M“ JESS‘E {EEG EEEfCéEN-C-Y AM} EF“ECTIVE NESS 0E UNE‘I’ERSITY EMSYRUCTION ~ F CEESTRAL AMERICAN CASE~ Thesis for the Degree of Ph. D MIIZIHIGAN SMTE UNWERSET‘.’ WEEK? 8. HEMiNK 19.5.7 e " LIBRARY Michigan Stew: University ,9.“ ,,. i w... THESlS ’I’-‘ ‘03“- $ This is to certify that the thesis entitled MEASURING EFFICIENCY AND EFFECTIVENESS OF UNIVERSITY INSTRUCTION-~A CENTRAL AMERICAN CASE presented In; Lynn David Hemink has been accepted towards fulfillment of the requirements for Ph.D. degree in_d_Q§_t.i_DE u 0 Major proflassor \ Date June 22, 1967 0-169 .d—r—fi —— - ABSTRACT MEASURING EFFICIENCY AND EFFECTIVENESS OF UNIVERSITY INSTRUCTION ~-A CENTRAL AMERICAN CASE-- by Lynn David Hemink History notes that colleges and universities have been reluctant to apply input-output analysis to university in- struction. The classroom has been perceived as incongruent with the input-output analysis. As a consequence, university administrators have been forced to allocate precious resources with a minimum of rational method and direction to insure max- .imum utilization in instruction. ' To ameliorate this existing shortcoming, the following thesis is developed: 1. Important aspects of higher education instruction are amenable to input-output analysis, involving the con- structs of efficiency and effectiveness. 2. Efficiency and effectiveness constructs are mod- erated by administratively controllable variables operating yithin the instructional environment. 3. When relationships are established empirically between administratively controllable variables and the efficiency-effectiveness constructs, a rational basis for administrative decision-making results. pg Lynn David Hemink A model of efficiency and effectiveness may be constructed that incorporates class enrollment, students presenting them- selves for final examination, and the number of students pass- ing a course. Two ratios of efficiency and one measure of effectiveness may be developed that isolate various phases in the instructional environment in which student inputs are being reduced. The national universities of Costa Rica, El Salvador, Guatemala, Honduras and Nicaragua illustrate the utility of the thesis. Data on all classes taught in the universities in the academic year 1962-63 formed the population. Data on six factors that could be affecting the efficiency-effectiveness ratios are gathered for the five universities. The six fac- tors that could be affecting the efficiency-effectiveness con- structs are: 1. The salary paid to the instructor in each class. 2. The size of each class. 3. The level of instruction as measured by the year in which a class is normally taken in a students program. 4. The type of instruction. (Lecture, laboratory, lecture— laboratory.) 5. The contact hours of the class with the instructor. 6. The number of classes taught by the professor during the academic year. Using Chi square and mean difference comparisons, it is pos- sible to determine relationships between the efficiency- effectiveness ratios and the six factors. The efficiency- It... (A . 0‘4: mil .9. t. V. (7" (I) - I HI ) .3. “‘l . "I Lynn David Hemink effectiveness constructs provide administrators with a.method for future decision-making, and relationships between the con- structs and the administratively controllable variables provide direction for future decision-making. Interrelationships between the ratios of efficiency and effectiveness are evidenced in data analysis which support the theoretically constructed relationships between these ratios. These relationships validate the model of efficiency and effectiveness as a practical method for analysis of instruction. Relationships between constructed ratios and the six fac- tors that are administratively controllable provide direction for future decision-making that is relevant only in Central America. Specifically, the following directions are indicated in Central America: 1. Lowest salaries paid to instructors correspond to highest efficiency and effectiveness ratios. However, the type of person receiving a salary under $250 equivalents may be a practitioner instead of an academic professor. 2. Class sizes smaller than 29 are more efficient than classes of 30-109, and classes larger than 110. The adminis- trative direction advocated, would be smaller class sizes. 3. Efficiency and effectiveness increase at successive levels of instruction. If this is a real difference and not the result of student self selection, administrators should attempt to ascertain means for holding more students at lower levels. a. Laboratory classes tend to be more efficient than Lynn David Hemink other methods of instruction. However, these courses are usually smaller and taken at upper levels of instruction where ratios are higher. 5. Contact hours beyond 2500 result in a rapid decrease in efficiency. Smaller classes may reduce contact hours. MEASURING EFFICIENCY AND EFFECTIVENESS OF UNIVERSITY INSTRUCTION --A CENTRAL AMERICAN CASE-- B3“ A .. Lynn D.” Hemink A Thesis Submitted to Michigan State University In partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY College of Education Department of Administration and Higher Education 1967 ACKNOWLEDGMENTS I wish to extend sincere appreciation to my doctoral committee chairman, Dr. Karl T. Hereford, for his generous assistance and guidance beyond the expected parameters. His interest and direction provided constant inspiration for an ever better product. I also extend my gratitude to the committee members, Dr. Walter F. Johnson, Dr. George M. Johnson, and Dr. John Useem, for their insightful comments and interest during the preparation and development of my thesis. I wish to thank all members of the faculty and staff of Michigan State Uni- versity whose cooperation made my doctoral program‘s rewarding and enjoyable experience. This thesis is dedicated to my family: To my wife, Ellen, for her patience, kindness, understanding, support and willing- ness to work, which made the completion of this doctoral pro- gram a mutual endeavor. To my mother and father whose devotion to the professionalization of education is a shining example to those truly interested in the education of man. TABLE OF CONTENTS ACKNOWLEDGEMENTS 11 LIST OF TABLES AND FIGURES . vi CHAPTER Page I NATURE, PURPOSE AND DELINEATION OF THE THESIS Statement of the Thesis Thesis Development Efficiency and Effectiveness Model of Efficiency and Effectiveness Time in the Efficiency-Effectiveness Model Variability of Efficiency-Effectiveness NUWUIMH I-' Concepts 9 Factors Affecting Efficiency and Effectiveness lO Purpose of the Dissertation 12 Overview of the Dissertation 13 II ANALYSIS EMPLOYED TO EXAMINE THE THESIS 14 Pepulation 14 Instrumentation 16 Procedure ’ 17 Design 19 Nature of Hypotheses 20 Limitations of Analysis 23 Currency Equalization 25 III EFFICIENCY-EFFECTIVENESS INTERRELATIONSHIPS 27 Institutional Effectiveness and Institutional Efficiency 27 Institutional Effectiveness and Instructional Efficiency 29 Institutional Efficiency and Instructional Efficiency 31 Internal Consistency 33 Time in the Efficiency-Effectiveness Ratios ~ 34 Institutional Effectiveness and Institutional Efficiency 35 Institutional Effectiveness and Instructional Efficiency 36 Institutional Efficiency and Instructional Efficiency 38 iii .q 4 H “II III Data Differences Between Models 40 Summary 45 IV FACTORS RELATED TO INSTITUTIONAL EFFECTIVENESS 46 Level of Expenditure and Effectiveness 49 Class Size and Institutional Effectiveness . 55 Level of Instruction and Effectiveness 59 Method of Instruction and Effectiveness 65 Contact Hours and Effectiveness 68 Number of Classes Taught and Effectiveness 73 Summary 76 V FACTORS RELATED TO INSTITUTIONAL EFFICIENCY 79 Level of Expenditure and Institutional Efficiency 80 Class Size and Institutional Efficiency 82 Level of Instruction and Institutional Efficiency 87 Method of Instruction and Institutional Efficiency 91 Institutional Efficiency and Contact Hours 95 Institutional Efficiency and Number of Classes Taught 100 Summary 102 VI FACTORS RELATED TO INSTRUCTIONAL EFFICIENCY 104 Level of Expenditure and Instructional Efficiency 105 Class Size and Instructional Efficiency 108 Level of Instruction and Instructional Efficiency 113 Method of Instruction and Instructional Efficiency 118 Contact Hours and Instructional Efficiency 121 Number of Classes Taught and Instructional Efficiency 126 Summary 128 VII ADMINISTRATIVE IMPLICATIONS--A CASE STUDY 130 Data Analysis by University 130 Data Analysis by Selected Faculty 134 Summary 139 iv .q-.' I .QO '-6. VIII SUMMARY AND CONCLUSIONS Bibliography Appendices A B C General Summary Conclusions Implications for Future Research 140 140 143 147 150 154 155 156 161 .n '4 N I C u‘C A I u‘[ l a ' H v‘b 7.: v‘j 11'- U“q Figure Figure Figure Figure 3.1 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 4.1 4.2 4.3 4.4 «PUOI'DH LIST OF TABLES AND FIGURES Model of Instructional Productivity Revised Model of Productivity Model of Efficiency and Effectiveness Time in the Efficiency-Effectiveness Model Percent of Institutional Effectiveness (A/E) by Level of Institutional Efficiency (Ex/E) Percent of Institutional Effectiveness (A/E) by Level of Instructional Efficiency (A/Ex) Percent of Instructional Efficiency (A/Ex) by Level of Institutional Efficiency (Ex/E) Percent of Institutional Efficiency Ex Eg, Instructional Efficiency §A Ex and Institutional Effectiveness A/E) by University Percent of Institutional Effectiveness /E) by Level of Institutional gEiciency (Ex Percent of Insti utional Effectiveness (A /E) by Level of )Instructional Ef iciency (A 1/Ex) Percent of Instructional Efficiency /Ex1) by Level )of Institutional gEiciency( Ex/ Institutional (EfEectiveness (A/E and A.l/E) Differences in Percent Percent Differences in Institutional Efficiency (Ex/E and Ex /E) Instructional Efficiency EA/Ex and AL/Exl) Differences in Percent Factors Related to Institutional Effectiveness (A/E) Institutional Effectiveness (A/E) by Level of Expenditure Level of Expenditure Corresponding to Highest Effectiveness (A/E) Percent of Institutional Effectiveness (Al/E) by Level of Expenditure vi '1) m oomwm CD 29 3O 32 33 36 37 38 41 43 44 48 49 52 53 NJ 9.; .5} .54 \l :1 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14 5.1 5.2 5.3 5.4 5-5 5.6 5.7 5.8 5.9 5.10 5.11 Percent of Institutional Effectiveness (A/E and A /E) by Class Size Institutiona Effectiveness (A/E) and Class Size (By Selected Faculty) Percent of Institutional Effectiveness (A/E) by Level of Instruction Percent of Institutional Effectiveness (Al/E) by Level of Instruction Percent of Institutional Effectiveness (A /E) by Level of Instruction Percent of Effectiveness by Method of Instruction Percent of Effectiveness by Method of Instruction (By Selected Faculty) Percent of Effectiveness by Number of Contact Hours Percent of Institutional Effectiveness by Contact Hours (By Selected Faculty) Percent of Institutional Effectiveness (A/E and A /E) by Number of Classes Taught 1 Frequency Distribution of Professors by Number of Classes Taught Level of Expenditure Displaying Highest Institutional Efficiency Level of Expenditure Displaying Highest Institutional Efficiency (By Selected Faculty) Percent of Institutional Efficiency by Class Size Percent of Institutional Efficiency, by Class Size (By Selected Faculty) Percent of Institutional Efficiency by Level of Instruction Percent of Institutional Efficiency by Level of Instruction (By Selected Faculty) Percent of Institutional Efficiency by Method of Instruction Percent of Institutional Efficiency by Method of Instruction (By Selected Faculty) Percent of Institutional Efficiency by Number of Contact Hours Percent of Institutional Efficiency by Number of Contact Hours (By Selected Faculty) Frequency Distribution of Professors by Number of Classes Taught vii 55 57 59 62 63 66 67 69 71 74 75 80 81 83 85 87 90 92 94 96 99 101 I rfild A). I f‘hv .(d It .U V. 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12 Percent of Instructional Efficiency by Level of Expenditure Percent of Instructional Efficiency by Level of Expenditure (By Selected Faculty) Percent of Instructional Efficiency by Class Size Percent of Instructional Efficiency by Class Size (By Selected Faculty) Percent of Instructional Efficiency by Level of Instruction Percent of Instructional Efficiency by Level of Instruction ~(By Selected Faculty) Percent of Instructional Efficiency by Method of Instruction Percent of Instructional Efficiency by Method of Instruction (By Selected Faculty) Percent of Instructional Efficiency by Number of Contact Hours Percent of Instructional Efficiency by Number of Contact Hours (By Selected Faculty) Frequency Distribution of Professors by Number of Classes Taught Percent of Instructional Efficiency by Number of Classes Taught Institutional Efficiency, Instructional Efficiency and Institutional Effectiveness viii 105 107 109 111 114 116 118 120 122 124 126 127 130 CHAPTER I NATURE, PURPOSE AND DELINEATION OF THE THESIS Recent history notes that colleges and universities have been reluctant to apply input-output analysis to university instruction. The classroom has been perceived as incongruent ‘with the input-output analysis that has proven to be successful in governmental agencies. As a consequence, university ad- ministrators have been forced to allocate precious resources with a minimum of method and direction for insuring that maximum resource utilization occurs in instruction. In order to ameliorate this shortcoming, the following thesis is advanced. Statement of the Thesis The thesis is threefold: 1. Important aspects of higher education instruction are amenable to input-output analysis, involving the evaluative constructs of effectiveness and efficiency. 2. These constructs (effectiveness and efficiency) are moderated by certain administratively control- lable variables that are operating within the in- structional environment. 3. When relationships are established empirically between the administratively controllable variables and the efficiency-effectiveness constructs, a ra- tional basis for administrative decision-making results. 1 The potential worth of the thesis lies in the utility and validity of the method of analysis herein proposed, and in the direction for administrative planning and action that is indi- cated by the method when applied in concrete instructional situations. Specifically, the method of analysis is designed to maximize the use of the limited resources available to higher education instruction. Thesis Development The typical model of instructional productivity can be stated thusly: FIGURE 1 MODEL OF INSTRUCTIONAL PRODUCTIVITY _ (E) + (I) --> (A) Given class enrollment (E), an instructional factor (I) is provided by the instructor and a number of students success- fully pass the course (A). The ratio of A/E represents the extent to which the professor has been successful in producing an output (A) in relation to a given input (E). This model assumes that the instructional factor (I) is the only mod- erating influence on the quantity of students who will pass the course (A). If this model is applied in concrete situations, however, there are factors other than the instructional factor (I) pre- sent in the institutional environment which affect the pro- ductivity of the class (A/E). In essence, losses in productivity that formerly have been attributed to instruction (I) may result from other factors. For example, in institutions in which the examination procedure permits the student to take the final examination at his discretion instead of upon con- clusion of coursework, the productivity of the instructor is difficult to assess. The examination procedure is an institutional policy decision that should be reflected in instructional productivity. It is necessary to construct a tmodel that delineates institutional variables that would other- wise have been attributed to instruction. In constructing a model that incorporates institutional factors such as the examination policy previously noted, another variable must be inserted in the structure of the model. If (Ex) represents those students who present them, selves for final examination in a class, a new model is possible. (Figure 2). FIGURE 2 REVISED MODEL OF PRODUCTIVITY (E) + (I) --) (Ex) --> (A) In this model, a class enrollment (E) and the instructional factor (I) produces examined students (Ex) from.which success- ful students (A) result. In essence, an enrolled student (E) must become an examined student (Ex) and pass the examination before becoming a successful student (A). The preceding model makes it possible to determine a variety of productivity ratios. For example, the ratio Ex/E ‘7- n‘1 st}: represents the ratio of students presenting themselves for final examination (Ex) divided by the class enrollment (E). A/Ex represents the ratio of students who pass the course (A) divided by those who were examined (Ex), and A/E is the ratio of total output (A) divided by the total input (E). The ratio Ex/E may be affected by factors other than instruction. For example, if an institution were to enroll all students on a full time basis with full course loads even if a student were only to attend one or two classes, this ~ variation would be exhibited in the Ex/E ratio and could not be attributed entirely to instruction. In this ratio, it is possible to avoid attributing low productivity to the professor that more properly resides in other phases of the institutional environment. This ratio represents a significant development in determining where productivity is being impaired. A/Ex is the ratio that more truly reflects instruction in an institution with a voluntary examination procedure. This ratio results from dividing the successful students by those examined. This ratio represents the relationship between the instructor and the academic ability of student an academician and is thereby a measure of instructional productivity. The ratio A/E is subject to a variety of institutional factors. The ratio of A/E will be affected by the instructor passing students (A), and also by administrative policies. For example, if the administrative policy is in effect that all students will be enrolled with a full complement of classes even though the student desires to attend only one class, the was: U... vi a ‘1 0. .3 2:. i. “ll. \Maar F C ( 7‘ n a“ per] ‘3 A/E ratio will reflect this procedure in part. Other factors :may also affect A/E, and it is thusly apparent that A/E is affected by factors in the institutional environment in addition to instruction. Efficiency and Effectiveness Productivity may be divided into efficiency and effective- ness. If efficiency is defined as the maximization of the useful return or output from any given input} and effective- ness is defined as the extend to which an intended outcome of a course of action becomes an observable and measurable outcome of that course of action? it is possible to construct a model of ratios noted as efficiency and effectiveness ratios, (Figure 3). FIGURE 3 MODEL OF EFFICIENCY AND EFFECTIVENESS . Input _ . _ Output I (E) |_, Ex/E ______, A/Ex _____, (A) nro ees ns u ona InStructional uccess u Efficiency Efficiency Students A/E 'Institutiona. Effectiveness 1James Buchanan, The Public Finances, (Hemewood, Illinois, Richard Irwin Co., 1965), p. 222. 2 EMrton Friedman, The Public Administration of Education in Central America, (IIME, E. Lansing, Michigan StéteIUniversity, , p. 7. The intended institutional outcome in enrolling students (E) is to produce successful students (A) in each class so that these students will ultimately be productive members of society. The extent of attainment of the institutional goal in each class is represented by dividing the number of successful students by the number of students enrolled (A/E). In essence, A/E is the extent to which a course of action has been effective in the institution. Whereas in previous models, A/E has represented the effectiveness of instruction, it is evident that the term insitutional effectiveness is more appropriate because it has been noted that many factors in addition to instruction may affect this ratio. Ex/E is not an effectiveness ratio because it does not incorporate the desired institutional outcome of passing 3 students (A). Ex/E represents the ratio of those examined (Ex) divided by the number of students enrolled in a class (E). This ratio (Ex/E) is an efficiency ratio because it represents the degree of maximization of an input that is necessary to achieve the desired institutional outcome (A). As noted pre- viously, this ratio may be determined by factors other than instruction. Hence, Ex/E is designated as institutional efficiency in recognition of the many institutional factors operating on the ratio. A/Ex is designated as instructional efficiency. The term instructional efficiency is employed in light of the major factor that affects this ratio. As noted previously, the 7 number of successful students (A) resulting from those examined (Ex) is determined by the instructor and the ability of the student. In essence, the instructional factor is clearly reflected in this ratio. A/Ex is regarded as a measure of efficiency because it does not incorporate total input. In- stead, it utilizes (Ex) which is only a partial input in the' institution. The difference between efficiency and effectiveness in this thesis may be stated thusly. Effectiveness represents the extend to which total inputs (E) reach the institutional objective as total outputs (A). Efficiency, on the other hand represents the dispersal and loss of inputs at substages in the educative process that contribute to the effectiveness of the institution but are not desired outcomes in and of themselves. In essence effectiveness illustrates that an intended outcome occurred. Efficiency illustrates how the outcome occurred. Time in the Efficiency-Effectiveness Model In those institutions which permit a student to be examined whenever he desires, it is noted that the length of time 3 necessary to pass each course is increased considerably. The societal impact of lengthened matriculation is evident in developing countries that are in dire need of educated people IIME STAFF REPORT, A Case Study, Academic Progress of University Students Universit ofISan Carlos of Guatemala, 1953, (IIME: E. IansIng, MIcfiIgan, I964, p. 42.) immediately. As the length of matriculation increases, the fewer people that are available to turn resources into economic production. The ideal procedure would thereby result in dis- tinquishing students progressing rapidly from those who are lengthening their matriculation. For the lengthening of time in enrollment represents a loss in productivity. The model of efficiency and effectiveness can be modified to distinguish students taking examinations immediately upon conclusion of a course from those that deferred taking the examination. The advantage of this modification in the model is that it is possible to ascertain the percentage of students who are progressing rapidly through an educational program from those who are lengthening their programs through failure of examinations or deferred examination. Figure 4 presents the efficiency-effectiveness model when the factor of time spent in matriculation is important. FIGURE 4 TIME IN THE EFFICIENCY-EFFECTIVENESS MODEL Input . Output (E) Exl/E Al/Exl I (A1) Enrollees nstitutiona Instructional Successfuf Efficiency Efficiency Students Al/E Institutional Effectiveness (Exl) represents those students presenting themselves for examination at the earliest opportunity. (A1) represents those students passing a course at the earliest opportunity. It is noted that the ratios Exl/E, Al/Ex1 and AI/E are designated in exactly the same terms as noted in the earlier model. It is thereby possible to compare ratios between models. To illustrate the use of the model in Figure 4, if (E) = 10, (Exl) = 5, and (A1) = 2, it is possible to deter- mine ratios reflecting the number of students that are passing a course at the earliest opportunity and thereby are progressing rapidly. With an (Exl/E) ratio of 5/10, it is evident that only 50% of a class enrollment took the examina- tion immediately upon conclusion of a course. If the (Ex) value was 8, it is possible to determine that three students had either failed the course previously, or had not taken the examination at the earliest opportunity. This distinction is of great importance in societies which need educated people immediately. Variability of Efficiency and Effectiveness Concepts Various interpretations of efficiency and effectiveness were reviewed for applicability to university instruction. 10 6 7 Eliot,“ Benson,5 Barnard, Riecken and Romans, Gage, 10 were reviewed and rejected because Etzioni,9 and Williams their definitions of these important terms were not amenable to the exigencies of university instruction. The defini- tions of Friedman and Buchanan possess utility for delimiting outputs in education that differ from most input-output analyses. Factors Affecting Efficiency and Effectiveness Having constructed measures of efficiency and effective- ness in which to evaluate input-output analysis in universities, it is necessary to determine the relationship of Specific factors and the measures of efficiency and effectiveness. “Charles W. Iliot, Education for Efficiency, (New York: Houghton-Mifflin Co., 1909), p. l. . 5Charles Benson, The Economics of Public Education, (Boston, Houghton-Mifflin Co., 1961), p. 351. 6Chester Barnard, Functions of the Executive, (Cambridge, Mass., Harvard University Press, 1939), p. 19. 7H. W. Riecken and George C. Homans, Psychological Aspects of Social Structures. Ine Lindzey, G. (ed.) Handbook of Social Psychology, Vol. II, (Reading, Mass., Addison-Wesley, 1954), p. 805. 8N. L. Gage (ed.), Handbook of Research on Teaching, (Chicago, Rand McNally andfiCo., 1963), p. 117. 9Amitai Etzioni Complex Organizations, (New York, The Free Press, 1961) pp. 71488. lOHarry Williams, Planning for Effective Resource Allocation in Universities, (WaEhington, D.C., American Council on Education, 1966), pp. 2-3. 11 Six factors are delineated that are administratively con- trollable and yet may affect instruction.ll These are: \ QQntr an 1. The level of expenditure paid to the instructor for teaching a class. This factor may be controlled by administrative determination of the salaries paid to professors. The salaries paid may to aligned to correspond with the salaries deemed to result in optimum efficiency and effectiveness. The class size of each class taught by each pro- fessor. The administration may limit enrollments in a class to conform to class sizes deemed to be most efficient and effective. The level of instruction of each class as determined by the year in which each class is normally taken in a student program. The curricular variable may be administratively controlled by offering specified courses or a sequence of courses at levels that provide optimum efficiency and effectiveness. The method of instruction utilized to teach each course. (Lecture, laboratory, lecture-laboratory.) The method of instruction may be somewhat controlled by an administrative policy which advocates usage of the optimum methods of instruction as evidenced in efficiency-effectiveness ratios. 11The instructor is not considered as an administratively ollable variable because of the possible connotation as :1mpingement on academic freedom. 12 The number of contact hours of the instructor with each class. This factor may be manipulated by altering the hours of instruction or the size of a class. Contact hours that are deemed to provide the greatest efficiency and effectiveness would be the ultimate goal in manipulation. The number of classes taught by each professor during the academic year. An administrative limitation may be placed on teaching loads that exceed the most efficient and effective class loads. These factors were chosen specifically because they may affect instruction in each class, and may be controlled by administrators. As a result of examining relationships be- tween the administratively controllable factors and the efficiency-effectiveness measures, it is possible to provide direction for future decision-making that reflects more (optimum usage of limited resources. Pharpose of the Dissertation The purpose of the dissertation is threefold: 1. To elaborate and examine the thesis in detail by applying it to concrete data. To postulate appropriate uses of the model of efficiency and effectiveness. To suggest possible direction for improving efficiency and effectiveness. 13 Overview of the Dissertation In this chapter the thesis was developed that input- output analysis is amenable with university instruction. A model of efficiency and effectiveness was constructed to pro- vide administrators with a method of determing the productivity of instruction. Six factors were identified that may provide direction for improving efficiency and effectiveness. In the chapters that follow, the utility of the model of efficiency and effectiveness will be determined, the factors affecting these measures will be examined in concrete situations and directions for improving efficiency and effectiveness will be discussed. Specifically, the following will be discussed in subsequent chapters: Chapter 2--Analysis Employed to Examine the Thesis Chapter 3--Efficiency-Effectiveness Interrelationships Chapter 4--Factors Related to Institutional Effectiveness Chapter 5--Factors Related to Institutional Efficiency Chapter 6--Factors Related to Instructional Efficiency Chapter 7--Administrative Implications-~A Case Study Chapter 8--Summary and Conclusions CHAPTER II ANALYSIS EMPLOYED TO EXAMINE THE THESIS The purpose of this chapter is to present the method of analysis employed in applying the efficiency-effectiveness model to a special population. Of particular importance were the population, instrumentation, procedure, design, the na- ture of hypotheses, and the limitations of the analysis. Additionally, the equalization of currency between countries is explained to insure understandable application throughout the remainder of the dissertation. Population The sources of data for analysis were the students and faculty of the five Central American national universities of Costa Rica, El Salvador, Guatemala, Nicaragua, and Honduras for the academic year 1962-63. Variables surrounding these data were enumerated as precisely as possible. As a result, the data utilized represent a comprehensive picture of students and faculty in the universities. There was no random sample; the population was finite and quantitatively measurable. The national universities of Central America were chosen to ascertain the utility of the thesis for several compelling reasons: 14. 15 1. Data collected by IIME12 were adaptable to illustrate the thesis. 2. The region is comprised of developing nations that need strong higher education institutions, hence the method advanced here may be directly beneficial to the administra- tion of such institutions. 3. Direction obtained from analyses would likely be acted upon by the Central American universities, thereby affording the basis for long-term field testing of the ideas and methods herein advanced. Having chosen Central America as the location in which to illustrate the thesis, limitations were necessarily ac- cepted as a result of administrative policies operating in the five universities. The most important of these are: l. The five universities under consideration enroll students as full time students, when in fact few students attend a full load or classes any semester. l2(IIME) Interuniversity program of the university of San Carlos of Guatemala and Michigan State University-~A research center conceived by Michigan State University and the University of San Carlos of Guatemala to conduct re- gional research and development in Central America. IIME invented proposals for improvements in Central America's educational systems. It engaged in the design of plans for educational improvement, and it offered its services to lend planning assistance to others. 13The fact that few students enroll on a full time basis in terms of class attendance does not reflect a lack of seriousness on the part of the students. Most students may be employed in addition to school attendance at rates that may exceed 44 hours per week. Employment may be necessary to support families in Central America and this does not diminish the seriousness of the students. 16 This administrative policy maybe interpreted as enrolling all students on a program basis instead of on an individual course basis as is found in the United States 2. The examination procedure is comparable to the traditional European system of examing students at their leisure and not necessarily at the conclusion of coursework. This administrative policy is utilized in many universities throughout the world. Instrumentation The initial inventory employed to gather the data for this, as well as other studies, was developed by the IIME staff. The principle inventory was the "pink sheet" located in Appendix A. The completed inventory was then transferred to a computer card for future use. These data were analyzed in a variety of ways for specific IIME publications, and were reduced to the extent that each section of each course taught in the five Central American universities during the academic year 1962-63 could be ascertained. From the initial bank of data, it was possible to modify the section analysis cards to form a new computer card which incorporates relevant measures of efficiency and effectiveness in addition to possible factors that could be affecting these institutional ratios. The result was a computer card designated as, ”Effieiency-Effectiveness Summary Card." 17 Procedure The efficiency-effectiveness summary card contained the following data: 1. The university, faculty, and department of the instructor giving final grades in each class. 2. A number to identify each professor. 3. The level of instruction of each course taught. For example, a course could be offered to those in their second year of study. This is a second level course. 4. The method of instruction of the course taught. In this way, it was possible to distinguish a course that is primarily lecture from one that is laboratory content from one that is a combination of lecture and laboratory instruction. 5. The hours of instruction a class received. This variable (H), refers to the total hours of instruction received by a class during a term by the professor giving the final grades. 6. The number of students enrolled in a class (E). 7. The number of students who presented themselves for examination in a course (Ex). 8. The number of students who presented themselves for examination immediately upon completion of the formal coursework (Exl). 9, The number of students who took the examination and 18 upon completion of the course and passed. (A1). 10. The total number of students who took the examination and were successful (A). 11. The instructional salary (level of expenditure) of the professor giving the final grades (C). 12. The number of classes taught by the professor during the 1962-63 year. With this information for each class, it was possible to develop all of the efficiency-effectiveness ratios men— tioned in Chapter 1. It was also possible to design an experiment to measure any existent relationships between the efficiency-effectiveness ratios (A/E, Ex/E, A/Ex, Al/E, AlExl, Exl/E) and: 1. Faculty load (HxE). This variable was derived by multiplying the hours of instruction by the number of stu- dents enrolled in a class. This variable was known as con- tact hours. 2. Level of expenditure (C). This figure represented the salary paid to the instructor giving final grades. Under conditions operant in Central America, the professor is paid by the Course, and (C) represented his salary for this class alone?!l Currency equalization will be discussed later in this lU'In all universities except University 5, some faculty were paid on a salary basis irrespective of the number of classes taught. The salary was divided evenly among the number of classes taught by a professor in this instance. 19 chapter. It will have a direct bearing on the meaning of - the level of expenditure. 3. Class size as expressed in enrollment (E). 4. Level of instruction of the course. 5. Method of instruction employed in the course. 6. Number of courses taught by the professor. 7. Relationships between any and all of the efficiency- effectiveness ratios. As an illustration, is there any re- lationship between institutional efficiency (Ex/E) and institutional effectiveness (A/E) in a given university? Design Using the ACT II (Analysis of Contingency Tables) pro- gram and the CDC 3600 at Michigan State University, bi- variate frequency distributions were developed for each of the five universities. A/E x Class Size E x Method of Instruction A/E x Level of Instruction A/E x Class Size A/Ex x A/Ex x Ex/E x Ex/E x Ex/E x A l/E £173 ><><>< Level of Instruction Method of Instruction Class Size Level of Instruction Method of Instruction Class Size Level of Instruction Method of Instruction The actual distributions were: A/E x Level of Expenditure A/E x Contact Hours A/E x A/Ex A/Ex x Level of Expenditure A/Ex x Contact Hours A/Ex x Ex/E Ex/E x Level of Expenditure Ex/E x Contact Hours Al/E x Level of Expenditure fi/E x Contact Hours fi/E x Exl/E 20 Al/Exl x Class Size Al/Exl x Level of Expenditure Al/Exl x Level of Instruction Al/Ex1 x Contact Hours Al/Exl x Method of Instruction Al/Ex1 x Exl/E Exl/E x Class Size Ex /E x Level of Expenditure Exl/E x Level of Instruction Ex%/E x Contact Hours Exl/E x Method of Instruction The ACT program performed the following operations on the designated tables: row and column means and standard devia- tions; percentages of each cell on the associated row, column and table totals; theoretical frequencies, cell contributions to table Chi square, table Chi square and degrees of freedom for test of independence; and product moment correlation coefficient. As an adjunct in the interpretation of data on the number of classes taught by a professor during the academic year, statistical means were computed for all of the professors in a university who taught one class, two classes...15 classes. Differences between means and standard deviations were then ascertained. In Summary, Chi square and mean differences were the techniques employed to determine any relationships between the isolated factors and the defined ratios. The utility of the model was determined by the relationship between efficiency- effectiveness ratios. Nature of Hypotheses The review of efficiency and effectiveness in Chapter 1 indicated that there are varying definitions that may be applied to these terms. Similarly, there is no theory that deals with 21 factors that may be affecting productivity in Central Ameri— can universities. While it is feasible to apply assumptions. from other nations and regions, there is no assurance that factors such as class size will affect production of success- ful students in precisely the same fashion. The resultant problem is that it was not possible to build statistical hypotheses that are grounded in prior research or in an existent body of theory. The problem was not insurmountable however. As Saupe states: ...Hypotheses are not absolutely necessary for research, even in doctoral research, to be respectable. While it is possible to argue that the significance (not statistical) of educational research can be measured by the extend to which it draws upon past research and a body of theory and contributes to it, it is still true that there is much virgin territory to be explored by educational researchers. The first study in an area can not be guided by research hypotheses, but it can result in the suggestion of fruitful hypotheses for further research efforts and thereby stimulate the development of a body of theory. The significance of this thesis lies in the generation of problem statements that culminate in the generation of hypotheses about Central American university efficiency and effectiveness. Joseph L. Saupe, A Plea for Order, College of Educa- tion Quarterly, (East Lansing: *Michigan State University, July, 1961), p. 4. If there is no theoretical scheme which applies to an area of investigation or if the topic for study represents a siz- able jump from such body of theory, the researcher may then have no specific hypothesis concerning the subject of his research. He has only a problem. He has no basis for guessing that there is not a difference among the variables he chooses to study. It would then be out of place for him to attempt to formulate hypotheses about the particular place of nature with which he is concerned. In keeping with the rather cogent opinion of Saupe, this thesis generates hypotheses about the following ques- tions: 1. What is the relationship between institutional effectiveness and level of expenditure, level of instruction, type of instruction, contact hours, class size and the num- ber of classes taught by the professor? 2. What is the relationship between institutional efficiency and level of expenditure, level of instruction, type of instruction, contact hours, class size and the num- ber of classes taught by the professor? 3. What is the relationship between instructional efficiency and level of expenditure, level of instruction, type of instruction, contact hours, class size, and the number of classes taught by the professor? 4. Is it possible to attribute any integrity to the theoretical postulations that have been asserted between the efficiency and effectiveness ratios? 16Ibid., p. 4. 23 Limitations of the Analysis While the limitations of the analysis were not extensive, they were rather important to the scope and possible outcome. 1. The focus of the analysis were those factors that can directly affect instruction within the univer- sities. No attempt was made to minimize the im- portance of factors of a non-instructional nature on the total efficiency and effectiveness of the institution, but these factors were not within the purview of present analysis. Therefore, the physical plant, administrative factors and other similar factors were not considered in the determina- tion of efficiency and effectiveness. The full level of expenditure per class was not always utilized because of the unique fashion in which each class may be divided among faculty members in Central America. In contrast to the system in operation in the United States, a pro- fessor may teach a class for only three class meetings and then another professor may teach for another period of time. Each person teaching a segment of a term, be it one hour/term or two hours/week, is paid at a rate that is commensurate with his title and his knowledge. It is therefore possible for 20 people to deliver the content of of a course given to a class and then to receive 24 20 different salaries. For the purposes of this analysis, the professor who assigned final grades (A), was the only person whose salary was designated as being significant in terms of the factors that affect the various ratios. Not all factors that could be affecting instruction were controlled within the framework of analysis. For example, it is entirely possible that the lack of ventilation in a class room may have a deleterious effect on the professor to the extent that it adversely affects his presentation. Ultimately, the efficiency- effectiveness ratios also could be affected adversely. Similarly, there are other factors that could be controlled. The intent in presenting relationships between efficiency-effectiveness and six factors was to illustrate what might be done with the efficiency-effectiveness model. No claim is made that these factors represent a significant deter- mination of comprehensive factors affecting efficiency and effectiveness in instruction in Central America. The influences of budget and societal influences were not considered in measuring the efficiency and effectiveness of the institution. Wortman considered the cost/productivity relationships in these univer- sities, and the societal influences may be too dis-' parate to quantitatively measure in this thesis. 25 5. Measures of efficiency and effectiveness and factors affecting these measures are categorized for sta- tistical analysis. The range of each category was arbitrarily delimited and no intimation should be made that these divisions are in any way rigid or significant. An entirely different method of dividing data would be equally useful. In summary, the delineated measures of efficiency and effectiveness were matched with certain factors indigenous to to the internal structure, organization and philOSOphy of the institution as limited above. Currency Equalization The countries in which the five Central American national universities are located, utilize different national currencies that are not comparable. In order to add meaning to these disparate entities, each level of expenditure (C) was con- verted into United States dollars at the official exchange rate as of 1962. One United States dollar is equal to: 6.625 Colones in Costa Rica 2.50 Colones in El Salvador 1.000 Quatzales in Guatemala 2.00 Lempiras in Honduras 7.1 Cordobas in Nicaragua The equalization of currency did not in any way suggest that interuniversity comparison of level of expenditure could be made without extensive qualification. The salaries paid as professors and the cost of living vary between countries ' . 9‘! 5H ’1. “AL -j th fill ”are I“? .1 I'll, I , § ‘ P‘ 'fl'u, ij‘ a). .1.5“‘. 26 and thereby necessitate a complete elaboration of inherent conditions before any trans-country comparisons would be meaningful. The reason for converting local currencies into United States dollars was to provide a familiar frame of reference for operation. It was this familiarity that is sought in conversion. In a concommitant fashion, it was recognized that there was a wide salary fluctuation between departments within the same institution. This is not an occurrence that is unique to Central America however, as the same practice is found in most universities in the United States. The law of supply and demand was applicable throughout as was the concept of higher salaries for particular professions like law and medicine. The implication to be gathered is that comparison of faculties between universities and also within universities would be misleading and may be discrepant. The local idiosyn- cracies and internal differences cannot be overlooked and as a result, local qualification is necessary before meaning- ful comparisons can occur. {..lclbu. ‘H V CHAPTER III EFFICIENCY-EFFECTIVENESS INTER-RELATIONSHIPS In this chapter the inter-relationships between the efficiency-effectiveness ratios are examined by university and also by ratio comparisons within these universities. The importance of inter-relationships cannot be minimized because the model of efficiency-effectiveness is predicated upon a series of assumptions that must be upheld if the model is to have practical utility. If no relationships were to be found when the model was applied, the model would have little utility in measuring the effects of factors upon the individual ratios. Therefore, inter-relationships between particular ratios are essential in determining the applicability of the theore- tical model. A To examine model applicability, efficiency-effectiveness data were distributed by university. Utilizing Chi square, it was possible to determine if a relationship between ratios did exist, and by utilizing mean scores derived from the Chi square framework, it was possible to determine the patterns of the relationship. Institutional Effectiveness (A/E) and Institutional Efficiency (Ex/E) Assumption. If the theoretical model is accurate, there should be a relationship between institutional efficiency as measured by (Ex/E), and institutional effectiveness (A/E). 27 28 Specifically, as Ex/E increases, A/E also should increase. The theoretical rationale behind this assumption is evi- dent. The number of examined students (Ex) normally cannot exceed the total class enrollment (E). However, (Ex) can approach (E) if most of the enrolled students (E) present themselves for examination (Ex), the closer that (Ex) comes to (E), in absolute numbers, the more closely (Ex) comes to being interchangable with (E). To illustrate in more concrete fashion, suppose that (Ex/E) (the measure employed for institutional effectiveness) in a given class is 4/10. This indicates that if all of those examined were to be successful (A), the maximum institutional effectiveness would result in an A/E ratio of 4/10. However, if for the same class the Ex/E ratio were to be 8/10, the maximum (A) value is now 8, and the maximum institutional effectiveness ratio of A/E would be 8/10. It is not necessary for maximal values to be employed in order to expect that the assumption will be supported. For example, in a class of 10 students, only four present them- selves for examination. Suppose that 50% of those examined were successful. Clearly, the effectiveness ratio (A/E) is 2/10. If eight students presented themselves for examination, the effectiveness ratio (A/E) would be 4/10. Data from the five Central American universities that were considered support the assumption that an increase in institutional efficiency (Ex/E), results in an increase in 29 institutional effectiveness (A/E). These data are presented in Table 3.1. TABLE 3.l--PERCENT 0F INSTITUTIONAL EFFECTIVENESS (A/E) BY LEVEL OF INSTITUTIONAL EFFICIENCY (Ex/E) Level of Institutional Efficiency Chi square Ex/E Ex/E Ex Significance University 0-50% 51-80% 81+% Level 1 25.5%* 49.9% ' 82.6% .01 2 25.5 56.7 92.5 .01 3 25.5 47.9 78.4 .01 4 25.5 49.1 89.2 .01 5 25.5 56.3 87.7 .01 , *Because of computer program limitations, raw scores were transformed into coded scores in order to process the data. As a result the minimum percentage that can be shown is 25.5% and the maximum is 95.5%. The pattern is not affected, however. In Table 3.1, Ex/E values are expressed in three groups: 1) 060% 2) 51-80% 3) 81+?» The A/E values were examined within each category of Ex/E, and are expressed in percent. Without exception, the percent of institutional effective- ness (A/E) increases as institutional efficiency (Ex/E) increases. For example, in University 1, A/E is 25.5% when the Ex/E is less than 50%. A/E rises to 49.9% when the Ex/E value is be- tween 51 and 80 percent and rises to 82.6% when Ex/E is greater than 81%. Institutional Effectiveness (A/E) and Instructional Efficiency (A/Ex) Assumption. If the model is applicable, there also should be a relationship between instructional efficiency (A/Ex) 30 and institutional effectiveness (A/E). Theoretically, as A/Ex increases, A/E should increase also. The rationale for this assumption is encompassed once again in the values of examined (Ex) and enrolled students (E). The successful student value (A), cannot be larger than examined value (Ex). However, it is theoretically possible that (A) could equal (Ex) if all those who were examined passed the examination. It becomes apparent that as A/Ex increases, E, being a fixed value, cannot fluctuate in the A/E ratio. If successful students (A) approaches examined students (Ex) in the A/Ex ratio, institutional effectiveness (A/E) will necessarily be high. As an illustration, let (A)=5, (Ex)=8, and (E)=10. In this instance, A/Ex=5/8 and A/E=5/10. If the (A) value were to be raised to 7, the A/Ex value equals 7/8 and A/E equals 7/10.‘ Hence, in theory, as A/Ex increases, A/E should in- crease also. To confirm these relationships, data from the five Central American universities were examined. Results are represented in Table 3.2 TABLE 3.2--PERCENT OF INSTITUTIONAL EFFECTIVENESS (A/E) BY LEVEL OF INSTRUCTIONAL EFFICIENCY (A/Ex) Level of Institutional Effectiveness Chi Square A/Ex A/Ex A Ex Significance University 0-50% 51-80% 81+% Level 1 29.5% 52.7% 82.9% .01 2 51.1 74.8 89.2 .01 3 29.1 52.3 75.7 .01 4 27.9 58.3 87.1 .01 5 45.5 72.7 89.5 .01 increi :0? e: 1; rm .1 he 1.; ani re ...tl -1 (:11. H Q ion 1‘ v’ ‘. (4 tion I): L .1 those r‘] H IRS: 31 In each university as instructional efficiency (A/Ex) increases, institutional effectiveness (A/E) increases also. For example, in University 2, when the ratio of instructional efficiency (A/Ex) is less than 50%, the institutional efficiency is 51.1%. However, when A/Ex is 51-80%, A/E becomes 74.8%, and reaches 89.2% when A/Ex is greater than 81%. The analysis of data thereby supports the assumption that as instructional efficiency (A/Ex) increases institutional effectiveness (A/E) increases also. Institutional Efficiency (ExZE) and Instructional Efficiency (A/Ex) Assumption. Theoretically, there should be no rela- tionship between institutional efficiency (Ex/E) and instruction- al efficiency (A/Ex). The rationale for the preceeding assumption can be illus- trated when the two ratios are analyzed. Ex/E is a ratio of those presenting themselves for examination (Ex), divided by those initially enrolled in a class (E). A/Ex represents those who passed (A) divided by those examined (Ex). In the Ex/E ratio, Ex is the varying quantity, and in the A/Ex ratio, Ex is the fixed quantity. Seeing that the common variable is not free to vary in both ratios, it does not appear likely that the two ratios should evidence a patterned relationship. Any relationship that is evidenced would indicate that an extraneous factor may be operating on both ratios. These re- lationships were examined in the five Central American insti- tutions and are presented in Table 3.3. 32 TABLE 3.3--PERCENT 0F INSTRUCTIONAL EFFICIENCY (A/Ex) BY LEVEL OF INSTITUTIONAL EFFICIENCY (Ex/E) Level of Instructional Efficiency Chi Square Ex/E Ex/E Ex/E Significance University 0-50% 51-80% 81+% Level 1 75. 1% 78.1% 84.7% .01 2 92.8 90.4 90.1 NS 3 74-5 79.6 77.5 NS 4 79.6 84.1 88.6 .01 5 94.0 89.5 91.0 NS The analysis of universities indicated a most interesting series of patterns. In universities 1 and 4 there is a con- stant increase in A/Ex as Ex/E increases. For example, as Ex/E rises from 0-50, 51-80 and 81+%, university 1, A/Ex rises from 75.1, 78.1 and 84.7%. The same relationships is evidenced in university 4. For these universities it appears that the higher the holding power of the class, as evidenced by Ex/E the higher the percentage of students who will pass (A). In raising possible reasons for the pattern noted in universities l and 4, it becomes necessary to speculate based upon current operating procedure in Central America. Because all students are registered as full time students, with a full load of courses, even though they may actually be attending only one or two classes, is it possible that those classes that are fully attended and examined are justly rewarded for their attendance? This possibility would appear to gain sup- port in the university data. The other universities exhibit a diversity of patterns 33 even though the Chi square values are not significant at the .05 level. University 2 exhibits a steadily declining percent in instructional efficiency (A/Ex) as institutional efficiency (Ex/E) increases. University 3 has the highest A/Ex percentage when Ex/E is 51-80%. University 5, on the other hand, shows a high A/Ex percentage when Ex/E is below 50%, a decline when Ex/E is 51-80%, and a slight increase when Ex/E is high (81+%). With a variety of patterns exhibited in the five univer- sities, (including contrary expectations) it could be stated that the data fail to support the assumption that institutional efficiency (Ex/E), and instructional efficiency (A/Ex) are un- related. Because of the nature of the efficiency ratios, this lack of support may indicate that an extraneous factor is apparent that is causing theoretically unwarranted patterns. This possibility is further explored in Chapter 6. Internal Consistency Another method of measuring the utility of the model is to compare the mean scores of A/E, Ex/E and A/Ex to insure that they follow a logical and meaningful pattern that is based upon the theoretical statements surrounding the model. The data in Table 3.4 indicate the mean scores for A/E, Ex/E and A/Ex for all classes in each university. TABLE 3.4--PERCENT 0F INSTITUTIONAL EFFICIENCY (Ex/E), INSTRUC- TIONAL EFFICIENCY (A/Ex), AND INSTITUTIONAL EFFECTIVE- NESS (A/E) BY UNIVERSITY University Ex/E A/Ex ALE 1 82.0% 80.8% 70.6% 2 90.1 90.7 86.8 a 77.5 gg.8 63.5 87.7 .3 80.8 5 _ _g 90.7 84.7 83.2 34 If the model is workable, the A/Ex ratio of instructional efficiency should be greater than the A/E ratio of institu- tional effectiveness because there is no university in which the examined value (Ex) equals the enrolled value (E). As evidenced by all universities, A/Ex is indeed larger than A/E.‘ In university 5, the difference between A/Ex and A/E is 1.5 percentage points. This would indicate that the Ex/E value should be high. The reason for this similarity is that the closer Ex is to E as represented by Ex/E, the smaller the difference between A/Ex and A/E. This pattern can be seen in university 3 where there is a low Ex/E and a 14.3 percentage point difference between A/Ex and A/E. Lastly, the A/E value for any university should be smaller than the Ex/E value because in no university does (A) equal (Ex). As evidenced in each university, this is in- deed the case. Time in the Efficiency-Effectiveness Ratios Time is a relevant factor in the production of educated persons, particularly in the developing nations. For example, in university 3, the average student invests 13.32 years to complete his required studies. In some fields, as many as 17 to 30 years are invested before completing curricula of 5 or 6 years duration.17 There is an apparent advantage to the university to examine students and to permit them to become successful at the earliest opportunity, consistent with good 17-. p. it., p. 2 35 .practice. The longer the time needed to educate a person, the smaller the productive contribution of that individual. One factor affecting the period of time to complete courses is the examination procedure itself. In Central America, it is not always necessary for the student to take his final examination in a course immediately upon conclusion of the coursework. The examination may be deferred for a period as long as 18 months under certain conditions, or-- as in university 4--the examination procedure may be divorced completely from course instruction. In this context, all of the efficiency-effectiveness ratios mentioned heretofore in this chapter have not differentiated students who took the examination immediately upon conclusion of the course from those who waited as long as 18 months. A second efficiency- effectiveness model can thereby be applied that encompasses only those students who are examined and pass at the earliest opportunity. Institutional Effectiveness (A/E) and Institutional Efficiency (Exl/E) Assumption. Let (Exl) represent the student who pre- sents himself for examination at the earliest opportunity. Let (A1) represent those students whpsuccessfully complete the examination at the first opportunity. As institutional efficiency (Exl/E) increases, institutional effectiveness (Al/E) increases. This assumption is based upon the deri- vation of the (Exl) and (A1) values. Those examined immediately (Exl) must come from the quantity (E). The closer that (Exl) 36 is to the value (E), the more students that are potentially successful (A1). The following data from the five Central American uni- versities indicate that as Exl/E increases, Al/E increases. TABLE 3. 5--PERCENT 0F INSTITUTIONAL EFFECTIVENESS (A l/E) EN LEVEL OF INSTITUTIONAL EFFICIENCY (Ex 1/E)l Level of. Institutional Efficiency Chi Square Exl/E Exl/E Exl/E Significance University 0-50% 51-80% 81f¢ Level 1 25.5% 47. 9% 86.2% .01 2 25.5 53- 9 89.2 .01 3 25.5 44.7 85.7 .01 4 25.5 49.5 89.2 .01 5 25.5 52.3 84.7 .01 Using university 1 as an illustration, when Exl/E is less than 50%, the mean value for Al/E is 25.5%. However, when Exl/E lies between 51% and 80%, Al/E rises to 47.9%. Simi- larly, when Exl/E is 81+%, Al/E averages 86.2%. The pattern is without exception in the five Central American universities. It would appear therefore that the data support the assumption: as institutional efficiency (Exl/E) increases, institutional effectiveness (Al/E) increases. Institutional Effectiveness (Al/E) and Instructional Efficiency (Al/Exl) Assumption. AS Al/Exl increases, Al/E increases. The rationale underlying this assumption is the same as that utilized in explaining the relationship between A/Ex and A/E. Since A1 cannot exceed Exl, the closer that Al comes to Exl, 37 the higher the absolute quantity of A1. Since enrollment (E) is a fixed value, the amount of successful students (Al) will determine the percentage of institutional effectiveness (Al/E), just as it was observed to determine the percent of instruction- al efficiency (Al/Exl). Since (Exl) and (E) are fixed, an in- crease in (A1) should produce an increase in both ratios (Al/Exl and Al/E). The data from the five universities (Table 3.6) confirm the relationships. TABLE 3.6--PERCENT 0F INSTITUTIONAL EFFECTIVENESS (A /E) BY LEVEL OF INSTRUCTIONAL EFFICIENCY (Al/Exl) Level of Instructional Efficiency 1/ /E 1/ Chi Square A Ex A x1 A Ex Significance University 0-50%1 51-80% 81+% 1 Level 1 25.5% 39.9% 79-6% ~01 2 25.5 57.1 87.4 .01 3 25.5 33.1 70.3 .01 4 25.5 42.3 84.7 .01 5 25.5 49.5 82.6 .01 With each successive increase in instructional efficiency, (Al/Exl) a corresponding increase may be observed in institution effectiveness (Al/E). Variation among the universities may be observed in the amount of increase in Al/E, however. For example, in university 2, an AL/Exl score of 51-80% corres- ponds to an average Al/E value of 57.1%. However, in univer- sity 3, in the corresponding category, the AI/E value is 33.1%. The data support the assumption that as instructional efficiency (Al/Exl) increases, institutional effectiveness (Al/E) increases also. 38 Institutional Efficiency (Exl/E) and Instructional Efficiency (A1/Exl) Assumption. There is no consistent relationship be- tween institutional efficiency (Exl/E) and instructional efficiency (Al/Exl). As stated previously with regard to the relationship between Ex/E and A/Ex, the variables within the variables within the ratios would not theoretically appear to be dependent one upon the other to the extent that a value in One formula could dictate a pattern in the other ratio. The Exl value is the common value in the ratios. The (E) value is previously established, so that the ratio of Exl/E is dependent upon the Ex1 value. But in the second ratio, A1 is the fluctuating variable and Exl is already established before the A1 value is known. It would appear that there should be no relationship between the values of institutional efficiency and instructional efficiency. As illustrated in Table 3.7 the data belies the theory. TABLE 3.7-~PERCENT OF INSTRUCTIONAL EFFICIENCY (A Exl) BY LEVEL OF INSTITUTIONAL EFFICIENCY (Ex E) Level of Instructional Efficiency ’Chi Square Exl/E Exl/E ExL/E Significance University 0-50% 51-80% 81+% Level 1 55.1% 66.1% 82.3% .01 2 89.5 87.7 89.2 .01 3 48.7 58.3 72.7 .01 4 74.7 75.1 86.8 .01 5 79.6 79.0 85.0 .01 39 In universities 1,3 and 4 there is a steady increase in instructional efficiency (Al/Exl) as institutional efficiency (Exl/E) increases. The percentages in this pattern vary widely, however, In university 3, when Exl/E is less than 50%, the Al/ Ex1 ratio is 48.7%. If hypothetical figures were used to illustrate this university based upon mean scores, the data indicates that an (E) of 100 yield of Ex value of 50% or 1 less. Assuming that the maximum value of 50% is employed, the Exl/E ratio would be 50% and the maximum Ex1 value would be 50. From these 50 students, the average Al/Ex1 would be 48.7%. This would mean that 24.4 students would be the maxi- mum expected to pass (A1) from an initial enrollment of 100. A different pattern is evidenced in universities 2 and 4. The Al/Ex1 percentage deClines in the middle Ext/E category. In university 2, for example, when Exl/E is less than 50%, Al/Exl is 89.5%, but when Exl/E is 51-80%, Al/Ex1 is 87.7%. As Ex increases to 81+%, the Al/Exl value increases to 89.2%. 1 The pattern exhibited by universities 2 and 4 raises a question as to why Al/Exl is high when Exl/E is low. Is it possible that in those instances where a high percentage of class enrollment (E), has been lost during a term, an attempt is made by the professor to salvage a greater percentage of those that have remained (Al/Exl)? If this is not the case, could a low Exl/E indicate that very few enrolled students actually were attending classes? If this be true, is Al/Exl a more realistic indication of class attendance than the Exl/E ratio? Because all students are enrolled as full time students 40 in Central America, the possibility does exist that a low Exl/E ratio may incorporate large numbers of "phantom" students as enrollees, whereas Al/Exl may reflect an Exl that more closely approximates the actual attendance.18 An analysis of data indicates that the assumption of no relationship between Exl/E and Al/Exl is not supported. Two distinct patterns emerge that are dissimilar in nature. Be- cause Exl/E and AL/Ex1 in fact seem to be related, the model is in no way affected. Instead it must be recognized that there may be a third factor that is affecting these ratios that was unanticipated. When time is an important factor in student success, it appears that the ratios_Al/E and Al/Exl and Exl/E have utility in Central America. Each ratio is able to isolate at a speci- fic point in time the efficiency and effectiveness in any class within an institution. It would thereby appear that the model, when applied to specific data is workable and useful. Data Differences Between Models One advantage in using the two models of efficiency and effectiveness is that it becomes possible to distinguish ratio differences between A/E and Al/E; Ex/E and Exl/E; A/Ex and Al/Exl. The importance in analyzing these comparisons stems from the very nature of the Central American need for an edu- cated populance. By comparing differences between those who 18Attendance records are not consistently maintained by the universities, hence the question may not be resolved directly. 41 passed at the first Opportunity with those who passed at a later date, it is possible to ascertain the percentage of students who are extending the time necessary to pass a course. To more clearly illustrate this advantage of the models, Table 3.8 contains the A/E and Al/E mean scores for each uni- Versity. The A/E value indicates the percentage of success- ful students (A) from those enrolled (E). Al/E represents those enrolled students who were successful at the first opportunity. TABLE 3.8--INSTITUTIONAL EFFECTIVENESS (A/E and Al/E) DIFFER- ENCES BY PERCENT University A/E A1/E Difference 1 70. 66.1 4. 2 86.8% 82.? 4.3% 3 63-5 53.9 9.6 4 80.8 77.5 3.3 5 83.2 73.3 9.9 By university standards in the United States, all the mean scores are relatively low. Without assigning a value to the rightness or wrongness to the pass-fail philosophy in Cen- tral America, it is possible to note that in universities 3 and 5, a large percentage of students fail to pass their examina- tions at the first opportunity. In university 3, the differ- ence between the A/E and the Al/E mean is 9.6%. This indi- cates that 9.6% of all successful students were retaking an examination that they had previously failed, or had not taken at the earliest opportunity. When the A/E percentage at univer- u~..__‘.._ it" for flaw 1 VV“ 0. 1’: J'\. ’r M“ «O E 42 sity 3 of 63.5% is examined, it appears that on the average, 36.5% must retake the examination or take it at a later date before they can ever hope to move on the more advanced courses. It has been noted also that students in this category account for only 9.6% of the successful total.- The time spent in achieving successful completion of courses and graduation is greatly lengthened as can be deter- mined rather dramatically in the preceeding examples. The full impact was elucidated in a case study by the IIME staff at the University of San Carlos of Guatemala. By utilizing the Index of Academic Achievement which is an expression of the number of calendar years spent by a student to complete one year of course work, the authors state, "If one adopts the definition that the only 'successful' full time student is the student who completes his courses in no more than the number of years scheduled in his course of study, there were only 466 successfu1.fu11 time students among the 5,036 who reen- rolled in 1963; 9.25 percent of the total".19 Similar comparisons can be made between Ex/E and Exl/E. These ratios exhibit the percentage differences between those being examined at the first opportunity and those being re- examined or prolonging the period before taking the examination. University differences are noted in Table 3.9. 1911MB staff report, op, cit., p. 47. 43 TABLE 3.9--PERCENT DIFFERENCES IN INSTITUTIONAL EFFICIENCY (Ex/E and Exl/E) University Ex/E Exl/E Difference 2 90.1 89.2 .9 3 77.5 66.1 11.4 4 87.7 84.7 3.0 5 90-7 83.2 7-5 The percentage difference represents those students who for one reason or another did not take the examination at the first opportunity, or were not successful at the first oppor- tunity. AS illustrated, there is wide variation in this figure between universities. It should be noted that the higher the percentage difference, the more students who have lengthened the amount of time necessary to complete a single course. When it is recognized that the percentage differences represent the average for each university, it becomes evident that universities l, 3 and 5 have a substantial percentage of students that are progressing, perhaps, unnecessarily slowly. This would tend to support the low rate of successful full time students noted in the IIME staff report. Probably the most significant comparison that can be made between the two models involves the means of A/Ex and Al/Exl' These ratios represent the percentage of those passed (A) out of those examined (Ex) and the percentage of those who passed at the first opportunity (A1) out of those who were examined at the first opportunity (Exl). If Al/Ex1 were signi- 44 ficantly lower than the A/Ex ratio, this might indicate that preference is given to those who have already failed the ex- amination or had foregone examination at an earlier time. ./However, the figures in Table 3.10 do not support this pos- sibility. TABLE 3.lO--INSTRUCTIONAL EFFICIENCY (A/Ex and Al/Exl) DIFFER- ENCES IN PERCENT University A/Ex Al/Exl Difference 1 80.8% 81.7% .9% 2 90.7 90.4 .3 3 7g.8 76.9 .8 4 8 .3 88.6 .3 5 84.7 85.3 .6 The greatest difference between A/Ex and Al/Ex1 is .9 of 1% in university 1. All universities exhibit very stable per- centages for both groups. Therefore, the data indicate no noticeable difference in percentages between those passing and those examined at either the first opportunity or at a later date. Moreover, the percentages of those passing to those ex- amined are quite high at either time. In university 2, for ex- ample, over 90% in either category will be successful. If nine out of ten of those examined are to be successful, it would appear that one of the problems of Central America is how to increase the absolute numbers of students who present themselves for examination. Once students present themselves for examina- tion (Ex or Exl), the probability of being successful (A or Al) is relatively high. 45 Summazy The purpose of this chapter was to present substantive data from the five Central American universities to support the constructed models of efficiency and effectiveness. Inter- relationships between ratios were suppOrted for all theoreti- cal assumptions with the exception of the two relationships between institutional efficiency (Ex/E) orCExl/E) and insti- tutional effectiveness (A/E) or (Al/Exl). In this instance, conflicting and unanticipated patterns were observed. The cause of the conflict may be a third factor or cluster of factors that are extraneous to both ratios. The utility of the model is not affected by this unanticipated pattern. The models isolate efficiency and effectiveness at selected stages of any course, thereby enabling an observer to ascertain at which points inputs are being reduced. By comparing ratios from one model to the other by univer- sity, it is possible to determine differences between students taking courses for the first time and those who are not. Be- cause there is little difference in either measure of instruc- tional efficiency within any of the universities, it would appear most advantageous to attempt to delineate those factors that may be limiting Ex in absolute numbers. For if a student is examined, it appears that the probability of being success- ful is relatively high. Moreover, there would seem to be no real justification to continue the multiple examination pro- cedure itself. Each university could gain materially in insti- tutional efficiency by limiting itself to a single exam at the conclusions of each course. CHAPTER IV FACTORS RELATED TO INSTITUTIONAL EFFECTIVENESS Six factors related to institutional effectiveness (A/E and Al/E) are examined in this chapter. As indicated in Chapter 1, these specific factors were selected because they fall within the ability of the institution to control administratively. According to the model, institutional effectiveness is the extent to which an intended outcome of a course of action becomes a reality. Within the context of higher education, institutional effectiveness is the extent to which an intended output (A or A1), becomes a reality from any given input (E). It appears axiomatic that the effectiveness of an insti- tution will be affected by factors that are inherent in the dynamic operation of the institution. By isolating Specific factors--and the effect that they have upon the effectiveness of the institution--it becomes administratively feasible to effect change that will make the effectiveness of the insti- tutional more congruent with institutional needs and objectives. Within the framework of the efficiency-effectiveness model, it is possible to spread the ratio of institutional effectiveness (A/E or Al/E) against the six factors using a Chi square technique. By comparing mean scores within each Chi square category, it is then possible to ascertain inherent 46 47 patterns in institutional effectiveness that might be re- lated to a given factor. This process will be illustrated by university and also by selected faculties within each university. The Specific factors to be applied in this chapter for relationships to institutional effectiveness are: 1. Level of educational expenditure, as determined by the salary paid to the instructor who gives the final grades to the students. 2. The size of the class, as determined by class enrollment (E). 3. The level of instruction at which the class is offered (i.e., first year, second year...eighth year.). 4. The principal method of instruction of the class (e.g., lecture, laboratory or combination). 5. The number of hours that professors are in instruc- tional contact with students, as measured by the hours of instruction H; multiplied by the number of students enrolled (E . (HxE). 6. The number of classes taught by the grading instructor. Data Analysis by University Data for each of the six administratively controllable variables were compared with the institutional effectiveness (A/E) values in each university. When Chi square values were determined in the comparisons, differences significant at the .01 level of confidence were found for each factor in at least one of the universities. Significant relationships 48 are summzrized in Table 4.1.20 TABLE 4.1--FACTORS RELATED TO INSTITUTIONAL EFFECTIVENESS (A/E) (X=.01 level of confidence)* Level of_ CIass Level of Method 6?: A/E Expendi- Enroll- Instruc- Instruc- University ture ment tion tion HxE 1 X X X X X 2 X X X 3 X X X X 4 X X X 5 X X X *Data for the five universities is found in Appendix B Level of Expenditure may be observed to be significantly related to effectiveness (A/E) in four institutions. Class enrollment is clearly related in each university as is (HxE), the number of contact hours with students. The other fac- tors are related in varying degrees. The table gives ample evi- dence of university differences, (iifii’ administratively con- trollable factors are differently related among the five universities.) The effects of each factor on institutional effectiveness (A/E) is discussed in subsequent sections of the chapter. 20Data on the number of classes taught by a professor were not subject to Chi square comparison and are not in- cluded in Table 4.1. Mean comparisons, which are not sub- ject to levels of confidence, were the analysis employed on the number of classes taught. Relationships were analyzed in subsequent portions of this chapter. 49 Level of Expenditure and Effectiveness (A/E) The relationship between institutional effectiveness (A/E) and level of expenditure is illustrated in Table 4.2. Five categories of expenditure are listed: $1—250, $251-500, $751-1000, $1001+. TABLE 4.2--INSTITUTIONAL EFFECTIVENESS (A/E) BY LEVEL OF EXPENDITURE (In Percent) A/E A/E* A/E A/E A7E Univer- Univer- Univer- Univer- Univer- Cost sity 1 sity 2 sity 3 sity 4 sity 5- 1-250 78.7% 90.7% 88.3% 91.0% 83.8% 251-500 67.3 85.6 63.5 88.0 83.5 500-750 69.7 85.0 60.3 74.2 83.5 751-1000 69.4 85.0 53.5 69.2 83.2 1000+ 70.9 82.6 65.1 74.5 77.8 wChi Square Significance Level .01 .01 .01 .01 .Ol NS: Not significant at .05. In universities l, 3, and 4 it should be noted that the highest effectiveness value (A/E) occurs when the level of expenditure paid to the grading professor is lowest. In University 1 when the salary paid is less than $250, effective- ness (A/E) is 7.8 percentage points above the next highest effectiveness ratio. In university 2, a salary less than $250 is 5.1 percentage points above the second highest effec- tiveness ratio. In university 3, a salary of less than $250 corresponds to an effectiveness ratio that is 23.2 percentage points above the next highest effectiveness ratio. In uni- versity 4, in the same category, the effectiveness ratio is 50 3.0 percentage points above the second highest category. The thesis could be advanced that: As the level of expendi- ture increases, the effectiveness ratio decreases. However, this tenet is not supported in the data. In fact, other than evidence that the lowest paid instructor has the highest effectiveness ratio, there is no consistent pattern. In universities l, and 3 the second highest effectiveness ratios occur when professors were paid over $1000 U.S. equivalents for a class. In universities 2 and 4, $251-500 dollar equiva- lents has the second highest effectiveness ratio. Because the highest effectiveness ratios in four univer- sities are found in the lowest category of expenditure, a variety of questions can be raised. Are the youngest, newest instructors being paid the least amount of money? Are these instructors more responsive to student needs and problems? Is a high effectiveness ratio in low cost courses an indica- tion of low instructor interest in discriminating the success- ful student from the unsuccessful? Are low cost courses con- centrated in selected faculties that have low class size? The answer to these questions may be found in the Central American practice of utilizing specialists or "practitioners" from the surrounding community to teach one or two classes. These practitioners may have no commitment to the evaluation process and therefore do not follow the usual pattern of the more established and permanent faculty. Based upon the analysis of data, the hypothesis can be generated that the salary level of the grading instructor is 51 not a compelling factor in determining institutional effective- ness. Indeed, institutional effectiveness may be achieved at different salary levels.21 Data Analysis by Selected Faculty In the four universities (1,2,3, and 4) where a relation- ship was noted between level of expenditure and institutional effectiveness, the faculties of Economics, Education, Pharmacy, Engineering, Dentistry, Medicine and Law were analyzed indivi- dually. In light of the noted relationship within these universities that a salary of 1-250 dollar equivalents is re- lated to the highest effectiveness ratio, Table 4.3 of level of expenditure figures within selected faculties and the category that exhibited the highest effectiveness (A/E) with .01 significance are delimited. The selected faculties do not comprise the total faculties within any of the universities. Rather, they represent those faculties that enroll the majority of the students in the institution and are somewhat similar across institutions. 21One should not conclude that lower salaries produce higher productivity. Rather, the Central American adminis- trators would be well advised to determine why low paid in- structors are highly productive in four institutions and higher salaried professors in the fifth. 52 TABLE 4.3--LEVEL OF EXPENDITURE CORRESPONDING TO HIGHEST EFFECTIVENESS (A/E) (By Selected Faculty) UnIVer- Econ- EdUca- Phar- Engineer- DentIs- *Medi- sity, omics tion macy ing try cine Law 1 NS NS NS NS NS NS NS 2 NS NS NS NS NS NS 3 NS NS NS $251-500 NS $l-250 NS 4 NS NS NS NS NS NS = not significant at .05 level of confidence. Significant relationships are observable in two of the selected faculties of university 3. The medical faculty did have the highest effectiveness ratio (A/E) when the level of expenditure was lowest (1-250 dollar equivalents). En- gineering had the highest effectiveness ratio at cost level of $251-500. Based upon the selected faculty data, it appears unwise to hypothesize that the highest effectiveness ratio occurs in those classes that pay low salaries. The lack of significant data is a hindrance in this instance. To generalize from two selected faculties within one university is not logical. Therefore, it appears unwarranted to judge the utility of the generated hypothesis that low level of expenditure in salaries may yield high institutional effectiveness (A/E) in selected faculties. Level of Expenditure and Effectiveness (Al/E) The relationship between institutional effectiveness (Al/E) and level of expenditure is reproduced in Table 4.4. 53 TABLE 4. 4--PERCENT OF INSTITUTIONAL EFFECTIVENESS (A l/E) LEVEL OF EXPENDITURE Institutional Effectiveness Al/E A E' Univer- Univer- Un ver- Univer- Univer- Cost sity 1 sity 2 sity 3 sity 4 sity 5 1-250 75 7% 86. 8% 86.8% 91.0% 73.9% 251-500 639 817 55.1 85.6 73.9 501-750 63.1 82. 6 49.9 71. 2 71.2 751-1000 66. 7 79.3 45.9 62. 3 77.8 1001+ 64. 7 76.0 51.5 694 71.8 Chi Square Significance Level .01 .01 .01 .01 NS NS = not significant at .05 level of confidence Universities 1, 2, 3, and 4 exhibit the pattern noted in A/E and level of expenditure in which the 1-250 dollar equivalent has the highest institutional effectiveness ratio. The difference in percent between the highest effective- ness ratio and the second highest ratio ranges from a low of 5.1% in university 2 to 41.7% in university 3. Once again it appears unwarranted to state that patterned relationship exists beyond the fact that the lowest cost in instructional salary appears to yield the highest effectiveness ratio. When level of expenditure and institutional effectiveness (Al/E) relationships are reduced to selected faculties, two selected faculties in university 3 achieved the desired significance level and in both faculties (Engineering and Medicine), the $1-250 level of expenditure did correspond to 54 the highest effectiveness ratio as measured by Al/E. However, the results observed in the small number of significant facul- ties would not appear to justify generating a hypothesis that relates level of expenditure and institutional effectiveness. Conclusions and Qualifications Regarding Level of Expenditure and Institutional Effectiveness Institutional effectiveness data (A/E and Al/E), when spread by level of instructor salary, exhibit a fairly con— sistent pattern.- With the exception of university 5, (A/E) and University 5 (Al/E), the lowest cost in dollar equivalents of $l-250 provided the highest institutional effectiveness ratios. However, when these data were further divided into selected faculties, this pattern was not evidenced in sufficient number to warrant hypothesis generation. A possible explanation for this apparent inconsistency may lie in the fact that the selected faculties do not re- present the entire faculty of any university. However, the faculties selected were those with sizeable enrollments and a fairly extensive curriculum offering. It might thereby be possible that the minor fields of study may have a higher ratio of institutional effectiveness because of smaller classes while lower salaries are paid to the professors because of the small student demand. An analysis of all faculties in an institution would provide additional evidence for this inconsistency. Consequently, there is no conclusively es- tablished relationship between level of expenditure and insti- tutional effectiveness measures (A/E and AL/E). m: \ Sr‘ not «an 7K\ PU hxv 2.. e .. 1A 3 \ .IV .v‘ AV 1 l \ 55 Class Size and Institutional Effectiveness To examine the relationship between class size and institutional effectiveness (A/E and Al/E), the data were divided for each university and for selected faculties with- in each university. to control class sizes. Therefore, for administrators to consider changes that may increase institutional effectiveness. 1-29, 30-109, As a result, class sizes of It is possible within certain limitations it would appear desirable and 110+ were analyzed to ascertain which class sizes may be deemed to be most effective. Data Analysis by University Each university exhibited a relationship between class Size and institutional effectiveness. In Table 4.5, insti- tutional effectiveness may be seen to decrease as class size increases. TABLE 4.5--PERCENT 0F INSTITUTIONAL EFFECTIVENESS (A/E and Al/E)BY CLASS SIZE l 2 3 4 5 Class University University University Univerisity Universit: Size A/E Al/E A/E Al/E A/E Al/E A/E Al/E AZE Al/l 1-29 76.9% 72.7% 90.1% 87.7% 71.8% 65.5% 85.0% 82.6% 87.4% 78.1 30-109 60.3 54.3 80.5 72.7 53.1 40.3 74.2 69.1 73.0 60.' 110+ 38.7 32.3 25.5 25.5 35.5 29.5 50.7 45.5 * * *NO classes taught in this class size category. a3 D 56 Regardless of the university or the measure of insti- tutional effectiveness, the pattern is consistent. The smallest difference between adjacent categories is 9.6 per- centage points between class size 1-29 and 30-109 in university 2. This indicates that there is an expected ten percent drop in institutional effectiveness as class size increases. And, the smallest difference between the smallest class size and the largest class size in any university occurs in university 5 where there is an average difference of 14.4 percentage points, and these categories are adjacent. The pattern is clearly established that the smaller class sizes exhibit higher institutional effectiveness. Data Analysis by Selected Faculty The full measure of the relationship between class Size and institutional effectiveness must be viable within faculties in order to have optimum utility. Some varia- tion in the pattern may be observed in Table 4.6. MG. ~ A» 57 TABLE 4.6--—INSTITUTIONAL EFFECTIVENESS (A/E) AND CLASS SIZE (By Selected Faculty) UnIVer- Univer- Univer- Univer- UnIVer- sity l sity 2 sity 3 sity 4 Sity_5 Economics 30-109 is ' highest .01 NS NS .01 Education .05 NS .01 Pharmacy * .01 NS .01 Engineering .01 .Ol .01 Dentistry NS .01 .05 Medicine NS 30-109 is highest NS NS Law .01 A/E in- .05 .Ol A/E in- creases creases as class as class size de- size in- creases creases *An empty cell indicates that the university does not offer the curriculum or that all class sizes were concentrated in one class Size. NS = Not significant at .05. .01 or .05 = Significance level and the noted pattern re- garding class size. In those selected faculties that exhibit significant Chi square vaers of .05 or .01 when paired with class size, four exceptions to the pattern exist. In Economics, university 1, the highest effectiveness ratio (A/E) corresponds to a class size of 30-109. In Law, university 2, and Law, university 5, effectiveness increases as class size increases. In Medicine, university 3, effectiveness is lowest in class Sizes of 30- 109. These differences in pattern suggest that caution must be exercised in concluding that a reduction in class sizes in all university curricula will increase effectiveness. In the 58 14 selected faculties that appear to follow the consistent pattern of decreasing effectiveness as class size increases, a reduction in class size may be warranted. Conclusions and Qualifications Regarding Class Size and Institutional Effectiveness These data appear to present a number of alternatives for Central American administrators. Of course, the limita- tions of financial resources and class space will impose restrictions upon the generated hypothesis, but it appears that a reduction in class sizes will increase institutional effectiveness in most faculties. If, for example, in university 3, class sizes of 30-109 were reduced to less than 29, the average institutional effectiveness would be expected to increase 16%. The resultant increase would also tend to decrease the number of students that would be repeating classes and taking up class space in the future. To be specific, suppose that a class in a faculty in university 3 that exhibits the noted pattern is reduced from 100 students to four classes of 25 students. In the larger class the expected effectiveness (A/E) value was 53.1%. But in a class Of 25, the A/E ratio is 71.8%, so 18.7% of the students could be expected to pass that would not have done so previously in the larger class. An 18.7% increase in institutional effectiveness could go far toward increasing the absolute numbers of graduates while hastening the completion 59 of programs and reducing the number of students repeating classes. This example illustrates that a reduction in class size can increase institutional effectiveness. To reduce class size, however, may incur substantial increases in re- quired levels of expenditure. When class size reduction can be coupled with a favorable salary commitment, however, in- creased productivity may become feasible as well as desirable. Level of Instruction and Effectiveness Relationships between institutional effectiveness (A/E) and the level of instruction at which classes are offered are presented in Table 4.7. First and second year courses are combined, third and fourth year courses form a second combination, and level 5, 6, 7, and 8 form the third category of level of instruction. TABLE 4.7--PERCENT 0F INSTITUTIONAL EFFECTIVENESS (A/E) BY LEVEL OF INSTRUCTION Level of Univer- Univer- Univer- Univer- Univer- Instruc- sity 1 sity 2 sity 3 sity 4 sity 5 tion A/E A/E A/E . A/E A/E 1 and 2 61.1% 85.0% 43.5% 80.2% 75.4% 3 and 4 80.5 89.8 65.8 80.2 86.4 5 - 8 82.6 87.7 81.1 84.1 92.5 Chi_Square Significance Level .01 - .01 .01 NS .01 NS = Not significant at .05. EEC Ha- VA;C .1, ( .Q a: bi 60 In universities 1, 3 and 5, effectiveness increases at each successive level of instruction. Moreover, the difference in effectiveness between level (1 and 2) and level (3 and 4) is greater than the difference between (3 and 4) and(5-8). If levels (1 and 2) are considered as "lower division" courses, while levels (3 and 4) are considered as "upper division" courses, it can be stated that institutional effectiveness (A/E) increases greatly at upper division levels. In university 3, for example, it can be observed that institutional effectiveness is 22.3% higher in the upper division than in the lower division, and in university 1 institutional effectiveness has increased 19.4% at the upper division. If levels (5-8) are considered "graduate level” courses, it appears that graduate level courses are slightly more effective than upper division courses, but the difference between graduate.level of effective- ness and that of upper division work is not as great as the differences between lower and upper division courses. Data Analysis by Selected Faculty When the preceeding data were divided into selected faculties, a variety of patterns were revealed. Institutional effectiveness was greater at each successive level of instruction in: 61 Univer- Univer- ’UhIVer- UniVer- Univer- sity 1 sity 2 sity 3 sity 4 sity 5 Law Law Law Engineering Engineering Engineering Economics ‘ ' Economics Economics Medicine Medicine ' Medicine Dentistry Dentistry Pharmacy Pharmacy Exceptions to the pattern are noted. In Education, university 1, institutional effectiveness (A/E) decreases at each successive level of instruction. In Economics and Dentistry, university 2, Economics, university 3, and Law, university 5, the institutional effectiveness average in levels (3 and 4) exceed graduate division effectiveness. In Dentistry, university 4, upper division institutional effectiveness is lower than that of both lower division and graduate division courses. Level of Instruction and Effectiveness (A1/E) Effectiveness (Al/E) and level of instruction appear to exhibit the same pattern noted in (A/E) effectiveness and level of instruction when data are distributed by university. University 4 is the only university that did not achieve a .05 level of significance. In universities l, 2, 3 and 5 at each successive level of instruction effectiveness (Al/E) increases. Specific increases are exemplified in Table 4.8. 62 TABLE 4.8-~PERCENT 0F EFFECTIVENESS (Al/E) BY LEVEL OF INSTRUCTION Level of Univer- Univer- Univer- Univer- Univer- Instruc- sity 1 sity 2 sity 3 sity 4 sity 5 tion A /E A /E A /E‘ A /E A /E l l 1 l l 1 and 2 55.9% 80.2% 36.3% 78.7% 63.1% 3 and 4 75.1 85.3 50.7 75.1 76.3 5 - 8 7900 8509 7306 7900 8707 Significance Level .01 .01 .01 NS at .05 .01 The greatest difference in effectiveness is observed between the lower division courses and other divisions. The greatest difference occurs in university 3 where a lower divi- sion average effectiveness of 36.3% is compared to the 50.7% average noted in the upper division. The graduate division in university 3 is more than twice as effective as the lower division (1 and 2). Data Analysis by Selected Faculty All five universities exhibit significant relationships when data are divided into selected faculties. Those facul- ties that achieved significant Chi square values of .05 are noted in Table 4.9. — 63 TABLE 4.9--PERCENT 0F EFFECTIVENESS (Al/E) BY LEVEL OF INSTRUCTION Univer- Levels Levels Levels sity (1 a 2) (3 8c 4) (5-8) 1 Education 84.4% 69.1% * Engineering 49.5 65.5 76.3% Law 45.5 65.5 82.0 Economics 40.7 67.0 76.9 2 Medicine 78.4 79.3 92.5 Dentistry 69.7 87.1 95.5 Engineering 75.5 83.5 90.4 Economics 52.7 85.6 61.1 3 Pharmacy 25.5 38.7 77.8 Medicine 44.3 54.7 85.9 Dentistry 30.7 53.1 70.9 Engineering 34.3 37.1 73.6 4 Economics 44.3 55.5 95.5 Dentistry 95.5 63.1 95.5 Law 60.7 74.5 83.5 5 Law 32.3 70.6 73.0 Dentistry 59.5 71.5 95.5 Pharmacy 38.7 78.4 95.5 Economics 3.9 73.9 95.5 *No classes offered at this level. Of particular importance are the mean differences ob- servable in university 3. Each faculty in univerSity 3 eVi- dences large differences in effectiveness percentage between undergraduate and graduate level courses. There is at least a 17.8% difference between these divisions in this university. The consistency of low effectiveness in each selected faculty may indicate a university policy regarding the passing of stu- 64 dents in undergraduate programs.22 Exceptions to the pattern of increasing effectiveness as the level of instruction increases are noted in Education, uni- versity 1, Economics, university 2, and Dentistry, university 4. In university 1, Education, effectiveness declines in the upper division. University 2, Economics tends to differentiate students in the graduate division more rigorously than noted elsewhere as the effectiveness average deClines 24.5% from upper division courses. Dentistry, university 4--as the course of study becomes increasingly clinica1--appears to fail more students in the upper division than in any other level. Conclusions and Qualifications Regarding_Effectiveness and Level of Instruction The overall conclusion that must be derived from the data is the mean effectiveness of the class increases at successive levels of instruction. Selected faculty data on both measures of effectiveness (A/E and Al/E) illustrate the same pattern. Within this pattern, there is a pertinent question that can be raised. University 3, selected faculty data indicate that undergraduate courses are much less effective than gradu- ate courses. Does this indicate a difference in grading philo- 22It is possible that undergraduates as a group are less able to meet university standards. Those that pass on to upper and graduate level courses have proven that they can meet the standards and hence their success rate as a group increases. The number of full time students at upper division levels may also be greater than at undergraduate levels. 65 sophy for graduate students, or have most of the incapable students been eliminated in undergraduate courses? If this pattern is continued in university 3, it becomes apparent that few students will have completed degree programs without retaking more than half of their undergraduate courses. A serious review of factors affecting the low effectiveness rate of undergraduate courses seems warranted in this university. Exceptions to the noted pattern also foster a significant question. Economics, university 2, appears to differentiate the upper division from the graduate division. It appears that the graduate division is not passing students at a rate com- parable to the upper division. Does this indicate a difference in the degree of difficulty of graduate work or is this faculty one in which those students that successfully complete undergraduate studies are not really prepared? If this be the case, it would appear that undergraduate instruction must be upgraded to better qualify students to complete their graduate level studies successfully. Data Analysis of Method of Instruction and Effectiveness (A/E) by University The method of instruction varies with the content of a course. Science courses, for example, may require laboratory periods as well as lecture sessions. Other courses may be limited inclusively to lecture, or seminar method. In medicine, for example, certain courses may be entirely laboratory practicum, 66 or clinical in nature. Three types of instruction are de- lineated in this analysis. Lecture and seminar courses are treated as a common type;23 laboratory, practice and clinical courses are a second type, and the third type of instruction is a combination of lecture and laboratory methods. An analysis of institutional effectiveness (A/E and Al/E) and methods of instruction by university indicate that only in university 1 is there exhibited a significant rela- tionship between institutional effectiveness and method of instruction. As noted in Table 4.10, laboratory courses appear to be the most effective. Lecture courses tend to be the next most effective, and combination (lecture-laboratory) courses, the least effective in this university TABLE 4.lO--PERCENT OF EFFECTIVENESS BY METHOD OF INSTRUCTION University 1 Method of Instruction A/E A1/E Lecture 58.5% 53.5% Combination 68.5 62.7 Laboratory 87.4 86.2 23Although this grouping of apparently dissimilar in- structional methods may offend the American educator, the actual classroom behavior of instructors in lecture and seminar courses does not vary materially. The lecture method is the common instructional mode. Seminars that involve sub- stantial independent study and discussion by students are rare in the universities under study. 67 Data from one university does not appear to warrant generating a hypothesis regarding the relationship between institutional effectiveness and method of instruction. Four universities exhibit nonsignificant differences between methods of instruction and therefore hypothesis generation is withheld pending review of selected faculty data. Data Analysis by Selected Faculty Significant Chi square values are evidenced in seven selected faculties in the five universities, Table 4.11 . The noted pattern of laboratory courses being the most ef- fective method of instruction is evidenced in all but the Faculty of Pharmacy, university 4, in which lecture courses are 30% more effective than laboratory courses. TABLE 4.ll--PERCENT OF EFFECTIVENESS BY METHOD OF INSTRUCTION Percent of A/E by Instructional Method University Lecture Combination Laboratory 1 Education 76.0% 81.1% 90.1% Economics ~76.0 41.5 95.5 3 Engineering 61.1 37.1 67.9 Medicine 70.6 83.2 95.5 Percent of Aq/E by Instructional Method 1 Education 73.6% 79.9% 90.1% Economics 68.5 41.5 95.5 3 Law 27.1 38.7 * Engineering 53.1 31.5 55.5 Medicine 53.9 74.2 95.5 Economics 66.7 35.5 * 4 Pharmacy 95.5 75.1 65.5 r— * No courses offered in this category. 68 Combination courses present an inconsistent pattern from the lowest effectiveness in Economics, university 1, and Engineering, university 3, to the highest effectiveness in Law, university 3. Lecture courses exhibit the highest effectiveness (Al/E) in Economics, university 3, and Pharmacy, university 4. Conclusions Regarding Method of Instruction and Effectiveness Based upon the selected faculty data, it appears possible to hypothesize that the method of instruction is related to institutional effectiveness. Laboratory courses exhibit the highest effectiveness, and it would be expected that this pattern would continue in most faculties. This pattern is not without variation however, and caution should be exer- cised before deciding that the hypothesis is congruent with practices in a particular selected faculty. Moreover, labora- tory classes may tend to be offered at upper division levels and enroll fewer students. Hence, the observed relationship between method of instruction and effectiveness may be an artifical relationship. Contact Hours (HxE) and Effectiveness by University Contact hours (HxE) represent the hours of instruction multiplied by the number of students enrolled in a course. Contact hours is a crude measure of instructor load in a course. and also a rough measure of the amount of interaction that the instructor may actually experience with class members. An 69 instructor in Central America may only teach a portion of the class periods allotted to a class and this ratio tends to allow for this practice. Data spread by institutional effectiveness (A/E and Al/E) and contact hours by university indicate the presence of a consistent pattern (Table 4.12). All universities exhibit a decrease in effectiveness as contact hours increase. At a given point in each university, the effectiveness percentage tends to drop more rapidly than previously noted. TABLE 4.12--PERCENT OF EFFECTIVENESS BY NUMBER OF CONTACT HOURS -Percent of A/E and Contact Hours Univer- Univer- 'Univera Univer- Univer- HxE* sity l HxE sity 2 sity 3 sity 4 sity 5 1-1000 80.8% 1-500 89.5% 69.1% 93.1% 84.4 . 1001-2500 74.2 501-1000 91.0 72.7 88.3 88. 2501-5000 59.5 1001-2500 87.7 71.2 77.5 86.5 5001+ 58.3 2501-5500 81.4 53.1 72.1 75.4 5501+ 71.2 45.1 ' 64. 74.5 Percent of AJ/E and Contact Hours 1-1000 78.8 1-500 88.0 63.9 92.8 81.4 1001-2500 70.0 501-1000 88.6 66.1 85.9 80.2 2501-5000 53.5 1001-2500 82.3 61.5 73.3 75.4 5001+ 49.1 2501-5500 73.6 42.7 66.4 66.7 5501+ 63.5 33.5 58.7 61.5 *In university 1, the categories of contact hours differ from the other universities. The distribution of contact hours in university 1 tended to be grouped heavily toward the middle categories leaving an inadequate number of classes at either extreme. To eliminate a Chi square value that was distorted by a small number of classes falling into the extreme categories, four groupings are employed in university 1, instead of five. 70 In university 1 the rapid decrease in instructor effectiveness occurs between 1001-2500 and 2501-5000 con- tact hours. In universities 2, 3 and 5, effectiveness ratios remain relatively stable before dropping rapidly beyond 2500 contact hours. The largest decrease in uni- versity 4 occurs beyond 1000 contact hours. The overall pattern for each university indicates a decrease in ef- fectiveness as contact hours increase. Additionally, there appears to be a point in each university where the decrease in effectiveness accelerates as contact hours increase. Contact Hours and Effectiveness by Selected Faculty Fifteen selected faculties in the five universities exhibit significant relationships between effectiveness measures and contact hours. Percentages of effectiveness for these faculties are shown in Table 4.13. Showopmo man» ca ompmda mommeo oz: 71 s H.Hm m.om m.mm s s H.Hm H.»m m.mm s seashsnm * m.mm m.mm H.m» m.om * m.mo «.mm o.Hm m.om cofiooozom m.mm m.mm m.m» m.mm m.mm m.mm :.m» ».sm m.mw m.mm oofisocoom m waHmmm>Hzp --uuuunuuus--uaumo. so pcoofimacwdo poz m.H: m.mm m.:» m.m» s oofieocoom ---u------u ..... mo. no pcooaeaewfio ooz * * ».wm 0.0» «.mm shposocoo h.mm 1.3m * t * m.mm >.mm * t * 3mg : waHmmm>Hzo m.mm ».mm H.mm ».mm m.»m m.mm m.m: H.mm m.mm m.mm somehoem H.»m m.mm m.mm m.mm m.mm «.4» m.H» e.»m m.mm m.mm oesofiooz s m.mm H.mm H.mm 6.0» * m.mm ».mm m.mm m.m» weanooeamcm m weHmmm>Hz= ».wm s.m» m.mm m.mm m.mm » mm m.m» m.mm m.mm m.mm murmureon m.mm ».mo H.m» o.mm o.mm m.mm m.m» o.mm ».mm ».om washoocflwcm m.mm m.mm m.mm :. t 0.0 o.m» :.:m :.m» * moasocoom m waHmmm>Hz= +Homm oomm-flomm oommuflooH oooHuaom omw-H +Homm commuaomm,commufiooa,oooH-Hom oomua mxm m.:» m.mm m.om m.mm m.s» m.H» m.mm rm.mm coapoosom ».ws H.mm m.mm m.mm am.me um.m» mm.mm um.mm wcfieoocdmem am.mm mm.mm ma.mm mm.am ---u---mo. es ecooflufism-e poz hhpofiecon H waHmmm>Hz= +Hoom oom-aomm commuaooa oooHaH +Hoon.oom--owm oonm-flooH oooaua . mam |m\l$ E L mmaom eoHaommmm q<20HeaeHemzH mo ezmommmuuma.s momee 13 72 As contact hours increase, the effectiveness ratios de- crease in the same manner as noted in the university data. There are a few irregularities in the selected faculty data. For example, effectiveness (A/E) in Engineering, university 2, increases from 90.7% to 93.7% before beginning a rapid descent to 45.5% when contact hours exceed 5501. However, these irregularities do not involve percentage differences large enough to note an exception to the pattern. In essence, the selected faculty data supports the pattern that as con- tact hours increase, effectiveness (A/E or Al/E) decreases. Conclusions and Qualifications Regarding Effectiveness and Contact Hours The relationship between effectiveness measures and contact hours warrants the generation of the hypothesis that: as con- tact hours increase, effectiveness ratios decrease. Also, when contact hours exceed 2500, effectiveness will decrease rapidly. It appears that a "breaking point” in effectiveness is reached at 2500 contact hours. In questioning the cause of this ”breaking point,” the answer may lie in either the number of hours of instruction offered in the course, or in class sizes. The generated hypothesis between class size and effectiveness would appear to supply part of the answer, because hours of instruction tend to be somewhat standardized by selected faculties in most universities. For example, 48 hours of instruction might re- present the standard number of hours of instruction for a 73 faculty. With standarized hours of instruction, it appears that the class size hypothesis may be reflected in the contact hour data: That as class size increases, the percentage of effectiveness declines. The possibility does exist that an instructor teaching one or two class periods is responsible for passing students. If this be the case, it would appear from the data that these instructors evaluate more rigorously as their con- tact with the class increases. This may indicate that instructors evaluating with little contact have little in- terest in the evaluation process. Either possibility can be supported in the data. Number of Classes Taught and Institutional Effectiveness In an academic year a faculty member may be called upon to teach a number of classes. In Central America the total number of courses for which an instructor is respon- sible may vary from one to fifteen classes in an academic year. By combining the institutional effectiveness ratios for all instructors teaching the same number of classes, it is possible to determine the institutional effectiveness ratio for the number of classes taught. These classes were then grouped to form three categories: 1-4 classes taught, 5-8 classes taught, 9-15 classes taught. By comparing the standard deviations and mean scores for each group of classes, it is possible to determine if any relationship exists be- tween institutional effectiveness and the number of classes 74 taught by an instructor in a calendar year. Table 4.14 represents the mean scores of institutional effectiveness (A/E and Al/E) for each university in each level of classes taught by a professor. At first glance, it appears that there are differences between institutional effectiveness and the number of classes taught by a professor. TABLE 4.14-~PERCENT OF INSTITUTIONAL EFFECTIVENESS (A/E and Al/E) BY NUMBER OF CLASSES TAUGHT Percent of A/E Classes Univer- Univer- Univer- Univer- Univer- Taught sity l sity 2 sity 3 sity 4 sity 5 1-4 72.4% 92.1% 65.1% 78.7 84.0% 5-8 65.8 95.4 57.6 84. 82.2 9-15 74.8 91.8 72.0 97.4 70.9 University Mean 70.6% 92.8% 62.9% 82.6% 83.4% Percent of A1/E 1-4 66.4% 81.1% 55.4% 71.2% 72.9% 5-8 59.9 85.2 46.0 83.0 73.0 9-15 68.1 83.5 63.6 97.3 58.0 University Mean 64.6% 83.3% 52.5% 79.1% 72.4% If teaching 1-4 classes in an academic year is considered a "light" teaching load, 5-8 as a "moderate" load, and more than nine courses a "heavy" load, it appears that the means scores differ with the number of classes taught by a professor. However, in most universities, those means that differ greatly from the university mean are based on a small number of pro- fessors. For example, in University 5, (Al/E), professors 75 teaching nine or more courses appear to be less effective than the university average. However, there is only one professor teaching more than nine courses in university 5. There are only seven professors teaching more than five courses in this university in a faculty of 181. The fre- quency of professors teaching each group of classes is noted in Table 4.15 TABLE 4.15--FREQUENCY DISTRIBUTION OF PROFESSOHSBY NUMBER OF CLASSES TAUGHT Classes Univer- Univer- Univer- Univer- Univer- Taught sity 1 sity» 2 sity 3 sity 4 sity 5 1-4 299 318 275 168 173 5'8 50 43 31 15 7 9-15 9 4 5 4 1 Total 358 365 311 187 181 It is evident that the number of profesSors in each university that are teaching five or more classes is too small to warrant the generation of hypotheses about effective- ness and the number of classes taught by the professor. Of interest, however, is the fact that few professors in the five universities teach what would be considered to be full schedule of classes in an academic year. This supports the findings of Friedman in which most faculty members in Central America may be employed in another position outside of the university.2u Few professors are engaged in research in Cen- 21+op. cit., p. 22. 76 tral America and hence thus possibility does not account for low teaching loads. Selected faculty data proves to be less conclusive than university data as the frequency distribution of faculty be- comes so small as to be inappropriate for analysis. There- fore, no hypotheses can be generated from faculty data. Conclusions and Qualifications Regarding Number of Classes Taught and Institutional Effectiveness The distribution of faculty members teaching a number of classes is not sufficiently varied to permit the generation of a hypothesis regarding effectiveness and the number of classes taught by a professor. Most faculty in Central America teach between one and four classes per year, which indicates that one course per semester is closer to the rule than the exception for a professor in an academic year. Summary Institutional effectiveness (A/E) or(A1/E) is clearly affected by certain administratively controllable factors within the central American universities and university faculties. In general, these conclusions seem warranted by the analysis: 1) Class size affects institutional effectiveness in ways that suggest that institutional effectiveness may be increased by planned reductions in large class sizes. 2) 3) 77 Instructor salary, in and of itself, is not generally a compelling factor in institutional effectiveness although it has been noted that professors receiving less than $250 for a class tend to be more productive. A systematic analysis of this pattern would appear warranted in the four universities in which this trend was evidenced. There is little selected faculty support, however. Effectiveness increases at successive levels of instruction from lower to upper to graduate divisions. The self-selection process--rather than improved instruction--may account for this phenomenon. Method of instruction is related to institutional effectiveness. In general, laboratory courses seem to be more productive than lecture courses. However, laboratory courses typically enroll few students at higher levels of instruction; hence, the generaliza- tion may need qualification. The number of contact hours of a professor with a class seems to be related to effectiveness. The suggested pattern is a decrease in effectiveness as contact hours increase. This pattern may reflect the trend noted in class size for the hours of instruction in a faculty may be standardized. 6) 78 The data surrounding effectiveness and the number of classes taught by a professor do not warrant hypothesis generation. The distribution of pro- fessors teaching a varied number of courses in each university is centered below four classes per academic year. Although 15 classes may be taught by a single professor, the frequency of these occurences does not warrant a statement of a relationship with effectiveness. CHAPTER V FACTORS RELATED TO INSTITUTIONAL EFFICIENCY In Chapter 4, the Six administratively controllable fac- tors were related to measures of institutional effectiveness (A/E and Al/E). The purpose of this chapter is to examine the relationship between the six factors and institutional efficiency. Institutional efficiency was defined in Chapter 1, as a measure of the maximization of input divided by output in the form of examinees (ex) divided by enrollees (E). The significance of institutional efficiency stems from the fact that the success- ful student (A) must come from the number of examined students (Ex). Therefore, a maximization of institutional efficiency will increase the input necessary to produce successful students. Two measures of institutional efficiency were employed. First, (Ex/E) in which (Ex) represents all students that were examined and (E) represents the total class enrollment. Second, (Exl/E) in which (Exl) represents students who presented them- selves for examination at the earliest opportunity, usually at the immediate conclusion of the course. Using Chi square and mean comparisons, it was possible to determine the relationships between measures of institutional efficiency (Ex/E and Exl/E) and the factors that were delineated in the previous analysis of institutional effectiveness. Data are presented by university and also by selected faculty within each university. 79 80 Level of Expenditure and Institutional Efficiency (Ex/E & Exl/E) An analysis of institutional efficiency (Ex/E & Exl/E) and level of expenditure data reveals that the highest measures of institutional efficiency in all universities are evidenced when salaries paid to the instructor are less than $250 dollar equiva- lents. Table 5.1 summarizes the highest institutional efficiency ratio in terms of dollar equivalent categories. TABLE 5.l--LEVEL OF EXPENDITURE DISPLAYING HIGHEST INSTITUTIONAL EFFICIENCY Highest Highest University Ex/E Exl/E 1 $ 1-250 $ 1-250 2 NS NS 3 1-250 1-250 4 1-250 1-250 5 NS NS NS = Not significant at .05 level of confidence Inspite of the evidence that all universities with signi- ficant Chi square values achieve the most institutional efficiency (Ex/E and Exl/E) when salaries paid to the grading instructor are lowest, selected faculty data (Table 5.2) does not exhibit this pattern with enough frequency to warrant the generation of a hypothesis. In university 1, the faculties of Medicine and Economics appear to follow the noted pattern. In university 3, faculties violate the pattern more often than not. Law, univer- sity 5, does not adhere to the pattern whatsoever. These facul- ties are the only faculties in which a significant Chi square value 81 was determined. The level of expenditure that corresponded to highest institutional efficiency is noted. TABLE 5.2--LEVEL OF EXPENDITURE DISPLAYING HIGHEST INSTITUTIONAL EFFICIENCY (Selected Faculty) Highest Highest University Ex/E Exl/E 1 Medicine $l-250 $l-25O 1 Economics 1-250 1-250 3 Engineering 251-500 251-500 3 Dentistry NS 251-500 3 Medicine NS 1-250 5 Law 251-500 NS NS 3 Not significant at .05 level of confidence Conclusions and Qualifications Regarding Institutional Efficiency and Level of Expenditure Based upon the analysis of university and selected faCulty data, there is no conclusive relationship between level of expenditure and institutional efficiency. The pattern evidenced in university data that low expenditure yields high in- stitutional efficiency is not supported in the faculty data to an extent that Justifies hypothesis generation. In the previous chapter the possibility was put forward that "practitioners" from the surrounding community may be affecting the effectiveness ratios. The possibility may exist that these people are also causing a distortion in the institu- tional efficiency data. If these practitioners are not committed 82 to evaluation, they may be affecting the effectiveness and efficiency ratio by following a pattern far different from the academician. If these practitioners are paid at a reduced rate, the indication would be exhibited in the university data. Within a faculty, their efficiency ratio may not be great enough to effect a significant Chi square value. However, as a collectivity in the university, these people might ex- hibit the noted pattern. It is possible that this is the explanation for the pattern in the university data whereby low level of expenditure results in high institutional efficiency. Class Size and Institutional Efficiency The relationship between institutional efficiency (Ex/E and Exl/E) and class size are examined by dividing data for each university and selected faculty within each university. It is administratively feasible for class size to be controlled and, therefore, evident relationships will enable Central American administrators to alter class sizes in accordance with standards of desired efficiency. Institutional efficiency was examined by categories of class sizes of 1-29, 30-109, and 110+, to determine which, if any, class sizes may be useful in maximizing efficiency. Data Analysis by University The analysis of class size data by university suggested that there is a consistent pattern throughout the five universities. 83 Without exception, as class size increases, institutional efficiency (Ex/E and Exl/E) decreases. The patterns are illustrated in Table 5.3. TABLE 5.3--PERCENT OF INSTITUTIONAL EFFICIENCY (Ex/E and Exl/E) BY CLASS SIZE Ex/E Class UnIVer- Univer- Unlver- ‘Uhiver- ‘Univer- Size sity l sity 2 sity 3 sity 4 sity 5 1-29 85.6% 91. 78.7% 90.1% 92.5% 30-109 76.6 88. 78.1 83.8 86.5 110+ 75.4 25.5 69.4 65.5 * Exl/E 1-29 80.8% 90.7% 72.1% 88.3 85.6% 30-109 67.9 86.5 59.9 79. 77.8 110+ 58.7 25.5 37.9 55.5 * *No classes offered in this category. An illustration of the pattern occurs in university 1. When class size is 1-29, institutional efficiency (Ex/E) averages 85.6%. It is, thereby, expected that 85.6% of all students en- rolled in classes with less than 29 students will present them- selves for examination. In a class with 30 through 109, the expected percentage of those presenting themselves for examination is 76.6%. This difference represents a loss of 9.0% as class size increases. Institutional efficiency as measured by (Exl/E) declines at a more rapid rate than (Ex/E) in all universities. The 84 indication is: As class size increases, the percent of stu- dents presenting themselves for examination at the earliest opportunity (Exl) declines rapidly. Subsequent enrollments in these courses may, thereby, be swelled by previously un- examined students. In university 3, average institutional efficiency (Exl/E) decreases from 72.1% in classes of less than 29 students to 37.9% where class size exceeds 110. This small percentage in large classes may result from massive 25 nonattendance, previously unexamined repeaters, or both. As evidenced in university 3, students being examined at a later time or being re-examined in classes of 110 is 31.5% of the total examinee percentage. Data Analysis by Selected Faculty The relationship between institutional efficiency and class size by faculty is presented in Table 5.4. It is evident from the analysis of these data that not all selected faculties follow the university pattern of decreasing institutional efficiency as class size increases. Five selected faculties deviated from the noted pattern. Economics and Law, university 2, established that as class size increases, institutional efficiency (Exl/E) is concerned. This poses the possibility that 25The arbitrary practice of considering each matriculant as a full time student when in fact, less than 30% may legitimately be considered full time, may account for nonattendance. 85 TABLE 5.4--PERCENT OF INSTITUTIONAL EFFICIENCY (Ex/E and Exl/E) BY CLASS SIZE (By Selected Faculty) Ex/E ’Univer- 7Univer- Univer- Univer- Univer- sity l sity 2 sity 3 sity 4 sity 5 Economics NS As class 30-109 is NS NS size in- ‘ highest creases, efficiency increases Education NS As class .01 size in- creases, efficiency increases Pharmacy NS NS .OI' Engineering .01 NS .01 Ikntlstry .05 .01 NS Medicine NS 304109 is NS NS highest Law NS As class .01 .01 NS size in- creases, efficiency increases Exl/E Economics NS .01 .01 NS .01 Education .05 ' As class .01 size.in- creases, efficiency increases Pharmacy .05 NS .OI' Engineerffig .05 .01 .05 Dentistry NS .01 . NS Medicine NS 30-109 is highest .05 NS Law NS As class ’NS .01 NS size in- creases, efficiency increases Blank space : Not offered or only 1 class size NS - Not significant at .05 .01 or .05 = Significance level and regular pattern 86 there were large classes of students repeating courses in this faculty. The percentage of students presenting themselves for examination after repeating or auditing a class in these large classes was very high in comparison to the smaller class sizes. However, the percentage of students being examined for the first time followed the more normal pattern of decreased institutional efficiency as class size increases. The Economics and Medical faculties in university 3 main- tained highest institutional efficiency (Ex/E) in class sizes of 30-109. However, the faculty of Economics maintained the expected pattern with regard to institutional efficiency (Exl/E). Again the possibility is posed that medium sized classes of students repeating or auditing courses were established, and a larger percentage of these students were examined than might normally be expected. Conclusions and Qualifications Regarding Class Size and Institu- tional Efficiency A The similarity evidenced in many selected faculties warrants the generation of the hypothesis: As class size increases, institutional efficiency (Ex/E and ExI/E) decreases. Pattern inconsistencies suggest that this hypothesis warrants applica- tion only in those selected faculties that appear to follow the pattern. It would appear, therefore, that this hypothesis has administrative application for at least fifteen faculties in the five universities. Class sizes of one through 29 are deemed 87 to be more efficient than larger classes. Class sizes be- tween 30-109 are more efficient than classes larger than 110. If increased institutional efficiency (Ex/E and Exl/E) is sought,a reduction in class size within the stated categories would appear to achieve the desired objective. Level of Instruction and Institutional Efficiency Level of instruction refers to the academic year in which a class is normally taken in a student's curriculum. Three levels of instruction were delineated by examination of data. Years (1 and 2) forms one level, years (3 and 4) another, and years (5-8) form the third group. As indicated in Table 5.5, data from universities 2 and 4 were not significant at the .05 level of confidence. TABLE 5. 5--PERCENT OF INSTITUTIONAL EFFICIENCY (Ex/E and EX l/E) BY LEVEL OF INSTRUCTION Ex/E Univer- Univer— Univer- Univer- Univer- Level sity l sity 2 sity 3 sity 4 sity45 1 and 2 78. 7% 89.2% 73.6% 87 1% 87.7% 3 and 4 86. 2 91.9 82.3 87. 4 91.9 5-8 86. 2 89.8 87.7 89 5 95.2 Significance Level .01 NS .01 NS ' .Ol Exl/E l and 2 69.7% 88.3% 53-9% 85.9% 77 5 3 and 4 82.9 90.7 67. O 82. 9 85. 5—8 83.2 89.8 79. 6 85.0 91.3 Significance Level .01 NS .01 NS .01 NS = Not significant at .05 88 The data in the preceeding table indicate institutional efficiency (Ex/E and Ex /E) increases at each successive level. The pattern is without :xception in universities l, 3, and 5. In university 1, as an illustration, institutional efficiency (Exl/E) in first and second year courSes averages 69.7%. In third and fourth year courses this percentage has increases to 82.9% and subsequently to 82.3% in years 5-8. The presence of this pattern tends to indicate that either students at higher levels are not subject to attritionam the same rate as at lower levels, or the enrollments are not in- flated with part time students to the extent as in lower levels. The likelihood of attrition differences does not appear great because a student does not normally "flunk out" in a Central American university. Instead, a course may be repeated or audited to prepare for re-examination. If attrition rates do vary during a term with the level of the course, it would appear that some factor other than the six under investigation are causing this phenomena. If students at higher instructional levels are predominately full time students, would not institutional efficiency differences between high and low levels of instruction actually reflect a loss in part time students? In essence, is it possible that institutional efficiency differences are measuring the extent of part time students in the university? If this be true, real differences between institutional efficiency with respect to full time students may not be extent. Since the full time student is most likely to be coveted it appears that Central American administrators would desire to ascertain if university 89 differences are real or artificial before recommending changes that may imbalance a functioning system. Data Analysis by Selected Faculty The analysis of selected faculty data (Table 5.7) by university indicates that the noted pattern in university data is evidenced. Of the 20 faculties in which relationships proved to be significant, only six exhibited patterns that were con- trary to the previously noted pattern of increasing institutional efficiency as level of instruction increased. In Education, University 1 and education, university 2, both measures of institutional efficiency decreased at successive levels of instruction. Institutional efficiency decreased in Medicine, university 2, from levels (1 and 2) to (3 and 4). Levels (5-8) exceeded all other levels in institutional efficiency. This would seem to indicate that if a student can reach years (5-8) in the medical curriculum, the probability Of being examined is greater than at earlier levels in his educa- tion. The same pattern is evidenced in Dentistry, university 4, and Dentistry, university 5. The efficiency measures of Economics, university 2, appear to differentiate third and fourth year students from(5-8 year students. The institutional efficiency (Ex/E) percentage decreases 27.8% on the average from 3 and 4 level courses to the (5-8) level courses. If (5-8) level courses are assumed to be graduate level courses, it appears that the Economics faculty may differentiate heavily those who should be examined out of those enrolled at the graduate level. in 4‘ ill- Ih-I IIQ. I :OI -|\~‘ fl 1‘! HI. 1 \ - F.\vi‘ 0.0» m.m0 m.mm m.m0 0.00 0.0» 30A 0.00 0.m» ».Hw 0:00H0H:me no: hnpme:on m.m0 3.0» ».0m «.mm 0.00 0.m» Senators 0.00 0.00 :.N» 0:00H0H:me no: mOHEo:oom m MBHmmm>HZD oenoauacman roe m.mm m.mm 4.40 3mg m.mm 0.4» m.mm m.m0 0.0» m.m0 ananapeoo m.m0 0.0» 0.mm 0.00 m.mm m.mm wOHEo:oom : MBHmmm>HZD 0.mw m.Hw m.mm 0.H0 0.00 0.3» m:H:oo:Hw:m 0:00HMH:me po: :.00 ©.m» H.»m so: m.wm 0.0» m.0» m.H0 ».:m :.Hw o:H0H002 0.0» m.Nm m.wm m.Nw m.:© H.H© homshmcm m.mw m.mw m.0» 0:00H0H:me no: mOHEo:oom m MBHmmm>HZD %, 0.00 N.N0 m.m0 0.00 N.N0 m.m0 :oHpmoscm m.m0 N.N0. 0.00 0:00Hmw:me no: w:H:oo:Hm:m .mm 0.40 0.00 m.m0 0mm H.Hw appeared: 0.00 ».Hw ».»w p:moH0H: H0 00: o:H0H002 H.Hm 3.00 0.m» ©.Hm 0. mm m.mm mOHEo:00m N NBHmmm>HZD u:moH0H:me no: R».:w m.»» m.H: 300 R».:w 0.0» 0.0: 0:00H0H:me no: w:H:om:Hm:m 0:0: RN.H» R».00 0:0: $0.0» RH.m0 :oapnosom H MBHmmm>HZD mum : 0:0 m N 0:: H mam # 0C0 m N can H HES: HENRI! m\Ham m\wm AO0H30mm couOOHom mmv ZOHBODmBmzH mo qm>mq 0m 2m\ am 0:0 N\xmv wozmHonmm H420HBDBHEm2H mo azmommmuum. m mqmHz= m.m0 m.00 0.0» m.H0 c 00. as pcsofiuacmam 0oz nofiaocoom * a 0.0» 3.0» 0.00 a a 0.0» 3.0» 0.00 anunapcon 0.m» 0.30 s s t 00. as pesoauacmam 0oz seq 3 waHmmm>Hzp 0.0» 0.00 ».30 m.m» 0.00 00. us ecmoaaacwam poz ooasocoom m.»m m.mm 0.00 0.m0 m.»0 00. no pesoauacwam 0oz hosseonm 0.m» 0.m» 3.»0 0. m 0.00 00. no ucooauaewam 0oz oeaoaoms 00. as ecsofiuacmam no: 0.m» 3.30 ».00 0.00 0.00 Etpnfiucmn 0.H3 0.H0 H.N» 0.00 t 0.0» 0.»» H.»0 m.N0 * 0:H:oo:H0:m m weHmmm>Hz= 3.03 0.0» 3.30 3.0» 0.0m 0.N0 0.»» 3.30 3.0» 0.0m mOHSo:oom 0.00 0.00 3.00 0.00 0.00 0.00 0.00 3.00 0.00 0.00 mppmecon 0.00 N.00 H.00 ».m0 N.N0 .00 R3.00 R».m0 $0.30 Rm.m0 wcHnmo:Hm:m m waHmmm>st + u 1H 0. pm acmOHhH: H0 902 0:H:00:H0:m H SaHmmm>H20 +H000 0000:3000 00001H00H 000H-H +3000 000013000 000013003 0003-3 nmxaxm M\wm H Aspasosm oouooaom any 0000: soaezoo mo mmmzaz mm Am\.xm ees m\nmv wozmHoHsmm qaonepeHemzH mo 9200000--0H.m mamas 100 Conclusions Regarding Institutional Efficiency and Contact Hours There is no consistent pattern in university or faculty data that would support the generation of a hypothesis. Faculty patterns are so diverse that it would be necessary to view the individual pattern of each faculty before deciding to alter existing procedures. In essence, administrative decision- making should be varied by faculty within each university to implement these results. Number of Classes Taught and Institutional Efficiency The data regarding institutional efficiency and the number of classes taught by the grading professor is subject to the same limitations noted in Chapter 4. In the five Cen- tral American universities, most professors teach less than five Classes in an academic year. The range in classes taught in the 1962-63 school year is from one to 15 classes taught, but with 80% of the professors teaching less than five classes differences in mean scores of efficiency are not subject to conclusive analysis. Table 5.11 represents the frequency distribution of the number of classes taught by professors in the five universities. The uneven distribution of faculty is evident. As a result of the uneven distribution of data, mean com- parisons are not warranted. No conclusive relationship between institutional efficiency and the number of classes taught by the grading professor can be determined. 101 TABLE 5.11-FREQUENCY DISTRIBUTION OF PROFESSORS BY NUMBER OF CLASSES TAUGHT Classes Univer- Univer- Univer- Univer- Univer- Taught sity l sity 2 sity 3 sity 4 sity 5 1-4 299 318 275 168 173 5-8 48 43 31 15 7 9-15 9 4 5 4 l Data Analysis by Selected Faculty The division of university data into selected faculties results in a frequency distribution that contains small values in the upper categories of classes taught. Because comparisons would be meaningless with small frequencies these data are not included in this chapter. Conclusions and Qualifications Regarding the Number of Classes Taught and Institutional Efficiency In Central America it appears that most professors teach on a part time basis. The number of professors teaching 1, 2, 3, or 4 courses in the academic years 1962-63 represent over 80% of the professors in the five Central American universities. The small number of faculty teaching varying numbers of classes does not Justify mean comparisons of the number of classes taught by the grading professor and institutional efficiency. Summary 102 Institutional efficiency (Ex/E) or(Ex1/E) is affected by administratively controllable factors in the Central American universities and university faculties. These conclusions seem warranted by analyses of data. 1. No conclusive relationship is evidenced between the salary paid to the grading instructor and institutional efficiency. The manipulation of instructor salary would not materially affect the institutional efficiency of the university. Class size affects institutional efficiency in ways that suggest that institutional efficiency may be increased by the systematic reduction in certain class sizes. 1 Institutional efficiency increases at successive levels of instruction, from lower to upper to graduate divisions. This phenomenon may result from student self-selection and more full time students in the upper and graduate levels. The method of instruction is related on institutional efficiency. Laboratory courses seem to be most efficient, but these courses are more apt to be taught in small class sizes at the upper and gra- duate levels of instruction. Hence, the generalization may have limited administrative utility in and of itself. 103 5. There is no conclusive relationship between faculty- class interaction as measured by contact hours and institutional efficiency. Individual selected faculties may heed the pattern evidenced in con- tact hours, but no generalization is warranted. 6. No conclusive relationship between the number of classes taught by the grading instructor and insti- tutional efficiency is evidenced. Over 80% of the professors teach little more than one course per semester on the average. The thesis was advanced that: Administratively con- trollable variables are related to institutional efficiency in ways that efficiency may be increased. Supportive data from the five Central American universities lend credance to this thesis. The next step would appear to be the delineation of interrelationships between the six factors and institutional efficiency. CHAPTER VI FACTORS RELATED TO INSTRUCTIONAL EFFICIENCY Instructional efficiency (A/Ex and Al/Exl) was designated in Chapter 1 to be the ratio of successful students (A and Al) to those examined (Ex). The term instructional efficiency was assigned to these ratios in recognition of the fact that the decision as to who will pass a course is determined principally by the evaluation of student achievement on the final examination. Therefore, the efficiency of the class is determined by those factors surrounding the instructor, and the instructional atmosphere and the ability of the student to transfer relevant concepts in the examination process. The relationships between two measures of instructional efficiency (A/Ex and Al/Exl) and factors that could affect instructional efficiency are examined in this chapter. Level of expenditure, class size, level of instruction, method of instruction, contact hours and number of classes taught by the instructor were spread by university and by selected faculties in order to ascertain significant Chi square values involving instructional efficiency. The mean scores within each Chi square cell were than analyzed to determine the direction as well as degree of significant relationships. 104 105 Level of Expenditure and Instructional Efficiency by Universigy The examination of level of expenditure and instructional efficiency is restricted by relationships that are not signi- ficant at the desired level of confidence. cates that universities l and 5 did not achieve the .05 Table 6.1 indi- level of confidence and instructional efficiency as measured by A/Ex in university 2 also fell short of .05 level TABLE 6.1-~PERCENT OF INSTRUCTIONAL EFFICIENCY (A/Ex and Al/Exl) BY LEVEL OF EXPENDITURE A/Ex Univer- Univer- Univer- Univer- Univer- Cost sity 1 sity 2 sity 3 sity 4 sity 5 1-250 84.7% 91.6% 90.1% 94. 84.7% 251-500 80.8 91.0 89.2 91. 85.0 501-750 78.7 90.4 73.3 85.9 84.1 751-1000 79.9 89.5 76.9 82.3 87.4 1001+ 79.9 87.7 76.9 82.6 82.0 Significance Level NS NS .01 .01 NS Univer- Univer- Univer- Univer- Univer- Cost sity l sity 2 sity 3 sity 4 sity 5 1-250 86.5% 91.6% 90.1% 94.6% 86.2% 251-500 82.0 90.4 88.3 91.9 85.3 501-750 79.0 89.8 72.4 85.3 85.3 751-1000 80.5 88.9 77.5 83.5 85.9 1001+ 80.5 85.9 76.3 83.5 82.0 Significance Level NS .05 .01 .01 NS 106 Those measures of instructional efficiency did evidence a significant relationship with level of expenditure present a consistent pattern. As the level of expenditure increases, measures of instructional efficiency decrease slowly. A contrary trend is noted in university 3 where a sharp de- crease occurs when the level of expenditure is 501-750 dollar equivalents. A slight increase occurs as level of expenditure exceeds 750 dollar equivalents. One noteable relationship requires comment. There is a relatively high level of instructional efficiency in evidence, in universities 2, 3 and 4 regardless of the level of expenditure. It appears clear that if a student presents himself for examination, his chances of success are quite high. Of course, particular selected faculties may not ad- here to this pattern, but the evidence for the university as a whole indicates that instructional efficiency is rather high in all institutions. Data Analysis by Selected Faculty Seven selected faculties exhibited significant relation- ships between instructional efficiency and level of expendi- ture. The pattern of relationships is varied in these faculties as evidenced in Table 6.2. 107 00o00000 03:0 :3 0000030 02* 00. 00 000o303003m 0oz ».00 t ».30 0.»» 0.00 303 00. 03 000o303003m 0oz * 0.00 0.»» 3.00 0.00 0o000000 0 0030000320 00. us peso303003m 0oz 3.30 0.00 0.30 0.00 0.00 000n300o0 3 0030000320 - 0.00 0.00 3.30 0.00 s 0.00 ».00 0.00 0.00 00300003000 3.0» 0.00 0.»» 0.00 0.00 0.0» 0.00 0.»» 0.00 0.00 oe3o3umz 0 0030000320 3.0» 0.00 0.30 3.00 _ 3.00 00. 00 0:0o303003m 0oz 000030000 00.00 00.00 00.3» 00.30 00.30 00.00 00.00 00.3» 00.30 03.00 mOHSocoom . 0 0030000320 +300330003-30»000»-3000 000-3000 000-30 +30030 0003-30»0 00»-3000 000-3003 000-33 330\33 xm\3 . 3 3 30030o30 oouoo3om 000 00003020030 00 30003 30 3 am\ 3 can xm\3v 0000303300 3320330000003 00 0200000-0.0 03033 108 In four faculties, the lowest level of expenditure ($l-250) exhibits the highest degree of instructional efficiency, but in two of those, Medicine, university 3 and Pharmacy, univer- sity 5, instructional efficiency equal to that in expenditure 1-250 is achieved at other points of level of expenditure. Therefore, the lack of pattern in selected faculties negates the hypothesis generated from the university data. Data pro- vides inconclusive evidence that as the level of expenditure increases, instructional efficiency decreases. Conclusions and Qualifications Regarding Instructional Efficiency and Level of Expenditure ~ University data supports the generation of the hypothesis that as the level of expenditure increases, instructional efficiency decreases. Faculty data is inconsistent and no hypothesis appears warranted. Selected faculty data tend to be most important in administrative decision-making and the importance of the hypothesis in university data should be minimized. glass Size and Instructional Efficiency At first glance, it would appear that no relationships should exist between instructional efficiency measures and class size. Instructional efficiency is derived from the num- ber of successful students divided by the number of students examined in a course, while class size reflects the enrollment 109 necessary to produce examined candidates. In essence, the question could be raised, why should class size affect the instructional efficiency ratios? While the question would appear to be warranted, data by university and selected faculties indicate that a relation- ship does exist. As illustrated in Table 6.3 by the five universities, as class size increases, instructional efficiency decreases. TABLE 6.3--PERCENT OF INSTRUCTIONAL EFFICIENCY (A/Ex and Al/Exl) BY CLASS SIZE A/Ex Class Univer- Univer- Univer- Univer- Univer- Size sity l sity 2 sity 3 sity 4 sity 5 1-29 85.0% 93.1 87.1% 90.4% 87.7% 30-109 74.8 85.6% 66.4 84.4 75.5 110+ 58.7 80.5 51.1 84.4 * SigfiIficance , Level .01 .Ol .01 .01 .Ol Al/Exl 1-29 85.9% 93.1% 86.5% 91.3% 88.3% 30-109 75.4 84.7 66.7 83.8 79.0 110+ 58.7 95.5 #9.9 84.4 * Significance Level .01 .01 .Ol .01 .Ol *No classes listed. The instructional efficiency (Al/Exl) patterns of univer- sities 2 and 4 provide the only exceptions to this hypothesis. In university 2, the instructional efficiency (Al/Exl) in class sizes above 110 is 10.8 percentage points above the average for 110 class sizes 30-109. In university 4, there is a .6% dif- ference between class sizes of 30-109 and classes above 110. Other universities follow a consistent pattern of de- creasing instructional efficiency as class size increases. Data Analysis by Selected Faculty Selected faculty data of the five universities establish 17 faculties with significant relationships. As illustrated in Table 6.4, 16 of the 17 faculties adhere to the pattern already noted. 111 .0035000 0320 :3 020500 030 0030.0350 no 0000030 oz: 0 * 0.33 0.30 * 0.3M 0.30 200030000 0 0.00 0.30 0 3.0 0.30 00000000 00. 30 00003030030 302 0 0.00 0.00 0030300: * 0.00 3.00 0 0.00 0.00 003200000 0 3330000320 0.00 0.00 0 0.00 0.30 0 303 3 3330000320 0.00 3.03 3.30 0.00 3.03 0.00 20000000 0.00 0.00 0.00 0.00 0.00 0.00 0030300: 0 0.00 0.00 0 0.00 3.03 000030000 0.3m 0.30 0.0m 0.30 0.00 m.mm 003300E3wcm 00. 00 00003030030 002 0.00 0.00 0.00 303 0 0.00 0.00 0 0.00 0.30 co3pmos00 0 3330000320 * 0.03 0.30 * 0.00 0.30 00300003000 * 3.00 0.30 0 3.30 0.00 003000000 0 3330000320 00. 00 00003030030 002 00.00 0.00 0.00 003300000 0 3.30 0.00 0 0.00 0.00 303 t 0.00 0.30 ¢ 0.00 3.00 003000000 0 00.00 03.30 0 00.00 00.30 00300003000 3 333000>320 +033 003-00 00-3 +033 003-00, 00-3 «N30 00330 3Nfl\33 xM\¢ 33333000 60000300 any 0030 00330 30 3300\33 000 00\3V 3020303000 3320330003023 00 3200000-3.0 03033 112 The faculty of Medicine, university 3, is the sole excep- tion. Class sizes of 30-109 appear to be less efficient than either extreme. Based upon selected faculty data and university data, it is apparent that the hypothesis can be generated that instructional efficiency decreases as class size increases. The question raised earlier regarding the reason for a relationship between instructional efficiency and class size may be answered in supporting data. The pattern evidenced by institutional efficiency (Ex/E) indicated that as class size increased, institutional efficiency decreased. This de- cline was rather gradual and the percentage difference between small and large classes was minimal. In absolute numbers, therefore, a class size of more than 110 would provide more students to be examined, on the average than a class of less than 29. In essence, in a larger class, the instructor is called upon to grade more final examinations than in a smaller class. Is it possible that the relationship between class size and instructional efficiency is more closely a reflection of the absolute numbers of examinations that must be evaluated in a large class? If this be true, additional support is garnered for smaller classes. Conclusions and Qualifications Regarding Instructional Efficiency and Class Size Class size appears to have an effect on instructional efficiency. As class size increases, instructional efficiency 113 measures decrease. The administrative implication derived from this pattern tends to favor smaller classes if greater student success is desiredfrom those examined. University data follow the established pattern and all but one of the selected faculties support the pattern. Level of Instruction and Instructional Efficiency Level of instruction refers to the academic year in which a_course is ordinarily taken in a student's curriculum._ Three levels of instruction were delineated for examination of data. Years (1 and 2) form one level, years (3 and 4) another, and years (5-8) form the third group.. Two measures, of instructional efficiency (A/Ex and Al/Exl) were Spread with the grouped levels by university and selected faculty to deter-— mine Chi square relationships. Data Analysis by University The analysis of data regarding level of instruction and inStructional efficiency is based on Table 6.5. Instructional- efficiency percentages are evidenced for each level of in- struction in each university, although university 2, (Al/Exl) and university 4, (A/Ex and Al/Exl) did not achieve the .05 level of significance. 114 TABLE 6.5--PERCENT OF INSTRUCTIONAL EFFICIENCY (A/Ex & Al/Exl) BY LEVEL OF INSTRUCTION A/Ex Univer- Univer— Univer— Univer- Univer- Years sity l sity 2 sity 3 Sity 4 sity 5 1 and 2 74.2% 89.2% 57.5% 88.6% 79.9% {3 and 4E ' 87.7 92.5 76.3 87.4 v 86.8 5-8 90.4 92.2 88.0 90.4 90.4 Significance . , NS at Level .01 .01 .Ol .05 .Ol A1/EXL 1 and 2 75.0% 89.2% 57.90 88.6% ' 80.8% g3 and A; 87.7 91.3 75.4 87.1 87.4 5-8 90.7 92.5 88.3 92.2 91.3 Significance NS at NS at Level .01 .05 .Ol ‘.05 .01 The pattern in each university with significant Chi square values is similar. Instructional efficiency inCreases at each successive level of instruction. In university 2, level (5-8) are .3 percentage points below level (3 and 4) in A/Ex, but this is a minor exception in the pattern. It is also noted that instructional efficiency measures do not dif-A fer greatly in any given grouping. For example, in group (1 and 2), university 1, the difference between measures of instructional efficiency is 1.5 percentage points. This is the greatest difference noted in any category. It is also noted that instructional efficiency differences between groups (3 and 4) and (5-8) are smaller than differences 115 between groups (1 and 2) and (3 and 4). This trend opens the possibility that instructors in lower level courses discriminate to a greater extent in determining which student will pass. It may be possible, however, that more unqualified students present themselves for examination in lower level courses, and that these students do not continue into upper level courses. While the latter possibility is usually evoked in Central America, the former possibility is not without significance, and may be operant in the universities. Data Analysis by Selected Faculty The pattern noted in university data is upheld in selected faculty data. Table 6.6 exhibits the data of 19 selected faculties evidencing significant Chi square values. 116 TABLE 6.6--PERCENT OF INSTRUCTIONAL EFFICIENCY (A/Ex and Al/Exl) BY LEVEL OF INSTRUCTION (By Selected Faculty) A/Ex ‘ Al/ExL . i 1 & 2 ) ( 3 & 4 ) (5-8) .(1 & 2) (3 &4) (5-8) UNIVERSITY 1 ' Law Not Significant at .05 65.5% 86.5% 90.1% Economics 63.5% 85.0% 90.4% 65.5 83.2 89.8 UNIVERSITY 2 Law 85.6 95.5 83.5 Not Significant at .05 Medicine 90.4 95.5 95.5 90.4 95.5 95.5 Engineering 83.2 90.7 92.2 82.0 88.0 93.4 UNIVERSITY 3 Economics 59.1 7.2 83.2 9.1 80.5 85.6 Pharmacy 46.7 1.4 93. 4.3 77.5 93.4 Medicine 70.3 76.3 94.6 70.3 76.3 94.0 Dentistry 63.1 76.9 86.8 63.1 72.4 85.0 Engineering 42.3 61.5 84.7 45.1 63. 88.3 Law 62.7 75.4 70.6 Not Significant at .05 UNIVERSITY 4 Economics 79.6 85.0 95.5 79.6 85.0 95.5 Dentistry 95.5 84.4 95.5 95.5 84.4 95.5 Law Not Significant at .05 73.0 92.5 95.5 UNIVERSITY 5 Economics 59.5 85.6 95.5 ‘62.3 87.1 95.5 Medicine 82.0 93.1 93.1 84.7 93.1 94.3 Pharmacy 82.9 95.5 95.5 85.6 95.5 95.5 Dentistry 74.8 89.5 95.5 Not Significant at .05 Law 70.6 82.9 91.9 67.9 82.9 91.9 Sixteen of these faculties evidence an increase in instructional efficiency at successive instructional levels. In Law, university 2, students in level (5-8) do not pass as often as students at lower levels. There is a 12.0% decrease from levels (3 and 4) in this faculty. The same patterns is exhibited in Law, university 3. Dentistry, university 3, shows a decline from levels (1 and 2) to (3 and 4): however, levels (5—8) increase once again. 117 The exceptions to the noted pattern warrant analysis. If instructional efficiency represents a measure of the extent to which faculties are discriminating adequate from inadequate stu- dents, is it possible that Law, universities 2 and 3, are evaluating to a greater extent at levels (5-8) then they were in levels (3 and 4)? The trend indicates that possibly two standards of evaluation are reflected in these data. One standard is used through level 4 and a more rigorous standard is applied in more advanced courses. Dentistry, university 4, appears to evaluate most rigorously at the (3 and 4) year levels 26 of instruction. Conclusions and Qualifications Regarding Level of Instruction and Instructional Efficiency The hypothesis has been generated that instructional efficiency increases at each successive level of instruction. This hypothesis has been substantiated by analyses of univer- sity and selected faculty data. Several pertinent questions were posed that attempt to locate the causes surrounding this relationship. T.o possible answers to this question were elucidated. 26 It should be noted, however, that data are limited to a single academic year. Comparison with performance in other years will further clarify the meaning and significance of these relationships. 118 Method of Instruction and Instructional Efficiency Three methods of instruction have been delineated in previous chapters. They are: the lecture, laboratory, and combination lecture-laboratory methods of instruction. In- structional efficiency measures and methods of instruction were divided by university and faculty data to ascertain pertinent relationships. Recurring patterns were used to generate hypotheses surrounding instructional efficiency and methods of instruction. Data Analysis by University Data spread by university (Table 6.7) indicate that universities l and 3 are the only universities sustaining a significant Chi square value between instructional ef- ficiency and method of instruction. This is consistent with previous data regarding institutional efficiency and methods instruction. TABLE 6.7--PERCENT OF INSTRUCTIONAL EFFICIENCY (A/EX & Al/Exl) BY METHOD OF INSTRUCTION A/LEX *UnIVer- Univer- Univer- Univer- Univer- Method sity_1 sity 2 sity 3 sity 4 sity 5' Lecture 81.4$ 91.0% 80.2% 88.9%' 84.1% Combination 74.5 89. 72.1 87.1 83.8 Laboratory 92.8 90. 78.7 88.3 91.3 Significance NSiat NS at NS at Level .01 .05 .01 .05 .05 AJ/EX'l Lecture #81.7 90.7% 79.0% 88}9% 85.3%? Combination 76. 88. 71.8 87.4 84.1 Laboratory 93.4 90. 80.8 89.5 91.3 Significance NS at NS at ’NS’at Level .01 .05 .Ol .05 .05 119 Universities 1 and 3 exhibit different patterns. In university 1, laboratory courses evidence the highest percentage of instructional efficiency. The lecture method is the second most efficient type of instruction, and combination courses evidence the least instructional efficiency. In university 3, the lecture methOd is the most efficient followed by the laboratory method and then the lecture-laboratory method. The paucity of significant data and the conflicting patterns evidencaiin the two universities with significant values, precludes the generation of a testable hypothesis. The only clearcut evidence at this Juncture is that the combina- tion method of instruction provides the lowest percentage of instructional efficiency. Data Analysis by Selected Faculty Selected faculty data (Table 6.8) indicate an inconclusive pattern. The lack of a large number of faculties with significant relationships is only a partial explanation for the inconclusive trend. 120 hpafiomm mane CH mmmpsoo Showmponma 02* * m.ms m.Hm mo. cc pcmoaeecwfim ecz coapccaem m weHmmm>Hzp * m.mm o.mw * m.mm o.mm moasocoam : SBHmmm>Hzp H.me m.m: m.mm m.ms H.me H.mm mafiecceflwem m weHmmm>Hzp m.mm .m.mw m.:~ m.mm «.mm m.:s ccaoacc: * am.mm afi.w~ mo. pm pcmcamacwam poz Bassoeoom mo. cc ecmoaeacmfim pcz * R~.mm aw.mm mcaecccawcm H WBHmmmEZD hLOpwponmq coapmcHnEoo manpooq hLOpmpopmq coHpmcHQEoo manpooq flxm\a< xm\< H H Ampasccm cceccacm amv oneosmemZH mo oomemz mm A am\ a new KM\¢V wozmHOHmmm gazOHeosmemZH mo ezmommmaum.m mamas 121 Of the six faculties, four employ only lecture and combination methods of instruction. In these faculties, the lecture method elicits the highest percentage of in- structional efficiency. In the two faculties that do em- ploy three methods of instruction, the pattern is irregular and inconclusive. Conclusions and Qualifications Regarding Instructional Efficiency and Method of Instruction The lack of conclusive evidence prevents the generation of hypotheses to describe a relationship between methods of instruction and instructional efficiency. There is a re- duction in the number of faculties that can be included in this category, as six faculties employ the lecture method in all courses. The reduction in usable faculties and the lack of faculties evidencing significant Chi square rela- tionships between method of instruction and instructional efficiency contribute to the inconclusive trends. Contact Hours and Instructional Efficiency As delineated in previous chapter, contact hours (HxE) result from multiplying the hours of instruction and the num- ber of students enrolled in the course. Earlier chapters also have indicated that contact hours may reflect the size of the class as the hours of instruction in most faculties are standardized. Resulting differences within faculties appear to reflect class size differences. 122 Data Analysis by University All universities exhibit at least one significant rela- tionship between instructional efficiency and contact hours. University 5 does not evidence a significant relationship be- tween instructional efficiency measure Al/Exl and contact hours. TABLE 6.9--PERCENT OF INSTRUCTIONAL EFFICIENCY (A/Ex and Al/Exl) BY NUMBER OF CONTACT HOURS ~ A/Ex Univer- ‘Univer- Univer- Univer- ’Univer- HxE sity 1 HxE sity 2 sity 3 sity 4 sity 5 1-1000 89.3% 1-500 92.8% 91.3% 94.3% 84.u$ 1001-2500 82.9 501-1000 93.4 86.2 93.7 86.8 2501-5000 74.2 100112500 89.5 80.2 85.6 87.4 5001+ 73.6 2501-5000 87.9 65.8 82.0 83.8 5001+ 84.1 58.3 81.7 73.0 Significance Level .01 .01 .01 .01 .01 “Maul Al/Exi UULVUL Univer- ‘UnIVer- Univer- ‘Univer- HxE sity 1 HxE sity 2 sity 3 sity 4 sity 5 1-1000 89.5% 1-500 92.8% 90.1% 94.3% 83.5% 1001-2500 84.1 501-1000 93.7 85.3 94.3 87.4 2501-5000 75.4 1001-2500 88.9 79.3 86.2 87.7 2 6 5001+ 71.8 2501-5000 86. 67.3 82-3 84~1 5001+ 82. 57.5 81.7 76.9 Significance NS at Level .01 .Ol .01 .01 .05 A rather consistent pattern is evidenced in the univer- sities. As contact hours increase, instructional efficiency decreases. In university 3, instructional efficiency (A/Ex) decreases 33.5% as contact hours increase from the lowest group (5501+). In other universities, the pattern is rela- 123 tively consistent, but the differences are less pronounCed than in university 3. Minor exceptions to the trend are evidenced in University 2 and university 5. In both universities, there is a slight increase in instructional efficienCy before the prolonged decrease occurs. This inconsistency appears to be of minor importance and does not affect the generated hypothesis that as contact hours increase, instructional efficiency decreases. Data Analysis by Selected Faculty In Table 6.10 it is noted that five of the faculties exhibit patterns of instructional efficiency that differ from the previously stated pattern. That as contact hours in- crease, instructional efficiency declines. 124 anommpwo was» CH mommmao 02* * :.ws m.mm m.mm * A * m.:> m.mm m.mm * somsncsm m.H: :.ms o.mw m.mc m.mm m.H: H.m> c.m~ m.mm m.mm mOHEccccm m weHmmm>Hzc m.mm :.me H.mm m.mo :.m~ pcccfioacwfim poz coacaoc: * m.os :.oc m.mm * pcmcfioacwfim ocz seemaecco m.ee o.ms m.mm w.mm e pecoaoacwfim eoz weaseccom : weHmmm>Hzp m.am ~.m: o.me .mm * H.om m.e: o.me s.am * wcaecccfiwcm m.em m.os c.os .ew o.ww m.~m m.cs m.:~ m.mm H.om soccemnm * * m.mm .om c.mm * * m.mm m.mo o.:o coacccsom H.mm :.me m.mm .ms m. m pccoaoacmam poz mcaeococm m weHmmm>Hz= m.mc ~.me m.mw .mm m.ao pcmcfloficwam poz mcfipcmcfiwcm m.mm m.mm 3.3m .mo s.mm ocmcaoacmfim ecz cchHocz :.me m.mc o.:m m.mm m.mc m.ms o.:m m.mm m.mo m.mm moasccocm m weHmmm>Hz= +Homm comm comm coca com +Homm comm comm coca com mesom -Hcmm -HcoH -Hom -H -Homm -HccH -Hom -H ecmpcoo m.mm m.mm H.mm c.Hm m.mm m.mm m.mm m.Hm supmfipcco m.mw m.cm m.mo m.mm camcficfiemam poz cofipccsom ac.mm Rm.me am.mm am.mo em.m: mm.mw am.mo am.mm wcaecccamcm H SBHmmm>Hzo +Hocm ooom comm oocH +Hcom occm comm coca mpsom aflonm naooa 1H, uHomN naooa 1H poopcoo 4xmfia xm\¢ <1, Bo