sincere appreciation to Antonio Benavides, Director of the Spanish organization at Cristo Rey, Lansing, Michigan; to Roberto Quiroz of the United Migrants for Opportunity (UMOI) Spanish organization in Lansing School District, Michigan; to Phil Hartman of Carman High School, Flint, Michigan, and to Michael Hughes of Mott Adult High School, Flint, Michigan, for their permission and cooperation for conducting this study among their populations. I am grateful to Dr. Graham E. Kerr, Buffalo, New York, and former Assistant Director of the Nigeria Diffusion Project, for providing the codebook and other documents and verbal information about the Nigeria - Ilewo data. I am indebted to Pilar Fernandez-Collado, Mr. & Mrs. Jairo Cano, Jose Chotquis, and Valbuena Sirio, for their help in translating into Spanish the English version of this study's questionnaire; and Juanita Adelman, Mr. & Mrs. Jairo Cano, and Gary Stahl for their help in conducting the Spanish interviews. My thanks to Crissy Kateregga for her help in print- ing and assembling this study's questionnaire as well as for mobilizing groups of coders. I am also very grateful to my African, American, and Canadian friends and fellow students at MSU and to relatives and their friends from Lansing Com- munity College for their help in coding this study's data. I am particularly thankful to Francis Ruvuna, James DinKelacker, Tim Mabee, and Mike Code for their assistance V with computer programming. I completed writing this manuscript when I was Assistant Professor of Communication theory and research in the Speech Department at Indiana State University, Terre Haute. I am very grateful to Dr. John C. Stockwell, Chair- man of the Speech Department, ISU, for providing funds to defray typing, copying, and mailing costs for the draft of this manuscript, and for making available Rachel Isabell, Senior Secretary, to type part of the draft. Thanks to Ruth Langenbacher for typing the final draft of this manu- script. My wife, Chris, and our sons Koliateker and Komorateker have individually and jointly contributed significantly to the completion of this study in more ways than can be stated. I gratefully acknowledge, with appreciation, their prayers, patience, and encouragement. vi TABLE OF CONTENTS Chapter Page I THEORETIC RATIONALE AND HYPOTHESIS . . . l A. INTRODUCTION. . . . . . . . 1 B. LITERATURE REVIEW . . . . . . 3 1. Literacy Correlates . . . . . 4 2. Literacy Function . . . . . 8 3. Skewed Regional Distribution of Literacy. . . . . . ll 4. Prior Conceptualizations of Literacy . . . . 15 a. Planning-Census- Type Definitions of Literacy. . . 17 b. Empirical- Type Definitions of Literacy . . . 21 5. Symbolism in Diffusion-Adoption Processes . . 29 a. General Types of Symbol Systems in Diffusion-Adoption Processes. 29 b. Specific Types of Symbol Systems Important in Diffusion- Adoption Processes . . . . 31 i. The Importance of Literacy in Quantitative Symbols and Concepts in the Dif- fusion- -Ad0ption Practices. . . . 32 ii. The Importance of Literacy in Non-Quantitative Sym- bols and Concepts of English Language in the Diffusion-Adoption Practices. . . . . 39 C. THEORETIC FRAMEWORK AND HYPOTHESES . . 42 1. Theoretic Relation Between Language ' and Behavior . . . 43 a. Human Ability to Acquire Language . . . . . . 43 b. Learning Language . . . . 44 c. The Role of Meaning in Human Communication . . . . . 47 vii Chapter Communication . . . . 2. Theoretic Relation Between Language and Literacy . . . . . 3. Theoretic Relation Between Literacy and Education . . . . . 4. Reconceptualization of Literacy . 5. Theoretic Hypotheses. . . . a. Nigeria - Ilewo . . . . b. USA - Michigan . . . i. Native English-speaking Group . . . ii. Native Spanish-speaking Group . . . . . II METHODOLOGY. . . . . . . . A. OPERATIONALIZATION OF VARIABLES. . 1. Variables in the Nigeria - Ilewo Data Set . . a. The Independent Variables . i. Education . . . . ii. Literacy: In Native Language (Yoruba). . In English Language . b. The Dependent Variables. . i. Awareness . . . . ii. Adoption . . 2. Variables in the USA - Michigan Data Set . . . . a. The Independent Variables . i. Education . . . . ii. Literacy in English and in Spanish . . . d. Human Linguistic Competence . e. The Role of Language in Human iii. Quantitative Literacy . Testing the QLIT Items for Reliability . Coefficients of Reliability Among Whites; LGRADE 12 Years . . Coefficients of Reliability Among Whites; LGRADE 13 Years . . Coefficients of Reliability Among Mexican Americans . . Coefficients of Reliability Among Black and Native Americans. . b. The Dependent Variables. . i. Knowledge . . . . ii. Adoption . . c. Development of Data Collection Instrument for the USA - Michigan Data Set . . . viii Page 48 51 57 61 64 68 68 69 69 7O 72 73 73 73 73 73 74 75 75 77 78 78 78 81 85 87 89 89 89 90 90 91 92 94 Chapter Page II (cont'd.) B. SAMPLING. . . . 100 1. Method of Selecting Respondents . . 100 2. Characteristics of Respondents . . 104 3. Limitations of the Samples . . . 107 C. DATA COLLECTION . . . . . 108 1. The Nigeria - Ilewo Data Set. . . 108 2. The USA - Michigan Data Set . . . 109 D. DATA PROCESSING . . . . . . . 112 E. METHODS OF ANALYSIS . . . . . 113 1. Multiple Regression Models . . . 113 a. Multiple Regression Models in the Nigeria - Ilewo Sample Space . . 114 b. Multiple Regression Models in the USA - Michigan Sample Space . 115 2. Other Statistical Analyses on the USA - Michigan Data Set. . . . 118 III RESULTS . . . . . . . . . . 120 A. RESULTS IN THE NIGERIA - ILEWO SAMPLE SPACE . . . . . 120 1. Results from Correlation Analysis . 122 2. Stepwise Multiple Regression for Awareness of Innovations . . . 125 3. Stepwise Multiple Regression for Adoption of Innovations. . . . 128 B. RESULTS IN THE USA - MICHIGAN SAMPLE SPACE . . . . 128 1. Results from the Native English- Speaking Group . . . . . . 130 a. Results from Correlation Analysis . . 131 b. Chi Square Tests Between LGRADE and the Propensity of Adopting Complex Innovations. . . . 133 c. Stepwise Multiple Regression for Quantitative Literacy . . . 136 d. Stepwise Multiple Regression for Knowledge of Innovations . . 136 e. Stepwise Multiple Regression for Adoption of Innovations. . . 139 ix Chapter III (cont'd.) 2. Results from the Native Spanish- Speaking Group . . . . . a. Results from Correlation Analysis . . b. Chi Square Tests Between LGRADE and the PrOpensity of Adopting Complex Innovations. . . . c. Stepwise Multiple Regression for Quantitative Literacy . . d. Stepwise Multiple Regression for Knowledge of Innovations . e. Stepwise Multiple Regression for Adoption of Innovations. . f. Results From One Way Analysis of Variance (ANOVA) for Knowledge (NOW) and Adoption (ADOP) of Technological Innovations Among Literacy Levels i. Results From One Way ANOVA for Knowledge (KNOW) of Innovations Among Literacy Levels. ii. Results From One Way (ANOVA) for Adoption (ADOP) of Innovations Among Literacy Levels. IV SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS A. SUMMARY . . . . . . . B. CONCLUSIONS . . . . . . . . C. RECOMMENDATIONS . . 1. Recommendations for Change Agents . . . . . 2. Recommendations for Future Research . . . . . . . FOOTNOTES . . . . . . . . . . . APPENDICES . . . . . . . . . . . x Page 141 142 144 147 147 150 152 154 155 158 158 159 172 172 173 177 180 Chapter APPENDIX: BIBLIOGRAPH A. B. C. LITERACY TEST. . . . . . TEST FOR KNOWLEDGE, ADOPTION, AND PROPENSITY OF ADOPTION OF TECHNO- LOGICAL INNOVATIONS AND FOR QUAN- TITATIVE LITERACY IN THE USA - MICHIGAN SAMPLE SPACE . . . RESULTS FROM PRELIMINARY TRIALS WITH DIFFERENT VALUES OF RESTRIC- TION PARAMETERS FOR FITTING THE REGRESSORS INTO THE PREDICTIVE EQUATIONS IN THE NIGERIA - ILEWO SAMPLE SPACE. . . . . . xi Page 180 182 201 203 Table 1. LIST OF TABLES World and Regional Distribution of Adults, Literate Adults and Illiterate Adults (aged 15 years and over) in Millions for 1960 and 1970 . . . . . . . Frequency and Percent of Respondents in Each Education Category in the Nigeria - Ilewo Sample Space . . . . . . Frequency and Percent of Respondents in Each Level of Literacy Category in Yoruba and English for the Nigeria — Ilewo Sample Space . . . . . . . Frequencies on Years Ago Respondent First Heard About Each of the Agricultural and Health Innovations in the Nigeria - Ilewo Sample Space . . . . . . Frequencies on Years Ago Respondent First Tried Each of the Agricultural and Health Innovations in the Nigeria - Ilewo Sample Space . . . . . . . . . Frequencies on Continued Use of Each of the Agricultural and Health Innovations in the Nigeria - Ilewo Sample Space . . . Distribution of the USA - Michigan Respon- dents on Last Grade of School Completed. Quantitative Concepts and Their Correspond- ing Number of QLT Items in the Instrument for the USA - Michigan Data Set. . . The Distribution of Respondents in the USA - Michigan Sample Space by Race and Sex . xii Page 12 74 75 76 79 80 81 97 102 Table Page 10. The Mean (K), Standard Deviation (S) and Standard Error of the Mean (8—) for the Variables in Nigeria - Ilewo Data Set . . . 122 11. Zero Order Correlation Matrix Among the Variables Used in the Nigeria - Ilewo Data Set . . . . . . . . . . 123 12. Results From Stepwise Multiple Regression for Awareness of Technological Innovations In Nigeria - Ilewo Data Set. . . . . . 126 13. Results From Stepwise Multiple Regression for Adoption of Technological Innovations in Nigeria - Ilewo Data Set. . . . . . 129 14. The Mean (K), Standard Deviation (S), and Standard Error of the Mean (S ) for the Variables in the Native Engligh-Speaking Group . . . . . . . . . . . 131 15. Zero Order Correlation Matrix Among the Variables Used in the Native English- Speaking Group . . . . . . . . . 132 16. Observed Frequencies and Chi Square Values for Tests of LGRADE with Four Simple/ Complex Innovations in Native English- Speaking Group . . . . . . . . . 135 17. Results From Stepwise Multiple Regression for Quantitative Literacy in the Native English-Speaking Group . . . . . . . 137 18. Results From Stepwise Multiple Regression for Knowledge of Technological Innovations in the Native English-Speaking Group . . . 138 19. Results From Stepwise Multiple Regression for Adoption of Technological Innovations in the Native English-Speaking Group . . . 140 20. The Mean (K), Standard Deviation (S), and Standard Error of the Mean (S ) for the Variables in the Spanish-Speafiing Group. . . 142 xiii Table Page 21. Zero Order Correlation Matrix Among the Variables Used in the Native Spanish- Speaking Group . . . . . . . . . 143 22. Observed Frequencies and Chi Square Values for Tests of LGRADE With Four Simple/ Complex Innovations in Native Spanish- Speaking Group . . . . . . . . . 146 23. Results From Stepwise Multiple Regression for Quantitative Literacy in Native Spanish-Speaking Group . . . . . . . 148 24. Results From Stepwise Multiple Regression for Knowledge of Technological Innovations in the Native Spanish-Speaking Group . . . 149 25. Results From Stepwise Multiple Regression for Adoption of Technological Innovations in the Spanish-Speaking Group . . . . . 151 26. The Mean of Knowledge of Technological Innovations for Spanish-Speaking Group . . . 154 27. One Way ANOVA Summary Table for Knowledge of Technological Innovations Among the Literacy Levels in the Native Spanish- Speaking Group . . . . . . . . . 155 28. The Mean of Adoption of Technological Innovations for Spanish-Speaking Group . . . 156 29. One Way ANOVA Summary Table for Adoption of Technological Innovations Among the Literacy Levels in the Native Spanish- Speaking Group . . . . . . . . . 156 30. Summary of Results Supporting and not Supporting the Theoretic Hypotheses in the Nigeria — Ilewo and the USA - Michigan Sample Spaces . . . . . . . . . 160 xiv Figure l. 2. LIST OF FIGURES Page Modes of Language Use and Their Respective Skills . . . . . . . . 59 A Schematic Ordering of Literacy Levels Among Adult Populations in Developing Countries . . . . . . . . . . 66 XV CHAPTER I THEORETIC RATIONALE AND HYPOTHESIS A. INTRODUCTION Literacy has been used extensively in the study of factors influencing knowledge and adoption of technological innovations. The results have been mixed but generally show positive correlations. For instance, Rogers with Shoemaker (1971) report, among other generalizations in diffusion studies, that 24 (63%) studies support but 14 (37%) studies do not support the generalization that "earlier adopters are more likely to be literate than are later adopters" (p. 357); 61 (76%) studies support but 19 (24%) studies do not support the generalization that "earlier adopters have greater knowledge of innovations than later adopters" (p. 374); 32 (74%) studies support but 11 (26%) studies do not support the generalization that "change agent contact is positively related to higher education and literacy among clients" (p. 381); 17 (71%) studies support but 7 (29%) do not support the generalization that "earlier knowers of an innovation have more education than later knowers" (p. 347), and 203 (74%) studies support but 72 (26%) studies do not support the generalization that "earlier adopters have more years of education than do later adopters" (p. 354). 1 2 The literature reviewed revealed two omissions in prior explications of literacy, which, to this author, have prevented a more thorough test of the relative impact of literacy on individual's knowledge and adoption of techno- logical innovations particularly in Developing Countries. The two omissions2 included: (a) literacy in quantitative symbols and concepts, and (b) literacy in English in those areas or communities where English is not the dominant language in daily discourse of the majority of the peeple. Literacy in quantitative symbols and concepts refers to the skill in the use of quantitative symbols and concepts. A quantitative symbol or concept derives from number,3 pro- portions of numbers (e.g., fractions, ratio, percent, etc.), and any statistical and mathematical structures and abstrac- tions (e.g., symmetry, transitivity, ordinality, cardinality, probability, etc.) which are used to express the quantitative or conceptual character of phenomena such as technological innovations. In contrast, a non-quantitative symbol or concept, by definition, includes that portion of spoken or written natural language which is devoid of the quantitative or con- ceptual character of number. Numerals may, however, be associated with a non—quantitative symbol or concept such as in denoting a technological innovation. In that case, the numerals so used merely denote the technologiCal innovation rather than the quantitative nature or number abstraction of 3 the innovation; e.g., Aphex 70. Accordingly, it was con- ceived that literacy in English, which refers to the ability to read and/or write the English symbol system, is a subset of non-quantitative literacy. In areas or communities where English is not the dominant language in daily discourse, other subsets of non-quantitative literacy would include literacy in native languages or dialects in those areas or communities. The purpose of this study therefore is to explore the relative potential effects of the two omissions noted above on predicting and explaining the knowledge and adoption of technological innovations. This exploration will involve a reconceptualization of literacy, the development of a quan— titative literacy test instrument, and the test of the recon- ceptualization of literacy in two different settings, viz.: in Nigeria - Ilewo and USA - Michigan. B. LITERATURE REVIEW There is a plethora of documented reports and expres- sions of scholars and leaders throughout the world on the impact of literacy4 on modernization and development vari— ables.5 A complete review of such literature would be superfluous (if not impossible) for the purpose of this study. Hence, this author will review in this section only some of the representative literature on literacy, and this will be done along four general headings: (a) Literacy 4 correlates, (b) Literacy function, (0) Skewed regional dis- tribution of literacy, and (d) Prior conceptualizations of literacy. 1. Literacy Correlates A cursory review of the literature on literacy indi- cates that literacy appears to affect the processes which manifest themselves in more modern attitudes and behaviors (see Lerner, 1958; Frey, 1964; Doob, 1961, 1965; Mendez and Waisanen, 1964; Lassey, et a1., 1965; Rogers and Herzog, 1966; Herzog, 1967; Wright, et a1., 1967; etc.). Schuman, Inkeles, and Smith (1967) found significant correlations in East Pakistan between literacy attainment and both the level of political identity and willingness to consider change. Lerner (1964) in Turkey, Rogers and Herzog (1966) in Colombia, and Rahim (1961) in Pakistan, all found highly significant correlations between literacy and exposure to mass media channels (radio) newspapers, and film) and aware- ness of new opportunities. Summarily, significant positive correlations have been observed between literacy and five indices of modern- ization:6 empathy (Lerner, 1958; Rogers and Herzog, 1966); achievement motivation (Rogers with Neill, 1966); cosmopo- liteness (Lerner, 1964; Rogers and Herzog, 1966); mass media exposure (Rogers, 1966; Lerner, 1963; Deutschmann, 1963), and political knowledge (Lerner, 1958; Rogers and Herzog, 1966). 5 Studies involving industrial labor productivity have also shown significant positive correlations with literacy. Investigations in the USSR show that elementary literacy attained during each year of primary schooling increases labor productivity by an average of 30 per cent, and that one year of formal education is twice as effective as one year of on-the-job training in terms of productivity (Adiseshiah, 1970). Desai and Punalekar (1971) studied the relationship between literacy and economic productivity of industrial workers in Bombay, India. In general, they found that compared to illiterate workers, literate workers: (1) consistently performed day-to-day factory duties more efficiently, (2) showed a far greater understanding of the production process and a more develOped sense of responsi— bility toward their work, (3) were more self-sufficient and more apt to join modern types of social organizations out- side the world of the factory, (4) considered themselves much more self-reliant in the important domestic and civil activities, and (5) were much better acquainted with the co- operative credit society and medical benefits. Hoiberg, Hysham and Berry (1974) sought to determine the neurOpsychiatric implications of illiteracy among the U.S. Navy recruits in the Naval Training Center in San Diego, California. They found substantially more discharges for neuropsychiatric reasons among the illiterate enlistees who had been assigned to an Academic Remedial Training 6 Division (ART) than in a matched control group of literate men who had not been assigned to ART. They concluded that the enlistee who needs academic remedial training is a four times greater neuropsychiatric risk to the Navy than is the literate. The investigators noted that the conclusions drawn over 20 years ago by Hunt and Wittson (1951) are still valid. That is, individuals who need academic remedial training continue to be a greater neuropsychiatric risk to the military than are literates. Recently, Stauffer, et a1., (1978) investigated the abilities of literates and nonreaders to recall and use in- formation from a national network television news program. The study involved 67 literates from a small, private college in suburban Boston and 61 adult basic education (ABE) students as nonreaders from Philadelphia Adult Basic Education Academy and the Adult Basic Learning Centers in Worcester and Brockton, Massachusetts. Among other findings, they observed that: (l) the literates recalled 55 per cent more stories than the nonreaders, (2) the literates gained 63 per cent more information from the news program than the nonreaders, (3) the two groups were virtually identical in their use of and Opinions about television news, and (4) among the ABE students, significantly higher knowledge scores were achieved by younger students with more formal education and higher reading achievement levels. 7 These findings led the investigators to conclude that the remarkable memories of nonliterates in tribal societies, which have been noted by many scholars (e.g., Riesman, 1956; Junod, 1927; Cole, et a1., 1971; etc.) should not be assumed on the part of nonliterates in a technological society such as the U.S. Secondly, they argued that despite observations of compensatory "common sense" develOped by functional illit- erates in technological societies, this characteristic may not extend to an ability to recall and use information from television news with the same efficiency as literates. The study indicated that these populations cannot obtain infor- mation from television with equal ease. The investigators speculated that one reason for this difference may be that the educational process that develops reading and writing skills also enhances the ability to de- code visual and oral information. Test scores of ABE stu- dents were positively related to higher levels of reading achievement and formal education. In addition, the research- ers reasoned that one other reason for the difference may be due to the difficulty of the language used by network news writers. Their analysis of random samples of transcripts of the newscast used in the study yielded a rating of "fairly difficult" (a category above "standard") on the Flesch Formula (Flesch, 1952) and a grade level rating of 13.0 on the Gunning Formula (Gunning, 1952). 8 Although these two techniques are an imperfect measure of the difficulty of spoken English, nevertheless the inves- tigators argue that the findings from the analyses of the transcripts suggest that the oral difficulty of television news (complex sentence structure, multisyllabic words, use of highly specialized vocabulary) may constitute a considerable problem for the functional illiterate. 2. Literacy Function Extensive use of literacy has generally been based on two interrelated convincing arguments for literacy. First, it is argued that if a person is not literate, he/she cannot access print information. Second, literacy has a profound consequence for the cognitive structure and, ultimately, on the communication behavior of the persons endowed with it. That is, literacy alters the individual's perceptions of the symbol-referent relationship. This second aspect of literacy is noted to be more important than just the mechanical ability to read and write (see Rogers with Svenning, 1969; Herzog, 1967; Doob, 1966; Burnet, 1965; Lerner, 1963). With literacy, change occurs in a number of mental abilities, such as a loss of eidetic ability, which becomes unnecessary according to Doob (1964) in his early research among Africans. Doob (1966) argues that eidetic imagery (a "photOgraphic" ability to remember stimuli) "must reflect a human ability which has survived from some earlier 0 «HA - .n-c‘ . . t!) ll. III u-u- ,_ r-.... .‘F A “44 (I) I u. u. A. 'u 4.. PA. (I) 9 evolutionary state and which has become virtually function- less in modern (literate) adults." According to Rogers with Svenning (1969), ". . . lit- eracy seems to be a key for unlocking more complex mental abilities. Whereas the illiterate is largely dependent on memorization of details, the literate individual is able to manipulate symbols, which allows counterfactual thinking. The ability to generalize through symbolization, the faculty of restructuring reality via the manipulation of symbols, and the ability to empathize with strange roles are all mental capacities that facilitate one's effective function- ing in a complex, rapidly changing urban-industrial world. Thus, one might view literacy as development of the funda- mental skills of reading and writing, which leads to or is accompanied by growth of a set of mental abilities that are necessary to modernization" (p. 71). This view of literacy has some support in research on literacy. For example, Carothers (1959), a psychiatrist, has reported a psychological impact of literacy among rural African tribes. Summarily, Carothers (1959) noted that in non-literate societies, no clear distinction is made between thought and reality. The spoken word is much more closely identified with reality for the non-literate and has, what Carothers termed a "majic power." What is heard and what is spoken is more important for the rural African than what is seen. The effect of literacy is to reduce this magic 10 efficacy of the word, to make words represent thought sym- bols, and thus to create a mental distinction between symbol and reality. This distinction enables literates to think in terms of symbols. Following Carothers' (1959) work, McLuhan (1962) claims that when use of one of the senses predominates, as does the aural among illiterates, the other senses become to some degree anesthetized. With literacy comes an arousal of the visual sense, thereby attuning the individual to both the audio and the visual messages being transmitted. Indeed, McLuhan's (1964) thesis that "the medium is the message" implies that the psychological impact or meaning of a mes- sage depends on the channel by which it is transmitted. Moreover, in formal education programs, reading experts have also noted the cognitive impact of literacy. Gray (1940), for example, has written broadly on the effects of learning to read in broadening one's outlook, deepening one's understanding, changing one's behavior, and stimulat- ing one's emotional and individual growth. Rogers (1969, p. 72), who has done extensive work on literacy programs in Developing Countries, has concluded that: "Literacy, . . . , contributes to the modernization process by (1) providing the means for print media exposure, (2) allowing the receiver to control the rate of message input, (3) facilitating the retrieval of print messages for delayed use, and (4) unlocking more complex mental abilities." Iru 'I dug... a 'LL. .- Cab. ‘n. O.“ ’v ‘Q. ‘1 he 11 Since literates seem better able to manipulate sym- bols and to think abstractly, one therefore expects literacy to have instrumental relevance in predicting and explaining peasant modernization. Literacy must be appreciated as an important facilitator of modernization, a process which re- quires the absorption and comprehension of a vastly increased amount of complex information. The individual who becomes literate has learned to learn for himself (Burnet, 1965:14). 3. Skewed Regional Distribution of Literacy Given the profound impact of literacy on modernization and development variables and on the cognitive structures of individuals, the disturbing fact, however, is that the dis- tribution of literacy rates between the major regions of the world is very skewed. That is, while the Developed Countries of the world enjoy high literacy rates, the Developing Coun- tries are severely plagued with a high incidence of illiter- acy among their populations of potential productive ages. Because of the general lack of credible statistics, it is difficult to assess confidently the extent of world literacy. However, the estimates by the Statistical Office of the United Nations (1977) are shown in Table 1 showing the breakdown by continent of the total number of adults, liter- ate adults, and the number and percent of illiterates in 1960 and 1970. These estimates show that in 1960 approxi— mately 735 million (39.3%) of the persons aged 15 years and above were illiterate. By 1970, the estimated percentage 12 one .mouom mo UHHQsmom m.oamoom OHDMHOOEoQ .mcwno mo Deansmom m.mHmoom moosaoxm .IIII. .pnma .mmaum ea moaumaumum oHHOB ..M.z .xuow 3oz .wcofiumz GDUflnD on» no moflmmo HMUHumHumum "mousom .Emcumfl> mo Deansmmm UHUMHOOEDQ HmEHom * .ouflu3 one omen anon on muflawnm on» we oocwmoo mm3 homnouaq fl m.oa H NH ma m.HH H a HH maammoo m.m ma mom Hmm m.m em mme ems mmm: a muonsm m.ma mum mmo hm~.H «.mm Nam ova mam enema m.m~ am mmfl mma m.~m ov mm mma aofiumaa enema m.H N mma HGH e.m m mma sma moaumaa nuuoz a.ma med Hm «ma o.Hm ama mm mmH aoflnma m.vm mma som.H em~.~ m.mm mma ama.a mem.a *eauoz w .02 w .oz measea measea muasea mussea muasea measea muaumufiHHH mumumuaq Hmuoe mumnmuaaaH mumumuaq Hence onomm oema coma .osma cam oeaa Hoe maoaaflaz an Inm>o can memos ma omens measea oDMHoUHHHH can muazoc «oumuouflq .muasod mo cowusnfluumfln Hmcowmmm can pauoz .H manna 13 had decreased to 34.2, but the absolute figure had increased to 783 million people. Meanwhile, the estimated number of literate adults (aged 15 years and above) in the world increased from 1,134 million in 1960 to 1,504 million in 1970. The majority of the illiterates are in Africa (81.0% in 1960 and 73.7% in 1970), Asia (55.2% in 1960 and 46.8% in 1970) and Latin America (32.5% in 1960 and 23.6% in 1970). The Developed Countries in North America, Europe and the USSR, and Oceania continued to enjoy high literacy rates throughout the same periods. UNESCO (1965) analysts have estimated that the in- crease in the number of illiterates in a given country is related to the illiteracy rate by a correlation coefficient of 0.55. That is, countries with high rates of illiteracy (70 per cent or more) have a propensity for increasing the absolute numbers of adult illiterates while countries with relatively low illiteracy rates (35 per cent or less) tend to lower both the rate and absolute number of illiterates. Since Developing Countries are the ones with high rates of illiteracy, these figures clearly demonstrate the magnitude of illiteracy problems which these countries face. Furthermore, the proportion of female illiterates generally exceeds that of males, often significantly. In at.least three countries-~Saudi Arabia, Somalia, and Yemen-- ‘the total adult female populations are reported to be 14 illiterate while in many others the figure is over 90 per cent (UNESCO, 1965). The International Institute for Adult Literacy Methods which was established by UNESCO and the Government of Iran in 1968 reported in 1974 that there are ". . . more than 800 million illiterates throughout the world . . . . Despite what has been done and what is being done, the number of il- literates is not decreasing. In fact, there are more illit- erates today than there ever have been and by the eighties they are likely to total more than 800 million" (p. 3). Although the figures indicate that illiteracy is most prevalent in Developing Countries, it must be noted that they do not show that by far the greatest number of unedu- cated come from rural areas in those countries where agri- culture forms the backbone of the national economy. It is strongly argued that the debilitating effects of illiteracy are very complicated but lucid: . . . it is precisely in the areas where il- literacy rates are highest (parts of Africa, Asia, and Latin America) that development lags farthest behind the rest of the world. Here we find lowest per capita income rates, most rapid increases in population, and least developed systems of communication and transportation. Illiteracy is a part of the vicious cycle that hobbles underdeveloped nations: without literacy, special skills cannot be taught; without special skills, agriculture cannot be modernized or industry developed; without a modernized agriculture and industry, production and income will not increase; without income, there are no re- sources to develop education and literacy. Literacy is viewed as a possible input to 15 alter the inertia of the system and break out of the cycle (Herzog, 1967, p. 2). Thus, from the above and similar assertions, one sur- mises that illiteracy prevents many of the segments of popu- lations of productive age in the Developing regions of the world from participating and enjoying the benefits of tech- nological advances in several fields such as agriculture, health, child welfare, industry, social development, etc., to mention only a few. Lack of ability to make use of technological knowl- edge in these fields means that many nations in the so-called Developing regions are not developing as rapidly as might otherwise be the case. "Gunnar Myrdal, in his important study of social and economic conditions in countries in Southeast Asia, Asian Drama, makes it quite clear that tech- nological development is greatly slowed down, and may even make little or no headway in raising the standards of life for the people in a number of countries in this area, simply because of the weight of illiterate numbers in the popula- tions" (World Education, 1970, p. 11). 4. Prior Conceptualizations of Literacy The purpose of this section is to review some of the available representative literature in which literacy has been explicated. From this review will derive a new and more rigorous explication of literacy. This new explication is based on the contention that the traditional one- .1 'v..v .‘...- .a no» 5"! :n.. vud , 5.. ‘u v” \LI 5- h u... '{A "v 2‘. u.“ (I) ,. n. I 16 conceptualizations of literacy have been inadequate since they have generally conceived literacy in terms of individu- al ability to read and/or write only in some natural native language, and have not attempted to provide uniform classi- fication of literacy levels. Although the measurement of literacy is much in vogue today among educationalists, scholars, and social-change engineers, its historical origins and interest seem to be unknown. That is, it is not clear when and why man first became interested in the measurement of literacy. However, in the United States, early interest in the measurement of literacy seems to have developed in associa- tion with immigration laws of this country. In his thesis on The Literacy Test for Immigrants, 1886 - 1917, Houdek (1957) attributes the early literacy test in the United States to Edward W. Bemis,-an economist, who proposed that the United States "Admit no single person over sixteen, and no man over that age who cannot read and write his own language" (Bemis, 1888, p. 263). Bemis argued that his pro- posal would help to maintain America's high standard of living and aid American labor by shutting out fifty per cent of the Polish, Hungarian, and Italian immigrants. Although most of the poeple who discussed the test during the early years of its history favored both a reading and writing test, the proposal was later modified to a simple test of reading skill, and as to the language in 17 which the immigrant was to prove his literacy, it was usual- ly agreed that it should be in "English or some other lan- guage" (House Report, No. 140, 1913). Level of literacy in terms of reading skill then became the key determinant of allowing immigrants from Europe to the United States. How— ever, Houdek (1957, pp. 4-5) states: Where these workers got the idea of an educational test is a matter of question, for Bemis supposedly only lectured on the idea as far back as 1887. Thus it seems quite possible that either someone else had promulgated the idea before him or that the idea had been in existence for some time, but had not become as popular as other restrictive and selective proposals. The questions of who initiated the idea of a literacy test and the purpose for which it was developed are not par- ticularly significant. However, it is important to know that over the years the measurement of literacy has been conceived in a number of different ways. This is important because, as it was noted earlier, literacy bears great utility in practical and research endeavors in moderniza- tion and deve10pment activities. Various literacy definitions seem to fall into two rather general categories:7 (a) planning-census-type definitions and, (b) empirical-type definitions. a. Planning-Census-Type Definitions of Literacy. A planning-census-type definition of literacy is here conceived as literacy by fiat of the interviewer or a self-report of 18 the interviewee on his/her literacy skills such as reading and writing. A fiat definition of literacy usually uses grades of school (years of school) with which to estimate an individual's literacy skills (reading and writing). As the name implies, planning-census-type definitions of literacy are usually applied during population census to get informa- tion quickly on literacy estimates for national planning. Some examples and related discussions of planning-census- type definitions follow. Discussing "The measurement of literacy in Pre- industrial England," Schofield (1968) reports that literacy was conceived as the ability to sign one's name. This method of literacy test was very much in use in pre-indus- trial England particularly " . . . when large numbers, or whole classes, of peOple were required to attest their ap- proval of a document by signing their names if they could, or if they could not sign by making a mark. These could be situations, analogous to a census, in which virtually every- one was required to attest his approval of a document" (Schofield, 1968, p. 319). According to Schofield (1968), the occasions on which this occurred in pre-industrial England included: (1) the Protestant Oath of 1624, which had to be taken by all males over the age of eighteen to the effect that they would "maintain and defend the true Reformed Religion expressed in the Doctrine of the Church of England against all Poperie l9 and Popish Innovations," (2) the Test Oath of 1723, promis- ing allegiance to George I and renouncing the jurisdiction of the Pope, which had to be sworn by everyone over the age of eighteen and, (3) the Anglican Marriage Register, which from 1754 contained entries of all marriages other than those of Jews, Quakers, and members of the royal family. This register was due to an act of Parliament of 1753 which accorded legal validity only to marriages registered in Anglican registers and signed by the parties and two witness— es. In 1837, other denominations were licensed to register marriages and a state system of registration was begun. There are, however, serious methodological drawbacks in the use of signatures and marks as testimony of literacy in wills, allegations and bonds for marriage licenses, and the deposition of witnesses in ecclesiastical courts (for details, see Schofield, 1968, pp. 320-325). In general, it is to be noted that the ability to sign one‘s name or to make marks for attesting approval of a document are conceptually very imprecise measures of lit- _eracy since they do not consider the extent to which the individual has acquired literacy skills (reading and/or writing). Schofield (1968) convincingly states: . . . historians have . . . made the problem worse for themselves by being imprecise as to what they mean by literacy. This has meant that the level of literacy skills con- sidered appropriate in any historical con- text has rarely been adequately specified. This is perhaps not surprising as it is 20 seldom easy to decide what this level should be. For example, in a discussion of the role of literacy in the history of politics, is the ability to write relevant? or is the ability to read sufficient, and if so to what level? enough to understand a simple handbill, or the works of Locke? For economic history the difficulties are even greater. For ex- ample, any assessment of the relationship be— tween literacy and industrialization entails decisions as to the levels of literary skills necessary to the introduction of the new techniques in agriculture and a wide variety of industries on the one hand, and to the re- placement of traditional patterns of consump- tion and the generation of a mass market demand on the other. At least for the English industrial revolution it would seem that these necessary levels of literary skills varied widely in different sectors of the economy. The meaning of literacy therefore changes ac— cording to the context, and it is the respon- sibility of the historian to specify the appropriate level of literary skills consist- ent with his understanding of the context (pp. 313-314). Other examples of planning-census—type definitions of literacy may be noted. Until the 1940 decennial census in the United States, illiteracy was determined by asking adults whether they could read and write. Later, literacy was defined as equiv— alent to having completed six grades of school (Rogers with Svenning, 1969). Harman (1970) reports that the U.S. Bureau of the Census defines illiteracy as "the ability to read and write a simple message either in English or any other lan- guage" (cited from Current Population Reports, 1963, p. 20). In the Colombian census, literacy is measured on the basis of an individual's ability to write his name. Other 21 national censuses determine literacy by asking individuals if they can read a newspaper and write a letter (Rogers with Svenning, 1969). Illiteracy is defined as inability to read or write in Portuguese in Angola and the Republic of Cape Verde; to read and write either French or Arabic in Chad; to read and write French in Gabon and Senegal; either to read or write Sesuto in Lesotho; both to read and write English in Swazi— land; persons with no schooling are defined as illiterates in Sudan, Uganda, Hong Kong, and Japan; both to read and write a simple letter in any language in West Malasia; and to read or write in any native language in Oceania (Cook Islands, Gilbert Island, Niue Island, and West Samoa) accord- ing to UNESCO's Statistical Yearbook (1976, pp. 43-59). Finally, there is a wide disparity in the age groups included in national rates of literacy. For example, Indo— nesia calculates its literacy rate for persons between 13 and 45 years of age; Cuba and Malaysia report literacy rates for those 10 years of age and over, and Bulgaria includes only people who are more than 15 years old (Rogers with Svenning, 1969). b. Empirical-Type Definitions of Literacy. Unlike the planning-census-type definitions of literacy which are by fiat and/or self-reporting, empirical-type definitions of literacy are here conceived as literacy measures which are usually task-oriented. That is, they usually have an 22 a priori set of measures or instrument through which the level of literacy of an individual may be determined by the individual's performance on the instrument. Consequently, they are more rigorous than the planning-census-type liter- acy "tests." They usually consider individual's reading and/or writing skills and understanding of what is read. Some consider even the actual application of what is read in programs currently known as "functional literacy." Some ex- amples of the empirical—type definitions of literacy are worth noting. More than four decades ago, Huse (1933) in discussing the reading needs of citizens of a democracy, gave vigorous emphasis to the importance of a clear grasp of the meaning of what is read. In his judgment, reading for understanding is to be contrasted with mechanical reading. It involves the translation of the meaning represented by the symbols into understandings that can be expressed in the reader's own words. Equally important is their translation "into terms of purpose, authority and validity" (p. 8). Unless this is done, "the public is the inevitable victim of fraud both commercial and literary," and "the mental life of the people may be corrupted" (p. 9). In Huse's view, a high level of capacity to translate is an indiSpensable requisite of a literate citizen. Compelled by their interest in the con— cept of functional literacy during World War II, the U.S. Army defined illiterates as "persons who were incapable of 23 understanding the kinds of written instructions that are needed for carrying out basic military functions or tasks" (Current Pppulation Reports, 1963, p. 23). A 1970 confer- ence on planning strategies for a national adult "right to read" movement decided that adult literacy assessments should be made independent of grade equivalents: The challenge is to foster through every means the ability to read, write and com- pute with the functional competence needed for meeting the requirements of adult living (Conference on Strategies for Gener— ating a National 'Right to Read' Adult Movement, Raleigh, North Carolina, 1970). UNESCO (1969) has been involved in a literacy teach- ing program which has been worked out to reduce the normal time of reading lessons by half using a computer which de- termines the frequency of words and syllables used by local workers. UNESCO reports one such program in a Brazilian mining company: The use of a computer in a Brazilian project is expected to cut by half the amount of time needed to learn to read. The CVRDC Mining Company, Brazil, which has started a function- al literacy programme for its staff with UNESCO- assistance, has used a computer to determine the frequency of words and syllables used by local workers. This literacy programme is closely linked to the technical promotion and vocational training of the staff. The computer has shown that the basic vo- cabulary of 2,300 words is made up of a total of 540 different syllables. Sixty per cent of the words use as little as 9 per cent of the syllables, meet 80 per cent of the speaking re- quirements. On the basis of these data, a teaching programme has been worked out which should cut the normal duration of reading lessons by half (p. 15). 24 Other UNESCO conceptions of literacy falling under the purview of empirical-type definitions of literacy may be noted. In 1951, a UNESCO committee conceived that a person is literate when he can "both read, with understanding, and write a short simple statement on his daily life" (Gillette, 1972, p. 22). In their interest in the concept of function- al literacy, another UNESCO committee came up with a defini- tion of literacy in 1962 when they stated: A person is literate when he has acquired the essential knowledge and skills which enable him to engage in all those activities in which literacy is required for effective functioning in his group or community, and whose attain- ment in reading, writing and arithmetic make it possible for him to continue to use these skills towards his own and the community's development (Gillette, 1972, pp. 23-24). In a Final Report of the Regional Workshop for Special- ists and Officials Concerned with the Preparation of Reading and Follow-up Materials in Asia, Bangkok, 25 November-13 December, 1968, UNESCO (1969) asserts that the Workshop felt the need for adOpting some workable standard of literacy in terms of three R's. The Worksh0p discussed the Literacy Scale used in Laos for drafting a work-oriented literacy pro- ject in Laos. The scale establishes six levels of literacy: LEVEL I l. Able to hold a pencil___ 2. Copies simple figures___ 3. Tells time by the clock___ 4. Writes one-figure numbers___ 25 LEVEL II LEVEL V 5. Adds and subtracts one___ 17. Adds, subtracts, 6. Writes his/her name___ multiplies, and 7. Reads separate letters___ divides with 3 8. Writes from dictation figures numbers with two figures___ 18. Reads seHEEnces word by word___ LEVEL III 19. Writes simple sentences 9. Adds, subtracts and multi- 20. Writes any___ plies numbers of two number___ figures___ 10. Writes separate letters___ LEVEL VI 11. Reads usual words___ 12. Writes from dictation 21. Knows geometrical numbers with 3 figures___ figures___ 22. Reads fluently___ LEVEL IV 23. Drafts a text___ 24. Does simple opera- 13. Knows the square, the tions in metric diameter system 14. Reads simpIe words___ '_—_ 15. Writes phonetically words___ 16. Writes from dictation numbers with 5 digits___ Number passed = = Level 4 Source: UNESCO (1969), Work-Oriented Functional Literacy; Reading and Follow-upiMaterials: Final Report of the Regional Work- shop for Specialists and Officals Concerned with the Preparation of Reading and Follow-up Materials in Asia, Bangkok, 25 November-l3 December 1968 (Appendix D). This scale was devised for testing individuals to be employed in factory work. The Workshop participants correct- ly observed that "Level VI, which is supposed to correspond to a level of proficiency equal to the level of a sixth grade school leaver does not consider the mastering of the 26 simple calculations needed for agriculture. The Workshop suggested that other countries might adapt this Literacy Scale-—taking into account the level of instruction required by a particular development programm -- or evolve a new one to suit their requirements . . . (pp. 6-7). As may be noted above, UNESCO's literacy measures seem to vary not only over time but also over space. Else- where, for instance, UNESCO (1971) uses what its experts call "Attainment Tests" for determining the level of literacy: These have been used, on the one hand, for evaluating the level of literacy proper, especially in four aspects: rapid calcula- tion, solution of easy vocational problems, understanding of another person's thoughts expressed in writing (e.g., a technical leaflet), and ability to express oneself in writing. On the other hand, and concurrent- ly, they are utilized in the Experimental World Literacy Programme for evaluating the knowledge acquired in the specific field covered by each programme deal - that is, technical and vocational knowledge and socio- economic knowledge of vocational relevance (P- 7). In his work on "Literacy and Community Economic Devel- opment in Rural Brazil," Herzog (1973) reports that "literacy, the dependent variable, was measured by the farmer's score on a 50-word oral reading test derived from the final lesson of an adult literacy primer used in Minas Gerais" (p. 332). In evaluating Pilot Projects, UNESCO has also made use of "Attainment Tests" in an attempt to determine literacy levels among the populations concerned. "These have been 27 used, on the one hand, for evaluating the level of literacy proper, especially in four aspects: rapid calculation, solu- tion of easy vocational problems, understanding of another person's thoughts expressed in writing (e.g., a technical leaflet), and ability to express oneself in writing. On the other hand, and concurrently, they are utilized in the Ex- perimental World Programme for evaluating the knowledge ac- quired in the specific field covered by each programme deal-- that is to say, technical and vocational knowledge and socio— economic knowledge of vocational relevance" (UNESCO, 1971, p. 7). In general, the following problems are to be noted in prior conceptualizations of literacy. First, prior measures of literacy depend both on the honesty of the respondent and his/her ability to assess ac- curately his/her own competence in reading and writing. However, in situations where individuals think that it is not acceptable to be illiterate, such as in Urban areas (see Burnet, 1965: ll; Freeman and Kassenbaum, 1956), or where they have little opportunity to maintain a former competence in reading and writing, as in peasant communities (see Rogers with Svenning, 1969; Singh, 1970; Ahmed, 1973), self-defined literacy is likely to be a relatively less accurate measure. Second, prior conceptualizations of literacy have generally been concerned with literacy only in some natural native language. However, in diffusion practices involving 28 technological innovations such as in the African and Asian countries, such measures may be inadequate in predicting and explaining knowledge and adoption of technological innova- tions since the native languages in those countries are fre- quently not associated with the technological symbols and concepts. The language of technology is frequently English in which the clients of technological innovations are gen- erally illiterate. This point will be elaborated in the next section. Note further that even the measures of liter- acy in native language per se have been imprecise since they do not provide a clear understanding of the extent to which the individual has acquired literacy skills in the language concerned. Third, prior literacy conceptualizations which have attempted to use quantitative symbols and concepts have been less rigorous since they have generally tended to as- sume that the individuals concerned already possess an elab- orate numbering system in their native languages, and that they can use it properly in the solution of their develop- ment problems. But this may be a far cry from reality as it will be shown in the next section. Finally, prior conceptualizations of literacy have generally not been uniform. That is, they have not attempted to provide one classification scheme or typology of literacy levels on which to array all people over time and space. Such a scheme would be more useful in predicting and explain- ing an individual's knowledge and ad0ption of technological 29 innovations particularly in Developing Countries. This study will provide later in this chapter (see section C. 4) a scheme which reconceptualizes literacy. 5. Symbolism in Diffusion-Adoption Processes This section is concerned with the general types of symbol systems in diffusion-adoption processes and with the specific types of symbol systems which are potentially critical in diffusion—adoption processes of technological innovations involving Deve10ping Countries. The latter symbol systems are concerned with literacy in quantitative symbols and concepts and with literacy in English which, as noted earlier, is a subset of literacy in non-quantitative symbols and concepts. Both of these aspects of literacy were identified earlier in this chapter as the two aspects of literacy which have generally been overlooked in diffu— sion practices. a. General Types of Symbol Systems in Diffusion- Adoption Processes. To determine the extent to which liter- acy in quantitative symbols and concepts and in English are both important in the diffusion-adoption processes in Devel- oping Countries, requires first the identification of they general types of symbol systems which may be present among the change agents and the clients of innovations in these areas. From these general symbol systems will be selected the specific types of symbol systems which are particularly germane to the diffusion and adoption of technolgoical 3O innovations in Developing Countries. Here, a change agent is to be conceived as a professional who influences innova- tion-decision in a direction deemed desirable by a change agency (Rogers with Shoemaker, 1971). An agricultural ex- tension worker is an example of a change agent. An agri- cultural department or ministry of agriculture may be the change agency. Writing from an African context, Mwanika (1978) has identified two types of symbol systems by relating them closely to the speech communities which use them for commun- icating about technological innovations. The two general types of symbol systems include, firstly the symbol system of the clientele community and, secondly, the symbol system of the change agents. The first symbol system comprises the natural native language, or that which is often crudely referred to as the "mother-tongue." There may be 1, 2, 3, . . . , N mother- tongues in a community where N expresses the cardinality of the mother-tongues. Developing Countries frequently have multiple languages and/or dialects. This is certainly the case in Africa the context which this author is writing from. The second symbol system involving the change agents may be subdivided into three different symbol systems. One such symbol system is the natural native language or mother- tongue of the change agent. Besides his/her mother-tongue a change agent may of course speak other native languages in 31 his/her area of jurisdiction. The second symbol system is the natural language in which the change agent was educated and trained. This is usually English although it may be some other European language such as French, German, or Russian. Since the change agents often have their own mother-tongues, a language such as English is an adopted natural language in this context. The third symbol system is concerned with the artificial language of mathematics. The change agents had to learn this language not only for communicating quantita- tive information, but more so for encoding, decoding, and communicating more precise information which is a fundamental characteristic of this language. Individuals endowed with this language can deal with (analyze and synthesize) complex relationships among the phenomena of their environment as will be demonstrated shortly. By virtue of their education and training, change agents are therefore expected to be more adept in this language relative to their clients. b. Specific Types of Symbol Systems Important in Diffusion-Adoption Processes. This section focuses on two of the three symbol systems identified above as related to the speech community of change agents. More specifically, it focuses on the quantitative symbols and concepts of the artificial language of mathematics and on the English symbol system. The importance of literacy in each of these sets of symbol systems in diffusion-adoption processes is to be discussed below. 32 i. The Importance of Literagy in Quantitative Symbols and Concepts in the Diffusion-Adoption Practices. Basically, there are two reasons for the importance of lit- eracy in quantitative symbols and concepts in the diffusion and adoption of technological innovations; viz., the speci- fications of technological innovations are quantitative in nature, and the economic decisions involved in the use of technological innovations are also frequently quantitative. With respect to the quantitative specifications of technolgoical innovations, it is to be noted that many im- portant technological innovations come in specific calibra- tions, formulations, rates of application, and so forth. An agricultural extension worker, for instance, very frequently recommends to a farmer to apply so many pounds/kilograms of a given chemical per gallon/litnaof water or per given sur- face area; to plant a given crOp at so many feet (inches) or meters (centimeters) between rows and between plants in a row; to plant a certain number of seeds per hole or per sur- face area where seeders or planters are not a suitable or available choice; etc., etc. To be sure, these recommenda- tions may be conceived as lessons which the extension workers teach the farmers. The farmers must learn them in order to apply properly the technological innovations to their farm enterprizes. However, since the quantity demanded of such an innovation as a chemical fertilizer is frequently a vari- able among farmers, it must then be understood that extension 33 lessons or recommendations are but sets of standards taught for the proper application of the innovation. Each farmer must determine his/her own fertilizer requirements using the rate of application as the standard, and the area of his/her farm which needs fertilizing. This may involve fractional amounts which may not be present in the native language of the clients as it will be documented shortly below. The importance of quantitative symbols and concepts in the economic decisions involved in the use of'technolog- ical innovations is based on the fact that "the reception given to a new idea is not so fortuitous and unpredictable as sometimes appears to be. The character of the idea is it- self an important determinant" (Barnett, 1953, p. 313). That is, the characters of innovations or "attributes of innova- tions,"8 according to Rogers with Shoemaker (1971, p. 13) are important predictors of the adoption of innovation. One important attribute of an innovation is relative advantage which refers to the degree to which an innovation is perceived as being better than the idea it supercedes (Rogers with Shoemaker, 1971, p. 133). This attribute has been found to be positively related to the rate of adoption of innovations (see Kivlin, 1960; Fliegel and Kivlin, 1962a, 1962b; Tucker, 1961; Fliegel and Kivlin, 1966; Patrini, 1966; Kivlin and Fliegel, 1967a, 1967b; Fliegel, et a1., 1968). 34 The degree of relative advantage is often expressed in economic profitability. However, there are ". . . a number of subdimensions of relative advantage: The degree of economic profitability, low initial cost, lower perceived risk, a decrease in discomfort, a savings in time and effort, and the immediacy of the reward" (Rogers with Shoemaker, 1971, p. 139). Quantitative comparisons become very important in de- termining the relative advantage of any innovation particu— larly in farm enterprize substitution or combination. Economists in general and agricultural economists in particu- lar would assert that a farmer who is thinking of substitut- ing one farm practice with a new one must naturally compare the relative advantage or economic profitability which is anticipated from the new practice to that observed in the old one. This, however, requires only a simple substitution decision. Otherwise, a farmer may (as is often the case) be faced with more complex comparative and substitution decisions. Such decisions may arise if a farmer is consider- ing combining certain farm enterprises in some way on the basis of current and future market prices for farm inputs in general, and for farm outputs (produce). For example, given three crOps, viz., corn, peanuts, and soybeans, a farmer may decide to grow, say, only two of these crops in a given cropping season. In this case, such a farmer will be faced with three different groups (pairs) of relative 35 advantages to compare and choose from. More specifically, the three possible pairs or combinations of crops for this farmer will be either to grow: 1. Corn and Peanuts, 2. Corn and Soybeans, or 3. Peanuts and Soybeans. Each crop in each of these pairs is a bundle of economic profitability, risk, and the effort to grow it. Hence, each of these pairs is really a bundle of bundles. In considering economic profitability, for example, this farmer will have to consider several variables: the current market prices for seed, labor, fertilizer, pesticides, etc., as well as the expected selling price for each produce to be able to determine which combination of bundles of crops to grow. To do this, he/she will have to apply some basic quantitative or mathematical structures or models of equal- ity/inequality (reflexive, symmetry, and transitivity) which will enable him/her to make the necessary comparisons. All such comparisons require that the farmer be literate in quantitative symbols and concepts. The preceding paragraphs have attempted to show the use of quantitative symbols and concepts in diffusion-adop- tion processes. In addition, it is to be noted that this use is not by default but rather by design. The natural numbering system (1, 2, 3, 4, 5, 6, 7, 8, 9, 0) is well known to have important practical and scientific utility in 36 our lives as Judd (1927, p. 107) states: The number system which the race has devel- oped is a complex of symbols and of rules of combination. Some mental effort is nec- essary for the mastery of the system itself. In so far as this is true, arithmetic is a content subject. Equally true is the state- ment that the number system is a means of arranging the facts of experience in such a way that they can be dealt withyprecisely although they are quite chaotic in their own quality and order of presentation. Be- cause the number system helps the individual to arrange his experience, it is the indispen- sable instrument of all science and of com- merce where facts must be dealt with not in a chaotic way but in such a way that relations are defined and clearly recorded (emphasis added). This arrangement of the facts of experience in order to process them more precisely is a quantitative behavior or "quantitative thinking" which "takes place when an individ- ual uses numbers in some way in dealing with the elements of a situation that lend themselves to mathematical analysis or description" (Grossnickle and Brueckner, 1959, p. 308). The natural number system is frequently assumed to be a universal language (see Kramer, 1970; Alcksandrov, et a1., 1969; Judd, 1927; Smith, 1923; Urban, 1939; Hogben, 1951; Dantzig, 1954; Menninger, 1970; Grossnickle and Brueckner, 1959; Langbehn, et a1., 1972; Cassirer, 1953; etc.). Unfortunately, however, not all people (particularly those from the rural areas in DevelOping Countries) may be able to use this language adequately. Evidence in the lit- erature indicates that some cultures and subcultures of 37 human populations have a very limited numbering system in their natural languages or dialects. Such people have de- velOped categorical (nominal) quantity labels or number an- alogues for expressing the "quantitative" nature of objects (see, for example, Dantiz, 1954; Cassirer, 1953; Weitheimer, 1967; Menninger, 1970). Menninger (1970) reports that "some primitive peoples have completely fused the number and the object into a single entity. The Fiji Islanders, for example, call 10 boats bola, 10 coconuts kgrg and 1000 coconuts saloro. Naturally this does not hold for any arbitrary number (such as 5 nuts or 23 nuts). . . . The examples given show that the primitive people of the Fiji Islands have no number sequence, at least not an extensive one, that has been consciously and clearly detached from objects and thus become abstract" (pp. ll-12). The Detroit Free Press (December 21, 1976) reported under the "Guinness World Records" column that "least number- conscious people are the Nambiquara of the North West Moto Grasso section of Brazil who lack any system of numbers. They do, however, have a verb which means 'they are two alike.'" Moreover, besides the presence of unintelligibility and limitation in number qua number among some of the cul- tures or subcultures of the world, some people of other cul- tures also lack in their native languages or dialects equivalent concepts for the number concepts of proportion 38 such as per cent, fractions, to say nothing of the decimals. This author, for instance, observed the absence of equiva- lent concepts for all the rational numbers (fractions) except for one-half in five Ugandan languages and dialects with which he is most familiar. In these languages, any measure of magnitude or capacity which is less than unity is always expressed as a "half" even if the actual measure may be greater or less than one-half. That is, the native speakers of those languages do not have equivalent concepts in their languages for the various rational numbers such as 1/3, 2/3, 1/4, 3/4, 4/5, 7/8, etc. Therefore, they do not seem to understand the fact that between any two distinct (different) rational numbers, no matter how close, there are infinitely many other rational numbers. Between 0 and l, for example, there are infinitely many new "units" of l/N since an arbitrarily large denominator (N) may be selected. In other words, an indefinitely small quantity (i.e., a quantity as small as you please) may be selected. In math- ematical jargon, this infinite number of new units refers to the concept of "density." That is, rational numbers are pretty dense or thick. Other evidence indicates that limitations or unintel- ligibility in number has some connection to level of literacy. For example, from a series of experiments in a pilot study among rural illiterate and semi-literate Africans in Zambia, Fuglesang (1969) it was observed that the conservation of 39 substance (mass), quantity, number, and area; the concept of a straight line; special representation, concept of horizon- tality and verticality, and elementary logical concepts, i.e., concept of class, all do not exist or are unstable in illiterates. According to UNESCO (1972), "As a general rule, illit- erates are vague and imprecise about measures of length, area, weight and time. When they know how to count, it is hardly likely to be more than a hundred. Even if they know to add, subtract, or even multiply (by repeated additions), they are normally unable to divide" (p. 2). Evidently, the linguistic handicaps in number noted above should not be construed as these populations' inabil- ity to learn or comprehend the more complete numbering system which may be found in common use in some other natural lan- guages. It is a matter of language, and a more advanced facility in the use of language. Given a medium which facil— itates the acquisition of literacy as a more advanced facil- ity in the use of language, such populations should be able to acquire a more elaborate numbering system, i.e., literacy in quantitaitve symbols and concepts. This author contends that such a medium is formal education. This point will be developed later in this chapter (section C73). ii. The Importance of Literacy in Non—Quantita- tive Symbols and Concepts of English Language in the Diffus- ion-Adoption Practices. The importance of literacy in non- quantitative symbols and concepts of the English language in 40 the diffusion and adoption of technological innovations is based on three reasons. First, the "objects of diffusion" (i.e., technological innovations) particularly those of great importance in most of the Developing Countries are generally cast in scientific symbols and concepts whose nonmenclatural mold derives from the English9 language. For instance, today's change agent's kit is impregnated with such tools as perenox, gammalin, Aphex 70, aldrin, dieldrin, fertilizer, hybrid seed, IUD (Intra-Uterine Devices), Ariana, AI (Artificial Insemination), baby bottle, and so on, and so on, ad nauseam. It will be a rare case for such technological innovations to bear the nomenclature from the native languages or dialects of the clients in most of the Developing Countries particularly those in Africa and Asia. Second, the change agents are generally college or university graduates who, as noted earlier, are educated and trained generally in the English language. Thus, they are expected to be literate in English--the language of technol— ogy. Through their education and training, the change agents are naturally expected to be familiar with the (1) conceptual or denotative meaning (nomenclature), (2) intui— tive meaning of the non-quantitative symbols and concepts and (3) to be familiar with the numbering system and, hence, the quantitative specifications and economic decisions associated with the technological innovations, and (4) intuitive meaning 41 of the quantitative symbols and concepts of the innovations. For example, given some quantity of a chemical fer- tilizer such as sulphate of ammonia ((NH4)ZSO4), the change agent should know: that sulphate of ammonia is a specific name which is given to a particular category of fertilizers. It is distinct, on certain physical and chemical features or properties from, say, a potassium sulphate (K2804) or a potassium nitrate (KNO3) fertilizer or, indeed, from any other technological innovations such as those which were stated above. Finally, unlike the change agents who are generally college or university graduates, the rural clientele audi- ences in Developing Countries are not only generally illit- erate in their native languages, as it was documented earlier, but most of them do not speak nor are literate in English. The rural clientele audiences are mostly limited to their native languages for interpersonal communication among themselves and with the change agents if the latter can speak the native languages or dialects in their areas of jurisdiction. The clients who speak English, if any, would of course be expected to understand, to some extent, and to communicate directly with the change agents by virtue of the linguistically shared symbol system. However, for those clients who cannot speak this language (usually the majority), the change agents will either cognitively transform and 42 translate the English symbol system they are endowed with into the native symbol system of the clientele (if they can speak the latter) or they will have to seek the services of an interpreter to be able to "communicate" with their target audiences. Like the acquisition of quantitative literacy, the acquisition of literacy in English can also be facilitated through the channels of formal education as will be dis- cussed in the next section. C. THEORETIC FRAMEWORK AND HYPOTHESES Following the preceding discussions, a more rigorous explication of literacy is needed. Such explication would include both literacy in quantitative symbols and concepts, and literacy in the non-quantitative symbols and concepts of English language if diffusion practices involving tech- nological innovations among the generally illiterate popu- lations are to be effective. Literacy itself is to be conceived as a behavioral aspect in the use of language by which persons endowed with it can access and manipulate symbols and concepts in a lan- guage for the messages they convey. Therefore, the theoretic explanations for reconceptualizing literacy as a conveyer of linguistic symbols and concepts must be concerned with the behavioral aspects of language. 43 l. Theoretic Relation Between Language and Behavior Many scholars have for many years theorized and studied the development of symbol systems by humans and the relation of these systems to overt behavior. The theoretic explanations in those considerations which are important in the ordering of variables in this study are those which account for: (a) the human ability to acquire language, (b) the learning of language, (c) the role of meaning in communication, (d) human linguistic competence, and (e) the role of language in human communication. This author contends that the theoretic explanations in these areas most appropriately account for the clients' differences in awareness or knowledge and adoption of or propensity of adopting technological innovations. Hence, those explanations will be the bases for this author's recon- ceptualization of literacy and, subsequently, the formulation of theoretic hypotheses. a. Human Ability to Acqgire Language. Scholars whose considerations have focused on human ability to acquire sym- bol systems (languages) contend that humans have a biolog- ically innate ability to learn language--not a particular language, but any language whatsoever (see, for example, Chomsky, 1965; Katz, 1966; Lenneberg, 1964, 1969, and McNeill, 1966, 1971). Moreover, even children who are isolated for long periods of time are able to acquire language with minimal effort (Lenneberg, 1964). 44 These scholars convincingly argue that language ac- quisition is possible only if some biological built-in mechanism that predisposes humans to speech is postulated. In addition, they posit that language acquisition is a pro- duct of these innate structures as well as maturation and experience. Thus, the language of a given speech community is the result of the individuals' innate structures and of their maturation and experience. These postulates imply that the mature speaker in a speech community has a highly distinctive and complex set of linguistic rules at his/her command. However, these rules are not so abstract that they cannot be acquired by the members of other speech communities. On the contrary, with appropriate training, the language rules in one speech com- munity can be acquired by the members of any other speech community. This implies that in diffusion practices involv- ing technological innovations among those populations whose dominant language is not English and/or whose native lan- guages do not have an elaborate numbering system, such pOp- ulations have the ability to learn the English and more elaborate number symbols and concepts given proper training. b. Learning Languagg. Those scholars who have been concerned with the development and use of symbol systems by humans and the relation of these systems to overt behavior have included: at one extreme, the behavioral theorists; somewhat in the middle, the mediational theorists, and at 45 the other extreme, the cognitive psychologists. The behaviorists (e.g., Watson, 1924, 1930; Bloom- field, 1933; Bousfield, 1953, and Skinner, 1957) contend that the behavior of any organism can be described and ex- plained in terms of the organism's response to the stimuli presented by the features of environment. That is, they look outside the actor (to the environment) for the explana- tory cues of behavior. Thus, this perspective explains the development and use of language and its relation to overt be- havior by using the classical stimulus-response (S.R.) or motion theory in which the stimuli are assumed to have a direct effect on the behaviors of the perceivers of the stimuli and discount inferring behavior from any state of consciousness or internal meaning. Skinner (1957), for instance, argues that all human behavior can be explained by examining the ways in which the behavior was first conditioned, and that future behavior is dependent on the ways that past behavior was rewarded when it occurred. To Skinner (1957), an acceptable theory of any behavior, including language, must be able to accurately predict the observable responses an individual will make to a particular set of stimuli. Thus, Skinner (1957) takes operant conditioning rather than cognitive processes as the basis for learning language. In operant conditioning, when the organism correctly responds to a stimulus, the organism is rewarded or reinforced for that behavior. 46 He conceives language as an utterance which is to be considered as an aspect of behavior in general. The lan- guage itself consists of functional units10 whose importance can be strengthened or weakened by the application of rewards. Such rewards increase the probability of repeating the be- havior until the behavior becomes firmly fixed; and, in case of symbols, until the symbols come to be firmly associated with the object or action as referents. In sum, Skinner's (1957) analysis views verbal be- havior as a way of controlling the environment and for class- ifying environmental events (objects and actions). That is, the speaker of a language attempts to control the behavior of the others for his/her benefit by using language. From the above discussions, it is to be noted that behavioral theorists recognize the relation between symbols and referents and between language and overt behavior. In diffusion practices involving technological innovations, this implies that the clients of such innovations must first be aware or have the knowledge (i.e., have learnt) of the symbol sets which denote the innovations (referents) if they are to adopt (overt behavior) any innovation at all. Note, however, that awareness or knowledge of symbol sets assumes that one has meaning for those symbol sets. For effective communication, it is important that the speak- ers have meaning for the symbol sets being used. The be- havioral perspective is, however, inadequate to account for 47 this importance since it discounts inferring behavior from any state of consciousness or internal meaning. Another theoretic framework is therefore needed to account for the importance of meaning in human communication so as to better understand the relation between language and behavior. 0. The Role of Meaning in Human Communication. Mediation theory, noted earlier, accounts for the importance of meaning in the development and use of language by human beings, and its relation to overt behavior. This theory was proposed by Osgood (1963) who offers a position different from Skinner's (1957) in that he considers both the expres- sion system and the content system--the meaning system of language. He postulated that meaning is an internal process, which is a learned relationship developed between an extern- al stimulus and an internal ("mediated") response state. This internal response itself stimulates internal behaviors which may then lead to overt behavior. Thus, Osgood's (1968) conceptions of language and its relation to overt behavior are founded on the stimulus-organ- ism-response (S.O.R.) or action theory which liooks inside the actor for the explanatory cues of behavior. These inside cues are assumed to pose intervening variables between the stimulus (referent of a symbol) and overt behavior. That is, mediation theory attempts to account for the different ways by which message recipients construe the world of their ex- perience (perceive and process messages from symbols) and, 48 subsequently, how they may respond (behave) to that world. In sum, mediation theory attempts to relate language and thought (meaning). It attempts to explain how humans learn the meaning of a word in relation to objects since, in addition to those stimuli and responses which are external and observable, there are those involving internal mental processes which occur as a result of perceiving words. The theory is also able to explain cases where one's response is so well-learned that one does not have to go through the com- plex procedures of re-learning. These cases involve com- plete internalization of meaning to the point where mediation is automatic. What mediation theory implies is that for effective communication to occur between the speakers, such speakers must share the meaning of the symbol system in use. In dif— fusion practices involving technological innovations, this implies that these practices will be effective only if the change agents and their clients share meaning for the symbol sets which denote the technological innovations. Note, however, that both the behavioral and mediation— al theories do not explain the extent to which a user of a language or languages has adequate competence to use the languages. Another theoretic framework is therefore needed to account for this. d. Human Linguistic Competence. One alternative to the behavioral and mediational theories is cognitive theory 49 which was noted earlier. Cognitive theories have been set forth by several psychologists who often adopt the work of Chomsky (1957, 1965, 1968) as a basis for their conceptions of language development and use. Cognitive theories differ from both the behavioral and mediational theories in that they focus on the things a speaker of a language would need to knew in order to use a language appropriately. That is, they focus on linguistic competence--one's knowledge of lan- guage, rather than on linguistic performance—-one's use of language. Cognitive theorists contend that humans think in kernel sentences which are stored and abstracted in their heads. For example, if one perceives in one's head the sentence, "Didn't the farmer plant corn?," cognitive theo- rists would argue that one stores the kernel sentence, "The farmer planted the corn." In addition, they would assert that somewhere along the line one must account for one's ability to reproduce the sentence as stated rather than just the kernel sentence form. Their explanation for this ability is that along with the kernel sentence, one abstracts and stores certain cues from the original sentence which remind one that it was negative, interrogative, and so forth. Humans are able to make these transformations by using spe- cific sets of transformation rules11 which are appropriate for the grammar of the language in use. Thus, prOper use of a language presupposes knowledge of the grammar of the lan- guage. 50 Linguists note that one's knowledge of the grammar of a language comprises the basic linguistic elements of pho- nemes (sounds) and morphemes (words, or parts of words as meaning units), i.e., vocabulary or lexicon, and the rules for combining words into sentences. These elements and the apprOpriate rules of combinations are noted to be important for effective communication and understanding. More succinctly: When you know a language you learn the sounds used in that language, the basic units of meaning, such as words, and the rules to combine these to form new sentences. The elements and rules constitute the grammar of a language. The grammar, then, is that we know; it represents our linguistic compe- tence. To understand the nature of language we must understand the nature of this intern- alized, unconscious set of rules which is part of every grammar of every language. Every human being who speaks a language knows the grammar. When linguists wish to describe a language they attempt to describe the language which exists in the minds of its speakers. There may of course be some dif- ferences between the knowledge that one speaker has and that of another. But there must be shared knowledge because it is this grammar which makes it possible for speakers to talk to and understand one another (Fromkin and Rodman, 1978, p. 9). In diffusion practices involving technological inno- vations, this implies that the change agents and their clients must, to some extent, share the grammar of the lan- guage or languages in the diffusion process if they are to communicate to and understand each other. In particular, they must, to some extent, share the words (vocabulary) 51 since these are usually the most important in diffusion practices involving technological innovations. e. The Role of Language in Human Communication. The postulates of innate structures for language and the behavior- al, mediation, and cognitive theories noted in the preceding paragraphs have been attempts to provide foundations for our understanding of human development and use of symbol systems. Implicit in those attempts is the concern for the presumed role which language plays in human communication. This sec- tion briefly discusses some of the theoretic considerations in this regard. The concern for language is based on the convincing argument that language is fundamental in all human discourses, and it is well-established that language is the "bearer of meaning" and a "medium of communication." That is, language functions not simply as a device for reperting experience (medium), but more so as a way (method) of defining experi- ence for its speakers (see Urban, 1939, p. 37). It ". . . is the means by which man symbolizes and orders aia concepts of Eifi universe" (Whatmough, 1956, p. 83; emphasis added). These conceptions of languageare mirror images of Sapir's (1931: 578) notion that language actually shapes the way in which we perceive, think, and therefore act. Specif- ically, he states: Language is not merely a more or less .systematic inventory of the various items of experience which seem relevant to the individual, as is so often naively assumed. 52 But also a self-contained, creative sym- bolic organization, which not only refers to experience largely acquired without its help but actually defines experience for us by reason of its formal completeness and be- cause of our conscious projection of its implicit expectations into the field of ex- perience. In this respect language is very much like a mathematical system which also records experience in the truest sense of the word, only in its crudest beginnings, but, as time goes on, becomes elaborated into a self-contained conceptual system which pre- visages all possible experience in accordance with certain accepted formal limitations. . . . Meanings are not so much discovered in experi— ence as imposed upon it, because of the tyrannical hold that linguistic form has upon our orientation in the world. The same position is taken by Sapir's most famous stu- dent, Benjamin Lee Whorf (1952: 5) who states: . . . that the linguistic system (in other words, the grammar) of each language is not merely a reproducing instrument for voicing ideas but rather is itself the shaper of ideas, the program and guide for the indi- vidual's mental activity, for his analysis of impressions, for his synthesis of his mental stock in trade. . . . we disect nature along lines laid down by our native language. The categories and the types that we isolate from the world of phenomena we do not find here because they stare every ob- server in the face; on the contrary, the world is presented in a kalcodoscopic flux of impressions which has to be organized by our minds--and this means largely by the linguistic systems in our minds. Sapir and Whorf's notions of language have generally come to be collectively known as "linguistic relativity hy- pothesis" or "Sapir - Whorf hypothesis." In short, the hy- pothesis posits that (1) without language we cannot think, 53 (2) language influences perception, and (3) language influ- ences thinking patterns. The individual's language is thus seen as the factor which determines the way the individual perceives the world. That is, language determines the com- munity's view of the world or, what Whorf (1956) refers to as the Weltanschauun. The implication is that if a speech community has de- veloped an elaborate language or extensive vocabulary (i.e., linguistic competence), peOple in such speech community will be able to perceive far more things in the world around them than will the speech community which has a limited or un— elaborate language or vocabulary. In diffusion practices involving technological innovations, this implies that those clients who have a more elaborate symbol system germane to diffusion will perceive more technological innovations than those with unelaborate symbol system. Other scholars have conceived human language as a signal system which influences human behavior. Pavlov (1927) , for instance, distinguished between what he called the first and second signal system. He noted that man has physical structures and reflexes similar to those of other animals. For example, man reacts to intense light by pupillary con- striction and to a sudden loud sound by the startle reflex; when man is severely threatened, his heart, blood-sugar levels, perspiration, and breathing all increase as adaptive devices. Pavlov called these built-in mechanisms the EEEEE signal system. 54 This first signal system functions on a biological level, but language also functions as a signal system. It enables man to regulate his own behavior or someone else's behavior. Language, then, radically changes behavior of a biological organism. In Pavlov's terms, it forms the second signal system. As Pavlov (1927) said, "The word created a second system of signals of reality which is peculiarly ours, being the signal of signals. On the one hand, numerous Speech stimuli have removed us from reality. . . . On the other, it is precisely speech which has made us human" (p. 357). Influenced by Pavlov's views, the Russian psychologist Luria (1961) set up an experiment to demonstrate that, with increasing language experience, the child comes to have in- creasing control of his/her own behavior--that words control behavior. In the initial stages of development, the child comprehends the meaning of words such as "go" and "stop" but they have no effect on his/her behavior. When he/she is told to "go," which in Luria's eXperiment meant to press a bulb, the child does so. But when the child is told to "stop," in Luria's experiment, the child continues to press the bulb. The command "stOp" does not control the child's motor behavior until the child has further experience. With increasing development, however, the child eventually learns to stOp, or to release the bulb, on com- mand. In this experiment, Luria emphasized the meaning of 55 words in invoking behavior. The implication for these findings is that meaning and the vocabulary (i.e., linguistic competence) are very important in human communication and, ultimately, for human behavior. In diffusion practices involving technological innovations, this implies that the potential clients of such innovations must first have the meaning and adequate vocabu- lary or symbol sets for the technological innovations if they are to ad0pt such innovations at all. Summarily, whatever else language is and what it does, it is in the first instance the tool for meaningful commun- ication between and among its users since it is only through communication that language comes into being and only via language that communication can occur. This author finds it superfluous to delve into this chicken-egg argument. Rather, he wishes to emphasize that language qaa language is meaning— less until it is used for communicating. This meaning de— rives in a speech or language community of its users whom it provides with habitual modes of analyzing the phenomena of their experience into perceptual categories (Hoijer, 1954). Hence, fundamentally, "it is the method of adaptation to and control of environment" (Urban, 1939, p. 31). Insofar as emphasis is on meaning or semantic struc— tures in vocabulary or lexicon, the philosophy of language is therefore necessarily grounded on the philosophy of sym- bolism since a symbol is the basic linguistic element which 56 bears meaning to its users. To talk of the differences be- tween or among languages is ipso facto to talk of differences in symbols. In other words, if there are no shared symbols between or among languages, such languages will be said to differ. Subsequently, users of such languages will be said to speak different languages. To the extent that languages differ significantly from each other, so would we expect to find significant and formidable barriers to meaningful, and effective communication (with respect to overt behavior) be- tween the speakers of such languages. These barriers are expected to prevail whenever members of one speech community attempt to communicate with those of another speech commun- ity for whatever reasons. As a medium of communication and method of directing the perceptions of its speakers, language must therefore have a set of shared symbols among its users. In diffusion practices involving technological inno- vations, this implies that communication barriers are to be anticipated between the change agents and their rural clients since these Speech communities generally differ in their linguistic repertoires with respect to the "language" of diffusion as noted earlier (section B. 5). Such barriers are expected to be reflected in low awareness or low knowl- edge and, subsequently, low adoption of or low propensity of adopting technological innovations among the potential clients of these innovations. 57 To establish the strength of the relationship between literacy and knowledge or adoption of innovations, some uni- form and precise conceptualization and operationalization of literacy is needed. As noted earlier, the conceptualization and operationalization have generally lacked both precision and uniformity. This suggests that a reconceptualization of literacy as a conveyor of linguistic symbols and concepts is needed. Such a reconceptualization would include the linguistic rep- ertoires (viz., quantitative and non-quantitative English symbols and concepts) along which change agents and their clients were noted earlier to differ. However, such a recon- ceptualization requires first specifying clearly the relation- ship between language and literacy. 2. Theoretic Relation Between Laaguage and Literacy Hitherto, the relationship between literacy and lan- guage has not been specified. An explication of this rela- tionship is necessary before considering how to reconceptu- alize literacy. This explication requires an understanding of how we use language gaa language and an understanding of the specific skills which are associated with this use. Following the previous conceptualizations of language, two modes may be identified at which language may be used. Firstly, we may use language at the iaagapersonal mode of communication. The specific skills associated with this mode include thinking/reasoning and internalization of 58 meaning. These skills are to be conceived as internal skills to the person as a processing system. The speaker of a language performs these skills using the symbols and concepts germane to the language the speaker has been able to acquire. Thinking/reasoning and internalization skills are, however, covert. That is, they are not easily observ- able behaviors. In addition, these skills are characteris- tic of all normal human beings. Thought is merely inaudible speech or, "talking with concealed musculature" (Watson, 1924, 1930). Secondly, and more importantly, we use language at the interpersonal mode of communication.. By definition, the interpersonal mode of communication implies exchanging mes— sages with those others who, by virtue of the shared symbol system, are or may be involved in the communication process. The specific skills which are associated with this mode in- clude speech (face-to-face or broadcast), reading, and writing. These skills are to be conceived as external skills. That is, they are overt or directly "observable" skills by which we can influence the behavior of others and, to be sure, by which others can also influence our behavior. These two modes of communication and their respective skills are summarized in Figure 1. Clearly, there is a transition from the intrapersonal to the interpersonal skills of using language. This trane sition is in effect a transition from how peOple construe 59 I Mode of Using Language Type of Skills 1. Intrapersonal Communication Thinking/Reasoning, Internalization of Meaning 2. Interpersonal Communication Speech, Reading, Writing Figure 1. Modes of Language Use and Their Respective Skills their world of experience to the ways by which they commun- icate about it. Consequently, the intrapersonal and interpersonal modes of communication are not independent; they are inter- related. This interrelationship, however, is recursive rather than non-recursive since the acquisition of external skills at the interpersonal mode of communication presuppos- es the acquisition of the internal skills at the intra- personal mode of communication. Moreover, to be able to perform the internal skills associated with the intraperson- al mode of communication, presupposes that the person has already acquired the symbols and concepts germane to a lan- guage as noted earlier in the review of the theoretic foundations of language. 60 However, although there is a transition from the intrapersonal to interpersonal mode of communication, never- theless the acquisition of speech, reading, and writing skills associated with the latter mode is to be conceived as an evolutionary rather than an automatic process. This evolution begins with speech by which man is able to make "audible" at the interpersonal mode the linguistic symbols and concepts which were before inaudible at the intrapersonal mode. This is a unique human ability which is well-estab- lished as the basic characteristic which distinguishes man from other animals. Through his innovative use of language, however, man has also been able to develOp more specialized skills of writing and reading by which he can respectively store and retrieve speech from symbols wiggled on surfaces such as parchment, slate, chalkboard, paper, tape, film, etc. The ability to store and retrieve speech from these surfaces implies yet more sophistication in the use of language; an ability which is a significant part of the scientific process, and in the diffusion of technology. Clearly, a special facility is needed for the speakers of a language to be able to store and/or retrieve speech from such surfaces. Literacy is here conceived to be such a facility. The acquisition of this facility therefore implies more sophistication in the use of a language by the persons who possess it. Here then lies the relationship between 61 language and literacy. Literacy facilitates the storage and/or retrieval of linguistic symbols and concepts from surfaces. 3. Theoretic Relation Between Literacy and Education Recall, it was noted earlier that humans have a bio- logically innate ability to learn language--not a particular language, but any language whatsoever.‘ In additiOn, it was observed that in diffusion practices involving technological innovations among those populations whose dominant language is not English and/or whose native languages do not have an elaborate numbering system, the postulate of innate linguis- tic structures implies that these populations have the abil- ity to learn the English and a more elaborate number symbols and concepts given proper training. Similarly, since there is a clear connection between language and literacy as noted earlier, these pOpulations have the ability to acquire literacy in English (non-quan— titative literacy) and literacy in number symbols and con- cepts (quantitative literacy) given proper channels or media for training. It is therefore important to know the channels or media through which a literacy facility may be acquired. It is here speculated that individuals' access to those channels and the extent to which they have been exposed to them, would explain their literacy levels (both in quantitative and 62 non-quantitative symbols and concepts) and, subsequently, their level of awareness or knowledge and adoption or pro- pensity of adoption of innovations. This author contends that formal education (years of school) is the channel or medium through which literacy may be acquired. That is, literacy may be acquired from the learning processes which are frequently provided for by the curricula of formal education institutions. From historical times, formal education institutions have always provided opportunities for developing skills in the so—called "three Rs" (Reading, 'Ritin', and 'Rithmetic), among other skills. Some people may of course acquire literacy through the non-formal education programs such as "continuing education," "adult extension education," "functional literacy," and so forth; or through informal channels such as friends, rela- tives, etc. However, these are generally subordinate lit- eracy channels. The main literacy channel is frequently to be sought in the modern formal education institutions. Unfortunately, however, not every modern man has the opportunity, nor the desire, to say nothing of the resources,’ to attend the modern education institutions. If this inequity is true, then we would expect the adult segments of populations in the Developing Countries to differ significantly in their levels of literacy (both quan- titative and non-quantitative) when they are arrayed on the‘ continuum of formal education. Individuals without formal 63 schooling, are expected to be illiterate in any natural lan- guage, unless they have had non-formal literacy training. In those areas where English is not the dominant lan- guage of discourse (e.g., Africa and Asia), individuals with some minimum of formal schooling should be literate in their native languages. Those with more years of schooling, where English is the official language, should be literate in both English and native language (i.e., biliterate). As the years of schooling increase, the probability of literacy in both languages (i.e., biliteracy) as well as quantitative literacy, increases. Since the English language uses a well developed numbering system, and many other languages do not (see section 5.b.i), we should expect individuals who are liter- ate in English to be relatively more SOphisticated in quanti— tative symbols and concepts compared to those literate only in native language or illiterate. From this line of reason- ing, it appears there is interdependence among years of formal schooling and literacy in the quantitative and non- quantitative symbol systems. In sum, it is expected that: 1. When individuals go to school, they acquire new symbol systems. 2. As individauls' levels of formal education increase, their chance of acquiring more efficient symbol systems (quantitative and 64 non-quantitative/English) increases. 3. Subsequently, the higher the individual's level of formal education, the more complex the individual's symbol system (language) becomes, and 4. Higher levels of formal schooling require more complex symbol systems; thus recogniz- ing the interdependence of literacy and schooling. Further, the acquisition of more complex symbol sys- tems is believed to be necessary to engage effectively in problem solving situations involving complex relationships such as those found in the diffusion-adoption processes in- volving technological innovations as noted earlier (section 5.b.i). Thus, the more sophisticated symbol users (i.e., higher literacy and school grade completed) will be more likely than less sophisticated symbol users to adopt or to have the propensity to adopt innovations which are more complex. 4. Reconceptualization of Literacy The preceding sections have developed materials to show the relation between: language and behavior, language and literacy, literacy and education, and the presumed con- nection between literacy and formal education on one hand, and awareness or knowledge and adoption or propensity of adoption of technological innovations on the other. 65 Clearly, a reconceptualization of literacy is needed. The reconceptualization proposed includes the addition of uniform measures of English as well as native language lit- eracy (in areas where English is not the dominant language of discourse, e.g., Africa and Asia), and the development of measures of quantitative literacy to be included in the overall assessment of levels of literacy. This would involve more precise indicators of levels of literacy within each of the three components of literacy (viz., native language, English language, and quantitative symbols and concepts), and it would eventually lead to a composite measure of the three components. The aim is to have at least interval levels of measurement for each of the indicators. At some point, it may be possible to establish a uni- form norm for saying that a person is literate (or not liter- ate) in any of those three components. When that is possible, a composite set of levels as shown in Figure 2 may be con- structed for classifying individuals on their literacy levels particularly in Developing Countries. The four levels of literacy in the below scheme are stated in an increasing order of language sophistication with respect to the number and types of symbols an individual is able to access and process at the level of literacy at which he/she is located. That is, individuals located on Level I access and process the least number and type of symbols. These are the illiterate individuals. Individuals located 66 TYPE OF LITERACY Literate Literate in in Literate Quantitative LITERACY Native in Symbols and LEVEL Illiterate Language English Concepts Level I X Level II X Level III X X Level IV X X X Legend: The "X" indicates the type of literacy associated with each level of literacy. Figure 2. A Schematic Ordering of Literacy Levels Among Adult Populations in Developing Countries on Level IV are able to access and process the greatest number and types of symbols at the interpersonal mode of communication involving reading and writing skills. Levels II and III are intermediate levels at this mode. However, individuals located on Level III (biliterates) are able to access and process more types of symbols than those individu- als located on Level II (Monoliterates). Thus, as one increases reading and/or writing skills with respect to native language, English and quantitative symbols and concepts, then one will be more able to deal with (analyze and synthesize) complex relationships such as among technological innovations in the diffusion-adoption process 0 67 It is therefore to be assumed that relative to indi- viduals located on the first three levels of literacy, indi- viduals located on Level IV can handle more effectively the more precise and fairly complex relationships among the phe- nomena in their environment. They are able to do this by virtue of their literacy in quantitative symbols and concepts which are natural means of arranging the facts of experience in such a way that such facts can be dealt with (analyzed and synthesized) precisely. Subsequently, since diffusion-adoption processes fre- quently involve complex relationships in which quantitative symbols are germane, the level of awareness or knowledge and adoption or propensity of adoption of technological innova- tions is expected to follow the order of the four levels of literacy noted above. That is, Level I literacy should be associated with the least level of awareness or knowledge and adOption or propensity of adoption of technological in- novations, and Level IV should be associated with the high- est level of awareness or knowledge and adoption or prOpensity of adOption of technological innovations. The level of awareness or knowledge and adoption or propensity of adOption of technological innovations should be inter- mediate on Levels II and III. However, Level III should be associated with higher awareness or knowledge and adoption or propensity of adoption of technological innovations compared to Level II literacy. 68 An effort will be made to obtain samples from the USA population which will include representatiVes8 83qu N88 88: ucme :3 8: 880:8 88H ~88 9.5:. A5 H.898 .538 338“ 880 Eu L8H... 8.an :88 ESSEH Auras H6989: ER .838 33 .838 8H8 84 lumum: g 96g 52% 95550an .HEHHSBEQH Newmuzv 8am many—Mm 9.6: I 3.8.32 93 5 95.. H 52mm: 93 Huh—dug; 05 no snow neg “~th armor—0&8“ out floor so mace—NE .m manna. 80 vmm «on com vmm vwm com vwm com vwm vwm com com com m: cm 2; 3H mg mg H3 Nm m2 m3 m3 cm H: mm» m3 3m Sm N: mmm mom mom Nmm mom wvm 3N mom m3 02 95.33 SAC 53. NGw mmauuon >mudm 8N8. ucmu: c: x0: 9.5on6 How: @800 059 :30 888.8 -38 8388 «.88 H8 L83 8:8 :88 85895 $88 Hum/BEH .35 .838 33 .838 .NH Lona: was; $548.. maHe<>QHzH HEP—BHUEOHN HNNNunzN 88m 29.8 96: I mwummfiz 93 Ca 953.385 5.38: can any»??? 99 no comm mo mm: ECNHCB co 8N8§h .o manna 81 of the respondents on this variable is shown in Table 7. Table 7. Distribution of the USA - Michigan Respondents on Last Grade of School Completed Last Grade of School Completed Frequency % 0 8 3.3 2 3 1.2 3 l .4 4 4 1.7 5 4 1.7 6 6 2.5 8 28 11.6 9 10 4.1 10 18 7.4 11 15 6.2 12 78 32.2 13 36 14.9 14 13 5.4 15 9 3.7 16 9 3.7 Total 242 100.0 ii. Literagy in English and in Spanish The respondent's level of literacy was de- termined by the Slosson Oral Reading Test (SORT) instrument (see Appendix A) which consists of a list of 200 words ranging from very easy words to more difficult ones. The "same" instrument was used to determine the level of literacy among Spanish speakers after translating the English words in the SORT instrument into "equivalent" Spanish words. In either type of literacy (Spanish 82 and English), the respondent's level of literacy was determined from the number of words he/she was able to read from the list. Biliteracy: Both the English and Spanish versions of the SORT instrument were admin- istered to the bilingual cases to determine their level of biliteracy (i.e., literacy in both Spanish and English). The score for biliteracy was derived as a multiplicative function between English and Spanish liter- acy scores. This method of score transfor- mation was based on some empirical and theoretical considerations. Research indicates that bilinguals are generally superior in their mental and crea- tive abilities compared to monolinguals. For example, bilinguals (EnglishfiFrench) out— perform their monolingual counterparts on all verbal and non-verbal tasks (Peal and Lambert, 1962); bilinguals (among English, Greek-American, Spanish-American, and Czech- American speakers) are more creative and score higher on non-verbal "uses" test than monolinguals (Jacobs and Pierce, 1966); bi- linguals (English-Spanish) outperform mono- linguals on all such tasks as object 83 constancy, naming, and sentences (Feldman and Shen, 1969); bilinguals (English- Spanish) have higher school achievement and self-concept than monolinguals (Del Buono, 1971); bilingual (English-Spanish) children have as high IQs as monolingual children if not higher ones in some re- spects (Gezi, 1974); bilinguals (Persian- English) have greater flexibility in writ- ing coordination compared to monolingual Persians (Hoosain, Atai, and Salili, 1975), and research on semantics and structure among monolingual and bilingual (English— French) shows that classical coding meas- ures (latency, reaction time, number of words, number of syllables, and inter- personal agreement) correlate with each other for both monolinguals and bilinguals. However, the intensity of meaning is in- versely correlated with the coding measures only for the monolinguals. For the bilin- guals, intensity of meaning is directly related to the coding measures (Russ, Gold, and Cherulnik, 1975). This led Russ and his colleagues to speculate that although coding measures do hold up between monolin- guals and bilinguals, they may mean different 84 things. Words that have intense meaning for bilinguals may be those that elicit a host of associations; while for the monolinguals, the semantically intense words elicit few associations. Given the theoretic connection between language and literacy noted in the previous chapter (see section C. 2), the superiority of bilingualism noted above implies superi- ority of biliteracy. Recall, biliteracy which includes literacy in English (i.e., Level III literacy) implies superior ability to construe the environment by those endowed with it. That is, as one increases reading and/or writing skills in native language (i.e., Spanish in this case) and in English, then one increases one's capacity to deal with (analyze and synthesize) the phenomena in one's environment. Moreover, there is also empirical evi- dence indicating that literacy in English is acquired faster if persons are already literate in their native languages. Modiano (1966), for example, found in her research in three Mexican tribal areas that persons who were already literate in their tribal iii. 85 languages acquired greater facility in reading comprehension in English. In such areas as Africa and Asia where English is not the dominant medium of discourse,13 this finding is to be expected since liter- acy in English presupposes literacy in native language in those areas. The acqui- sition of literacy in English in those areas implies mere extension of the liter- acy skills which one already has in the native language. Therefore, in light of the above re- search evidence and conceptualizations, the multiplicative algorithm was found appropri- ate for transforming more precisely the pattern of Spanish and English literacy scores into the cardinality of a biliteracy score. In addition, this algorithm gives the flexibility of deriving a cardinality from two or more numbers whose referents are so different (as is the case with Spanish and English literacy scores) that direct addition of such numbers would not be very meaningful. Quantitative Literacy:l4 The author con- structed items (see Appendix B, pp. 192-199) 86 with which to measure this variable along four dimensional skills in which numbers and/ or number concepts are used: 1. Respondent's knowledge of the function- al characteristics (ordinality and card- inality) of number, 2. Respondent's knowledge of the standard units of measures and weights which are in common use in his/her area, 3. Respondent's computation skills, and 4. Respondent's intuitive use of numbers and number concepts. Like the SORT instrument, the quantita- tive literacy test (QLT) items were also translated into Spanish to facilitate admin- istration to the Spanish-speaking respond- ents. A composite score was derived for each respondent from the items designed to measure the four dimensional skills stated above. It was conceived that the higher the respondent's score on the QLT items, the higher would be his/her level of quantita- tive literacy. The QLT items were tested for reliability as described below. 87 Testing the QLIT Items for Reliability If an instrument is valid, it is reflecting primarily the characteristic which it is supposed to measure, with a minimum of dis- tortion by other factors, either constant or transitory; thus there would be little reason to investigate its reliability--that is the extent to which it is influenced by transi- tory factors (Sellitz, et a1., 1959, p. 166). The above paragraph is both informative and instruc- txive. Firstly, the paragraph is informative in that it sgmecifies the close connection between validity and relia— txility, and, in general, to the extent that a measure is un- realiable, it lacks validity.15 Secondly, the paragraph is iristructive in that if the validity of a measure is not kuiown, it is necessary to investigate the reliability of the measure. Since no earlier studies had been concerned with the CXJnstruct of quantitative literacy, it was not known if the 453 items had satisfactory validity for measuring this con- Steruct. It was therefore necessary to test for the reliabil- ity of these items in their measure of quantitative literacy. In this study, the 48 items designed to measure the cOnstruct of quantitative literacy were tested for their re- 1iJability among the study's respondents who were divided in- t£> fbur groups for the purpose of this analysis: (1) Whites (bu = 83) whose last grade of school completed (LGRADE) was equal to or less than 12 years of school, (2) Whites (N = 58) Whose last grade of school completed was equal to or greater b‘nj L... .‘. Ln '4- Hi 88 than 13 years of school, (3) Chicanos or Mexican Americans (N = 61), and (4) Black and Native Americans ( N = 38). In each of these four groups, the 48 items were tested through the Split-half model of the reliability analysis routine to test for the equivalence of measures. Four re- liability coefficients were computed in each group: First, a correlation coefficient between the two halves each of which had 24 items was computed. This co- efficient measures the extent to which the two halves meas- ure the same thing, i.e., quantitative literacy. Second, the Spearman-Brown split-half coefficient was generated to measure how reliable the scale would be if the two equally reliable halves were combined into one. Third, the Guttman split-half coefficient was computed. This is similar in form to the Spearman-Brown coefficient except that the Guttman split—half coefficient does not nec- eSsarily presume equal reliabilities or equal variances. Finally, coefficient alpha was computed for both halves. This is a maximum likelihood estimate of the reli- ability coefficient if a parallel model is assumed to be true. That is, the model in which the items or tests are aSiaumed to have the same true score variances over a set of objects (i.e., quantitative literacy) being measured, and the same error variance over replications. The following paragraphs report the results of these four coefficients in each of the four groups of respondents S tated above . 89 Coefficients of Reliability Among Whites; LGRADE 512 Years (N = 83) In this group, the correlation coefficient between the two halves (scale form) is .712 which is very high.16 The Spearman-Brown split-half coefficient is .932 which is significantly high. The Guttman split—half coefficient (.831) generated is also significantly high. The coefficient alpha computed for the first and second half of scale forms is .888 and .889 respectively both of which are significantly high and stable. Coefficients of Reliability Among Whites: LGRADE213 Years (N = 58) The correlation coefficient between the scale forms in this group is .334 which is rather low. The Spearman- Brown split-half coefficient (.501) , and the Guttman split- half coefficient (.472) are moderate. The coefficient alpha for the first and second scale form is .564 and .696 respec- tively both of which are fairly high and rather stable. C‘OSfficients of Reliability Among Mexican firicans (N = 61)— In this group, the correlation coefficient (.746) be- tween the scale forms is very high. The Spearman-Brown sPlit-half coefficient (.854) is significantly high, and the Guttman split-half coefficient (.851) is also significantly high. The coefficient alpha for the first and second scale fOrm is .897 and .844 respectively both of which are signif- j~<=antly high and stable. 90 Coefficients of Reliability Among Black and Native Americans (N = 38) The correlation coefficient (.814) between the scale forms in this group is significantly high. The Spearman— Brown split-half coefficient (.898) , and the Guttman split- half coefficient (.894) are both significantly high. The coefficient alpha for the first and second scale form is .893 and .856 respectively both of which are significantly high and stable. Clearly, the four reliability coefficients related to these groups generally indicate relatively high consist- ency in the scale items in their measure of the construct of quantitative literacy. Any variation in the scores therefore is to a greater extent to be attributed to variation among the respondents rather than to the scale. The items on the QLT were selected to have "face Validity"; i.e., the items were assumed to be relevant in measuring the selected dimensions of quantitative literacy. A factor analysis of the QLT data showed the items presumed to be measuring the same concepts and operations generally lOaded together on the expected factors.17 b. The Dependent Variables. These included guanti- &ve literacy, knowledge of selected technological inno- Vations, and adoEtion of those innovations. While quanti- tlative literacy is primarily considered a predictor and eJP‘Iplanatory variable for the knowledge and adoption of technological innovations, it is used as a dependent 91 ‘variable in some of the computations to determine the extent ‘to which it can be predicted by the other independent vari- .ables. Since operationalization of QLT has already been ciiscussed, only the operationalization of knowledge and .adoption will now be presented. i. Knowledge of selected technological innova- tions in health, Co-operative Extension and Social Service Agencies was derived from responses to the following items: 1. Respondent had heard of the innovation. Yes = 1; No = 0 2. If yes, where had he/she heard about it? Any correct source = 1; no source or wrong source = 0 3. Give a meaning of the innovation. Correct answer = 1; wrong or don't know = 0 4. Where would the respondent go for in- formation about the innovation? Any correct source : 1; no source or wrong source = 0 5. Under which of the listed conditions would he/she consider using the inno- vation? Correct answer = 1; wrong or don't know = 0 In addition, a sixth item was asked for cardiopulmonary resuscitation (CPR): Do you think you know enough to do CPR if someone needed it?: Yes = 1; No = 0. Each respondent's score of knowledge de- rived from the summation of his/her ii. 92 responses on the items as described above. Adoption measures were derived for the inno— vations by asking respondents about their use of each. Four of the technological in- novations offered two Options for adoption ‘ --a simple version and a complex version. These were scored zero if the respondent had never used either of the two versions; 1 if the respondent used the simple version, and 2 if the respondent used the complex version. For the other innovations, re- spondents were scored 0 if they had never used it; 1 if they used it sometimes, and 2 if they used it most of the time. As may be noted in the questionnaire (Appendix B), the structure of the items for the four complex innovations was somewhat different; and in the analysis pertaining to those four innovations, the propensity of adopting complex innovations is taken as the dependent measure. The data were ob- tained by providing the respondents with a situation in which they were to choose either the simple or the complex version of the technological innovation. These four technological innovations were: DIABKITS 93 (two versions of kits for testing for dia- betes); DRUGFORM (two forms of administer- ing a drug for an illness); TECHCPR (two techniques of emergency for CPR), and FOOD- FORM (two forms of food preparation for the same food product). One form of each of those innovations required more precise measurements and was thus considered more complex than the other form (see items 7, 8, 16, and 35 in Appendix B, pp. 187-189, 191. These innovations offering the simple- complex choice were provided on the assump- tion that individuals who are less SOphiSti- cated in the use of numbers and number concepts (i.e., low in quantitative liter- acy) would shy away from adOpting the more complex technological innovations. Hence, they would adopt the simpler versions of the innovations which did not require pre- cise measurements on their part. The more quantitatively SOphisticated individuals were expected to adOpt the more complex types of technological innovations which re- quired precise measurements and, ipso facto, in which numbers and number concepts were used. 94 The respondents' level of adoption de- rived from the summation of their scores on the adoption indicators as described above and in footnotes 18 and 19. The complete text of the knowledge, adoption, and quantitative literacy items, and biographical information for the instru- ment used in the USA - Michigan sample may be seen in Appendix B. c. Development of Data Collection Instrument for the USA - Michigan Data Set. Generally, the construction of the instrument for the USA - Michigan sample space involved the steps described below. First, the author reviewed the literature on mathe- matics and statistics to supplement his basic training and personal experience in these areas. From these areas, a number of specific quantitative concepts derived from four Operational concepts of intuition and logical thinking: (1) Series in which a group of events are related by order of occurrence. The specific quantitative concepts or number prOperties considered included ordinality--cardina1ity, commutativity, and invari- ance or conservation of number. (2) §g£§_which are concerned with intuitive cate- gories for classifying objects or events. As Langbehn, et a1., (1972) assert, "intuitively, we 95 think of a set as a collection of things which are separate and distinct from other things. Thus, given something, we know whether it belongs to that set or it does not" (p. 113). Set is a name which is given ". . . for an aggregate, ensemble, or collection of things that are combined under a certain criterion or according to a certain rule. The concept of a set arises by an abstraction. By considering a certain collection of objects as a set, we disregard all the connections and relations between the various objects that make up the set, but we preserve the individual features of the objects" (Aleksandrove, et a1., 1969, p. 5). The specific quantitative concepts considered included inclusion/exclusion, isomorphism (one-to-one correspondence) and the axioms of equality, viz., reflexitivity, symmetry, and transitivity. (3) Proportionalipy which is concerned with the rela- tion between quantities or cardinalities such that if one quantity varies another varies as a multiple of the first. The specific quantitative concepts in- cluded fraction, ratio, and percentage. (4) Matrices which facilitate the simultaneous classification of or operation upon the relation be- tween elements in two ways as defined by the rows and columns of a matrix (see Weiss and Yoseloff 1975, p. 260). The specific quantitaitve concepts 96 considered included simultaneity in classification through the mathematical operations of addition, sub- traction, and division. Second, items were developed and compiled to measure the four dimensions of quantitative literacy stated earlier in this chapter. These dimensions were manipulated along the specific quantitative concepts which derived from the concepts of series, sets, proportionality, and matrices as stated above. Third, after compiling the quantitative literacy items, the author checked with some mathematicians, statisti- cians, and similar others for the correct responses on the QUTitems. Appropriate changes were made on the wording and structure of some of the items whenever this was necessary. The quantitative concepts stated above and their correspond- ing QLT items in the instrument are summarized in Table 8. Fourth, a number of community agencies were contacted for information on their current innovations for the clients they serve. The agencies included health, Co-operative Ex- tension, and Social Service Agencies, inter alia. From these contacts and documents secured from the agencies, ten innovations were selected to be included in the instrument. In selecting the innovations, effort was made to en- sure that the types of innovations to be included in the in- strument represented a range of innovations from the very simple to the more complex innovations. The more complex 97 Table 8. Quantitative Concepts and Their Corresponding Number of QLT Items in the Instrument for the USA - Michigan Data Set Basic Quantitative Specific Quantitative Item # (see Concepts of Operation Concepts Appendix B, pp. 182-199 1. Series a. Ordinality: la,b; 2a,b; b. Cardinality: c. commutativity: d. Invariance or conservation of number 2. Sets a. Symmetry: b. Transitivity: c. Inclusion- Exclusion: d. Isomorphism (one- to-one correspond- ence): 3. Proportionality a. Fraction: b. Percent: c. Ratio: 4. Matrices a. Simultaneous operations 13a; 14a; l6a,b; 3a,b; 10; ll; lZa-c; 20e; 22a-e; 23a-e l7 7; 13b; 14b 8a-c; l9 5; 6; 15a-c 20c; 21a-c l8; 20d 20a-b innovations were distinguished from the simple ones in that the former imposed a demand for more precise measurements on the part of the potential adopter than did the simpler inno- vations. It was assumed that the more numerate (i.e., 98 quantitatively sophisticated) individuals would be those with higher levels of education. It was therefore expected that individuals with higher levels of education would be more likely than not to adopt or to have the propensity to adopt the more complex innovations compared to the individu- als with lower levels of education; the latter category of individuals would tend to shy away from the complex types of innovations in which numbers are used. For this purpose, items were included in the instru— ment to measure individual's propensity to adopt certain types of innovations which were described in the instrument. This was done by describing a situation of adoption in which the respondent was required to choose adopting only one of two types of technological innovations described, one of which was a simple innovation and the other a more complex one which required use of numbers in some way. To these rather contrived innovations--eight of them (see items #7, 8, l6, and 35, Appendix B, pp. 187-189, 191) were added ten simpler innovations which included Consumer Reports, Dollar Watch, Tel-Med, Consumer Credit, Expanded Nutrition Program, Diabetes test, Cardiopulmonary resusci- tation (CPR), Project Health, the Michigan Winter Heating Bill, and Meal Planning. Hence, this range of technological innovations was expected to discriminate more sensitively the knowledge levels and the adOption behavior among the populations in question. 99 Fifth, the items designed to measure respondent's knowledge and adoption levels for the stated ten simpler innovations and adoption propensity for the eight more com- plex innovations were then developed. For the simpler inno- vations, the items included were those assumed to be more sensititve and complete in measuring the important dimen— sions of knowledge and adoption as discussed in the opera- tionalization of variables in this sample space (also see Appendix B, pp. 182-200). Sixth, the instrument was then put together using both the QLT items and the knowledge and adoption items, and items for biographical information. These were then all translated into Spanish. Seventh, the instrument was then pretested among ten Spanish speakers from a Spanish Speaking Senior Citizens organization in Lansing, and among six students from Michigan State University. Following this pretest, further word changes were made whenever necessary to improve the sensitivity of the instrument. In addition, the open—ended questions which asked for the respondent's reasons for his/her answers on some of the QLT items were deleted since the questions did not seem to provide useful data for analysis. This deletion also reduced the amount of time and energy of the respondent in completing the instrument. This saving in time and ener- gy was particularly necessary for two reasons. First, 100 reading problems were anticipated among some of the respond- ents in this study. Secondly, except for the MSU students who were to be given extra credit for their participation in this study, the rest of the respondents merely volunteered to participate in the study. Consequently, it was necessary to keep the final version of the instrument as short as possible. After the preceding preliminary precautions, the final version of the instrument was then printed in both English and Spanish. B . SAMPLING The theoretic hypotheses to be tested in this study were stated in the last section of the first chapter of this dissertation. The preceding section of the current chapter, (1) operationalized the variables in this study's two sample spaces of Nigeria-Ilewo and the USA - Michigan, and (2) re— ported the reliability test of the QLT items. The purpose of the current section is (l) to discuss the method of select- ing this study's respondents, (2) to discuss the characteris- tics of these respondents, and (3) to point out the limitations of the samples. 1. Method of Selectinngespondents This is an exploratory study whose main goal was to obtain valuable insights which may lead to further investiga- tions. As mentioned earlier, some of the observation units 101 (respondents) in this study merely volunteered to participate :in this study; others were given an incentive of extra credit :in an introductory communication course for them to partici- pate in this study. In the Nigerian village of Ilewo, the respondents Inerely volunteered to participate in the study. Every male over the age of twenty was interviewed throughout this vil- lage as part of the Nigerian Diffusion Project at Michigan State University with the Economic Development Institute of the University of Nigeria at Enugu as a cooperating institu- tion. The survey involved 364 male cases altogether. At the time of the survey, Ilewo was a re—settlement and pre- dominantly farming village. The respondents for the USA - Michigan sample space derived from: 1. 4. the Spanish Re-entry (school dropouts) students of English as a second language and the Spanish Speaking Senior Citizens Organization under the auspices of Cristo Rey in Lansing; the Spanish Re-entry students of English as a second language in the United Migrants for opportunity (UMOI) organization within the Adult Basic Education (ABE) program of the Lansing School District; two high schools in Flint, viz., Carman High School and Mott Adult High. These samples in- cluded demographically heterogeneous ABE students, and Michigan State University (MSU) students who were enrolled in an introductory communication course. These data sources provided a total of 242 useful cases for analysis. Their distribution by race and sex is Shown in Table 9. 'ITEilale 9. 102 The Distribution of Respondents in the USA - Michigan Sample Space by Race and Sex RNIEANDEHDKOFIESRGEENTS Blwflc Imndcan Nathma Vflfite IJPflA .Hmnican Amnjcan Amaicai .Nmnicmi Oflmm Tbtfl. EKDIHCES F M. F M F M F M F D4 (Ixfismolky Ikeixmzy EStuflyms 5 5 10 Ekxuor Citizens 6 4 10 IJNDI: Re- eymxy Stukxms 4 6 10 (ZanmulHiQh School 4 l 6 8 3 3 36 21 82 Dininmhflt IHigh 7 ll 2 l4 3 21 5 63 ldSU 4 2 _l;_ _1;_ 42 16 l 67 Total 15 14 24 37 7 3 99 42 l 242 This investigator located the first four data sources through friendship and professional networks. After locating them, he then contacted the respective administrators for their permission and cooperation for this study to be done in their institutions. Like in the Nigeria - Ilewo sample space, the respond- ents from the first four data sources named above merely vol- unteered to participate in the USA - Michigan part of this 103 study. The respondents from Michigan State University were, however, promised and given extra credit for their participa- tion in this study. This extra credit was to be added to each respondent's potential grade in the introductory com- munication course in which the respondents were enrolled at the time of this study. This arrangement served as an incen- tive which attracted more respondents with college level education. Thus, in both sample spaces (Nigeria - Ilewo and USA - Michigan), a non—random sampling method was applied to obtain the needed measurements (observations) from these two sample spaces. This method was conceived as the most appropriate sampling technique for obtaining a large amount of observa- tions for valuable insights in a volunteer situation where "probability sampling either may be too expensive or lead to fewer such insights" (Blalock, 1972, p. 527). Measurements were therefore taken from all the obser- vation-units who volunteered in the sample spaces to partic- ipate in this study. Because of the volunteer situation, each individual ipso facto had an equal chance of being in- cluded in the study sample. Accordingly, it is also to be assumed here that there is independence of selection within the sample spaces in the sense that the choice of one indi- vidual or groups of individuals has no connection on the choice of another individual or groups of individuals to be included in the sample spaces. 104 Mbreover, it may also be assumed that the variables being studied are random variables in these sample spaces. That is, these variables can assume any of the possible values of random variables. Hence, each variable "as a random variable, the probability of distribution of a sample observation is identical with that of the population of measurements--the random variable under consideration" (Chou, 1972, p. 270). Given these assumptions, the findings which derive from the test of this study's hypotheses are generalizable to populations which are made up entirely of the individu- als who are relatively homogeneous with respect to the vari- ables or characteristics being studied here. 2. Characteristics of Respondents Both the Nigeria - Ilewo and USA - Michigan sample space have observation-units (respondents) or groups of ob- servation-units (respondents) who are representative of this study's population with respect to the variables being studied. The Nigeria - Ilewo sample space has observation- units whose level of formal education is a continuum ranging from no years of school to thirteen and above years of school. Following the conceptualizations in the first chapter with respect to this variable, it is expected that years of school will discriminate these respondents with respect to their levels of literacy in native language 105 (Yoruba) and in English and, ipso facto, with respect to their levels of awareness and adoption of technological in- novations in agriculture and health. Moreover, since the observation-units in this sample space came, as it was stated above, from a re-settlement and predominantly farming village at the time of the survey, it may be assumed that these observation-units had farming ex- perience. The agricultural innovations should therefore be assumed to be relevant to the observation-units particularly since the agricultural innovations included in the survey were those which were related to the type of agricultural systems in the village. The health innovations were also those which were related to the health practices in a rural setting. It may therefore be assumed that any discrimina— tion on awareness and adoption levels is to be attributed to differences among the observation-units with respect to their levels of education and, ipso facto, to their differ- ences in levels of literacy rather than to the relevancy of the innovations in the survey. The USA - Michigan sample space has observation-units whose level of formal education is a continuum ranging from no years of school to college or university level of educa- tion. Again, following the conceptualization in the first chapter with respect to this variable, it is assumed that this variable will discriminate among the observation-units with respect to their levels of literacy in native language 106 (Spanish), literacy in English,18 and literacy in quantita- tive symbols and concepts and, ipso facto, with respect to their knowledge and adoption of technological innovations in health, c00perative extension, and social service programs. The college or university observation-units were in- cluded in this sample space to provide a complete array of individuals on the continuum of formal education variable. Thus, this inclusion provides useful comparative bases for the study variables such as levels of literacy (both quanti- tative and non-quantitative) and knowledge and adoption of technological innovations when these observation-units are arrayed along the continuum of formal education. Moreover, those adult populations who have little or no formal education in this sample space, are rather similar to the majority of the adult populations in the Developing countries with respect to this variable. It is therefore expected that they have a pretty limited level of quantita- tive literacy. This, therefore, makes them apprOpriate, although not the preferred, populations for the tests on the relationships between quantitative and non-quantitative literacy on one hand and knowledge and adoption of techno- logical innovations on the other. Their similarity to "Third-World" populations makes the findings of the measures in this study generalizable to those populations. 107 3. Limitations of the Samples The Nigeria - Ilewo sample space, as stated earlier, involved 364 cases which were all male. Thus, this sample space has a sex bias. In areas where women traditionally make the decisions with respect to family farm operations and/or health practices, this male bias might lead to biases in the observations particularly with respect to measures of awareness or knowledge and adoption of technological innova— tions in agriculture and/or health. A limitation in the USA - Michigan sample space may also be observed. The observation-units in this sample space are frequently exposed to messages of technological in— novations in health, Cooperative extension, Social services, and so forth disseminated by some of the highest technolog- ical mass media developments in the world. These mass media have included television and radio which generally do not require reading nor writing skills to acquire knowledge for decision making with respect to the technological innovations. Thus, these mass media channels have potential to contribute additional explanation in the variance of knowledge of tech— nological innovations beyond that explained by literacy among these observation-units. 108 C. DATA COLLECTION l. The Nigeria - Ilewo Data Set As stated earlier, every male over the age of twenty years was interviewed in 1966 throughout Ilewo village as part of the Nigerian Diffusion Project which was administer- ed at Michigan State University with the Economic Development Institute of the University of Nigeria at Enugu as a co— operating institution. The interviews were carried out by a team of trained interviewers using the Project's question- naires. The survey involved a total of 364 male cases. The schedule had several indices together with ques- tions regarding the awareness and the use of technological innovations in agriculture and health as well as communica- tion behavior, farming operations, achievement motivation, empathy, fatalism, interpersonal trust, personal ratings in the village, occupational aspirations, education, literacy (in local language—Yoruba and in English), plus several sociometric and demographical questions. From this schedule, this investigator selected the variables of interest for the current study. The variables which were selected and their respective response categories were stated earlier in this chapter under the operationaliza- tion of variables. These data are very important in investigating the potential effects of literacy in English on the awareness or knowledge and adoption of technological innovations as a 109 new dimension of literacy in diffusion practices in parts of the world where English is not the native language. The data also provide a very important comparative basis on the relationship between literacy in English and awareness or knowledge and adOption of technological innovations since African languages and dialects do not have a common linguis— tic root with the English language compared to the Spanish language whose native speakers are included in the second part of this study. 2. The USA - Michigan Data Set The data in the USA - Michigan sample space were col— lected in the Spring and Summer of 1978 using the question— naire or instrument described earlier in this chapter (also see Appendix B). The questionnaires were either distributed to the respondents to fill them out at their own leisure, or they were administered in group situations and on a one-to-one basis whenever either method was the necessary possibility. This investigator superviSed all the groups involved in the data collection process. In addition, he picked up the completed questionnaires from each of the study sites where the questionnaires had been left for completion at a later time. For MSU students, this researcher made a brief appear- ance in each of the sections of an introductory Communication course to solicit for students to participate in his study 110 for extra credit. After a brief announcement and explana- tion, this researcher left copies of a memo with each instruc- tor for the students who needed extra credit to pick them up from him/her for extra details such as purpose of the study and schedules for participating in the study. Participation in this study meant filling out the study's questionnaire. Each participant was given 0.05 extra credit per hour of participation following the Department's regulations. Those students who needed extra credit reported to a room as stated in the memo and according to the schedule. The researcher introduced himself and reiterated the purpose of the study and the details concerning extra credit. The purpose of the study was stated to be in line with the mes- sage addressed to each participant on the cover page of the questionnaire (see Appendix B) with the addition that several other people (both in-school and out-of—school) were partici- pating in this study. The researcher then invited questions related to the study from the participants, and after answer- ing them (whenever there were some), he then passed out the questionnaires to be completed at that sitting. The Slosson Oral Reading Test (SORT) was administered to each participant before leaving the room. This segment of respondents yield- ed 67 completed questionnaires. In case of the High School and ABE students in the two Flint schools, the researcher delivered the questionnaires to the Principal of each school involved. Through their lll cooperation, the questionnaires were administered by the teachers who normally conducted the classes in which the vol- unteering students were enrolled. These students filled out the questionnaires during their leisure time. Important details in administering the questionnaires and the SORT instrument were given to the principal of each school to pass on to his teachers involved in the collection of the data. In one school in Flint, 88 questionnaires were return- ed out of the 150 distributed. Of these 88 questionnaires, six questionnaires had very few items completed. Hence, these six questionnaires were not very useful; they were dis- carded, thus, leaving the total of 82 useful questionnaires from this particular school. From the other Flint school, of the 70 questionnaires distributed, 63 well-completed ques— tionnaires were returned. The data from the three Spanish groups: the Re-entry students, and Spanish Speaking Senior Citizens organization, and the United Migrants for Opportunity (UMOI) who had read- ing problems were gathered in group situations with the help of some of the Spanish-speaking students and their spouses in MSU, and sometimes with the assistance of the Spanish- speaking persons working among these populations. From these groups derived 20 well-completed questionnaires. In addition, some questionnaires were distributed to those Spanish Speaking Senior Citizens who reported no reading 112 problems and had volunteered to participate in the study. They took the questionnaires with them to complete at their leisure. In this case, only the SORT instrument was admin- istered at the place of the meeting. Of the 50 question- naires distributed to this group, only 10 well-completed questionnaires were returned. Following the above analysis, the investigator was thus able to get altogether 242 completed questionnaires (67 from MSU, 82 from one Flint school, 63 from the other Flint school, and 30 from the Spanish-speaking groups). D. DATA PROCESSING After collecting the data, this investigator did a content analysis of the reSponses to the open-ended items of knowledge of technological innovations. From this con- tent analysis, he derived the common categories for coding the responses to the open-ended items. Using these categories and the response categories of the non-open-ended items in the instrument, this investi- gator constructed the codebook for the study. A c0py of this codebook may be obtained from this investigator. After printing the codebook, this researcher then proceeded with the coding of the data. The coding was done on opscan computer sheets by groups of coders whom this researcher trained to do the coding. These groups included the students who were doing independent study with this 113 researcher, students who needed extra credit toward their grades in an introductory course in the Department of Com- munication at MSU, friends in MSU and relatives and their friends from Lansing Community College. When the coding was over, a data deck was then punch- ed from the opscan computer sheets using the computer facil- ity of the Evaluation Services at Michigan State University. Using the MSU computer interactive system, this in- vestigator proceeded to clean the data deck for analysis. He had the assistance of two friends with computer program- ming experience. E. METHODS OF ANALYSIS This study's data were analyzed through multiple re- gression, and correlations which the regression routine pro- vided, Chi square (X2) tests, and one way analysis of variance (ANOVA). 1. Multiple Regression Models Since functional relationships have been specified among the variables in this study, multiple regression models were formulated relating the independent and depend- ent variables in each of the two sample spaces. Specifically, the following regression models were tested in each sample space using the stepwise mode of this analytic routine; in each model, B is the so—called constant term parameter. It 0 expresses the value of the intercept that the dependent 114 variable Yi takes on when the value of each independent vari- able Xi in the model is set to zero; Bi is the regression coefficient associated with the ith independent variable. That is, it is the slope of the regression line and it indi- cates the change in the mean of the probability distribution of the dependent variable per unit increase in an independ- ent variable, and Bi is the residual term which is associ- ated with the ith dependent variable. It expresses the difference between the observed value of Yi and the corre- sponding fitted or predicted value Yi: a. Multiple Regression Models in the Nigeria — Ilewo Sample Space. In this sample space, the following models were tested with the number and order of variables fianlly appearing in the equation determined by the stepwise routine: 1. Y1 = B0 + Ble + B2X2 + B3X3 + El where Y1 = level of awareness (AWARE) of technological innovations, X1 = level of literacy in native language (YORUBA), X2 = level of literacy in English (ENGLIT), and X3 = last grade of school completed (LGRADE). 0 1X1 + BZX2 + B3X3 + B4X4 + E2 where Y 2 level of adoption (ADOP) of technological innovations, x1 level of literacy in native language (YORUBA), 115 X2 = level of literacy in English (ENGLIT), X3 = last grade of school completed (LGRADE), and X = level of awareness (AWARE) of 4 I I I technological innovations. b. Multiple Regression Models in the USA - Michigan Sample Space. For purposes of analysis, this sample space was divided up into two groups which included the native English-speaking group (N = 169) and the native Spanish- speaking group (N = 61) as noted earlier. The following models were tested in the native English-speaking group, with the number and order of vari- ables finally appearing in the equation determined by the stepwise routine: 1. Y1 = B0 + Ble + BZXZ + El where Yl level of quantitative literacy (QLIT), X1 = last grade of school completed (LGRADE), and X2 = level of literacy in English (ENGLIT). ll. Y2 = BO + Ble + BZX2 + B3X3 + E2 where Y2 level of knowledge (KNOW) of technological innovations, xl last grade of school completed (LGRADE), X2 = level of literacy English (ENGLIT), and X3 = level of quantitative literacy (GLIT). 116 111. Y3 = BO + lel + B2X2 + B3X3 + B4X4 + E3 where Y3 = level of adoption (ADOP) of technological innovations, X1 = last grade of school completed (LGRADE), X2 = level of literacy in English (ENGLIT), X3 = level of quantitative literacy (QLIT), and X4 = level of knowledge (KNOW) of technological innovations. The following models were tested in the native Spanish-speaking group, with the number and order of the variables finally appearing in the equation determined by the stepwise routine: 1. Y1 = B0 + Ble + BZXZ + B3X3 + B4X4 + El where Y1 = level of quantitative literacy (QLIT): X1 = last grade of school completed (LGRADE), X2 = level of literacy in English (ENGLIT), X3 = level of literacy in native lan- guage--Spanish (SPANLIT), and X4 = level of literacy in both English and Spanish-biliteracy (BILIT). 11. Y2 = B0 + Ble + BZX2 + 33x3 + B4X4 + BSXS + E2 level of knowledge (KNOW) of technological innovations, where Y2 >< ll last grade of school completed (LGRADE), 117 X = level of literacy in English ENGLIT), X3 = level of literacy in Spanish (SPANLIT) X4 = level of literacy in both English and Spanish (BILIT), and X5 = level of quantitative literacy (QLIT). 111. Y3 = Bo + lel + B2X2 + B3X3 + B4X4 + BSXS + 36X6 + E3 where Y3 = level of adoption (ADOP) of technological innovations, x1 = last grade of school completed (LGRADE) , X2 = level of literacy in English (ENGLIT), X3 = level of literacy in Spanish (SPANLIT), X4 = level of literacy in both English and Spanish (BILIT), X5 = level of quantitative literacy (QLIT), and X6 = level of knowledge (KNOW) of technological innovations. The multiple correlation squared obtained in computing the regressions will be used to determine the variance account- ed for by all the regressors in the equation for each of the hypotheses involving regressions. The additional increment added when each regressor is added to the predictive equation will be used to determine the additional variance explained by addition of each regressor to the first one, and each 118 subsequent one, fitted in the predictive equation. The t-test will be used to test for statistical sig- nificance of the individual partial regression coefficients when all regressors are included in the equation; and the F- test will be used to test for statistical significance of the joint effect of the regressors at each step in the step— wise analysis. One advantage of this method with stepwise regression is that in cases where there is high multicolline- arity and the individual contributions of the regressors to the variance in the dependent variable are not significant and the joint contribution is significant, it suggests that the independent variables may be indicators of an underlying concept which produces a major portion of the variance in the dependent variable. The data in the zero order correlation matrix will be used to determine the strength of the relationship hypothe- sized between LGRADE and each of the other variables used in the study (Hypotheses l, 4, and 8). 2. Other Statistical Anaiyses on the USA - Michigan Data Set Two additional statistical tests will be performed on selected variables in the USA - Michigan data set to provide additional information which the multiple regression tests cannot give. Chi square tests will be performed on the distribution of frequencies between last grade of school completed (LGRADE) 119 and the propensity of adopting complex technological innova- tions in both the native English-speaking and Spanish-speak- ing groups. One way analysis of variance (ANOVA) will be performed on the significance of the means of knowledge (KNOW) and adoption (ADOP) of technological innovations among the constructed levels of literacy (LITLEV) in the native Spanish-speaking group. CHAPTER III RESULTS The preceding chapter operationalized this study's variables, and described the methods and techniques for sampling, data collection, and data processing. The first chapter concluded with the list of theoretic hypotheses which this study was designed to test in the two sample spaces of Nigeria - Ilewo and the USA - Michigan. The purpose of the current chapter is to report and discuss the results in the analyses which were performed in both of these sample spaces. A. RESULTS IN THE NIGERIA - ILEWO SAMPLE SPACE As stated in the preceding chapter, the data set of this sample space was analyzed through the stepwise mode of multiple regression on two regression models and the corre- lations which the regression routine yielded. The statistical multiple regression models tested were stated in the previous chapter (see section E. l. a). The two models may be re-stated by substituting the Y's and X's with the real variables, and by omitting the error term from the equation. The respective mathematical models that derived are as follows, with the number and order of the 120 121 variables in the final predictive equation determined by the stepwise routine: a. Aware = B + B YORUBA + B ENGLIT + B LGRADE 0 1 2 3 b. ADOP = BC + BlYORUBA + BZENGLIT + B3LGRADE + B4AWARE These models were analyzed via the current Version of the SPSS Subprogram Regression (SPSS volume 7.0) at Michigan State University. Four restriction parameters associated with stepwise multiple regression routine were imposed on each model to fit each regressor into the predictive equation. The four parameters included: (1) NSTEPS; the maximum number of steps, (2) FIN; which specified the minimum F value to enter a regressor into the equation, (3) TOL: which speci- fied the minimum tolerance level to enter a regressor into the equation, and (4) FOUT; which specified the maximum F value to remove a regressor from the equation. The basic as- sumption of these restriction parameters is that at each step of stepwise regression analysis, the regressor which makes the greatest increment to R2 (the coefficient of de- termination) is entered into the equation provided the F ratio associated with it exceeds the critical F value (FIN) for fitting such a regressor into the equation. Equivalent- 1y, it is the regressor which has the highest partial corre- lation with the dependent variable, after having partialled the regressors already in the equation. Three sets of values for each of the four restriction parameters were tried in fitting the predictors of each model 122 into the predictive equations. The results of these prelim- inary trials are reported in Appendix C. To delve into a more complete exploration of the rel- ative predictive power of the regressors in each model, the regressors were allowed to fit liberally into the predictive equation by using the default values for the four parameters; that is, (l) NSTEPS = number of regressors in the model, (2) FIN = .01, (3) TOL = .001, and (4) FOUT = .005. The mean (i) standard deviation (S) and standard error of the mean (SE) for each variable in this data set are shown in Table 10. Table 10. The Mean (i), Standard Deviation (S) and Standard Error of the Mean (SE) for the Variables in Nigeria - Ilewo Data Set (N=364) VARIABLE ‘x s s— X YORUBA 1.096 1.607 .08 ENGLIT 1.064 2.166 .11 LGRADE .613 1.141 .06 AWARE 72.779 25.808 1.35 ADOP 30.321 21.218 1.11 The results from the correlations and regression analyses are reported in the following sections. 1. Results from Correlation Analysis (N=364) As noted above, the regression analysis routine yield- ed the intercorrelations among the variables in the data set. 123 Table 11 presents the zero order correlations among the vari- ables in the data set. Table 11. Zero Order Correlation Matrix Among the Variables Used in the Nigeria - Ilewo Data Set (N=364) YORUBA ENGLIT LGRADE AWARE ADOP YORUBA 1.000 ENGLIT .853* 1.000 LAGRADE .859* .894* 1.000 AWARE .247* .254* .255* 1.000 ADOP .101 .096 .149* .697 1.000 * significant; P <:.025 (.05, two-tailed) The following observations are to be noted on the pattern of correlations in Table 11. First, except for the correlation between adoption (ADOP) of technological innovations and literacy in native language (YORUBA), and between ADOP and literacy in English (ENGLIT), all the other correlations are significant (_P_ < .025) . Second, all the variables have positive intercorrela- tions. That is, they all vary directly with each other. Third, there is a very strong correlation between awareness (AWARE) and adoption (ADOP) of technological inno- vations compared to the correlations of YORUBA, ENGLIT, and LGRADE with AWARE. The correlation of ENGLIT and of LGRADE with AWARE is moderately low and about the same. YORUBA has 124 the weakest correlation with AWARE compared to the correla- tion of ADOP, ENGLIT, and LGRADE with AWARE. Fourth, the correlation between LGRADE and ADOP is very low in absolute terms. However, it is stronger than that of YORUBA and ENGLIT with ADOP. The correlation be- tween YORUBA and ADOP is stronger than that between ENGLIT and ADOP but both correlations are very low. Finally, there is a very strong correlation between YORUBA and ENGLIT, and between each of them with LGRADE. Following the conceptualizations on the theoretic relation between formal education and literacy (see Chapter I; section B. 3), a direct relationship was hypothesized be- tween LGRADE and: YORUBA: ENGLIT: AWARE, AND ADOP. The data supported that hypothesized prediction (Hypothesis 1). As noted above, the data show that there is a significant (P <3.025) direct relationship between LGRADE and each type of literacy, and between LGRADE and the awareness (AWARE) and adoption (ADOP) of technological innovations. However, note that while these correlations are all significant, the correlation between LGRADE and AWARE (.255) is moderately low and that between LGRADE and ADOP (.149) is very low; the correlation between LGRADE and YORUBA (.859), LGRADE and ENGLIT (.853) are all very strong as noted earlier. These very strong correlations imply that in predicting and explaining variations in the awareness (AWARE) and adoption (ADOP) of technological innovations, any of these three 125 variables (i.e., LGRADE, YORUBA, and ENGLIT) will explain most of the variance, and adding the other two will not con- tribute much additional variance explained. Hence, in anticipation of the stepwise multiple re- gression analysis, it is to be noted that the strength of these correlations will affect the relative predictive power of these variables when they appear together in the regres- sion models. Each regressor would explain only minimal additional variance on the criterion variable (i.e., AWARE or ADOP). The following stepwise multiple regression analy- ses attempt to determine the variance for each regressor in each of the two models which were tested partialling out the variance each regressor shares with the other regressor(s) already in the predictive equation. 2. Stepwise Multiple Regression for Awareness of Innovations (N=364) In the stepwise analysis to identify the most effi— cient predictors of awareness of technological innovations in this sample, it was found that nearly all of the explained variance was extracted at the first step by LGRADE. As may be noted in Table 12, LGRADE explained 6.5 percent of the variance in awareness. The other two regressors explained only aboutaihalf of one percent of the additional variance. This result could be expected with the high level of multicollinearity among the independent variables. When all the regressors are forced into the equation under the relaxed 126 m How u uchHNquHN 82 u mz NNNxN N .HNN\N .NNN\H u we N NN. v N .N 38883.. N NNN.NN NNN.H HNN.NN Hucmum .89 8 NNN. HNN.H HNH.H NNN.N HNN. NNN. NNN. 88» N2 NNN. NNN.H NNH.H H.388 Nz NNN. NNN.N NNN.N 88 N HNN.NH NNN. NNN. NNN. .388 HNH.H N NNNHNN NNN. NNN. NNN. 888 H m l NH 8805 mm m 33.8.3 mwpm NNN. H \N .tm u 3N m 2880 Nm 83332 H8 88 083 .. 8832 HH..N 9039025 80830583. «0 88¢ng How 93 mg gages: mmfifimpm Eofim gamma . NH magma 127 default values noted earlier for the regression routine, the overall F value continues significant at each step; how- ever, the variance explained has been dispersed among the regressors and none produces a significant t for the partial regression coefficients. These findings show support for Hypothesis 2, which states that LGRADE, YORUBA, and ENGLIT, or some subset of the three, will explain significant aware- ness variance in the awareness of the technological innova- tions. The order of fitting may be noted in Table 12, suggesting that LGRADE has the highest partial correlation with AWARE, followed by ENGLIT, and YORUBA in that order. Kerlinger and Pedhazur (1973, p. 296) state that it is difficult, if not impossible to untangle the variance accounted for in a dependent variable and attribute portions of it to individual independent variables which are highly correlated with one another. One possible explanation in the present analysis is that the variables are indicators of a single underlying concept (variable). This is consistent with the position developed in the rationale in Chapter I where the inter- dependence of formal education, language and literacy were discussed, as well as their relationship to other behaviors. This underlying concept might be termed "symbol proficiency." 128 3. Stepwise Multiple Regression for Adoption of Innovations (N=364) As with the equation for predicting awareness, the first variable extracted accounts for nearly all of the ex- plained variance in predicting adoption. That variable is awareness of the innovations, accounting for 48.6 percent of the 50.5 percent of the variance explained by all four vari- ables in the equation. Thus, the remaining three variables add only about 1.9 percent to that explained by awareness of the innovations, as may be noted in Table 13. As in the preceding regression equation, when all the regressors are forced into the equation under the relaxed default values, the overall F continues significant at each step of the analysis. In this equation, however, three of the four t-tests of partial regression coefficients are significant. It is concluded that the subset of AWARE, ENGLIT, and LGRADE explain significant variance in predict- ing adoption with this sample of respondents, recognizing that the additional increment of variance explained by the regressors entering second and third in the analysis was very minimal. Thus, these findings show support for Hypoth- esis 3. B. RESULTS IN THE USA - MICHIGAN SAMPLE SPACE Recall, this sample space was subdivided into two groups for the purpose of analysis. The two groups included the native English-speaking group and the native Spanish- 129 8 H8 8 H8883 uoz u mz N .m n8 m. 38383 n N NNN} N .NNNxN .HNNxN .NNN\H u 8 N NN. v .H PH H8883... NNN.N NNN.N NNN.HH- 38889 8 NNN.H- NNN.H NHN.H- 8mg m NNN.N NNN.H NNN.N 883 N NHN.N- NNN. NNN.N: 9388 N NHN.NH NHN. NNN. «NNN.HN NNN. NNN. HHN. ESE N HNN.N NNN.N NHN.HHu 38809 N NNN.N NNN.H HNN.N 883 N NNN.N- NHN. NNN.H- 888 m NNN.NH NHN. NNN. NNNNNNH NNN. NNN. NNN. .88 N NNN.N- NNN.N HNN.HH- 38809 N NNN.N: HNN. NNNf NHH—NH m NNN.NH NHN. NNN.N «NNN.NNH NNN. NNN. NNN. 86% N NNN.N- NNN.N NHN.HH- 38809 N NNN.NH HHN. NNN.N «NNN.NNN NNN. NNN. NNN. E H N .N 88885 NHH m 3838 88 NNN. H \N HE u 3m 8 H385 Nm 8333 H8 88 083 .. 8.832 5 83998 8H8H888 No 8398 Ho...H 83888 3888 888m .58 838 .NH .3an.. 130 speaking group. As in the Nigeria - Ilewo sample space, the same mode of the regression routine with the same restriction para- meters was used in analyzing the data sets of the two groups in the USA - Michigan sample space. This section reports the results from the analysis of the three regression models which were tested in each of the two groups and from the correlations which the regression routine provided. In addition, this section will also report the results from the Chi square (X2) tests in the two groups, and from the analy- sis of variance (ANOVA) in the native Spanish—speaking group with respect to selected variables as noted earlier (Chapter II; section E. l. b. 2). 1. Results from the Native English-Speaking Group (N=1693 The three regression models which were tested in this group were stated in Chapter II (section E. l. b). By sub- stituting the Y's and X's with the real variables, and omit- ting the error term from the equation of each model, the respective mathematical models may be restated as follows with the number and order of the variables in the final pre- dictive equation to be determined by the stepwise routine: a. QLIT = B + B ENGLIT + B LGRADE; b. KNOW = B + B1 QLIT 0 l 2 0 2 LGRADE + B3 ENGLIT; C. ADOP = B0 + I KNOW + B2 LGRADE + B3 QLIT + B ENGLIT. + B 4 131 The mean (i), standard deviation (S), and standard error of the mean (SE) for each variable in the data set of this group are shown in Table 14. Table 14. The Mean (i), Standard Deviation (S), and Standard Error of the Mean (3;) for the Variables in the Native Engligh-Speaking Group VARIABLE i S SE ENGLIT 123.698 59.309 4.56 QLIT 29.231 11.211 .86 LGRADE 11.917 2.120 .16 KNOW 26.308 12.380 .95 ADOP 11.195 2.671 .21 The following sections report the results from the correlation, Chi square, and regression tests. a. Results from Correlation Analysis (N=169). The zero order correlations among the variables in the data set for this group which are presented in Table 15, support the Hypotheses 4a through 4c,(that as LGRADE increases, there will be increases in ENGLIT, QLIT, and KNOW). It will be noted that Hypothesis 4d (ADOP will increase as LGRADE in— creases) is not supported. Hypothesis 4e will be discussed later. The following observations are to be noted on the correlations presented in Table 15: 132 Table 15. Zero Order Correlation Matrix Among the Variables Used in the Native English-Speaking Group ENGLIT QLIT LGRADE KNOW ADOP ENGLIT 1.000 QLIT .638* 1.000 LGRADE .648* .553* 1.000 KNOW .215* .276 .201* 1.000 ADOP .064 .099 .101 .126 1.000 * Significant; g <<.025 1. Except for the correlations between ADOP and each of the literacy types, and between ADOP and LGRADE and KNOW, the rest of the correlations are significant (3 < .025). 2. All the variables have positive intercorrelations. That is, they all vary directly with each other. 3. The correlation between KNOW and each of the literacy types and LGRADE although significant are moderate- ly low. 4. The correlation between LGRADE and ENGLIT (.648), LGRADE and QLIT (.553), and QLIT and ENGLIT (.638) are all very strong. Following the above findings, the data support the direct relationship between LGRADE and each of the two liter- acy types (i.e., ENGLIT and QLIT); but not between LGRADE and KNOW, and LGRADE and ADOP as predicted. Note, however, that although a direct relationship was found between LGRADE 133 and KNOW, and LGRADE and ADOP, the strength of the relation- ship found between LGRADE and KNOW (.201) is moderately low and that between LGRADE and ADOP (.101) is not significant as noted above. Finally, note that the very strong correlations be- tween LGRADE and ENGLIT (.648), LGRADE and QLIT (.553) and QLIT and ENGLIT (.638) are large enough that any of these three variables relative to the others may not explain much additional variance. As noted in the analysis of the Nigeria - Ilewo data set, each of the strongly correlated regressors will explain only minimal additional variance on the criter- ion variable beyond that explained by the others. The following stepwise multiple regression analyses attempt to determine the variance for each regressor in each of the three models which were tested partialling out the variance each regressor shares with the other regressor(s) already in the predictive equation. b. Chi Square Tests Between LGRADE and the Propens- ity of Adopting Complex Innovations (N=169*). Recall, it was noted in Chapter II (section A. 2. b. ii) that items were * Note that the grand total of cases in each of the four adoption situations is less than 169. This is because the non-use or omissions category was later omitted from the tests since the hypothesis to be tested compares propensity of using either the simple or complex innovation. The ex- clusion of this category necessitated computing a new Chi square value. This was done on the Hewlett Packard #67 calculator. 134 included in the instrument for the USA - Michigan data set to measure the respondent's propensity to adopt sets of simple and complex options on four technological innovations which were described in the instrument. The technological innovations included a simple and complex version of: (l) kit for testing the presence of sugar in urine for diabetes (DIABKITS); (2) form of using a drug for a health problem (DRUGFORM); (3) technique of cardiopulmonary resuscitation (TECHCPR), and (4) form of using a food product (FOODFORM). The Chi square test of independence was used to test for the relationship between last grade of school completed (LGRADE) and the propensity of using the complex option of the above four sets of technological innovations (Hypothesis 4e). LGRADE was collapsed into five categories of years of school (viz., 0-3, 4-6, 7-9, 10-12, and l3-16). The Chi square routine showed that for the native English-speaking sample, no cases fell in the 0-3 and 4-6 LGRADE ranges for the four adoption situations noted above. Thus, the routine did not include the two LGRADE ranges in the analysis. The same five LGRADE categories will be used in the native Spanish-speaking group to test for the same relationships as noted above. The following paragrpahs and Table 16 report the results from the Chi square tests in the native English-speaking group. With three of the four innovations, the Chi squares were statistically significant, supporting the hypothesis 135 Table 16. Observed Frequencies and Chi Square Values for Tests of LGRADE with Four Simple/Complex Innova- tions in Native English-Speaking Group LGRADE RAN GE Row Chi square 7-9 10-12 13-16 Total INNOVATION Values DIABKITS Simple 22 53 32 107 Complex _3_ 25 32 60 12.23* df = 2 Column Total 25 78 64 167 DRUGFORM Simple 24 71 57 152 Complex _9 _6 _5 11 2.04 df = 2 Column Total 24 77 62 163 TECHCPR Simple 12 15 7 34 Complex 14 61 56 131 13.88* df = 2 Column Total 26 76 63 165 FOODFORM Simple 21 44 23 88 Complex _5 34 .49 ‘12 15.27* Column Total 26 78 63 167 df = 2 * Significant; E <:.05 that those with more formal education have a higher propens- ity to adopt the more complex innovations. An inspection of the data shows the inverse relationship between LGRADE and adoption of the simple version of the innovation; and the direct relationship between LGRADE and the adoption of the complex version of the innovations. 136 c. Stepwise Multiple Regression for Quantitative Literacy (N=169). As may be noted in Table 17, most of the explained variance (40.7 percent) was extracted in step 1 by ENGLIT; and that in step 2 of the regression, the addi- tion of LGRADE into the equation explained an additional in- crement of 3.4 percent of the variance. Thus, the two regressors jointly explain 44.1 percent of the variance in quantitative literacy. The partial regression coefficients for both ENGLIT and LGRADE were found significantly differ- ent from 0 by the t-tests at the second step of the regres- sion analysis where both regressors are in the predictive equation. These findings show support for Hypothesis 5. The strong intercorrelations among the regressors concerned result in the first regressor entered into the predictive equation extracting a major portion of the vari- ance, with lesser amounts attributable to that which was entered next into the equation. d. Stepwise Multiple Regression for Knowledge of Innovations (N=169). The data in Table 18 show that QLIT was extracted first in the stepwise analysis and explained 7.6 percent of the variance. Under the relaxed default values noted earlier, LGRADE fitted into the equation at step 2 and ENGLIT was added at step 3. Those two variables together added less than half of one percent to the variance explained at step 1. Table 18 shows that the three regress- ors together explain 8.0 percent of the variance in KNOW. 137 m mom w ufiofldaNN n N NEE N .SQH u 8 N we. v m “N 5,8338%... NE. NNN.N NNN.N fiafiN :82 N NNN.N m8. HNN.H «NNN.NN N8. HNN. NNN. ENE N NNN.N NS. NNN. sag m NNN.N NNN.N NNN.NH AucNuN :82 N 08.: do. H2. «NNN.N: EN. EN. NNN. 96% H m I N 9.8865 mm m 33.3, mBN N8. A \8 3.... N. nm N 3390 mm 3633: 955 mad—00mm .8398. gflmz may 5 8833 903888 6N SNNNBNNN 03332 $863N scum 83.6.8 .2 6369 1'38 l\ N mom 0. quoEENNN uoz u E N .m you w. pomoflNENN u N NNN N 63} .53} u No N No. v o N 6.868888... oNN.~ NNN.N NNN.NH Econ—882 Nz NNN. NNN. 8o. :33 so. go. NNN. 988% N2 NNN. NNN. NHN. g N NNN.N o3. NNN. Ego m NNN.N NNN.N NNN.NH 380988 Nz NNN. NNN. 3N. «NNN.N moo. NB. NNN. 8.83 N NNN.N N8. NNN. .36 m NNN.N NNN.N NNN.NH 386.883 Nz ohm. HNN. Nom. «25.2 NB. NB. NR. 930 N N I .N 08898 «N m 3888, NNN.N N8. A \8 to u nN m 3885 mm 0303.2 986 mfixoommnfidmfi 9382 05 5 mgoflmg 30330506 mo $03392 mom 86mg many: mmflfimpm 88m magma . ma magma 139 The high intercorrelations among the regressors con- tinue to create difficulties in attributing the contributions to the dependent variable by the individual regressors. This is further complicated by the small amount of the total var- iance explained by the regressors. From these data it appears that QLIT is a somewhat stronger predictor of knowledge of technological innovations in this situation than were the other two regressors. Thus, these findings show support for Hypothesis 6. e. Stepwise Multiple Regression for Adoption of Innovations (N=169). In the analysis of this model, the multiple correlation data and the regressors added at each step under the default values used are shown in Table 19. As was evident from the correlation matrix discussed earlier, the regressors explained practically none of the variation in adoption with this sample. One of the difficul- ties may have been the lack of relevance of the innovation to this sample of respondents nearly a half of whom were college students. The mean score for adoption was 11.195 out of a maximum possible of 30. This puts the mean at the top of the bottom third of the range. From the data in Table 19, it is apparent that Hy- pothesis 7 was not supported since no one of the regressors in the model extracted a statistically significant amount of the variance in adoption. m “out”. uoNoflEoNN 062 u Nz NNH\N N .NNH\N .NNN\N .Noflxa u No N NN.”. o “oomoNNNooNN poo NN N 140 NNN.N. NRA NNN.N ficfimooov N2 8%: NNN. NNH... NNN. so. NNN. NNN. 95% N2 NNN. NNN. H2. 950 N2 HE. NNH. 8o. 8&3 Nz NNN.H NS. NNN. 52M N NNN . 5 SN . a HNN. N fiqBNSQ N2 o8. NNN. oNo. HNN.H 8o. mmo. NNH. 930 N2 NNN. SH. «8. mafia Nz NNN.H NZ. NNN. 52 m NNN.N NNH.H NNN.N 358.968 Nz oooé NNN. NNN. HNN.H N8. «mo. NS. Egg Nz N84 NNH. NNN. 302x m NNN.NN NNN. NNN.NH Bafimooov Nz NNN.H NNH. NR. SEN NS. NS. NNH. 392 H m l .m $855 mm m wanmflnm> mmum NNN. A \o to N nm N H8890 mm 033% 988 33881385 9382 0:.» sun mayo; 1.53383 mo :03 How 83meQO 0.3.32 mmfifioum Em magma . ad 0389 141 2. Results from the Native Spanish-Speaking Group (N=6l) As in the native English-speaking group, the same three statistical models were tested in the native Spanish- speaking group with the exception that the models in the latter group included literacy in Spanish (SPANLIT) and literacy in both Spanish and English; i.e., biliteracy (BILIT) among the regressors. Recall, the three models to be tested were stated earlier (Chapter II; section E. l. b). When the Y's and X's are substituted with the real variables, and the error term is omitted from the equation of each model, the respective mathematical models may be restated as follows with the number and order of the variables in the predictive equation to be determined by the stepwise routine: QLT = B + B1 0 BILIT + B2 LGRADE + B3 ENGLIT + B4 SPANLIT; KNOW = B0 + Bl QLIT + 32 BILIT + B3 LGRADE + B4 ENGLIT + B5 SPANLIT; ADOP = BO + B1 KNOW + B2 QLIT + B3 BLIT + B4 LGRADE + B5 SPANLIT. The results from the regression test of these models as well as from the correlations which the regression routine provided will be reported in this section. In addition, this section will report the results from the Chi square tests between LGRADE and the propensity of adopting complex technological innovations as in the native English-speaking group, and from the analysis of variance (ANOVA) for knowl- edge (KNOW) and adOption (ADOP) of technological innovations by literacy levels (LITLEV). 142 The mean (X), standard deviation (S), and standard error of the mean (SE) for each variable in the data set of this group are presented in Table 20. Table 20. The Mean (N), Standard Deviation (S), and Stand- ard Error of the Mean (S ) for the Variables in the Spanish-Speaking Grofip VARIABLE Y s s? SPANLIT 114.885 46.493 4.56 ENGLIT 46.131 40.206 5.15 BILIT 5665.033 6167.704 789.72 QLIT 19.492 10.540 1.35 LGRADE 7.639 4.363 .56 KNOW 18.262 5.986 .77 ADOP 11.066 2.394 .31 a. Results from Correlation Analysis (N=6l). The zero order correlations among the variables in this group's data set are shown in Table 21. The following observations are to be noted on the cor- relations shown in Table 21. 1. Except for the correlation between ENGLIT and SPANLIT, KNOW and SPANLIT, KNOW and QLIT, KNOW and LGRADE, and ADOP and each of the other variables, the rest of the correlations are significant (F <=.025). 2. All the variables have positive intercorrelations. That is, they all vary directly with each other. 143 Table 21. Zero Order Correlation Matrix Among the Variables Used in the Native Spanish-Speaking Group SPANLIT ENGLIT BILIT QLIT LGRADE KNOW ADOP SPANLIT 1.000 ENGLIT .199 1.000 BILIT .561* .801* 1.000 QLIT .272* .463* .493* 1.000 LGRADE .282* .650* .461* .371* 1.000 KNOW .185 .283* .299* .244 .004 1.000 ADOP .079 .091 .013 .034 .205 .178 1.000 * Significant; g < .025 3. Though the correlations between KNOW and ENGLIT, KNOW and BILIT, QLIT and SPANLIT, and LGRADE and SPANLIT are all significant, each of them is moderately low. 4. The correlations of QLIT and ENGLIT, QLIT and BILIT, LGRADE and BILIT, and LGRADE and QLIT are moderately high; while the correlations between BILIT and SPANLIT, BILIT and ENGLIT, and LGRADE and ENGLIT are quite strong, especial- ly BILIT with ENGLIT at .801. Following the above findings, the data support the hypothesized direct relation between LGRADE and each of the four literacy types--SPANLIT, ENGLIT, BILIT, and QLIT: but not between LGRADE and KNOW, and LGRADE and ADOP. Note that although the correlations between LGRADE and KNOW, and LGRADE and ADOP are in the hypothesized direction, they are not statistically significant (g <:.025). In addition, note that 144 though all the relationships are in the expected direction, their strengths vary as noted above. The lowest correlation is between LGRADE and KNOW (.004), and the highest is between LGRADE and ENGLIT (.650) with respect to the hypothesized direct relationships. Note particularly that the correlations between LGRADE and BILIT (.461), LGRADE and QLIT (.371) , QLIT and ENGLIT (.463), and QLIT and BILIT (.493) are moderately high while those between LGRADE and ENGLIT (.650), BILIT and SPANLIT (.561), and BILIT and ENGLIT (.801) are all very high. These significant correlations imply that each of these vari- ables will explain only minimal additional variance on the criterion variable when they appear together in a regression model. b. Chi Square Tests Between LGRADE and the Propensity of AdOptinngomplex Innovations (N=61*). As in the native English—speaking group, the Chi square test of independence was also used in the native Spanish-speaking group to test for the relationship between last grade of school completed LGRADE) and the propensity of using the complex varsion of the four sets of technological innovations noted earlier. * Note that the grand total in some of the above con- tingencies is less than 61. This is for the same reason as noted earlier in the footnote for the Chi square tests in the native English-speaking group. A new Chi square value was also computed in the current group using the Hewlett Packard #67 calculator. 145 The results from the Chi square tests in this group are reported in Table 22. Note that unlike the English—speaking group, the cur- rent group has all the categories of LGRADE except for the highest category (i.e., 13-16 years of school); none of the subjects in the Spanish-speaking group had reached this level of education. Note also the small number of respond- ents in the 7-9 and 10-12 LGRADE ranges. In this group, only one of the four innovations (viz., FOODFORM) produced a significant Chi square when comparing LGRADE with the simple versus the complex versions of the innovations. It was therefore concluded that the propensity of using the complex FOODFORM increases as LGRADE increases as had been hypothesized. The data did not support the hy- pothesized direct relationship between LGRADE and each of the other three complex versions of innovations (Hypothesis 89) . Note, however, that 2/3 of the Spanish-speaking group were below the seventh grade level of education, while none of the English-speaking sample were below seventh grade level. From this perspective, it seems encouraging to pur- sue the study of the relationship between LGRADE and complex— ity of innovations. The following stepwise regression analyses attempt to determine the variance contributed by the independent vari- ables in predicting quantitative literacy (QLIT), knowledge 146 NN.”. m NuoNoHNHoNHN ¥ oN N NH NN NH Hmuoa oeoHoo N n No II III. II. II II NNN.NH NH N N N N onmaoo NN o OH NH NH NHNEHN zmomooom HN N NH NN oN HNuos oesHoo N u NN .I. .II. In. It. .I. NN.N NH N N N N xNHNEou NN N N NH NH NHNEHN mmomoma NN N NH NN NH Hmooa osoHoo N n No .1. III. II. II. II NN.N N o o N H onosoo NN N NH NN NH oHosHN zmomoomo ON N NH NN oN Hmooe osoHoo N H No :1. III. II. II. II. NN.N NH N N N N onosou NN N NH NN NH NHNEHN NBHNNNHN NNoHN> HNHoa NHuoH NIN NIN Nuo onaN>ozzH mumsvm H50 30m mozmm mowumz :N mcowum>ocaH memEoU\mHQENm usom nuflz mnflmoq mo mumwa mom mmoam> mnmovm Ncu tam mowocmuqmum oo>uomno .mm OHQNB 147 of innovations (KNOW), and adoption of the innovations (ADOP). c. Stepwise Multiple Regression for Quantitative Literacy (N=6l). Table 23 shows that BILIT was fitted first into the predictive equation, and it accounts for most of the variance in QLIT, viz., 24.3 percent. LGRADE explains an additional 2.7 percent, with virtually nothing additionally explained by ENGLIT and SPANLIT. As may be seen in Table 23, when the regressors are forced into the equation by the relaxed default values noted earlier, they are fitted in the order: BILIT, LGRADE, ENGLIT, and SPANLIT. The hypothesis that some subset of the regressors will explain significant variance on quantitative literacy is supported when BILIT and LGRADE explain 27 percent of the variance as indicated by the F test. The t-test, however, shows that only BILIT yielded a significant 8 for literacy and that occurred in steps one and two. d. Stepwise Multiple Regression for Knowledge of Innovations (N=6l). In predicting knowledge as in predicting quantitative literacy, BILIT is fitted first into the pre- dictive equation by the stepwise analysis routine. It does not explain as high a proportion of the variance in KNO -- about half of the variance in this case, viz., 8.9 percent of the total of 18.7 percent of the variance explained by all the regressors in the equation. As may be noted in Table 24, LGRADE, ENGLIT, and SPANLIT add approximately two 148 m.nom u HonNNHoNHN Hoz n Nz NN\N N NN\N .NN\N .NN\H u «N N No... N “N HomoHNHoNHNN NNN.N NNN.N NNN.HH Huomumooov Nz HoH. NNN. oNo. NNNH.N ooo. HNN. oNN. NHHzmmN Nz NNN. NNN. NHN. aHHuzm Nz NNN. HoN. NNN. NuNNuH Nz NNN.H NNN. NNN. BHHHN N NNN.N NNN.N NNN.NH . Hunmumooov Nz NNN. NNN. NNH. NNNo.N Hoo. HNN. oNN. anqozm Nz NNN.H NNN. NNN. NNNNNH Nz NNN.H NNN. NNN. NHHHN N mmo . m mmm . m mma . NH $83808 Nz NNN.H NNN. NNN. «HNN.NH NNN. NNN. NHN. NNNNNH N NNN.N NHN. NNN. aHHHm N NHH.N NHN.H NHN.NH Huomunooov N NNN.N NNH. NNN. NoNN.NH NNN. NNN. NNN. NHHHN H N I. m NNNNHONH Nm _m NHQNNHN>_ oNpN NNo. H \ SE N nN m H888 NN oHoflHE moons NoNxmoNN Inmflqmmm £362 5 0.333 $33.33 How sang 0.3.33 mmng spam 3.38m .mm 0.33. 149 m How u £83356 82 u m2 “m you 283358 u m mm\m a 0qu .B\m .mm\~ 5m} n we a me. a unmoflafifl mm~.m mah.~ mm~.va Aucmpmcaov mz com. mom. mm». «mmm.m «Ho. hma. mmw. saga mz mow.a awn. mmm. aaqzmmm m «oo.~ mmv. mom. quuzm m Hmm.~- ¢¢~. mum.: mn¢muq mz mum.n 5mm. ~o~.u quum m mum.m oqm.~ HmH.mH Aucmumcpov mz m~¢.H own. mvm. «Hmm.~ omo. mha. ma¢. aHazmmm m w¢o.~ «mv. «mm. quuzm m www.mu mqm. m¢m.u mn¢muq mz m¢m.u mam. mma.u quHm ¢ m-.~H Hm¢.fl ~oa.ma Auqmumcoov mz mm¢.a mmm. pmm. «vam.m mmo. mea. omm. angzm m2 mmm.Hu mmm. hov.u mn¢muq mz ems. How. mea. aanm m mmo.ma ma¢.H mnm.na Aucmumcnov mz mmm.au Hma. «mm.u *mbm.m mmo. NHH. mmm. mn¢mmn m ma~.~ mma. hem. Hagan m mmm.ma moo.H mam.ma Aucmumcnov m 5mm.~ Hma. cam. gmms.m mmo. amo. mam. quHm H m I m mmmmug mm m 33.3. > 03m m8. A \o 0E u nm m ago mm 3033: 5 9336.5 30330508. 00 $033 you gag 033d: «mafia :80 33mg 90.5 5308503 9302 05 . vm 0.309 150 to three percent additional variance explained, while QLIT at the last step of the analysis adds about one percent of the variance explained. While the t-test of the regression coefficients in the step 5 analysis show nonsignificance for the t-test for BILIT, this seems inapprOpriate since the two variables entered later into the equation and explaining a smaller percentage of the variance produced significant t's for the coefficients. Apparently, as more variables are entered in the equation, there was a partitioning of the variance away from BILIT in the relatively small pool of variance being explained by all variables. From the data in Table 24, it is concluded that Hy- pothesis 10 is supported, since BILIT alone explains signif- icant variance at the first step and all the regressors to- gether explained significant variance in the knowledge of innovations, although the additional increments added at steps 2-5 are very small. e. Stepwise Multiple Regression for Adoption of Innovations (N=61). As in the previous regression models, the stepwise regression routine forced into the predictive equation, under the default values noted earlier, all the regressors which were specified for the current model. Table 25 shows that the routine fitted LGRADE first into the predictive equation followed by KNOW, BILIT, SPANLIT, QLIT, and ENGLIT in that order. All the six 151 m you 0 unmoflfiaflm “.02 u mz . V vm\m w .mm\m .mm\v .hm\m .mm\~ ~mm\.n fl HG a mo M "unmoflwfldmwm 30: 0H m... how . m mow . a SH . m 323989 mz vmm. m3. 3o. mBJ So. 2:. Km. 932m mz mam. I 3m . .2: .I EH40 m2 omm. 3.". .So. 82% m2 Sm . I mma . H3 . I 9am mz mom . a cum . 3m . 52 m2 mmm . H moa . >3 . E m was . m Hum . H mvm. m 33989 mz mmmf Nvm . ooaf mom . H So. o3. mmm . BHAO mz omm . «.2. . hmv . HHS/2mm mz mmméI mow. mmmf Em m2 mmné 3m. ova. zbzm m ooo.~ «mo. «3”. BE m m3 . m omm . H vma. m 383ch9 m2 mam . >2. . How. wmo .H moo. m3 . vmm . 9.524% mz nmmAI mom. mmmf 833 m2 3&4 nmm. mmm. 302M m mao.m m3. mma. a v mmmK mmHJ omm.m 3:030:09 mz Pm. .T vmm . m: .I mmo . m mmo . mmo . mam . 930 m2 mmhé vmm. 5mm. g m mmo.~ 2.0. 3H. "Ramon m mbod v34 mamd 3539.08 mz ~34 mom. mow. mmm.m .80. m3. Hum. 572 m2 mmoé moo. NHH. g m mum . 3.. 3o . mom. oa 330908 mz oomé one. N3”. mmm.m m3. «3. mom. g H m l .m mmmwuocH m m manmng mmum m8. :5 0E u. pm m 33.80 mm N 29.03. 980 962.63 05 5 30305.95 3.3395305 mo QO3§ How :03meQO 3&3? 0m§m EOE 0330mm . mm 0.309 152 regressors together explained approximately 10.7 percent of the variance in ADOP. LGRADE alone explained the largest amount of variance explained at any of the steps by any of the regressors. Thus, Hypothesis 11 was not supported. f. Results From One Way Analysis of Variance (ANOVA) for Knowledge (NOW) and Adoption (ADOP) of Technological Innovations Among Literacy Levels (N=50*). Four levels of literacy (LITLEV) were derived to match the four theoretic levels of literacy discussed under Figure 2 (see Chapter I, section B. 4). The author wished to determine whether or not knowledge (KNOW) and adoption (ADOP) of technological innovations are related to the four levels of literacy derived. The four levels of literacy were derived by arbitrar- ily imposing some constraints on the respondents' literacy scores in Spanish (SPANLIT), English (ENGLIT), and quantita- tive symbols and concepts (QLT). A series of constraints were tried to get a relatively meaningful distribution of this group's 61 cases along the four literacy levels. The investigator at first tried to use the scores in the first and third quartile of SPANLIT, ENGLIT, and QLT as shown below. * This is not equal to the total of 61 cases in this group becuase 11 cases could not meet the criteria imposed in deriving the four levels of literacy. 153 Level I: SPANLIT i 64; ENGLIT 5 64, and QLIT : 10 Level II: SPANLIT 3 145; ENGLIT _<_ 64, and QLIT _g 10 Level III: SPANLIT 3 145; ENGLIT _>_ 145, and QLIT 3 lo, and level IV: SPANLIT _>_ 145; ENGLIT _>_ 145, and QLT _>_ 30 Unfortunately, however, these constraints were too conservative; 55 cases could not be placed on any one liter- acy level. Level III was lost in that no case could be located on it. Level I, II, and IV had 2, 3, and 1 cases respectively. These frequencies are too low for meaningful discrimination in variance among literacy levels. Consequently, the above set of constraints were re- laxed as shown below. Level I: SPANLIT _<_ 64; ENGLIT 3 64, and QLIT _<_ 10 level II: SPANLIT _>_ 100; ENGLIT _>_ 64, and QLIT _<_ 10 level III: SPANLIT 3 100; ENGLIT 3 100, and QLIT 5 10, and level IV: SPANLIT 3 100; ENGLIT _>_ 100, and QLIT 3 26 Under these constraints, 52 cases could not be placed on any one literacy level, and Level III was again missing. Level I, II, and IV had 2, 6, and 1 cases respectively. Once again, the investigator felt that these frequencies were too low for meaningful analysis. Hence, the constraints were further relaxed as shown below. Level I: SPANLIT: 80; ENGLIT 5 80, and QLIT: 25 Level II: SPANLIT 3 81; ENGLIT _<_ 80, and QLIT_<_ 25 Level III: SPANLIT :_81; ENGLIT Z 81, and.QLIT g 25, and Level IV: SPANLIT _>_ 81; ENGLIT 3 81, and QLIT 3 26 154 Under these sets of constraints, 9 cases were on Level I; 33 on Level II; 2 on Level III, and 6 on Level IV. These were used in the analyses; Level III and IV were com- bined into one level (Level III) which thus resulted in a total of 8 cases for the new level. The following para- graphs report the results from the ANOVA for knowledge (KNOW) and adoption (ADOP) of technological innovations among the three literacy levels. i. Results From One Way ANOVA for Knowledge (KNOW) of Innovations Among Literacy Levels—(N=SUT' Table 26 shows the means of knowledge of technological innovations for the number of cases in each of the three literacy levels. Table 26. The Mean of Knowledge of Technological Innovations for Spanish-Speaking Group N Cases Mean Level I 9 16.56 Level II 33 17.94 Level III 8 21.13 Grand Mean = 18.20 The summary statistics from the analysis of variance are presented in Table 27. 155 Table 27. One Way ANOVA Summary Table for Knowledge of Technological Innovations Among the Literacy Levels in the Native Spanish-Speaking Group SOURCES OF VARIATION SS df MS F Main Effects LITLEV 95.024 2 47.512 1.208 Residual 1848.976 41 39.340 Total 1944.000 49 39.673 Multiple R2 = .049; Multiple R = .221; ETA = .22 The F test was not significant (F = 1.208; df = 2 & 47; g <<.05). This implies that the means of knowledge of technological innovations presented in Table 26 do not differ significantly. In other words, individuals with the three different levels of literacy do not differ significant- ly in their knowledge of technological innovations. This is further indicated by the very low R2 (.049). It was there- fore concluded that there is no statistically significant relationship between level of literacy and knowledge of innovations. Thus, the data did not support Hypothesis 12a which predicted a direct relationship between level of liter- acy and knowledge of technological innovations. However, although this hypothesis did not have sta- tistical support, note that the means for knowledge of innovations were in the expected direction (see Table 26). ii. Results From One Way ANOVA for Adoption (ADOP) of Innovations Among Literacy Levels (N=50) 156 Table 28 presents the means of adoption of technolog- ical innovations for the number of cases in each of the three literacy levels. Table 28. The Mean of Adoption of Technological Innovations for Spanish-Speaking Group N Cases Mean Level I 9 10.22 Level II 33 11.06 Level III 8 11.63 Grand Mean = 11.00 The summary statistics from the analysis of variance are shown in Table 29. Table 29. One Way ANOVA Summary Table for Adoption of Technological Innovations Among the Literacy Levels in the Native Spanish-Speaking Group SOURCES OF VARIATION SS df MS F Main Effects LITLEV 8.691 2 4.345 .731 Residual 279.309 41 5.943 Total 288.000 49 5.878 Multiple R2 = .030; Multiple R = .174; ETA = .17 The F test was not significant (F = .731; df = 2 & 47; g < .05). This implies that the means of adoption of techno- logical innovations presented in Table 28 do not differ 157 significantly. That is, individuals with the three differ- ent levels of literacy do not differ significantly in their adoption of technological innovations. The extremely low R2 (.030) further indicates that there is no significant difference among the means of adoption of technological in- novations among the three literacy levels. It was therefore concluded that there is no significant relationship between level of literacy and adoption of technological innovations. Thus, the data did not support the theoretic Hypothesis 12b which predicted a direct relationship between level of literacy and adoption of technological innovations. However, although this hypothesis did not have sta- tistical support, note that the means for adoption of tech- nological innovations relatively tend toward the expected direction (see Table 28). CHAPTER IV SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS This study's research problem, rationale, and related theoretic frameworks and hypotheses were specified in the first chapter of this dissertation. The methods of approach- ing the study's research problem were described in the second chapter, and the results from the statistical analyses were reported and their respective interpretations were presented in the third chapter. The purpose of the current chapter is to summarize the results and to suggest some areas for practical and further research considerations. In this respect, the chapter is organized along three general parts, viz., (a) summary, (b) conclusions, and (c) recommendations. Summarily discussions for each of these three parts follow. A. SUMMARY ‘ This study's basic purpose was to explore the rela- tive potential contributions of quantitative and non-quanti- tative (English) literacy in predicting and explaining knowledge and adoption of technological innovations. 158 159 The study's data derived from two sample spaces, viz., Nigeria — Ilewo, and the USA - Michigan. The latter sample space was subdivided into the native English-speaking group and native Spanish-speaking group for the purposes of analysis. The hypotheses tested and the results of those tests are summarized in Table 30. B. CONCLUSIONS One major conclusion to be drawn from this study is that there are generally strong direct intercorrelations among the measures of literacy and last grade of school completed. Theoretically, this implies that in predicting and explaining variations in knowledge and adoption of techno- logical innovations, any of those measures could generally be sufficient. The results from the stepwise multiple re- gression analyses tended to support that conclusion. When sets of measures of literacy and last grade of school com- pleted appeared together in the regression models, the meas- ure which fitted first into the predictive equation extract- ed most of the variance explained such that each of the remaining measures in the set explained only minimal addi- tional increments of variance on the criterion. Ideally, one would expect the regressor fitted first into the predictive equation to yield a statistically 160 .0000me Imam mflfion 0H0 80m :0 350088 15820009: 0.8 8&3 a .52 0mg 03089m . 00008.5, 0050308 05 mo #00: 5H8 mmg .00: 06 >08. 380385 03 . wage 48> 080088 5388 3 Inflow >05 . 8388 0fi fi 0.83 90000.58 .30 00:3 80:08.85. 38 05 no 000.3995 :0 >300: 050.308 g 83 . u .d mdficofifiamu 30H Ems $03.6 3mm . u 8 03.... 1:388.» 33 3000.809 omuhonwmdm 3mm . u .3 mgoflgmu mcouuM >8> Umuhonwwsm 8mm . u .d mfificoflflmn 0:800 30> wmuuonwmsm 028305.85 8030393000 mo Ammgv 000.20% .0 08 . 130sz 8380 fi 86033 .0 . Emumofl 0008qu 930: 5 80000.3 .6. . Age 03298 H8218 «0 88.8 0.6.3 .6 "0000305. ugflmawm 8.8.308 H33 605 mo 0003.90 0300 Ho 9.80meme 50m @5303qu 9.3 . msoflgocfi M 8383808 no 88$ 83086 90 08.36680 5 u m age flag fi 80033 .o 020 . Aimee 0085a 930: fi 8333 .n . Energy 033850 Hoofim mo 0068 0.6.3 .0 .8888, 0800888 2388 :5 890 no 003.6 0500 Ho mHowmemmu wwnfi 05330.0. 05 . 0839,88.” 80.339500» .05 Ammgv mg 0:» @586an CH u m 0000805 mcofl0>oc§ 8330:5000 mo 385 8398a 6 08 .8000005 mcoflggfl 833850.... mo A9235 000.903“ .0 “880005 $0028 .330 5 88.83 .6. “0000805 33985 mm0smc0a 02.30: 5 30.803 .0 "mo H93 05 H 5060088 Age 80308 Hoofim 00 008m 003 um u m gmmamzflmélgg .H g EBBM .82 3006008 g 800% 035m 58002 .. «B 0:» 0:0 0.03 . 3.832 0.0 5 mmumfiommm 030085. 06. manomgm no: 0:6 980808 $380 .45 gm .8 038. 161 . 2.050200 05005. 5 00000000 .0 03. u 0 000000000 0:0 m . 28:65 00000900 08:00 00 00000 005 .0 "00:00.55 0:00 55:00.0 5598 :03 00000000 00 03:00:00 I005 0000000000 900 90.05300 0:0 .00 £000 . EH49 50.000: 030000000005 95000000: 5 "mm 000000 05 0:800:00 00000090 0000005 200308 .0 000000 56 0500000000 00000090 0:0 .00000000.." 0858. .m 0:00 500.0000 00: 000000 00 00000090 002 N0000505 0:06:00 .0 09.550 05 0500000500 00009550 “00000005 BHMmfio .H "0 00 8.000 0809.00 0 90.000000 .00 55088.0: 05. .0 33. n my 0:005 0:0 00000005 0800025 00:50 00: 3005050000 00000090 002 000500080000 00 88$ 0500000 00 00000 .0 flow. n .d 0.30 “00000005” 0:300:65.“ 18500000 35 30000002 00000950 0000590000 00 $0200 00000305. 00 0050: .0 Ammm. u 5 “00000005 @280000000 98900 >00> 000009900 3.39 06080: 0300000530 00 00:03 .0 800. u 3 “00000005 0200500000 0:800 08> 00000090 $002.00 00505 5 0000005 00 0000.0 .0 "00000005 0 2000030 00000900 000.00 00 00000 0005 m: u 0 900m $00035 03002 .0 mummm a; 268E I £8 .HH mg E835 902 g \QHHMOmmDm “0.00.83 on 0.309 38. u .3 9:8,." “mmmmmuofi 903mg .353 pa: Eioflmflfim Eugm uoz 3033868 mo Eozmv Eomag no .253 .m 35m. u 5 9.5m «mmmmwnufiu [8338 23: 33mg Eyegm $39 >083: “53$?qu mo H93 6 :8. u .3 Em .8338 :3: 33382 @8qume 5mg": 3.55 @8333 no H93 .0 8mm. n 5 92m “mmmmmnofi .8338 magnum 33882 nmfiomgm SEE €295 S... @833 no H98 .0. 3mm. u .5 mflm «mmmmwgfl Icoflflwu 33 33mg cmflomgm gammy swam fi 6832 mo H93 .m "mmmwmuofi Aggy @3398 m o mmfim umfl mm "mm vaflmnfiémmm gfimz .m .mcoflgefi 2 .ufiouémfim mflfion HSHEHEEB «0 ABE mmuflg .c m 35 mum. no: €832 Em . 3.39 3&3: Qflflflfiso .o 63 aflcaccmamnfi fl 802 ERQm uoz “Aaafiv fifimfi fi momumufl .n XHZmSV @3398 Hoofim mo 895 umS .m "moan—wk, unmoamdhmwm Gaga Ixm H33 awn» mo uwmnsm 96m Ho muommmnmwu noon 9:528 m5 . 283993 1.6328 ufimu mo 88s 8% m5 5068a 5 "pm .ufioflflmflm 30.5% 88 . Eng @883 933330 .0 mmufi Sm Samoflfimflm Em . A5623 flag 5 monumufl éw aflcmnammofi «A 950 cfiufigm Igv @3960 Hoofim mo mofim umS .m "gang unmowmgwm 5.3m L8 .2”? .55 mo ummnsm 9.0m Ho 38% 8% @3538. m5 .maflggfi flange .633 .563 figgmfiuflvflmfi “ m g g .82 mag \QEMBBm TQLGOOV om manna dog—£50 gfloflvmfim 05 35 cuppa Hommwumwu umufl 93 was 953 can...» 4 996 um panama? mmB ummu .m m5 £5 .m mmum um unmoflflnmfim #0: mg ummpuu man. k. .ucwo A353 mflfioh 88 mi .339 E33 mpflflflqgo .m 3.. Samoflaflm >33ch can. . flame @flfiflflm 6 :85 mums 95% a 8% . €328 fiflmfi 5 moflmufl .o «.953 .8333 mfi 5 . gang swam 5 6&33 .0. 8m; Eammflmmu Sm :23 cBuanm . Agv 32950 888 mo 83m 33 .m "85% ufioudaflm fiflm Ixm .23 52.3 no ”539m 95m Ho muommmnmwn 9a 3.5038 m5 . 803995 H8838 .68 «0 A53 @355. as» 33063“ 5 "Sm $83353 mums 3:28 .88 in 3 353%? wig? . Efle Emuflflflm 6 H mm; «.88 503% mfi 5 can . $5sz 53% fi @333 .o 983 muommmuowu Adm 533 wflhommbm . figmmv 2mg 5 momHmuH .n . Agv @339an 80:8 mo moflm umfl .m "855 unmoflflcmwm 5398 H33 Emfi m0 #396 9.0m Ho wHOmmmnmwn “Sou mgaom map 3.39 bfififl 93335 830% 5 ”am .833 dB ufioflflmfim vfiuoamsm .8mmmuufi EoBoB é unmoflflamflm no: 989mm 8 muggm #02 new .mmmmmuofi mng .m unmoflmwdmunm won 333 .EU Engm uoz Emma; Egg .N unmoflflcmflm U.8 883 do cmfiomgm uoz “83805 9555a 4 ”mo EHOM x3960 m mafiaoam mo >33on m5. .0 33. u .5 £83 95 .8388“ 983.“.593 £353 U.oc madmoflmflfim Eugm 3033333 no Agv 8382 no H93 .m E gonad .82 anon»: \nflmommpm TFUGOOV om manna . :ofloflé @3898 5 mam mummy: m5. Enogm uoz .mmmmmuocw 9.03% 3033858 no A83 8382 .n 4 . £03088 can . 88%..” magma—EVE samuognmfimummfimumfi nmflongm uoz 3033863 no £0sz mmnflgé l . .mo H93 Q5 ma .8m8uofi Sag $33 no H92 93 mm " m .moqmg .msoflgafi pamofldaflm fiflm 333953 mo ABEV £83392 2% Ixm 3508p mHOmmeme 08m . 3.39 333 93335.5 .m 5m 93 00 Ho: #58me . ABHHHE monumuflwm .0 Imam baggfiufi fl 802 oBuogm uoz . amaze fiflmfi 5 @833 .o . 32mg 3mg fi REES .n . Aggy 0339.8 Hoonom mo 8.3m 993 .m "woe—mg van-baa? finamxm 3.3 :03» m0 #886 3.0m Ho gamma :9" 5m 053038 93 .mgflgoqfi 30.33. .9303 no 88s 8335 mfi 330%?“ 5 u: 9% 588m .82 g \nflmomgm A6 .ucoov cm 0.35.. 165 significant relation to the criterion since the regressor so fitted extracts most of the explained variance. The data indicated that this was so for some but not for all the models. In those models where there was no statistically significant relationship between the regressor fitted first into the predictive equation and the criterion, this implies that the amount of variance explained by the regressor was too small. When there is the lack of significance on the regres- sors fitted first into the predictive equation, it may be due to specification errors in formulating the appropriate regression equation. Such errors may, for example, be due to the omission of a relevant regressor and/or incorrect use of the functional form of the regression equation. If the omitted regressor is correlated with the regressors specified for the model, the estimators of their partial re- gression coefficients will be biased and inconsistent. Thus, the t or F test of significance will not be valid, since such a test will tend to accept the null hypothesis more frequently than is justified by the given level of signifi- cance. The specification error may also have occurred in this study due to incorrect functional form of the character- izing equation. The models tested in this study assumed linear relationships among the variables in the equation. However, although this is a basic assumption in regression analysis, nonetheless it must be noted that the relationships 166 in the models may, following Kmenta (1971, pp. 391-405), be nonlinear with respect to the variables but linear with re- spect to the parameters estimated (i.e., intrinsically linear models), or may be nonlinear with respect to both the variables and the parameters (i.e., intrinsically non- linear models). Such nonlinear or curvilinear relationships may be some form of power functions of polynomials. Thus, where the functional relationship is nonlinear, the assump- tion of linearity is expected to limit the real explanatory power of the regressors. This limitation results in the variance explained by the regressors being statistically not significant. The data indicated two models in which a regressor fitted first showed significant variance explained on the criterion at the first step but not on the last step in which all the regressors were forced into the predictive equation under the default values of the restriction para- meters. This involved biliteracy in the models predicting quantitative literacy and knowledge of technological innova- tions in the native Spanish-speaking group. Apparently, as more regressors were entered in the equation, there was a partitioning of variance away from biliteracy in the rela- tively small pool of variance being explained by all the regressors. This partitioning of variance is to be expected since the regressors were generally intercorrelated strongly. 167 This partitioning of variance may also explain why some of the regressors fitted after the first step of the stepwise regression analyses were individually not signif- icant. When most of the explained variance is extracted by the regressor fitted first into the predictive equation, there was only minimal additional increments to the explain- ed variance by the regressors fitted into the equation at the subsequent steps of the regression. In the Nigeria - Ilewo sample space, last grade of school completed, literacy in English and literacy in native language (Yoruba) were not individually significant although jointly they were significant in explaining variance in the awareness of tech- nological innovations. Individually, awareness of techno- logical innovations, last grade of school completed, and literacy in English are significant but literacy in native language (Yoruba) is not significant in explaining variance in adOption of technological innovations, while jointly, the four regressors are significant. In the native English-speaking group, literacy in English and last grade of school completed are both individ- ually and jointly significant in predicting quantitative literacy. Individually in the stepwise regression, quanti- tative literacy explains significant variance in the knowl- edge of technological innovations but last grade of school completed and literacy in English do not, but jointly the three do. Knowledge of innovations, last grade of school 168 completed, quantitative literacy, and literacy in English are neither individually nor jointly significant in explain- ing variance in the adoption of technological innovations. In the native Spanish-speaking group, biliteracy, last grade of school completed, literacy in English, and literacy in native language (Spanish) are individually not significant when they are all present in the equation at the last step of the analysis in which quantitative literacy is predicted, but the four regressors are jointly significant. Individually, biliteracy, last grade of school completed, and literacy in English are significant but literacy in Spanish and quantitative literacy are not significant in pre- dicting the knowledge of technological innovations. However, jointly, the five regressors are significant. In predicting the adoption of technological innovations, last grade of school, knowledge of innovations, biliteracy, literacy in Spanish, quantitative literacy, and literacy in English are neither individually nor jointly significant in predicting the adoption of technological innovations. One possible explanation for the lack of individual and joint significance among the regressors predicting the adoption of technological innovations in both the native English-Speaking and native Spanish-speaking groups may be due to specification errors of the types stated earlier. However, one other possible explanation may be due to lack of relevance of the innovations to some of the respondents 169 in the samples. The native English-speaking sample, for instance, had nearly a half (i.e., 66 of 169) of the cases who were college students who may have lived in school resi- dential halls. Such innovations as Tel-Med, Project Health, Expanded Nutrition Program, Michigan Law for winter heating bills and Meal Planning Guides may not be apprOpriate for students living in school dormitories where these services are usually provided. A closer examination of the data from stepwise regres- sion analyses revealed the following observations. First, when literacy in English and in non-English native language (i.e., Yoruba and Spanish), are among the regressors in the models predicting the awareness/knowledge and adoption of innovations, and quantitative literacy, in 4 out of 5 models literacy in English is a better predictor than is literacy in non-English native language. It is therefore to be con- cluded that literacy in English is generally a stronger pre- dictor of awareness/knowledge and adoption of technological innovations and quantitative literacy than is literacy in non-English native language. Secondly, when quantitative literacy and literacy in English are among the regressors predicting knowledge and adoption of technological innova- tions, in 3 out of 4 models, quantitative literacy is a better predictor than is literacy in English. It is there- fore to be concluded that quantitative literacy is generally a stronger predictor of knowledge and adoption of 170 technological innovations than is literacy in English. Finally, a closer examination of the data from stepwise re- gression analyses in the native Spanish-speaking group re- vealed that in 2 out of 3 models in which biliteracy, literacy in Spanish, and literacy in English are among the regressors for quantitative literacy, knowledge and adoption of technological innovations, biliteracy is a stronger pre- dictor than either literacy in Spanish or literacy in English. It is therefore to be concluded that biliteracy is generally a stronger predictor of quantitative literacy, knowledge and adoption of technological innovations than either literacy in Spanish (native language) or literacy in English. On the tests for the relationship between last grade of school completed and the prOpensity of adopting complex technological innovations, the data indicated that 4 out of 8 situations involving the propensity of using a simple or complex technological innovations, there is a direct rela- tionship between last grade of school completed and the pro- pensity to use a complex technological innovation. Since, this is a 50-50 situation, no conclusion could be made con- fidently. However, it is to be noted that 3 of the 4 posi- tive relationships stated above were observed in the native English-speaking group whose last grade of school completed ranged from 7 to 16 years, while the fourth was observed in the native Spanish-speaking group whose last grade of school completed ranged from 0 to 12 years with about 2/3 of them 171 171 having less than 7 grades of school completed. One may therefore conclude that last grade of school completed is generally related positively to the propensity of using com- plex technological innovations. Further work in this area would be helpful to see if the hypothesized relationship can be more clearly supported. Recall that complex technological innovations were by definition those which required precise measurements in that they required use of numbers on the part of the potential adopters of these innovations. This conclusion therefore implies that the propensity of using technological innova- tions with precise measurements generally increases as last grade of school completed increases. Finally, on the tests of relationship between the three constructed levels of literacy and knowledge and adOp- tion of technological innovations, the data indicated no significant relationship by the ANOVA. That is, there is no significant direct relationship between these levels of literacy-and knowledge or adoption of technological innova- tions. One possible explanation for these findings may be due to the small number of cases on some of the literacy levels. Such small number of cases may not have provided a strong base for adequate discrimination in variance. Although the differences among the means of knowledge and adoption of technological innovations were not large enough for statistical significance, note that the means 172 were in the expected direction of the levels of literacy. That is, the knowledge and adoption of technological innova- tions increase slightly as the level of literacy increases. C . RECOMMENDATIONS This study has some considerations for change agents and for future research. The following lines of thought are particularly recommended for consideration. 1. Recommendations for Change Agents First, since the relationship between quantitative literacy and literacy in native language (Spanish) is moder- ately low, and the relations of quantitative literacy to literacy in English, biliteracy, and last grade of school completed are very strong, change agents are recommended to begin (or continue) English literacy programs in their liter- acy education campaigns. It is expected that this would facilitate the acquisition of quantitative literacy in those areas where English is not the native language or the common medium of discourse. Finally, since the propensity of using complex tech- nological innovations generally increases with the level of formal education, it implies that the change agents' knowl- edge of the level of formal education among their clients and knowledge of the level of complexity of the technological innovations to be diffused among the clients is generally useful. Such knowledge will facilitate the packaging and 173 targeting of messages concerning the technological innova- tions. This recommendation implies judicious practice of audience and content analysis by the change agents. 2. Recommendations for Future Research Probably the most immediate need is to investigate the status of specification errors in the models tested in the current study. Such investigation would, for instance, consider including into the models new variables as well as testing for the functional form of the characterizing equa- tions as noted earlier. This may provide the variables and their functional forms which account for greater amount of variance on the knowledge and adoption of technological innovations thus identifying optimal conditions for diffusion of technolog- ical innovations. The most interesting variables for consideration in future models may include, for instance, leadership style and level of income. It is here speculated that leadership style and level of income are strongly related particularly to the adoption of technological innovations. Radical or counter culture groups and/or authoritarian clan leaders and/or village or community chiefs or opinion leaders as well as level of income may facilitate or impede the adop- tion of technological innovations beyond that explained by literacy or formal education. This study did not control for these variables. 174 As for the functional form of the characterizing equations, the current study did not test for the non-linear relationships. Hence, in the absence of information on the alternative forms of functional relationships, the linear equations tested in this study may lack parsimony for de- scribing the study's data sets. What is still needed there- fore is to investigate for the alternative non-linear rela- tionships, and where these relationships are found, to linearize the equations before performing the appropriate regression analyses. Techniques are available for testing for non-linearity and for linearizing the equations (see, for example, Kmenta, 1971, pp. 451-472; Kelejian and Oates, 1974, pp. 92-103 and pp. 167-175; Kerlinger and Pedhazur, 1973, pp. 208-218; Namboodiri, Carter and Blalock, 1975, pp. 150-156, pp. 173-174, pp. 186-187, p. 194 and pp. 600-605; Harris, 1975, pp. 233-236, and Kim and Kohout, 1975, pp. 368-373). The current study measured literacy in quantitative symbols and concepts only among native English-speaking and native Spanish-speaking populations. The latter population, however, speaks a language which has a common linguistic root with the English language. It is not yet known to what extent literacy in quantitative symbols and concepts is im- portant in predicting and explaining knowledge and adoption of technological innovations in those populations whose native languages or dialects do not have a common linguistic 175 root with English or any one European language. A study is needed to investigate this aspect of literacy. Such a study would also consider if there are specific quantity descrip- tors used by these populations for communicating quantitative or number concepts, and then build these into the instrument to achieve more reliable and valid observations. A study along this line may have considerable application not only for change agents but also for international agencies, organ- izations and multinational corporations which frequently dif- fuse highly quantified technological innovations with precise formulations and rates of application. What also seems needed is further refinement of the quantitative literacy test (QLT) instrument which the current study developed. Some of the items used to measure some of the dimensions of quantitative literacy in the current study used diagrams (see Appendix B, pp 194-199). Some alternatives could be to use objects instead of diagrams or to use both diagrams and objects as manipulations for the same dimensions of quantitative literacy. It may be the case that individu- als with different levels of literacy differ in their ability to manipulate given quantitative concepts with objects rather than with diagrams. A study is needed to investigate these alternative operationalizations. In addition, further refinement of the current QLT instrument could be achieved through validation. This could be done by administering the instrument to several samples 176 taken from different populations, and/or testing for its correlations with known quantitative or reasoning tests such as mathematical assessment tests, abstract reasoning, and so forth. Finally, further refinement of the current QLT instrument could be attained by testing it over-time. Four reliability coefficients on the QLT items were noted in the second chapter (sectiona A. 2. a. iii). The coeffi- cients generally indicated relatively high consistency among the scale items in their measure of the dimensions of quan- titative literacy. However, to determine more appropriately the stability of these items essentially calls for over-time studies which allow us to have a more careful estimate of the reliability of the instrument (see Heise and Bohrnstedt, 1970; Werts, Joreskog and Linn, 1973; Wiley and Wiley, 1970, 1974). Finally, while the use of the SORT instrument in this study was a move toward uniformity of measuring non-quanti- tative literacy, the study did not consider the respondents' comprehension of what they read. Hence, a measure is still needed which involves more interpretation of what is read as that may influence decision making about technological innovations. 177 FOOTNOTES 1This study uses Solo and Rogers' (1972) conception of De- veloping Countries or Less Developed Countries (LDCs) as ". . . those with relatively lower levels of per capita income, literacy and education, production, etc." (p. 87). Solo and Rogers (1972) use the United Nations' arbitrary classification of less developed nations as including all those of Latin America, Africa, and Asia, with the excep- tion of Japan, South Africa, Australia, and New Zealand. According to Solo and Rogers (1972, p. 87), "development refers to the type of change that produces higher per capita incomes and levels of living through more modern production methods and improved social organization" (see also Rogers with Svenning, 1969, pp. 8-9). 2In his extensive literature search, the author contacted, among others, Professor Everett M. Rogers at Stanford Uni- versity, who has worked extensively on literacy particular— ly in Developing Countries. No diffusion study could be located in which quantitative and English literacy have been studied; nor could one piece of research be located which has been done on diffusion in the most recent years. 3Following Munroe (1963), a careful distinction is to be made between the symbols "number" and "numeral." "A number is an abstract idea. A numeral is a mark or set of marks used to denote such an idea. That is, a numeral is the name of a number" (p. 11). 4Reports and expressions of scholars and leaders throughout the world on the significance of literacy on modernization and development may be gleaned from several of the UNESCO literacy newsletters; see for example, Literacy: A News- letter, December 1969; 1970, No. 1; April 1970, No. 2; October 1970, No. 4; 1971, Third and Fourth Quarter, and 1972, First Quarter; UNESCO, 1963; UNESCO, 1973. 5The modernization and development variables are discussed in the section on the correlates of literacy, and for their critiques, see footnote No. 6. 6Herzog (1967) has critiqued in his dissertation the litera- ture on the five indices of modernization. According to Herzog (1967): Em ath is regarded as an ability to ident- ify with and symbolically participate in a new unfamiliar role. Achievement motivations is the desire to succeed, apart need orie: 3959 from 399E abet func 7Thes pra cen cat 10 A H rn ll 12 WHBr-fs‘b 178 apart from social pressure, in order to gratify a personal need to do so. Cosmopoliteness is defined as a positive orientation toward an urban mode of life. Mass media ex- posure is defined as being in the audience for messages from newspapers, radio, television and cinema. Political knowledge is defined as possession of basic information about one's region and country, so as to enable one to function as a citizen. 7These categories of literacy definitions are not necessari- ly mutually exclusive for it is conceivable that in research practices an investigator could use both the planning- census-type and empirical-type measures. Hence, these two categories are used here as general categories for purposes of explicating the problems with prior conceptualizations of literacy. 8According to Rogers and Shoemaker (1971, p. 137), there are five attributes of innovations: a. Relative advantage: The degree to which an innovation is perceived as being better than the idea it supersedes. b. Compatibility: The degree to which an innovation is perceived as consistent with the existing values, past experiences, and needs of the receivers. c. Complexity: The degree to which an innovation is per- ceived as relatively difficult to understand and use. d. Trialability: The degree to which an innovation may be experimented with on a limited basis, and e. Observability: The degree to which the results of an innovation are visible to others. 9The focus here is on English, although the scientific sym- bols and concepts may appear in the languages of other industrialized nations. loSkinner's (1957) functional units of language include: mands, tact, echoic, textual, intraverbal, and autoclitic verbal behaviors (see Skinner, 1957 for details). 11Details of transformation rules may be gleaned in Chomsky (1957, 1965). A detailed critique of transformation rules and generative grammar is presented in Lyons (1974). 12Although the Nigeria - Ilewo survey operationalized what was conceived as knowledge of technological innovations, this author found the measures rather inadequate. The measures were of awareness rather than knowledge of techno- logical innovations. Consequently, he reconceptualized knowledge for awareness which is only one aspect of 13 14 15 16 17 18 179 knowledge. This author will provide a more adequate operationalization of knowledge in the next section (2. b. i) for the USA - Michigan sample space. This is based on the assumption that the native language or dialect uses the same alphabet as the English language. However, where this is not the case, it is conceivable that an individual may be literate in English without being literate in his/her native language or dialect, e.g., Arabic, Hindi, Persian, etc. Note that quantitative literacy is treated as a dependent variable in first regression model in each of the two groups in this sample space. This is only in general because an exception may be noted. Selltiz, et a1., (1969), for instance, have noted that "When the estimate of reliability consists of split-half equivalence coefficient, low reliability does not necessar- ily detract from validity; paradoxically, it may even in- crease validity. In order for split-half equivalence to be high, all items of the test must be highly correlated; that is, they must all provide a measure of essentially the same characteristic or of characteristics that vary together. To use the technical term, they must be homogen- eous. But for some purposes, a test that taps a number of different characteristics may be more valid than one that measures a single characteristic" (p. 178). The criteria for describing the magnitude of the reliabil- ity and correlation coefficients was arbitrarily set at .70 or greater for very high or very strong; .50-.69 as high or strong; .35-.49 as moderately high; .25-.34 as moderately low; and .24 or less as very weak or very low. The results form factor analysis of the QLT items can be obtained from the author on request. Note that with the exception of the Nigeria - Ilewo sample space and the Spanish-speaking group, English is to be conceived as native language in the USA - Michigan sample space. APPENDIX A Literacy Test Adapted from SLOSSON ORAL READING TEST (SORT) APPENDIX A Literacy Test READING IEVEL Plaxment NANE DATE last First Mbifle DATE OF BIRIH EXAMINER List A (20) List B (40) List C (60) List D (80) List E (100) 1. see 1. with 1. game 1. safe 1. harness 2. lock 2. friends 2. hide 2. against 2. price 3 . mother 3 . came 3 . grass 3 . smash 3 . flakes 4. little 4. horse 4. across 4. reward 4. silence 5. here 5. ride 5. around 5. evening 5. develop 6 . can 6 . under 6 . breakfast 6 . stream 6 . promptly 7. ‘want 7. was 7. field 7. empty 7. serious 8. come 8. what 8. large 8. stone 8. courage 9. one 9. bump 9. better 9. grove 9. fbrehead 10. baby 10. live 10. suddenly 10. desire 10. distant 11. three 11. very ll. happen 11. ocean ll. anger 12. run 12. puppy 12. farmer 12. bendh 12. vacant 13. jump 13. dark 13. river 13. damp l3. appearance 14. down 14. first 14. 1unCh l4. timid l4. speeChless 15 . is 15 . wish 15 . sheep 15 . perform 15 . region 16. up 16. basket 16. hope 16. destroy l6. slumber 17. :make 17. fbod l7. fOrest 17. delicious 17. future 18. ball 18. road 18. start 18. hunger 18. claimed 19. help 19. hill 19. heavy l9. excuse 19. common 20 . play 20 . along 20 . station 20 . conplete 20 . dainty 180 181 List F (120) List G (140) List H (160) List I (180) l. cushion 1. installed 1. administer l. prairies 2. generally 2. importance 2. tremor 2. evident 3. extended 3. medicine 3. environment 3. nucleus 4. custom 4. rebellion 4. counterfeit 4. antique 5. tailor 5. infected 5. crisis 5. twilight 6 . haze 6 . responsible 6 . industrious 6 . memorandum 7. gracious 7. liquid 7. approximate 7. whimsical 8. dignity 8. tremendous 8. society 8. proportional 9. terrace 9. customary 9. architecture 9. intangible 10. applause 10. malicious 10. malignant 10. formulated ll. jungle ll. spectacular ll. pensive ll. articulare 12. fragrant 12. inventory 12. standardize 12. deprecate l3. interfere 13. yearning 13. exhausted l3. remarkably l4. marriage 14. imaginary 14. reminiscence 14. contrasting 15. profitable 15. consequently 15. intricate 15. irrelevance 16. define l6. excellence l6. contemporary l6. supplement 17. obedient 17. dngeon l7. attentively l7. inducement 18. ambition l8. detained' 18. compassionate 18. nonchalant l9. presence 19. abrmdant l9 . crxnplexion 19. exuberant 20. merchant 20. compliments 20. continuously 20. grotesque List J (200) l. traverse 2. affable 3. compressible 4. excruciating 5. pandemoniun 6. scrupulous 7. primordial 8. chastisement 9. sojourn 10. panorama ll. facsimile 12. auspicious 13. contraband l4. envisage 15. futility 16. enarmoured l7. gustatory 18. decipher l9. inadequacy 20. simultaneous I VFWfib APPENDIX B Test for Knowledge, Adoption, and PrOpensity of AdOption of Technological Innovations and for Quantitative Literacy in the USA - Michigan Sample Space APPENDIX B Test for Knowledge, Adoption, and Propensity of Adoption of Technoloqical Innovations and for Quantitative Literacy in the USA - Michigan Sample Space MICHIGAN STATE UNIVERSITY Department of Communication, and Department of Resource Development East Lansing, Michigan 48824 DEAR PARTICIPANT: We are currently engaged in a research project to determine the use of resources in various communities. It is our cpinion that adult persons are in a good position to provide accurate information on the extent to which resources are being used in a community. Therefore, we are asking you to provide answers to the questions asked in this booklet. This information is needed for planning effective communication programs for community development and for improving teaching methods. Your co- operation therefore will greatly contribute to the success of this project. We need answers which ygu yourself can provide. For this reason, you are requested to provide your own ideas on the basic questions on the use of resources as described in this booklet. Your answers to these questions are confidential, and will be used for planning and teaching purposes only. Please answer all the questions. Your cooperation is most appreciated. Thank you. 182 183 PLEASE RESPOND TO THE FOLLOWING QUESTIONS AS FAST AS POSSI- BLE. BE ASSURED THAT THE ANSWERS TO THESE QUESTIONS MAY DIFFER FROM PERSON TO PERSON. HOWEVER, WE ARE INTERESTED IN ALL OF THESE ANSWERS. SO PLEASE ANSWER ALL THE QUESTIONS. CHECK OR STATE AN ANSWER WHICH YOU THINK BEST EXPRESSES YOUR FEELINGS. 1. Have you ever heard of Consumer Reports? a. Yes b. No 2. If so, where did you hear about them, or who told you about consumer reports? 3. What does a Consumer Report mean to you? 4. If you needed information about Consumer Reports, where would you go for such information? 5. Under which of the following conditions would you con- sider using a Consumer Report?: a. When I need information about a recipe. b. When I want to buy a new car. c. When I need information on a health problem. 6. How often do you use Consumer Reports?: a. Most of the time. b. Sometimes. c. Never. 7. Have you ever heard of a Dollar Watch program? a. Yes b. No 8. If so, where did you hear about it, or who told you about a Dollar Watch program? 9. What does Dollar Watch mean to you? 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 184 If you needed some information about Dollar Watch, where would you go for such information? Which of the following statements is true in budgeting money?: a. A budget is concerned with income only. b. A budget is concerned with expenditure only. c. A budget is concerned with both income and expenditure. How often do you budget your money? (check one only): a. Most of the time. b. Sometimes. c. Never. Have you ever heard of a Tel-Med program?: a. Yes b. No If so, where did you hear about it, or who told you about Tel-Med? What does Tel-Med mean to you? If you needed some information about Tel-Med, where would you go for such information? Under which of the following conditions would you con— sider using Tel-Med services? (check one only): a. If I need to find out someone's telephone number. b. If I need some information on a health problem. c. If I need some information on income tax returns. How often do you use Tel-Med services? (check one only): a. Most of the time. b. Sometimes. c. Never Have you ever heard of Consumer Credit in your community?: a. Yes b. No 20. 21. 22. 23. 24. 25. 26. 27. 28. 185 If so, where did you hear about it, or who told you about Consumer Credit? What does Consumer Credit mean to you? If you needed some information on Consumer Credit, where would you go for such information? Which of the following credit sources provides more reliable information for comparing Consumer Credit costs? (check one only): a. A credit supplier who states finance charges as an Annual Percentage Rate on the balance of the loan. b. A credit supplier who states finance charges based on your having the use of the loan for the full year though the balance of the loan gets smaller each month. How often do you use Consumer Credit? (check one only): a. Most of the time. b. Sometimes c. Never. Have you ever heard of Expanded Nutrition Program? a. Yes b. No If so, where did you hear about it, or who told you about Expanded Nutrition Program? fir What does Expanded Nutrition Program mean to you? If you needed some information about Expanded Nutrition Program, where would you go for such information? 29. 30. 186 Under which of the following conditions would you consider using Expanded Nutrition Program? (check one only): _ a. When I need information on meal planning. b. When I am over twenty years of age. c. When I need information on Consumer Credit. How often do you use the services of Expanded Nutrition Program? (check one only): a. Most of the time. b. Sometimes. c. Never. --PLEASE CONTINUE-- 187 PLEASE RESPOND TO THE FOLLOWING QUESTIONS AS FAST AS POSSI- BLE. BE ASSURED THAT THE ANSWERS TO THESE QUESTIONS MAY DIFFER FROM PERSON TO PERSON. HOWEVER, WE ARE INTERESTED IN ALL OF THESE ANSWERS. SO PLEASE ANSWER ALL THE QUESTIONS. CHECK OR STATE AN ANSWER WHICH YOU THINK BEST EXPRESSES YOUR FEELINGS. 1. Have you ever heard of a condition called diabetes? a. Yes b. No 2. If so, where did you hear about it, or who told you about diabetes? 3. What does diabetes mean to you? 4. If you needed some information about diabetes, where would you go for such information? 5. Under which of the following conditions would you have your blood tested for diabetes? (check one only): a. If you are getting overweight. b. If you take too much sugar. c. None of the above. 6. Have you ever had a blood test for diabetes? a. Yes b. No 7. Suppose you were considering to buy one of the two kits described below for testing the presence of sugar in your urine. And suppose the price of these kits is the same; which of these two kits would you buy if you need- ed one of them? (check one only): a. One kit contains strips of cellulose and a color chart. To test for sugar in urine, you moisten a strip of cellulose with urine and then compare the dipped end of the cel- lulose with the color chart. The greater the amount of sugar in urine, the deeper will be the color of the cellulose strip. b. Another kit contains small glass tubes, a color chart, and tablets of a chemical re- agent. You mix 3 drops of urine with 10 drops of water in a small tube. Then, you 10. ll. 12. 13. 188 add one tablet of the reagent to this mixture. Wait 15 minutes. Then, compare the color of the mixture with the color chart. The greater the amount of the sugar in urine, the deeper will be the color of the mixture in the tube, and you can estimate the exact quantity of sugar in urine. Suppose you wanted to buy one of the two drugs describ- ed below, and suppose the two drugs cost the same amount of money. Which of these two drugs would you buy if you needed one of them? (check one only): a. One type of drug comes in pills which you can take by mouth with some water. You are required to take two pills three times a day. b. The same drug may also be taken with a new type of injection which does not cause any pa'n. This drug comes in a small kit which contains a syringe, different drugs to be mixed together, and instructions on how to use the drug. You must mix the right amount of each drug in the kit and then inject the mixture into a particular part of your body by yourself. Have you ever heard about CPR (Cardiopulmonary Re- suscitation)? a. Yes b. No If so, where did you hear about it, or who told you about CPR? What does CPR mean to you? If you needed some information about CPR, where would you go for such information? Under which of the following conditions would CPR be necessary? (check one only): a. When a person has too much sugar in his/her urine. ’ b. When there is unconscious victim of a heart arrest. c. When a pregnant woman loses her appetitie for food. 5‘ 14. 15. 16. 17. 18. 19. 20. 21. 189 Have you ever done CPR? a. Yes___ b. No___ Do you think you know enough to do CPR if someone needed it? a. Yes___ b. No___ There are many techniques of emergency life support for resuscitation of the unconscious victims who have only stopped breathing though the heart continues to beat. Which of the following two basic life support emergency techniques would you use if you were to at- tempt rescuing an unconscious victim? (check one): a. The unconscious victim is made to lie flat on his abdomen on a hard surface. The rescuer then presses the victim's back thus forcing the victim's abdomen against his diphragm, compressing the lungs and causing expiration. b. The unconscious victim is made to lie flat on his back on a hard surface. The rescuer then tilts the victim's head backward to open the victim's airway. Breathing may be restored by blowing hard into the victim's mouth. This is repeated every five seconds. The rescuer blows four quick lung inflations into the victim. Have you ever heard of Project Health? a. Yes b. No If so, where did you hear about it, or who told you about Project Health? If you needed some information about Project Health, where would you go for such information? What does Project Health mean to you? Who is eligible for Project Health services? (check one only): a. Every one over twenty-one years of age who is on a Medicaid Card. b. Every one under twenty-one years of age who is on a Medicaid Card. c. Every one who is on a Medicaid Card without regard for age. 22. 23. 24. 25. 26. 27. 28. 29. 30. 190 How often do you use the services of Project Health? (check one only): a. Most of the time. b. Sometimes. C. Never. Suppose your doctor prescribed for you a drug which needs to be kept out of direct sunlight or places where it could be very warm (above 80°F). Which of these places would be ideal for storing such a drug? (check one only): a. A small storage space in a cabinet above the stove. b. A small wide-mouth thermos. c. A car baggage compartment. Have you ever heard of a Michigan Law providing $38 million to help low income and elderly citizens pay their Winter heating bills? a. Yes b. No If so, where did you hear about it, or who told you about this law? If you needed some information about this law, where would you go for such information? What is the maximum amount of money which would be paid to persons eligible under this law? (check one only): a. $100. b. $200. c. $300. d. $400. How often have you made use of this law? (check one only): a. Most of the time. b. Sometimes. c. Never. Have you ever heard of the Meal Planning Guides? a. Yes b. No If so, where did you hear, or who told you about Meal Planning Guides? 11......‘3. ‘17 1 gr. =3 31. 32. 33. 34. 35. 191 If you needed some information about Meal Planning Guides, where would you go for such information? What does a Meal Planning Guide mean to you? Which of the following lists have a complete list of the basic food groups? (ckeck one only): a. Milk and milk products, meat, vegetables and fruits, breads and cereals. b. Milk and milk products, meat, beans, leafy vegetables, apples, oranges. c. Meat, vegetables and fruits, breads and cereals, beans, lemons, cabbage. d. Vegetables and fruits, breads and cereals, milk and milk products, oranges. How often do you plan your meals to include the basic foods in the list which you checked in #33? (check one only): a. Most of the time. b. Sometimes. c. Never. Suppose you were to buy one of two forms of the same food product whose price is the same. Which one of the following two forms of a similar food product would you buy if you needed one of them? (check one only): a. One food product is pre-mixed and pre-cooked and canned. This food requires only warming before it is served. b. The same food product as above may be pre- pared from different ingredients which are mixed according to a recipe. For this pur- pose, you will need: -one cup of all-purpose flour, -two teaspoons of baking powder, -half teaspoon of salt, -half cup of milk, -two tablespoons of egg, -two tablespoons of liquid shortening, and -two tablespoons of sugar. --PLEASE CONTINUE-- 192 THE FOLLOWING QUESTIONS DEAL WITH SPECIFIC SITUATIONS TO BE COMPARED. PLEASE ANSWER ALL THE QUESTIONS TO THE BEST OF YOUR UNDERSTANDING BY CHECKING ONE CHOICE IN EACH QUESTION WHICH YOU FEEL BEST EXPRESSES YOUR ANSWER. l.a. If John is taller than Pete, and Pete is taller than Jane, who is the tallest of the three? (check one only): a. John b. Pete c. Jane d. Not sure b. Who is the shortest of the three? (check one only): a. John b. Pete c. Jane d. Not sure If Jimmy prefers basketball to football, and he pre- fers football to swimming, which sport does Jimmy like most? (check one only): a. Basketball b. Football c. Swimming d. Not sure b. Which sport does Jimmy like least? a. Basketball b. Football c. Swimming d. Not sure (ckeck one only): If swimming requires three times more effort than football, and football requires five times more effort than basketball, which sport requires the most effort in this situation? (check one only): a. Swimming b. Football c. Basketball d. Not sure b. Which sport requires the least effort in this situa- tion? (check one only): a. Swimming b. Football c. Basketball d. Not sure 4. On visiting a foreign country, you find that for the price of a twelve-pack of any beer, you can get a "fifth" of a bottle of any type of whiskey. But you could get three bottles of any type of wine for the price of two— twelve packs. In this situation, which of the following statements is correct? (check one only): 193 a. A "fifth" of whiskey costs more than three bottles of wine. b. A "fifth" of whiskey costs the same price as three bottles of wine. c. A "fifth" of whiskey costs less than three bottles of wine. d. Not sure. 5. Suppose you are a member of a tennis club in which there are both men and women. Which of the following is true? (check one only): a. There are more men than members in this club. b. There are more members than the total number of men, gr the total number of women. c. The total number of all men and all women together is less than the number of all members in this club. 6. Suppose a customer gave you a set of dimes for a tip, and another customer gave you another set of dimes. If, for some reason, you displayed both these sets of dimes on some flat surface as shown here: First set: Second set: Are these two sets the same? a. Yes b. No 7. 194 If the dimes in #6 are re-arranged as shown here: First set: 00. Second set: 0 00 Do you still have the same two sets of dimes as in #6? a. Yes b. No We always classify the things we see or feel into cate- gories which are similar or different from each other on certain features. We give labels to those categories so that we may be able to talk about them. For instance, you understand what I mean when I say "Mike has a farm on which he keeps different types of animals such as cows, horses, pigs, and chickens. Use the following diagrams to answer the questions which follows: ii. iii. 10. 11. 12. 195 a. Which diagram is correct when I say, "All cows are animals?"____ b. Which diagram is correct when I say, "Horses are not cows?"___ c. Which diagram is correct when I say, "Some horses are male and some are female?"___ Suppose you are considering making a choice between two alternatives: a. To take-up a job offer which pays $500 a month, or b. To go to school where you will pay from your savings about $500 a year for two years for tuition and fees. Which choice would you make?___ Which of the following messages contains the most in- formation? (check one only): a. A tossed coin turns up heads. b. The railroad crossing is closed. c. My wife gave birth to a baby girl. d. The number of my bus ticket ends in digit 7. e. Not sure. Imagine a stadium with several people attending, say, a football game. How many pe0ple must there be in such a stadium so that there will definitely be at least two persons with a common birthday? (check one only): a. 365 people. b. 366 people. c. 367 peOple. d. Not sure. Suppose you have a square card "A" (remember: all sides of a square are equal). You have a series of nine other cards all of which are of the same width as card "A" but they differ in length such that: Card "B" is two times longer than card "A," Card "C" is three times longer than card "A," Card "D" is four times longer than card "A," Card "E" is five times long than card "A," Card "F" is six times longer than card "A," Card "G" is seven times longer than card "A," Card "H" is eight times long than card "A," Card "I" is nine times longer than card "A." a. How many cards do you have altogether? a. b. Not sure b. How many cards like card “ " can you make with card "C"? a. b. Not sure 196 c. If a card like card "A" is cut out from card "H," which card in the series will be similar to the remaining portion of card "H"? a. b. Not sure 13. If you took cards A,B,C,D,E from the set described in #12, and then formed a new series shown here: A B C D E a. Reading from left to right, how do you de- scribe the location of card "D" in this new series? a. b. Not sure b. Is this location different from the location which this card occupied in #12? a. Yes NO 14. If you re-arranged the series of cards in #13 to form a new series as shown here: 15. 16. 197 What location does card "D" occupy in this new series? a. b. Not sure Is card "D" still of the same length as before? a. Yes b. No Given the following diagrams: iii. iv. Which diagram implies: a. One The size of the following diagram is shown in a series b. Three c. Four of decreasing order: Draw in the space marked "E" the next diagram which this series can take on. Draw or state the last diagram which this series must be reduced to if you continued reducing the size of the series of these diagrams. Draw or state here: 17. 18. 19. 20. 198 Suppose you gave your son four apples and two oranges. To be fair, you give the same type of fruit to your daughter. But she prefers organges to apples. So, you decide to give her four oranges and two apples. If an apple weighs the same as an orange, does your daughter have the same amount of fruit as your son? a. Yes b. No On listening to a radio ad, you hear that a certain brand of coffee is 97% caffein-free. But moments later, an ad for another named brand of coffee says that the second brand of coffee is 3% caffein. Do these two brands of coffee have different amounts of caffein? a. Yes b. No John and Mike are farmers who are neighbors. John's plot of land is 100 yards long and 100 yards wide. Mike's plot is 200 yards long and 200 yards wide (see the diagrams below): John's plot: 100 yards 100 yards Mike's plot: 200 yards 200 yards Which of the following statements is true? (check one only): a. Mike's plot is two times bigger than John's. b. Mike's plot is four times bigger than John's. c. Mike's plot is six times bigger than John's. A 4-H member has three speckled rabbits and four gray rabbits for his project. At any time, all of these rabbits are either seated or running in the yard en- closed by wire-netting which this 4-H member built for his animals. Use the following box to answer the questions which follow: 21. 22. 23. 199 Seamairabbius Rmmfingznflxfits Spadded rabbits A B Gnnr maints C D a. Which letter or letters represents running speckled rabbits? a. b. Not sure b. Which letter or letters represents total gray rabbits? a. b. Not sure c. What fraction are speckled rabbits? a. b. Not sure d. What 18 the ratio of speckled rabbits to gray rabbits? a. b. Not sure e. If five rabbits are running, how many are seated? a. b. Not sure For the following pairs of fractions, which of the two is greater: a. 5/8 or 7/8? a. b. Not sure b. 3/4 or 3/5? a. b. Not sure c. 3/4 or 4/5? a. b. Not sure How many: a. Ounces in one pound? a. b. Not sure b. Pounds in one ton? a. b. Not sure c. Yards in one mile? a. b. Not sure d. Quarts in one gallon? a. b. Not sure e. Pints in one gallon? a. b. Not sure How many: a. Millimeters in one centimeter? a. b. Not sure b. Millimeters in one meter? a. b. Not sure c. Milligrams in one gram? a. b. Not sure d. Grams in one kilogram? a. b. Not sure e. Cubic centimeters in one liter? a. b. Not sure --PLEASE CONTINUE-- 200 To complete this questionnaire, please fill out the blanks below. This information is needed for the analyses. 1. 3. 10. 11. 12. Today's date 2. Time: AM PM Name Address Telephone # Age:____ 5. Sex: Male___ Female____ Marital status: Married___ Single___ Divorced___ Separated___ Employment: Employed___ Unemployed___ Race: Asian American Black American Mexican American Native American White American Other (please specify) Are you (or have you been) an Adult Basic Education (ABE) student? a. Yes b. No If so, what is (was) your highest student classification in ABE? If you are going to school now, what is your present school classification? (check one only): a. Freshman e. Graduate b. Sophomore___ f. Other (specify) c. Junior d. Senior Where do you go to school? If you are not going to school now, what is the last grade of school you completed? APPENDI X C Results From Preliminary Trials with Different Values of Restriction Parameters for Fitting the Regressors into the Predictive Equations in the Nigeria - Ilewo Sample Space APPENDIX C Results From Preliminary Trials with Different Values of Restriction Parameters for Fitting the Regressors into the Predictive Equations in the Nigeria - Ilewo Sample Space The first set of parameter values included 6 NSTEPS, 6.63 FIN, .01 TOL, and 6.00 FOUT. These values were too con- servative. Only one step of stepwise multiple regression analysis was possible in both the first and second model in which the awareness and adOption of technological innovations were respectively predicted. For the first model, literacy in English was the only regressor in the equation. Literacy in Yoruba and last grade of school completed were both not in the equation. For the second model, the awareness of tech- nological innovations was the only one in the equation. Literacy in Yoruba, literacy in English, and last grade of school completed were all not in the equation. The above values were then relaxed to 6 NSTEPS, 3.00 FIN, .001 TOL, and 2.00 FOUT. However, these new values did not change the results for the first model noted above. As for the second model, three steps were possible in which the awareness of technological innovations, literacy in English, and last grade of school completed were fitted into the equa— tion at the first, second, and third step respectively. Once again, literacy in Yoruba did not enter the equation for the second model under the new restrictions. 201 202 Finally, the following values were tried: 6 NSTEPS, 2.0 FIN, .001 TOL and 1.50 FOUT. However, this set of re- strictions produced the same results as the second set of restrictions noted above. In preference for the default values of the four parameters, no further trials were made. The default values permitted a complete exploration for the relative explanatory power of the regressors in each model. BIBLOGRAPHY Adiseshiah, Malcolm S. 1970 Let my country awake; the human role in de- velgpment: Thoughts on the next ten years (Paris: UNESCO). Adult Movement 1970 Conference on strategies forggenerating a national "Right to Read" (Raleigh, North Carolina: Adult Movement). Ahmed, Mushtag 1973 "Retention of literacy skills in adults" Literacy work: Functional literacy pilot project: Vol. II, No. 4, May 1973 (Teheran, Iran: International Institute for Adult Literacy Methods; UNESCO and Government of Iran). Aleksandrov, A.D., A.H. Kolmogrov, and M.A. Lavrent'ev (eds.) 1969 Mathematics: Its content, methods and meanings, Vols. 1-3. Translated from Russian by S.H. Gould and T. Bartha (Cambridge, Mass.: MIT Press). Barnett, Homer G. 1953 Innovation: The basis of cultural change (New York: McGraw-HillY. Bemis, Edward W. 1888 "Restriction of immigration," Andove Review IX. Blalock, Hubert M., Jr. 1972 Social statistics (New York: McGraw-Hill Book Company). Bloomfield, L. 1933 Language (New York: Henry Holt). Rousfield, W.A. 1953 "The occurrence of clustering in recall of randomly assigned associates," Journal of General Psychology, 40: 229-273. 203 204 Burnet, Marry 1965 ABC of literacy (Paris: UNESCO). Carothers, J.C. 1959 "Culture, psychiatry, and the written word," Psychiatry. Cassirer, E. 1953 Substance and function and Einstein's theory of relativity (New York: Dover Publications, Inc.). Chomsky, N. 1957 Syntactic structures (The Hague: Mouton & Company). 1965 Aspects of the theory of syntax (Cambridge, Mass.: MIT Press). 1968 Language and mind (New York: Harcourt Brace Jovanovich, Inc.). Cole, M., J. Glick, and D. Sharp 1971 The cultural context of learning and thinking (New York: Basic Books). Cou, Ya-Lun 1972 Probabilityiand statistics for decision making (New York: Holt, Rinehart and Winston). Dantzig, Tobias 1954 Number: The language of science; a critical survey written for the cultured non-mathema- tician (New York: The Macmillan Company). Del Buono, Xavier Antonio 1971 The relationship of bilingual/bicultural in- struction to the achievement and self-concept of seventh grade Mexican—American students. An unpublished dissertation (East Lansing, Michigan: Michigan State University). Desai, A.R., and S.P. Punalekar 1971 "Relationship between literacy and economic productivity of industrial workers in Bombay: A sociological analysis," in Literacy work: Educational development and education for socioeconomic development, Vol. 1, No. 2 (Teheran, Iran: International Institute for adult methods; UNESCO and Government of Iran). Detroit Free Press 1976 1TGuinness world records" (Tuesday, December 21). 205 Deutschmann, Paul J. 1963 "The mass media in an undeveloped village," Journalism Quarterly, 40: 27-35. Doob, Leonard W. 1961 Communication in Africa (New Haven: Yale University Press). 1964 "Eidetic Images Among the Ibos," Ethnology, 3: 357-363. 1965 Attitude scale items suggested for studies of modernization (New Haven, Conn.: Yale Univers- ity, Department of Psychology). 1966 "Eidetic imagery: A cross-cultural will—o'- the-wisp?" Journal of Psychology, 63: 13-34. Feldman, Carol, and Michael Shen 1969 "Some language-related cognitive advantages of bilingual five year olds," ERIC (ED 031307). Flesch, R. 1952 How to test readability (New York: Harper & Brosf). Fliegel, Frederick C., and Joseph E. Kivlin 1962a Differences among improved farm practices as related to the rate of adOption (University Park, Pa.: Agric. Exp. Sta. Res. Bul. 691). 1962b ”Farm practice attributes and adoption rates," Social Forces, 40: 364-370. 1966 "Attributes of innovations as factors in dif- fusion," American Journal of Sociology, 72: 235-248. Fliegel, Frederick C., et a1. 1968 "A cross-national comparison of farmers' per- ceptions of innovation as related to adoption behavior," Rural Sociology, 33: 437-440. Freeman, H.G., and C.G. Kassenbaum 1956 "The illiterate in American society: Some general hypotheses," Social Forces, 34: 371-375. Frey, Frederick W. 1964 "The politicization of the peasant." Un- published paper (Cambridge, Mass.: Center for International Studies, MIT). 206 Fromkin, Victoria, and Robert Rodman 1978 An introduction to language, 2d ed. (New York: Holt, Rinehart and Winston). Fuglesang, Andreas 1969 Communication with illiterates: A pilot study of the problem of social communication in develgping countries, Part I (Lusaka, Zambia: The National Food and Nutrition Commission). Gillete, Arthur 1972 Youth and literacy; you've got a ticket to ride (Paris: UNESCO). Gezi, Kal 1974 "Bilingual-bicultural education: A review of relevant research," California Journal of Educational Research, Vol. 25, No. 5: 223-239. Gray, William S. 1940 "Reading and factors influencing reading efficiency," in William S. Gray (ed.), Read- ing in general education (Washington, D.C.: American Council on Education). Grossnickle, Foster E., and Leo J. Brueckner 1959 Discovering meanings in arithmetic (Phila- delphia, Pa.: The John C. Winston Company). Gunning, R. 1952 Technique of clear writing (New York: McGraw- Hillf. Harman, David 1970 "Literacy programs in the world," Harvard Educational Review: A special issue, Literacy in America, Vol. 40, No. 2, May. Harris, Richard J. 1975 A primer of multivariate statistics (New York, New York: Academic Press). Heise, D.R. 1969 "Separating reliability and stability in test- retest correlation," American Sociological Review, Vol. 34: 93-101, February. Heise, D.R., and G.W. Bohrnstedt 1970 "Validity, invalidity and reliability," in E.F. Borgatta and G.W. Bohrnstedt (eds.), Sociological Methodology (San Francisco: Jossey-Bass). 207 Herzog, William A., Jr. 1967 Literacy training and modernization: A field experiment; techniCal report number 3. Project on innovations in rural societies (East Lansing, Michigan: Department of Com- munication, Michigan State University). 1967 The effect of literacy training on modern- ization variables. Unpublished dissertation (East Lansing, Michigan: Department of Com- munication, Michigan State University). 1973 "Literacy and community economic development in rural Brazil," Rural Sociology, Vol. 38, No. 3: 325-337, Fall. Hogben, Lancelot 1951 Mathematics for the million (New York: W.W. Norton and Company, Inc.Y. Hoiberg, Anne, C.J. Hysham, and N.H. Berry 1974 "The neuropsychiatric implications of illit- eracy: 20 years later," Journal of Clinical Psychology, Vol. 30, No. 4: 533-535, October. Hoijer, H. (ed.) 1954 Language in culture (Chicago: University of Chicago Press). Hoosain, Rumjahn, Parvin Atai, and Farideh Salili 1975 "Mirror tracing by monolingual and bilingual Iranian children," Perceptual and Motor Skills, Vol. 41, No. 2: 531-537, October. Houdek, John T. 1957 The literagy for immigrants: 1886-1917. M.A. Thesis TEast Lansing, Michigan: Michigan State University). House Report 1913 Number 140, 68th Congress; 2 sessions, December (Washington, D.C.: U.S. Congress). The reference to literacy testing. Hunt, W.A., and C.L. Wittson 1951 "The neuropsychiatric implications of illiter- acy," U.S. Armed Forces Medical Journal, 2: 365—369. Huse, H.R. 1933 The illiteracy of the literate (New York: Appleton-Century Co.i. 208 International Institute for Adult Literacy Methods (IIALM) 1974 Literacy work, V01. IV, No. l, July-September (Teheran, Iran: IIALM; UNESCO and Government of Iran). Jacobs, John F., and Marnell L. Pierce 1966 Bilingualism and creativity, ERIC (ED 021848). Judd, Charles 1927 Psychological analysis of the fundamentals of arithmetic (Chicago: University of Chicago Press). Junod, H.A. _ 1927 The life of a South African tribe, Vol. 2 (London: Macmillan). Katz, J.J. 1966 Thegphilosophy of language (New York: Harper & Row)l Kelejian, Harry H., and Wallace E. Oates 1974 Introduction to econometrics: Principles and applications (New York: Harper & Row, Publishers). Kerlinger, Fred N., and Elazar J. Pedhazur 1973 Multiple regression in behavioral research (New York: Holt, Rinehart, and Winston, Inc.). Kim, Jae-On 1975 "Factor analysis," in Norman H. Nie, et al., (eds.), Statistical Package for the Social Sciences (SPSS), 2d ed. (New York: McGraw- Hill Book Company). Kim, Jae-On, and Frank J. Kohout 1975 "Multiple regression analysis: Subprogram regression," in Norman H. Nie, et al., (eds.), Statistical Package for the Social Sciences (SPSS), 2d ed. (New York: McGraw-Hill Book Company). Kivlin, Joseph E. 1960 Characteristics of farm practices associated with rate of adoptiOn. Unpublished disserta- tion (University Park, Pa.: Pennsylvania State University). Kivlin, Joseph E., and Frederick Fliegel 1967a "Differential perceptions of innovations and the rate of adoption," Rural Sociology, 32: 78-91. 209 Kivlin, Joseph E., and Frederick Fliegel 1967b "Orientations to agriculture: A factor analysis of farmers' perceptions of new practices," Rural Sociology, 33: 127-140. Kmenta, Jan 1971 Elements of econometrics (New York: MacMillan Publishing Company, Inc.). Kramer, Edna E. 1970 The nature and growth of modern mathematics, Vol. I & II TNew York: A Fawcett Premier Book). Langebehn, George J., Thomas G. Lathrop, and Carl J. Martin 1972 Fundamental concepts of mathematics (Columbus, Ohio: Charles E. Merrill Publishing Company). Lassey, William R., et a1. 1965 "The consequences and meaning of literacy for attitudes and social behavior," (San Jose, Costa Rica: Programa Inter-Americana de Informacio Popular), Mimeo. Lenneberg, E.H. 1964 "A biological perspective of language," in E. H. Lenneberg (ed.) New directions in the study of language (Cambridge, Mass.: MIT Press). . 1969 "On explaining language," Science, 164: 635- 643. Lerner, Daniel 1958 The passigg of traditional society: Modern- izing the Middle East (Glencoe, Illinois: The Free Press). 1963 "Toward a communication theory of moderniza- tion," in Lucian W. Pye (ed.), CommunicaEion andppolitical development (Princeton, N.J.: Princeton Univeristy Press). 1964 "Literacy and initiative in village develop- ment," Rural Developmee} Research Report (Cambridge, Mass.: MIT Center for Inter- national Studies). Luria, A.R. 1961 The role of speech in the reulation of normal and abnormal behavior (J. Tizard, ed., Oxford: Pergamon). 210 Lyons, John 1974 Noam Chomsky (New York: The Viking Press). McLuhan, Marshall 1962 The Gutenberg Galaxy (Toronto: University of Toronto Press). 1964 Understanding media: The extensions of man (New York: McGraw-Hill). McNeill, D. 1966 "Developmental psycholinguistics," The Genesis of language (Cambridge, Mass.: MIT Press). 1971 "The capacity for the ontogenesis of grammar," The ontogenesis of grammar: A theoretical eymposium (New York: Academic Press). Mendez, D.A., and Fred Waisanen 1964 "Some correlates of functional literacy." Paper presented at the ninth Congress of Latin American Psychology (Miami, Florida). Menninger, Karl 1970 Number words and number symbols: A cultural history of numbers TCambridge, Mass.: MIT Press). Modiano, Nancy Docia 1966 Readingpcomprehension in the national language: A comparative study of bilingual and all- Spanish approaches to reading instruction in selected Indian schools in the Highlands of Chipas, Mexico. UnpuBliShed dissertation (New York University). Munroe, M.E. 1963 The language of mathematics (Ann Arbor, MiChigan: The University of Michigan Press). Mwanika, Thomas O. 1978 "Effects of complex technological symbols and quantitative concepts on the diffusion and adoption of innovations for rural development: Generalization and methodology." Preliminary paper submitted to Michigan State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy (East Lansing, Michigan: Department of Communication, MSU) . 211 Namboodiri, H.K., Lewis F. Carter, and Hubert M. Blalock, Jr. 1975 Applied multivariate analysis and experimental designs (New York: McGraw-Hill). Osgood, C.E. 1963 "On understanding and creating sentences," American Psychologist, 18: 735-755. Pavlov, I.P. 1927 Conditioned reflexes: An investigation of the psychological activity of the cerebral cortex TLOndon: Oxford University Press). Peal, Elizabeth, and Wallace E. Lambert 1962 "The relation of bilingualism to intelligence," Psychological Monographs: General and Applied, Vol. 76, No. 27. Petrini, Frank 1966 Upplysningsverksambetes effekt I, innovation- ernas spridningstakt--The rate of adoption of selected agricultural innovations; Rept. 53 (Uppsala, Sweden: Agricultural College of Sweden). Rahim, S.A. 1961 Diffusion and adoption of agricultural prac- tices: A study of communication, diffusion and adOption of improved agricultural practices in a village East Pakistan (Comilla, East Pakistan: Pakistan Acad. for Rural Development). Riesman, D. 1956 The oral tradition, the written word, and the screen image (Yellow Springs, OH: Antioch Press). Rogers, Everett M., and William Herzog 1966 "Functional literacy among Columbian Peasants," Economic Develepment and Cultural Change, 14: 190-203. Rogers, Everett M. with Ralph E. Neill 1966 Achievement motivation among Columbian peasants: Diffusion of innovations research project, No. 5 (East Lansing, Michigan: Department of Communication, Michigan State University). Rogers, Everett M., with Lynne Svenning 1969 Modernization among peasants: The impact of CommunicatiEn (New York: Holt, Rinehart and Winston). 212 Rogers, Everett M.,with F. Floyd Shoemaker 1971 Communication of innovations: A cross-cul- tural approach (New York: The Free Press). Russ, Raymond C., Joel A. Gold, and Paul D. Cherulnik 1975 "Semantics and structure: A comparison be- tween monolingual and bilingual subjects," The Journal of Social Psychology, 96: 163-172. Sapir, Edward 1931 "Conceptual categories in primitive languages," Science, 74: 578. Schofield, R.S. 1968 "The measurement of literacy in pre-industrial England," in Jack Goody (ed.), Literacy in traditional societies (Cambridge, England: Cambridge University Press). Schumann, Howard, Alex Inkeles, and David S. Smith 1967 "Some social psychological effects and non- effects of literacy in a new nation," Economic Development and Cultural Change, XVI, No. 1, October. Selltiz, Claire, Marie Jahoda, Morton Deutsch, and Stuart W. Cook 1959 Research methods in social relations (New York: Henry Holt & Company, Inc.). Singh, Ram N. . 1966 Characteristics of farm innovations associated with the rate of adoption (Guelph, Canada: Ontario Agricultural Extension Education, Rept. No. 14). Singh, T.R. 1970 "Literacy in two Indian villages," in Literacy Discussions: Literaey Research, Vol. I, No. 3 (Teheran, Iran: International Institute for Adult Literacy Methods, UNESCO and Government of Iran). Skinner, E.F. 1957 Verbal behavior (New York: Appleton-Century Crofts). Smith, D.E. 1923 The history of mathematics, Vols. I & II (New York: Dover Publishing Company, Inc.). 213 Solo, Robert A., and Everett M. Rogers (eds.) 1972 Specht, David A 1976 Inducing technological change for economic growth and development (East Lansing, Michigan: MiChigan State University Press). "SPSS: Statistical package for the social sciences version 6 users guide to subprogram reliability and repeated measurements analysis of variance," in Computer Laboratory, User Information Center, Michigan State University, User Guide Supplement: SPSS Revisions with local modification (East Lansing, Michigan: Computer Laboratory, User Information Center, MSU). Stauffer, Joh, Richard Frost, and William Rybolt 1978 "Literacy, illiteracy, and learning from tele- vision news," Communication Research, Vol. 5, No. 2, April. Tucker, Carlos F. 1961 UNESCO 1965 1969 1969 1970 1970 1970 1970 Prediction of rate of adoption from character- istics of farm innovations. M.S. Thesis (Columbus, Ohio: Ohio State University). Statistics of illiteraey (Paris: UNESCO). Work-oriented functional literacy: Reading and follow-up materials; final report of regional workshop for specialists concerned with the preparation of reading and follow-up materials in Asia, Bangkok, 257November, 13 December, 1968 (Bangkok: UNESCO, Regional Office for Education in Asia). Literacy: A newsletter, December (Paris: UNESCO). Literacy, 1967-1969: Progress achieved in literacy throughout the world (Paris: UNESCO). Literacy: A newsletter, No. 1, January (Paris: UNESCO). Literacy: A newsletter, No. 2, April (Paris: UNESCO). Literacy: A newsletter, No. 4, October (Paris: UNESCO). 214 UNESCO 1971 "Evaluation of pilot projects," Literacy: A Newsletter; First Quarter (Paris: UNESCO). 1971 Literacy: A newsletter; Third Quarter (Paris: UNESCO). 1971 Literacy: A newsletter; Fourth Quarter (Paris: UNESCO). 1972 Literacy: A newsletter; First Quarter TParis: UNESCO). 1972 Literacy; 1969-1971: Progress achieved in literacy throughout the world (Paris: UNESCO). 1973 Practical guide to functional literacy: A method of training for development (Paris: UNESCO). 1976 "Doing the job: Implementation," The experi- mental world literacy program: A critical assessment (Paris: UNESCOY. 1976 Statistical yearbook (Paris: UNESCO). United Nations 1977 World statistics in brief (New York, New York: Statistical Office of the United Nations). Urban, Wilbur M. 1939 Language and reality: The philosophy of lan- guage and principles of symbolism (New York: The Macmillan Co.). ' U.S. Department of Commerce, Bureau of the Census 1963 Current population reports, Ser. P-20, No. 45, and Ser. P-23, No. 8 (Washington, D.C.: Department of Commerce, Bureau of the Census). Watson, J.B. 1924 Psychology from the standpoint of a behavior- ist (Philadelphia, Pa.: Lippincott). 1930 Behaviorism (New York: Norton). Weiss, N.A., and M.L. Yoseloff 1975 Finite mathematics (New York: Worth Publishers, Inc.). 215 Weitheimer, Max 1967 "Numbers and numerical concepts in primitive peoples," A source book of gestalt psychol- cgy_(London: Routledge & Kegan Paul, Ltd.). Werts, C.E., K.G. Joreskog, and R.L. Linn 1973 "Identification and estimation in path analysis with unmeasured variables," American Journal of Sociologyy 1469-1484. Whatmough, Joshua 1956 Language: A modern synthesis (New York: St. Martin's Press). Whorf, Benjamin L. 1952 Collected papers on metalinguistics (Washing- ton, D.C.: Department of State, Foreign Service Institute). 1956 Language, thought and reality (Cambridge, Mass.: MIT Press). World Education 1970 Workshop on literacy, population, and family planning education (New York: World Educa- tion). Wiley, D.E., and J.A. Wiley 1970 "The estimation of measurement error in panel data," American Sociological Review, Vol. 35. 1974 "A note on correlated errors in repeated measurements," Sociological Methods and, Research, Vol. 3, No. 2, November. Wright, Peter C, et al. 1967 The impact of a literacy program in a Guatemalen ladino peasant community. Report submitted to U.S. Agency for International Development (Tampa, Florida: The University of South Florida).