(i,— J ——‘w v '. 1'97". : Q. I” 'f ' 'm :2’_ I . ' ".'," H >", ,‘"" ,\,I,I ." , ' : 1': r ,!6,;"3:‘.I E. lfi" "":'":'>:r::"1 _ 1 : ~ : I“ Ir. Iii-1:3 L‘ ’: INT“ " "5"?1..':-"' In. mg»: I? ‘5“ 5‘ A I: I?‘ "'-.§%‘ .. Y .4" \ W9 '1 '71:} ‘3‘? ."V'"(l\l N,.. jfn:.‘: “if,“ u’w'z‘ "Pig-2.). 1:1:- '::.‘.".,I .. ',;~I I '| ..,,:':1 . ~ «I 9; ‘ ‘ 1,1,1 $1.42 '-. ,. -'. _ 4;»; «.4, .4 J fix..." ' ’ “ ““"z’w 5x4' - I It. i‘I ‘- "I o', "’ 3., Y I}, II. I" ”*1 ." 2,;12‘" , ‘v ‘. , I . *0- .. ' a g _ ‘x'k‘vgztst ""{V'V\‘y‘ I,‘ ,I, ,. '-: ~\‘-, -,‘,:‘:‘.:w ,, '6" \ . .4 . , j ‘5‘. ‘ -I’ ,:.. .T ‘ ' I I. ' 'I .l I l ‘ I ' ‘ I ' ‘1 ' I‘ll , '. 4 II- 1, ,_ I ,,"-."..'I’I"‘ , 'I II ' ' "v ‘ _ ,, .. a, . I,I, ' . I ‘l,' , , ,I" ' . ‘, '- I 1 ' ‘ ' ' I. .IU' “ ”I , ' , . '4, , , ,. _ I'll”! ' ' n‘ ... n ‘3 ' “"1 .I_:-$\::.' " :'4 .' ,.' .' 1. ' "' ' “"3: " ' ' ' ' .I:O" ' “I . '1 l' V ' I ' H 1‘ I. , I I ., l , . I _I .h, .._ 'I‘.I I .,. M23; ,,. I, "'II L"', 'Ifltu'm'I 1 4 I .J I I, I 1, y, ,, ,' ‘ " ' ,1 4‘ 1‘ H. .1: 1 1 I}},',,',',",;,IIJ,IIK'fi‘mo' . [II/l / K/ 07/ Ill/WWI 3 1293 01101 0752 LIBRARY Miclrirznn State University This is to certify that the thesis entitled AN APPLIED BEHAVIOR ANALYSIS OF GENERALIZATION IN THE CLASSROOM SETTING presented by Michael Barcroft Medland I has been accepted towards fulfillment of the requirements for jPh.D. degreein' Educational Psychology WP mot Date 4/26/78 0-7830 ’1 J JL'f-l 0 I 2000 © Copyright By MICHAEL BARCROFT MEDLAND 1978 AN APPLIED BEHAVIOR ANALYSIS OF GENERALIZATION IN THE CLASSROOM SETTING BY Michael Barcroft Medland A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling Personnel Services and Educational Psychology 1978 ABSTRACT AN APPLIED BEHAVIOR ANALYSIS OF GENERALIZATION IN THE CLASSROOM SETTING BY Michael Barcroft Medland The failure to consistently promote stimulus, response, and time generalization has been a major block in the use of behavior analysis principles and techniques for the solution of social problems. The present research analyzes why consistent promotion of generalization has proven so difficult in applied settings. An examination of the literature indicated that inadequate programming-for-generalization and the research designs in use were the explanations generally given for the failure to obtain generalization; the programming techniques and experimental designs were believed to in- hibit generalization. These and related concerns were explored with the result that three issues were identified. First, the ABAB reversal design was thought to promote discrimination as opposed to generalization. Second, the concern was that the use of expected rewards (i.e. rewards that the subjects can plan on coming if their behavior meets a specified criteria) also promoted discrimination. Michael B. Medland And third, the present language of generalization appeared inadequate to facilitate studies taking place in social settings. The present research investigated the three issues by using a multi-element/multiple-baseline design in two classrooms. The design allowed generalization to be studied in four closely related settings in each classroom. Both teachers and students were observed across multiple behaviors. The primary behaviors examined were the types of consequences used by the teacher and the on-task and off-task behaviors of the students. Expected rewards were used in the first of two interventions. The results revealed that both generalization and discrimination occurred for teachers and students relative 'to the intervention setting, buttfluaconcepts of behavioral contrast and induction fit the situation and the results with greater rigor than did generalization and discrimina- tion. The results indicated that the ABAB reversal design and the expected rewards explanations were issues that developed out of an inadequate conceptual framework and, thus, could not be resolved. This prompted an examination of the differences between applied research and the experi- mental research where the present conception of generaliza- tion developed. Four differences were noted: the complex- ity of the response contingencies, the complexities of Michael B. Medland past history, the point at which the generalization test is made, and the reciprocal interaction between contin- gencies. Out of the examination of these differences it was concluded that generalization was a problem of behavioral induction and, thus, productive research on generalization depends on the answers to two questions: How can behavior continue to be maintained or changed given initial in-' duction? And how does induction occur for an organism who is in a different spacial/temporal location from where the intervention took place? The answer to the first was outlined in terms of the reciprocal interaction between organism and environment. The answer to the second necessitated asking two questions about the contingencies of reinforcement. They were: How do contingencies of reinforcement interact within the repertoire of an organism? And how are contingencies of reinforcement arranged within and between segments of an individual's history? The last chapter delineated some preliminary answers and, thus, set forth a framework for the analysis of generalization. TO POP ii- ACKNOWLEDGEMENT S In appreciation of their assistance, I would like to thank Walter Hapkiewicz, Thomas Stachnik, Robert Craig, and Robert Winborn. iii Chapter I II TABLE OF CONTENTS INTRODUCTION AND REVIEW OF THE LITERATURE A. THE CONCEPT OF GENERALIZATION B. PROGRAMMING FOR GENERALIZATION l. The Extent of the Gen- eralization Problem 2. The Techniques Used to Program for Generalization a. The manipulation of ante- cedent conditions b. The manipulation of tar- get and target-related behaviors c. The manipulation of consequence conditions d. Contingency manipulation: package techniques C. EXPERIMENTAL DESIGN AND GENERALIZA- TION RESEARCH l. The ABAB Design 2. The MB Design D. RATIONALE FOR THE RESEARCH PROCEDURES A. BEHAVIORAL DEFINITIONS 1. Teacher Behavior 2. Student Behaviors iv Page l3 l4 17 20 23 28 29 31 36 39 40 41 46 Chapter III Page B. EXPERIMENTAL DESIGN 48 1. General Description and Logic 48 2. Experimental Phases 51 a. Prebaseline 51 b. Baseline 51 c. Game 53 d. Game plus Teacher Attention 53 C. EXPERIMENTAL PROCEDURES 54 1. Setting and Subjects 54 a. Setting 54 b. Teachers 54 c. Students 54 2. Observation and Recording Procedures 55 a. Observers: training and reliability 55 b. Observational techniques 55 3. Intervention Procedures 59 a. Group contingency game procedure 59 b. Teacher attention pro- cedure 63 4. Teacher Training Procedures 64 a. Game training 55 b. Attention training 65 5. Summary and Schedule of Experimental Procedures 67 RESULTS 69 A. OBSERVER RELIABILITY 70 B. STUDENT BEHAVIOR 71 C. TEACHER BEHAVIOR 93 Chapter Page IV DIschSION 106 A. TERMINOLOGY 106 B. EXPERIMENTAL DESIGN 108 c. EXPECTED REWARDS 111 D. THE DIFFERENCES BETWEEN BASIC ‘AND APPLIED RESEARCH ON GEN- ERALIZATION 112 1. Complexity of Response Contingencies 115 2. Complexity of Past History 116 3. The Point at Which the Gen- realization Test is Made 117 4. The Reciprocal Interaction of Contingencies 118 V A FRAMEWORK FOR THE ANALYSIS OF GENERALIZATION 122 A. THE INTERACTION OF CONTINGENCIES OF REINFORCEMENT 123 1. The Single Contingency 123 2. Alternate Sets 126 3. The Interaction of Con- tingencies and the Concept of Identity 130 B. THE ARRANGEMENT OF CONTINGENCIES OF REINFORCEMENT 131 1. Arrangements of Contingencies 135 2. Examples of Arrangements 141 REFERENCES 161 APPENDICES A Teacher and Student Behaviors 173 B Student and Teacher Recording Forms 181 C Reinforcement Menues 5th Grade and 2nd Grade 183 vi Chapter Page APPENDICES (continued) D Game Procedures 192 E Teaching/Learning Moded used in Teacher Training 202 vii Figure 2.1 2.3 3.1 3.7 3.8 3.10 LIST OF FIGURES Multi-element/multi-baseline design across subjects, settings, and be- haviors Experimental settings examined in the multi—element/multi-baseline design. Research schedule Percent change from baseline in on- task behavior for 2nd grade subjects across phases and settings Percent change from baseline in on- task behavior for 5th grade subjects across phases and settings On-task behavior for subject 2 across phases and settings On-task behavior for subject 3 across phases and settings On-task behavior for subject 7 across phases and settings On-task behavior for subject 10 across phases and settings The mean percentage of consequence be- havior for each teacher relative to their baseline performance across inter- vention phases and settings The mean percent of continuation and change behavior across phases and settings Teacher 1 change attention as the per- centage of total behavior across phases and settings viii Page 50 52 68 76 80 87 87 89 9O 96 99 102 Figure Page 5.1 The arrangement of contingencies for the Guttman and Kalish (1956) non- discrimination generalization experi- ment. 143 5.2 The arrangement of contingencies for the Blough (1967) experiment 144 5.3 A comparison of two ways to test for generalization 145 5.4 The arrangement of contingencies for the Hanson (1959) peak shift experi- ment 147 5.5 A comparison of two ways to train for generalization 148 5.6 The arrangement of contingencies for the Renold's (1961) induction and contrast experiment 150 5-7 Schematic diagrams of the four types of behavioral interaction . 153 5.8 The arrangement of contingencies for the Jacquet (1972) schedule induced licking experiment 155 5.9 The arrangement of contingencies for the present experiment 158 ix Table 3.1 LIST OF TABLES The mean percent of classroom-time spent in independent-study by student, experimental phase, and setting for each grade (classroom) The mean percent of on-task behavior for independent-study-time by subject, experimental phase, and setting for each grade (classroom) The mean number of 10 second intervals per hour containing a continuation or change comment by the teacher which were related to the subject across experimental phases and settings Mean percent of class time each teacher spent engaged in instruction con- sequence, and no-response behaviors across settings and experimental phases Page 72 75 84 94 CHAPTER I INTRODUCTION AND REVIEW OF THE LITERATURE Applied behavior analysis has encountered a major problem: the failure to consistently promote stimulus, response, and time generalization. Since the first rec- ognition of the problem (Baer, Wolf, and Risely, 1968), it has been discussed increasingly in the literature, and during the last few years major research has been under- taken to find a solution (e.g., Stokes and Baer, 1977; Conway and Bucher, 1977). Research preceding that proposed here has taken two main thrusts. The first is based on the belief that the failure to promote generalization rests with the re- search designs used by behavior analysts, the second on the belief that adequate "generalization programming," using behavior principles, has been left largely untapped. Solving the design problem would require an alteration in research methodology. Solving the pro- gramming problem would require the rigorous application of behavior principles: the programmers would have to "program for generalization." The two tactics are not mutually exclusive. Im- proved research designs must evaluate unequivocally the l data produced by the programming, and the programming must apply behavior principles that would promote stimulus, re- sponse, and/or time generalization whenever such generaliza- tions would be useful. The prOposed research will attempt to solve the generalization problem by both logical and experimental analysis. The logical analysis will focus on the explana- tions for the failure to promote generalization. The re- mainder of this chapter will 1) examine the concept of generalization, 2) review the criticism of research de- signs and programming, and 3) frame a rationale for the research proposed. The discussion of generalization and the review will provide a background for the rationale. A. THE CONCEPT OF GENERALIZATION Applied behavior analysts consider generalization in terms of two main categories: stimulus generalization and response generalization. When conditioned behavior is emitted under stimulus conditions different from those present in training, the term stimulus generalization is applied. There are two types of stimulus generalization. One type of stimulus generalization, referred to as response maintenance or resistence to extinction, is the maintenance of behavior in a given setting after inter- vention has been terminated (Kazdin, 1975 a ; Wahler, 1969). The "given setting" can be anything from the setting in which the intervention took place but has since been removed, to a setting quite distant from the intervention point (e.g., from the classroom to the home setting). The maintenance of behavior is seen as gen- erlization across stimulus conditions because behavior change in one situation (e.g., a classroom in which a "game" intervention is in effect) may generalize to another situation (.e.g, the same classroom after the program is withdrawn). The maintenance of behavior is considered the most important issue after behavior change has been achieved (Jones and Kazdin, 1975; Kazdin, 1977a Koegel and Rincover, 1977). The second type of stimulus generalization, re- ferred to as transfer of training, is a transfer of behavior from the program intervention situation or setting to another in which no program has been in effect. Unlike the response maintenance situation, the inter- vention program is in effect when the behavior change occurs in the non-program setting. Transfer of training is the concern of most applied treatment interventions in that the behavior change achieved in one setting (e.g. ward or classroom) was trained for the purpose of being utilized in another (e.g., community or home) (Kazdin, 1977; Marholin, Siegel, and Phillips, 1976). The categories of response maintenance and trans- fer of training can often go together in an intervention program. For example, in most intervention programs there could be a two-fold goal, to ensure that behavior change is continued when the individual is removed from the program (maintenance) and placed back into a non- training setting (transfer). Until recently this has not been a major concern for applied behavior analysts (Kazdin, 1977; Marholin et a1., 1976; Stokes and Baer, 1977). But notice the distortion that has taken place: maintenance is considered in a never-was-a-treatment en- vironment and transfer is considered after the termination of the intervention. Response generalization, often referred to as response induction, is said to occur when changes in re- sponse classes never directly conditioned covary with changes in a reinforced response class (Kazdin, 1975a; Reynolds, 1975). The covariation can occur "within" (Garcia, 1974) or "outside" (Cooke and Apolloni, 1976) the intervention setting. Within setting response induction, for example, would involve a change in mathematics per- formance (e.g., number of problems correct) when on-task during mathematics (e.g., Looking at or attending to materials) is being reinforced. Outside setting response induction would involve a change in reading performance during reading period when mathematics behavior is being reinforced in mathematics period. In both cases response induction refers to the change in the probability of a response that exists in the repertoire of the individual. Within and outside response induction has been little explored but is of primary interest to behavior analysts. Response induction is also closely related to "shaping". The concern is for developing a response not in the repertoire of the individual but which can be progressively approximated by using differential re- inforcement on the response variations that occur. In this way a response class is induced which was not in the individual repertoire. Although response induction, as related to shaping and differential reinforcement, is important for applied behavior analysis, it has not been the concern of those interested in generalization in the applied setting. The concept of reSponse induction is not without its difficulties. The difficulty seems to pivot about the concept of response class with the topics of response differentiation and response variability closely involved. For example, if one has not specified a functional re- sponse class (i.e., one thatijscapable of measuring be- havior change that could be due to the intervention manip- ulations) one may reach conclusions that are grossly in- accurate. Goldiamond (1975), in a long and complex paper, advocates the use of "alternate sets" of response classes. By alternate sets he is referring to the de- finition and analysis of classes related to the target response class. Tflmaidea appears to be one of delineating response classes in terms of a universal set (all possible behaviors in the situation of interest are in some way included), sets (the classes normally specified as target, plus, the alternate sets which remain and fill out the universal set),and subsets (the classes involved in a micro-analysis of the response sets). Al- though difficult to implement, the present research attempted it. Conway and Bucher (1977) clarify the relationships between the types of generalization: response maintenance, transfer of training, and response induction. They ask, "What behavior was changed, under what stimulus conditions, and for how long a time." They portray the relationships in a three-way matrix: // / Target V/////Durable BEHAVIOR TIME Nontarget Immediate Training Other STIMULUS CONDITIONS Besides the relationship between response maintenance (time on the matrix) and transfer of training (stimulus conditions on the matrix) that are mentioned above, the relationship of response induction (behavior on the matrix) to maintenance and transfer are apparent. One can see, for example, that induced behavior within or outside the training setting can be of short or long term duration. The ability to assess such three way changes (i.e., across behaviors, settings, and time) would re- quire implementing Goldiamond's ideas on alternate sets. It is important to explicitly state that the forms of generalization defined above are dependent variables: the maintenance, transfer, and induction of behavior are all dependent on the variation of antecedent and/or consequence stimulus conditions. In other words generalization is, in theory, explainable, predictable and controllable. It is the problem that applied be- havior analysts are having in explaining, predicting, and controlling generalization that has prompted the present research. Some of the research problems seem to be con- nected to terminology. Although Conway and Bucher (1977) appear to put terminology in a nice concise package, a close examination of the package leaves one with the B. PROGRAMMING FOR GENERALIZATION The programming for generalization approach to the generalization problem has asked and attempted to answer two basic questions: to what extent is the generalization problem being investigated and how does one program for generalization? Closely related to the above questions and of special interest is knowledge about the success or failure rate of attempts to program for generalization. At the outset it is important to realize that the cate- gorization of programming into success or failure is extremely tenuous. First of all no researcher has as yet specified to what degree generalization must take place in order to say that the procedures used were,without qualification,successful. In general when researchers have investigated generalization intfiuaapplied setting, they have taken any situation which happens to be avail- able even though it may not be ideal. Second, one wonders at what point the research should be categorized as programming for generalization. When researchers investigate some form of generalization, can one say that they are programming for generalization? From the perspective of applied research with its emphasis on social relevance, one would perhaps give a spontaneous yes. But from the perspective of the re- searcher with more than a sprinkling of curiosity and limited resources, one's answer may even be a spontaneous no. It has only been very recently that the researcher has come to the point of specifying research activities with the needed detail in regards to generalization (e.g. Kazdin, 1975b; Koegel and Rincover, 1977). It is with the above considerations that a reader of a review and/or a reviewer must observe generalization research to date. The review that follows will pivot around the extent to which a generalization problem exists and the techniques used for programming. Within both contexts the success or failure of programming will be examined. The coverage of the review will be across much of applied behavior analysis but will only detail examples related to the immediate research areas (i.e., teacher and student behaviors in an educational setting). 1. The Extent of Generalization Research Generalization research has recently been reviewed by a number of researchers (Conway and Bucher, 1977; Keeley, Shemberg, and Carbonell, 1976; Marholin and Siegel, 1977; Marholin, Siegel, and Phillips, 1976; Warren, 1976; Stokes and Baer, 1977). Several of these address them- selves to the question regarding the magnitude of the generalization problem in applied behavior analysis. Stokes and Baer (1977) found only 270 behavioral studies relevant to generalization. Of these 120, some 44%, were classified by them as contributing to a 10 "technology of generalization." Although they did not clearly distinguish between "relevant" and "contribute", they appear to mean in the former case any study that in an even ancillary way considered examining generalization (i.e., a two week post check by telephone as an examina- tion of maintenance). In the latter case they appear to be concerned with some form of "rigorous" examination of generalization (i.e., a three to six month follow-up using observational procedures). Out of the thousands of studies published in the last ten or so years, to have only 270 consider generalization seems like an extremely small proportion and 120 seems even less so. Warren (1976) focused on the Journal of Applied Behavior Analysis because it is the least practitioner oriented, and reviewed generalization research for the years 1973 to 1976. He found that 94 of the 159 studies contained in the journal could be viewed as immediately contributing to an applied behavioral technology and were connected with a real—life setting on real-life problems. Of these, 53 of the studies (56% of the sample) failed in any way to address the issue of generalization. Of the remaining 41 studies he found that 28 (30% of the sample of 94 studies) directly measured transfer of train- ing,20 (21% of the sample) directly measured maintenance, and 27 (29% Of the sample) directly measured some form of induction. 11 Finally, Keeley, Shemberg, and Carbonell (1976) examined not only the extent to which researchers tackled the problem of generalization but the extent to which researchers have been successful. They reviewed the operant studies published in Behavior Therapy: Behavior Research and Therapy, and the Journal of Applied Behavior Analysis for the years 1972 and 1973 which involved manipulating environmental consequences to alter human behavior. Thus, this qualification included operant self-control but excluded modeling studies. Of the 146 projects which met their criterion, seventeen (11.6% of total 146 analyzed) were concerned with long term maintenance (6 months or more). Three of these were called "unqualified successes." There were four which were categorized as "qualified successes" which meant that changes were in the appropriate direc- tion, but serious methodological problems precluded a clear interpretation of outcome. Two were reported as "mixed successes" which meant that in only some of the subjects or behaviors were changes significant in the traditional or absolute sense. The remaining eight studies "represent six single case reports, and two multiple studies, which lack systematic, objective data, precluding meaningful interpretations of outcomes." Transfer of training was investigated in fifteen studies (10.3% of total). Seven were considered failures 12 and four were successes. The other four investigations included a qualified, a marginal, and two mixed successes. Only one of these studies examined transfer over the long- term. Thirteen of the studies (8.9% of total) reported short-term response induction. Of these two were judged successes, one a qualified success, three as mixed successes, one a marginal success, and six as failures. None of the studies specifically measured long-term induction. The conclusions reached by these reviewers are in general agreement. Keeley et a1. (1976) state: The present results suggest that re- searchers stop flooding the literature with demonstration studies of the obvious and sometimes trivial. Continued presentation of reversal designs demonstrating nothing more than stimulus control seems unwarrented...Research establishing boundary conditions of the approach, whatever these are, is sorely needed. Also, researchers must accept the responsibility of studying complex clinical problems in complex settings. Warren (1976) agrees but his conclusions are framed by a different perspective: Recommending to behavior analysts that they measure for generalization is not a novel behavior. But the discovery of how little they do it, even in the most applied situation is disheartening...[and] while the costs of generalization assessments can be quite high, their value to the field should make them worth it. 13 The above results, no matter how conservatively they are approached, lead to the conclusion that there is in fact an extensive generalization problem. If a technology of generalization is developing as Stokes and Baer (1977) contend, they were correct to call it an "implicit" and "embryonic" one. 2. The Techniques Used to Program for Generalization The techniques used by the implicit and embryonic technology of generalization have been given various names and positions of importance (Cf. Kazdin, 1977a; Marholin et al., 1976; Stokes and Baer, 1977; Sundel and Sundel, 1975). Researchers utilizing the techniques have two points at which they can attempt to induce gen- eralization. They can program the treatment environment or they can program the non-treatment environment. They can of course do both if the economic, social, and/or political resources are in their favor. But no matter where they choose to intervene, they have to manipulate in some manner the components of the contingencies of reinforcement at their disposal. The manner in which they manipulate a component or components, define a technique. For the most part, the techniques used are consistent with the experimental analysis of behavior. The purpose here is to give these techniques an even closer relationship to the work done in the ex- perimental analysis of behavior. This will be done by 14 relating the techniques and the various studies which are representative of their implementation to the three terms of a contingency of reinforcement. These include the antecedent stimulus (SD), the response (R), and the consequent stimulus (SR). a. The manipulation of antecedent conditions. Generally this is the area called stimulus control. The techniques used here are most readily employed in the treatment setting and are most applicable to programs aimed at promoting transfer and short-term maintenance. There are two primary categories of techniques. The first, often called programming common stimuli or train- ing common exemplars (Cf., Marholin et al., 1976; Stokes and Baer, 1977), is aimed at associating the target be- havior with a broad range of cues. The use of peer trainers or tutors has proved effective because they are also in the non-treatment environment (e.g., Johnson and Johnson, 1972). Various objects that have a high probability of appearing in the non-treatment setting can be used as part of the treatment setting (e.g., Allen, 1973; Griffiths and Craighead, 1972; Rincover and Koegel, 1975). This type of antecedent control has been called con- texual control. Multiple trainers have been used to promote transfer across individuals. This is especially prev- 15 alent in language training (e.g., Garcia, 1974; Stokes, Baer, and Jackson, 1974). The placement of “significant others" in a prob- lem setting has been shown effective in altering children's behavior (Meddock, Parsons, and Hill, 1971; Peterson, Merwin, Moyer, and Whitehurst, 1971; Peterson and Whitehurst, 1971; Redd and Wheeler, 1973). The concern is not for the use of the adult as a reinforcing agent but for the adults function as a discriminative stimulus. For example, the entrance of a significant other into aproblem classroom could stop inappropriate behavior and set the occasion for appropriate behavior that may be part of a reinforcing contingency within the setting. In principle, if the adult is there long enough and the reinforcing contingency is strong enough, the appropriate behavior may persist in the setting when the adult is absent. Thus, transfer and maintenance occur. The problem with these techniques is that it is in fact difficult to tell if sufficient exemplars or common stimuli have been used (e.g. Allen, 1973; Stokes, et al. 1974; Rincover and Koegel, 1975). Without the availability of reinforcement in the non-training setting, the transfer that may be achieved would be short-lived The second category of antecedent techniques has been referred to as "instructional control" (Marholin et al., 1976) and "rule-governed behavior" (Skinner, 1969). In 16 this case the antecedent stimulus represents a contingency of reinforcement that exists in the environment (i.e., the rule as an antecedent stimulus displayed to the in- dividual represents all of or some fraction of the three terms for contingency of reinforcement). The rule can be represented verbally (e.g. Madsen, Becker, and Thomas, 1968; Steinman, 1970a, 1970b; Wilcox, Meddock, and Steinman, 1973) or modeled (Bandura, 1965, 1969; Frederiksen, Jenkins, Foy, and Eisler, 1976; Gladstone and Sherman, 1975; Martin, 1975). Tflmaoperation of rule governed techniques are dependent on a preceding history of re- inforcement for imitating and instruction following. If such a history has occurred, the rule will set the occasion for the rule related behavior. I O'Leary, Becker, Evans, and Saudargas (1969) and Packard (1970) have clearly demonstrated that in- structions that have not had a history of reinforcement for being followed do not in fact alter behavior of school children. Medland and Stachnik (1972), on the other hand, have shown that once rules are given a history of rein- forcement for being followed, they can at a later date exert some degree of control over the behavior of elementary school children. But the use of explicit instructions that detail the where, when, and how of a trained behavior has also been shown to perform a discriminative function (i.e., 17 no transfer to other settings). Horton (1975) trained two elementary teachers to use behavior-specific praise. They were told in detail where, when, and how to use it. The result was no transfer to other periods of instruction. Lovaas, Koegel, Simmons, and Long (1973) explored transfer and maintenance of behavior that has been modeled, imitated, and reinforced. They found that the autistic children transferred but that maintenance was dependent of a supportive, reinforcing post treatment setting. b. The manipulation of target and tapget-related behavior. The researcher who wishes (or has no other choice) to induce generalization by manipulating response classes is confronted with all of the problems related to response induction (i.e., the definition of response class, response variability, response diversity, and response differentiation) that were outlined in the terminology section above. It can even involve selecting a behavior that is not a primary target behavior but could lead to the development and increased probability in the non-treatment setting of a response class of primary interest. The decision one makes regarding the classification of the behaviors of interest as instances or response classes will determine if one is going to speak of generalization of these behaviors in terms of transfer of training or response induction. 18 But pragmatically there are three classes of techniques for inducing generalization through the manipulation of target or target-related responses. The first class involves the selection of behaviors that are compatible with or entry behaviors to other behaviors that are reinforced in the non-training environment. Baer and Wolf (1970) have used the term "behavioral trap" to define this area of techniques. These methods have been used to promote general social behaviors (e.g., Allen, Hart, Buell, Harris, and Wolf, 1964; Altman, 1971; Buell, Stoddard, Harris, and Baer, 1968; Cooke and Apolloni, 1976; Hauserman, Walen, and Behling, 1973; Hingtgen and Trost, 1966; Whitman, Mercurio, and Caponigri, 1970), language development (e.g., Martin, 1975; Merchenbaum, 1969; Schumaker and Sherman, 1970; Wheeler and Sulzer, 1970), and academic behaviors (e.g., Hay, Hay, and Nelson, 1977). Hay et a1. (1977) examined the change in elemen- tary student on-task and academic behaviors depending on which of the behaviors was targeted. They found that academic contingencies had a reliable positive effect on on-task contingencies but the latter did not reliably change academic performance. Their findings were con- sistent with the literature (e.g., Ayllon and Roberts, 1974; Kirby and Shields, 1972; Ferritor, Buckholdt, Hamblin, and Smith, 1972). 19 The second class of techniques used to induce generalization through the manipulation of target or non-target behaviors istflueselection and training of behaviors that are incompatible with the non-treatment target behavior (i.e., a behavior that is not physically possible if the trained behavior is omitted). A number of studies have reported the decrease in disruptive be- haviors that were not under direct contingency control (e.g., Ayllon and Roberts, 1974; Burchard and Tyler, 1965; Marholin, Steinman, McInnis, and Heads, 1975; Winkler, 1970; Winett and Roach, 1973). The third class of response manipulation techniques involves training a response class in which the gen- eralization of that reSponse class is reinforced (Stokes and Baer, 1977). The studies in the area of "creativity" utilize this technique (e.g., Goetz and Beer, 1971, 1973; Fallon and Goetz, 1975; Goetz and Salmonson, 1972; Holman, Goetz, and Baer, 1976; Parsonson and Baer, 1977). They have in general found transfer, induction, and maintenance. Mothers and teachers have also been trained by generalized response techniques (e.g., Herbert and Baer, 1972; Parsonson, Baer, and Baer, 1974). Parsonson et a1. (1974) taught teachers to administer consequences by using a wide range of examples and non-examples for both the appropriate and inappropriate behaviors of retarded pre- schoolers. They found that not only did the behavior of 20 the teachers transfer but was still maintained after several months. The obvious merit of the above types of training techniques are that one can judiciously select target behaviors that can promote generalization by virtue of their relation to non-training contingencies. The obvious drawback to the effective use of the techniques is the lack of available knowledge with regard to the interrelatedness of behaviors. c. The manipulation of consequence conditions. The focus in this section will be upon the manipulation of positive consequence conditions. The punishment technique (e.g., response cost, time out, reprimands) will not be examined. In general the aim of the positive techniques is tied to long-term maintenance once transfer is accomplished. There are five techniques that are generally available. They include: (1) intermittent reinforcement (2) delay of reinforcement (3) establishing social stimuli as reinforcers (4) employ non-treatment personnel as reinforcing agents (5) employ vicarious consequences. The use of intermittent reinforcement has been shown to be effective in maintaining stable patterns of behavior (Ferster and Skinner, 1957). But it can also 21 be seen as promoting transfer. If the subject cannot dis- criminate in which setting the reinforcement will occur, behavior may come to be displayed in all of them (Stokes and Baer, 1977). For example, Schwarz and Hawkins (1970) reinforced a child for good-posture, absence of face- touching and appropriate voice loudness as recorded on video tape. Mathematics and spelling sessions were shown back-to-back but only mathematics period behaviors were reinforced. The result was that spelling behaviors changed in desired directions also. In terms of the maintenance of behavior, several researchers have called attention to the fact that further research is needed with regard to human subjects and the use of intermittent re- inforcement (Marholin et al., 1976; Stokes and Baer, 1977). Delay of reinforcement procedures have been used in two ways. The first involves the delay between response and consequence (e.g., Schwartz and Hawkins, 1970), and the second increases progressively the time between token reinforcement and backup reinforcement ex- change (Marholin et al., 1976). Using the second delay procedure, O'Leary and Becker (1967) found that token reinforced behavior did not decrease. The establishment of social stimuli (i.e., praise, smiling, touching) as reinforcers has been used extensively in work with the profoundly retarded (e.g., Lovaas, Freilag, Kinder, Rubenstein, Schaeffer, and Simmons, 1966; 22 Long, 1969; Walher, 1969). Social stimuli have been shown to be effective in most cases (e.g., Broden, Bruce, Mitchell, Carter, and Hall, 1970; Pinkston, Reese, and LeBlanc, 1973). But a number of studies have not found social stimuli reinforcing (e.g., Lovaas et al., 1966; Quay and Hunt, 1965; Walker, 1969). The fourth procedure for enhancing transfer and maintenance is in the use of non-treatment personnel as reinforcing agents. Parents, teachers, siblings, and peers have been employed in this capacity as behavioral change agents. The training of these individuals has received a great deal of research. Reviews have been done by O'Dell (1974) and Patterson (1971) with several other general reviews giving it considerable attention (e.g., Conway and Bucher, 1977; O'Leary and O'Leary, 1976). Re- searchers have shown the effectiveness of parents (e.g., Hawkins, Peterson, Schweid and Bijow, 1966; Patterson, McNeal, Hawkins, and Phelps, 1967), teachers (e.g., Madsen et al., 1968; Phillips, 1974), and peers (Bailey, Timbers, Phillips, and Wolf, 1971; Siegel and Steinman, 1975; Solomon and Wahler,1973; Surratt, Ulrich and Hawkins, 1969) in successfully modifying the behavior of deviant children. The employment of vicarious consequences is a newly developed technique. Although studies have re- ported behavior changes of subjects who have not been 23 directly exposed to the reinforcement contingencies (e.g., Bolstad and Johnson, 1972; Broden et al., 1970; Patterson, 1974; Tracey, Briddell, and Wilson, 1974), it was not until Kazdin (1973a, 1977b) began his systematic studies that one could clearly consider reinforcing the behavior of one individual as a way to change the behavior of another in the applied setting. The subject in this case is influenced by antecedent (stimulus) control but the re- searcher makes this occur by manipulating consequence con~ ditions for another subject. Thus, the technique is classified in the consequence manipulation category. To classify the technique in the consequence category is a moot point. What is not moot is that the subject must have a past history of rule following in order for the treatment to be effective. d. Contingengy manipulation: package techniques. By "package technique" it is meant that the researcher uses a number of component techniques to obtain behavior change and promote generalization. For example, in teacher training, Horton (1975) used instructions (ante- cedent conditions), modeling (antecedent conditions), graphic displays (antecedent conditions), and intermittent reinforcement (consequence conditions) to promote behavior change and investigate transfer. He was in fact using a package of techniques all discussed above. 24 Games, tokens, and contracts are three of the most popular package techniques used in behavior analysis re- search. As implemented to date, they all have one element in common: expected rewards. Expected rewards are of interest at this point because they are seen as powerful methods for rapidly changing behavior but at the same time are seen as techniques that impede the maintenance of be- havior. If one is programming-for-generalization, the ability to change behavior rapidly in the intervention setting is important but it is as equally if not more im- portant to maintain the behavior change once the interven- tion has been terminated. Thus, it seems important that such a problem is resolved. Essentially, expected rewards can be defined as preintervention statements that describe the contingencies of reinforcement under which an individual will be operating during an intervention. For example, in classroom game procedures, the specification of the game's rules delineates the contingencies; in token economies the rules for token reception spell out the contingencies; and in behavioral contracts the rules of the contractual arrangement repre- sent the contingencies. Expected rewards are, thus, rules or contracts representing contingencies. In all these cases the contingencies are specified prior to the instatement of the intervention. 25 At least two studies have concerned themselves with expected rewards and the findings were that expected re- wards led to faster extinction than unexpected rewards. The authors concluded that a prestatement of contingencies clearly restricted the maintenance of focal behavior (Resnick, Forehand, and Peed, 1974; Kazdin and Polster, 1973). For some the resolution of the problem of expected rewards seems to deal with the relationship between con- tracted rewards and intermittent reinforcement, one of the basic procedures for the facilitation of behavior maintenance. The question for these researchers has become one of how do you make explicit the how much and the whats of a contingency contract and still maintain unexpected rewards? According to Kazdin (1975a), "intermittent rein- forcement can be readily incorporated into all programs" and that "it is important to make the schedule of reinforce- ment increasingly intermittent", but how one goes from one explicitly negotiated contract to another, where the second requires more behavior for the same reward, is not explained. A number of writers have attempted to develop techniques but few of these have been evaluated or have been put in a contract context (Schaefer and Martin, 1966; Krumboltz and Krumboltz, 1972; Walker and Buckley, 1974; Macht, 1975). 26 Levine and Fasnacht (1974) have taken another approach. They have discussed expected rewards in the con- text of intrinsic and extrinsic reward. They warn that expected rewards may preclude generalization and mainten- ance because they are extrinsic rewards, rewards perceived by the subject as being external to the behavior being rewarded. The argument they present is that intrinsic rewards, rewards that are connected to or are derived from the behavior being rewarded, would foster generalization and maintenance of behavior. Some examples of extrinsic rewards would be pay, promotion, fringe benefits, etc. Intrinsic rewards would include those over which the in- dividual has a high degree of self-control and are a part of the activity itself (Notz, 1975). Theproblem with connecting intrinsic and extrinsic rewards with expected rewards is that one turns to internal mechanisms for ex- planations and,thus, loses contact with what can be manip- ulated: the contingencies of reinforcement. One obvious way to resolve the problem of expected rewards would be to discontinue their use, but there are two reasons for not doing so. First, there are the positive aspects of expected rewards, one of which is allowing individuals to check their progress against an established standard, and another is the immediate effect expected reward procedures have on target behaviors. Second, the specification of an expected reward is for the 27 most part a statement of an instructional objective and, as such, represents the application of a foundational con- cept in educational technology. Both contain the condi- tions under which behavior is to occur, the behavior, and the performance standards for the behavior. Expected rewards also specify the consequences. To restrict their use in the area of applied behavior analysis would be un- warranted given the evidence to date. The above unresolved matter of expected rewards will likely remain a mystery until it is seen by the applied researcher as resulting from a larger underlying issue. The issue phrased as a question is: what con- tingency is being established? Skinner put it as follows: Arranging contingencies of operant re- inforcement is often confused with des- cribing them. The distinction is as important as that between contingency shaped and rule-governed behavior (1974). This emphasizes that the contingency established by a researcher is not necessarily the one the subject responds to. The behavior may be of the same form but the controlling variables are different. Thus, they are different operants as are contingency shaped and rule-governed behaviors with the same tOpography. Beiser (1974) suggests that the contingency that is established might be called the strategy of contracting and not the target behavior of interest. The differences in the controlling variables between the two behaviors 28 appears to be related to the formation of a discrimination in which it is only reinforcing to respond when contractual arrangements in the form of games, tokens, contracts, etc. are available. If this is the case, faster extinction would occur during the non-contract situations. The pre- sent research attempts to assess the extent to which ex- pected rewards are involved in the establishment of dis- criminated behavior which is detrimental to the maintenance of behavior. C. EXPERIMENTAL DESIGNS AND GENERALIZATIONS The discussion of the experimental design explana- tion will focus on the two major designs used in applied behavior analysis research: The ABAB reversal design (ABAB design) and the multiple baseline design (MB design). The ABAB design involves the switching of baseline and intervention conditions. The MB design staggers the introduction of interventions for different behaviors, settings, and/or individuals. The variations of each design simply involve the use of more interventions and/or the use of various staggering strategies across varying numbers of behaviors, settings, and/or individuals. The experimental design critics have outlined a number of inadequacies in the ABAB and MB designs. It is the purpose of this section to examine and then 29 evaluate their comments in relation to research on gen- eralization. l. The ABAB Design The ABAB design and its extensions (e.g., A-B-A- B-BC reversal design) functioned for an extensive period as Applied Behavior Analysis' main approach to the deter- mination of functional relations resulting from the manipulation of the independent variable and the changes occurring in the dependent variable. But as applied research progressed from an emphasis aimed at demonstrating functional relations between basic behavioral principles and behavior to one of developing a technology of behavior, it was expressed by many that the ABAB design was inadequate for the task (Hensen and Barlow, 1976; Kazdin, 1973b, 1975a; Kazdin and Bootzin, 1972; Marholin et al., 1976). Kazdin (1973b, 1975a) cites two defects in the ABAB design. In the first he states that the abrupt change in procedures that characterize the design, re- duce the probability that generalization will take place. The abrupt changes, it is said, provide cues that lead to a discrimination between reward and non-reward conditions. Tennov (1976) adds that such cues are increased when expected rewards (e.g., tokens, games, and contracts) are utilized. Essentially the argument is that the ex- 30 perimental design is in principle operating like a dis- crimination training session. The result is a preclusion of generalization. The second charge of inadequacy is that of multiple- treatment interference. Multiple-treatment interference exists when prior treatments are not erasable and, thus, does not allow the behavior to return to baseline when the reversal condition is reinstated (Campbell and Stanley, 1966). Such interference is often called a carry-over effect (Hensen and Barlow, 1976). A number of researchers have encountered this problem (e.g., Hawkins, Peterson, and Bijou, 1966; Medland and Stachnik, 1972). The point of importance is that the multiple- treatment or carry over effect is in fact a description of non-reversibility which is generalization. It is this occurrence of generalization that is believed to break up the possibility of clearly issuing a cause and effect statement about the effectiveness of the intervention (independent variable) on the behavior of interest (de- pendent variable). Because one must be able to make clear cause and effect statements about the intervention before generalization can be usefully evaluated, the non-reversibility of behavior precludes the use of the design in the study of generalization. This argument is easily dismissed if one realizes that the thinking be- 31 hind it is related to seeing the ABAB design in terms of igpra—subject replication. If one uses the design in terms of inter-subject replication, the design can be used to assess cause and effect, and generalization. Marholin et a1. (1976) has stated that the return-to- baseline condition is actually a transfer probe. This is true only if one sees the ABAB design as an inter- subject replication procedure. Although thesmultiple-treatment argument can be dismissed, the discrimination argument cannot. The extent to which the charge is true needs to be evaluated. If it is in fact true, the number of designs available to study generalization have been essentially reduced by 50%. 2. The MB Design The MB design problems have been pointed out by Kazdin (1973, 1975a, 1975b, 1975c), Leitenberg (1973), and later by Hensen and Barlow (1976). The major con- cern with the design as related to generalization is the problem of interdependentcn:interrelated behaviors, situations, and/or subjects. Leitenberg (1973) describes the interrelated behavior problem as follows: "If general effects on multiple be- haviors were observed after treatment had been applied to only one, there would be no way to clearly interpret the results. Such results may reflect specific therapeutic effect and subsequent response generaliza- tion, or may simply reflect nonspecific therapeutic procedure under investigation" (p. 95). 32 Of importance for the MB design is that the interdependence problem is a description of the occurrence of generaliza- tion. The absence of a specific effect is assumed to suggest that some extraneous event, rather than treatment, may have led to changes across several behaviors. The re- sult is that cause and effect statements cannot be made. Kazdin (1975c) further outlines the problems with the MB design by pointing out the inconsistencies in the design's basic assumptions. He states: ...the design appears to depend upon two conflicting assumptions about the influence of experimental events: (1) Any confounding influences (extraneous events occurring in time) will affect more than one of the be- haviors; and (2) Any nonextraneous events (treatment) will only affect the specific behavior for which it is introduced... If the behaviors selected in a multiple- baseline design are independent (un- correlated), it may be that either extra- neous events or the intervention would only have specific effects (i.e., alter one be— havior at a time). If the behaviors are correlated, either extraneous events or the intervention might produce generalized effects (1975c). The result, Kazdin (1975c) points out, is that when a study shows that the onset of an intervention for one behavior is associated with changes in another be- havior as well, it is as equally plausible to attribute the changes to some extraneous events or to the treat- ment. With these generalized changes, the problem is not 33 being able to decide whether extraneous or intervention events produced the changes. Again, no clear cause and Three recommendations are given: (1) collect data across baselines (e.g., behaviors, individuals, and situations) which are especially likely to be independent, (2) use as many baselines as possible, and (3) introduce a temporary "reversal" for one of the behaviors to assess the role of the intervention (Kazdin, 1975c) . The first is of little help to the applied researcher because of the possibility of interrelatedness in the behaviors of interest. The applied setting dOes not often allow such a selection of behaviors under the best of conditions. The second recommendation does not seem to have any logic behind it. For example, if you had ten baselines and with one intervention four behaviors changes, you still have the same dilemma: the rival extraneous events could still be the cause and not the intervention. The third recommenda- tion, the use of a temporary reversal phase, puts one right back with the discrimination problem of the ABAB design. The only conclusion left to the investigator in- terested in establishing reliable techniques for assessing generalization is that even if generalization is achieved, the results cannot be used to demonstrate that the tech- niques employed are in fact causally related to the out- comes. 34 What both of the above analyses have overlooked is that the occurrence of interrelated behaviors, situations, and/or subjects is a description of the occurrence of gen- eralization. Viewed in this way interrelatedness as a prob- lem is dismissed in the same way that the carry-over effect was dismissed for the ABAB design: by seeing the design from the perspective of ipEEE-subject replication and not intra-subject replication. Taking such a perspective leads to the ability to issue a causal statement even if the specific determining variables cannot be assessed. The determination of specific controlling variables leaves a new problem: the one of going from reliability to gen- erality (i.e., going from a point of knowledge that allows. one to repeat a phenomenon to the point where the necessary and sufficient conditions for the production of the phenomenon is known). One has in effect replaced one problem with another.But like having the right question, having the right problem is critical. Going from reliability to generality is a difficult task. Its importance was noted by Sidman (1960) when he pointed out that it indicates the growth of a science. In summary, the multiple-treatment and the inter- relatedness problems of the ABAB and MB designs respectively are outgrowths of going from description to explanation. When it is said that multiple-treatment effects or inter- relatedness of behaviors, situations, and/or subjects is 35 occurring, one is in fact describing the occurrence of generalization and not explaining it as related to the ex- perimental designs. What makes the problem conceptually difficult is that the multiple-treatment effects and interrelatedness problems are rival explanations to treat- ment effectiveness in terms of going from reliability to generality which is related to the degree of specificity of the causal statements made. Viewing the designs from an inter-subject replication perspective allows one to make causal statements regarding the assessment of generaliza- tion. It is only when the designs are used to both examine the occurrence of generalization and replicate the find- ings within a single subject that problems develop. The design problem that remains for examination is the discrimination charge made against the ABAB design. The resolution of the other design problem left a new problem: that of going from reliability to generality. The ability to make such a move is dependent on more than just the available experimental designs but depends as much as anything on the conceptual framework within which the researcher works. It is at this point that the prob- lems of terminology discussed earlier become important; for it is terminology thatf'OrmS the elements of any con- ceptualization. 36 D. RATIONALE FOR THE PRESENT RESEARCH The relationship between the design, programming for-generalization, and conceptual problems mentioned in the introduction and detailed throughout the rest of the chapter comes down to this: programming attempts to over- come all variables that could counter generalization; but the experimental designs and conceptual issues, respectively, can establish situations that impede the development of generalization and/or cloud one's analysis of it. The result is an anomaly: the attempts at securing generaliza- tion meet with both success and failure. It seems evident that there are designs that structure situations in such a way as to prevent or hinder programming techniques from success. The possibility of failure can be due to the form of the environmental structure set up by the investigator as in the ABAB design and the related discrimination problem, or the structure of behavior as in the interrelatedness of behaviors prob- lem. .Both are associated with determining specific causal effects. Allied with design problems, there are the expected reward programming techniques that are thought to impede generalization, and terminological issues that could impede the analysis and synthesis of findings. The result is a conflict between programming for and the evaluation of gen- eralization which leads not only to continued anomaly, but 37 to an unnecessary restriction in the range of usefulness to which applied behavior analysis could be put. Such a possibility necessitates a resolution. One starting point for the eventual resolution could be an experimental in- vestigation of the various programming, terminological, and design problems. This is the purpose of the present re- Search. The form of the present investigation involves a complex design that could both preclude and foster gen- eralization depending on the perspective one takes on the environmental arrangements it utilizes. The interven- tions employed were set up so that there was an increase in reinforcement control with each additional intervention. The complex design and hierarchical interventions were selected because they logically allowed one to examine the point at which programming could overcome design structure that generally would preclude generalization and/or the point at which weak programming would be facilitated by design factors. In the last case a non- conflict, harmonious program and evaluation design exist, and generalization results. The research started with only the basic design worked out. The interventions were instated when the data and applied setting allowed it to happen. In summary, the researcher realized that another attempt to program for generalization would only add a 38 plus or minus to the attempts to program for generalization. An attempt to discover why such failures or successes occurred was of far greater interest. The first step in solving such a problem involved reviewing the elements of the controversy. The present chapter outlined them. The next step was to design an experiment like the one outlined above and then carry it out. How that was done and what was found is delineated in the following chapters. CHAPTER II PROCEDURES Three ingredients were seen as necessary if not sufficient conditions for exploring the problems related to generalization as outlined in Chapter I. The first of these considerations involved behaviors that could be recorded reliably over extended periods of time but yet complex and sensitive enough to assess any possible behavioral changes within and between experimental settings. The second necessary aspect involved teacher instated procedures that could be trained successfully in a short time and economically managed by the teacher once in operation. The third necessity was the need for an experimental design that was complex enough to allow for the exploration of as many possible behavioral changes as can take place given as few interventions as possible. The primary behaviors selected for the teacher were teacher delivered consequences and non-consequence teacher behavior. For the student they were on-task and non-task. The behaviors for both teachers and students were further subdivided in order to do a microexamination 39 40 of the changes due to the teacher instated intervention. The teacher directed interventions selected involved group contingencies via a classroom game procedure directed at increasing student on-task behaviors. A second intervention procedure employed teacher adminis- tered attention (as a consequence behavior). The experimental design utilized was a multi- element/multiple baseline design across behavior (students and teachers) and settings (mathematics and reading). The multi-element component of the design, because it alternates intervention periods within a setting (e.g., mathematics period), allowed for further assessment of setting (stimulus) generalization. Each of these conditions will be outlined in detail below. A. BEHAVIORAL DEFINITIONS The purpose of this section is to outline the logic for and the relationships involved in the various student and teacher behaviors utilized in the present research. Appendix A contains detailed definitions and examples of behaviors for both students and teachers. The primary goal in the establishment of the behavioral definitions was to develop behavioral cate- gories that were universal sets of teacher and student behaviors. The aim was to categorize the sum total of the behavior emitted by either students or teachers into sets relative to the target behaviors. The sets would 41 then be broken down into subsets so that the change or non-change in a behavior could be examined at a micro- level. This approach is analogous to the examination of response differentiation as done in the experimental analysis of behavior. The terminology problems discussed in Chapter I guided this analysis. The general categories and their relationships are outlined below. 1. Teacher Behavior The target behavior for teachers was a set des- cribed as "teacher delivered consequences". The alternate set was described as "non-consequence behavior." This two set combination incorporated in it all the teacher behavioral repertoire in the classroom (the universal set). Each of these two sets was further divided into subsets. The teacher delivered consequence set was sub- divided in three different ways: in terms of type, category and direction. The "type" set contained two subsets: continuation or change. A "continuation con- sequence", as it was called, related to comments or actions by the teacher that had the double function (like in a chain of behavior) of being both a reinforcement and a discriminative stimulus for the student to continue the behavior he or she was performing at that particular time or related to some past activity which the student had engaged in. 42 The "change consequence" related to comments or actions by the teacher that had the double function of being both the presentation of an aversive stimulus and a discriminative stimulus for the student to change the behavior he or she was performing at that particular time or related to some past activity which the student had engaged in. The "category" set of consequence behavior con- tained four subclasses: academic, on-task, general attention, and physical contact. An "academic" consequence was related to comments by the teacher pertaining to the students academic behavior "I like the way you solved that word problem, Mary," was an example of the academic category. The "on-task" consequence behavior was concerned with comments by the teacher related to the students being task oriented. "Jim, thank you for staying in your seat during math," was an example of the on-task category re- lated to continuation consequences. The "general attention" category of consequence behavior was concerned with comments by the teacher that related to what the teacher thought of the student as a person or what the student did outside of class. Two examples would be, "Mary, you have a great smile," and "Zelda, your soccer goal in yesterday's game was terrific." Both fell within the continuation consequences category. 43 The "physical contact" category was concerned with the contact the teacher gave to a student. A pat or hug were considered examples related to the continuation con- sequence; a slap or kick were related to the change con- sequence. The third set of teacher consequences, the "direc- tion" set, was concerned with who the teacher was inter- acting with. The "direction" set was made up of three subclasses: class, group, and individual. If the teacher was interacting with the class, it was marked as such. When a group was being addressed, it was marked as such. If the teacher was instructing an individual, it was thus recorded. The "individual" set was broken down into targets and non-targets. The targets were given a re- cording number in each class and, thus, when a group was addressed with a target in it, the direction of attention a target received was recorded. The non-consequence set of teacher behaviors was subdivided in two ways: in terms of category and direc- tion. The direction subset exactly matches that outlined above in the teacher consequence section. The "category" set contained subsets of instruction and no response. The "instruction" category involved behavior by the teacher directed at shaping a student's performance before or during a task activity. "There are four new words in today's vocabulary builder ... now say then with me ...," is an example. 44 The directions and comments by the teacher were also included in the instruction category. Directions re- lated to statements by the teacher that gave the student an academic task to do. One example related to the class was, "Everyone, please read pages 95 to 110." Comments were statements concerned with granting permission to do a task, general information questions not related to in- struction, and statements directed at some non-academic event that was to take place. "Yes, you can go to the office when you are finished," was an example of a comment. The "no response" category included times when the teacher was not interacting with the students. This included times when the teacher was out of the room, sitting at the desk alone, etc. For recording simplicity the instruction, directions and comments categories where collapsed into one category. Because the ten-second recording intervals employed could contain more than one category, a hierarchial re- cording technique was employed. For example, if an interval contained instructions and an on-task comment, the interval was marked on-task. In the hierarchy no response gave way to instructions which in turn yielded to general attention, on-task, and academic categories. The three forms of consequences almost never overlapped. The general purpose of the hierarchy was to emphasize the target behavior set of consequences. The hierarchy can be pictured as follows: 45 Academic/On-Task/General Attention l}bonsequences Instructions W ) No ReSponse gnon-consequences J The relationship of sets to subsets of teacher behavior can be represented as follows: Consequence Non-Consequence Universal Behavior Behavior Set Of Teaching Behavior V V Continu- Change Instructions No ation Behavior Response Behavior andior and/or Academic On-task General Attention Physical Contact and/or $ Individual (Target and Individual (Target and Non-target) Group (Target and Non-target) Class Non-target Group ( Target and ( Non-target) l Class 46 With behaviors thus classified, a detailed analysis of change due to a particular intervention was possible. 2 . Student Behaviors The target behavior for students was a set des- scribed as "on-task". The alternate set was described as "non-task" behavior. These two sets encompassed the entire repertoire of student behaviors. Each of these two sets was further divided into two or more subsets. "On-task" was divided into the subsets of in- dependent-study and teacher-directed. Independent-study on-task was defined when the student was working alone or with one Other Student (without teacher present) and he or she was following the teacher's directions and/or classroom rules. Examples would include events like working mathematics problems during mathematics time. Non-examples would include sitting looking out the window or drawing during mathematics period. Teacher-directed on-task behavior was defined when the student was working with the teacher on a one-to-one basis or as part of a teacher-directed group of students. Examples included reading alone with others in the group with the teacher or reading silently per teacher instruc- tions with teacher present. The "non-task" set of behaviors was divided into four subsets: independent off-task, teacher-directed off-task, movement, and out-of—the-room. "Independenthflfitaskvwas 47 indicated when the student was not following directions or classroom rules while working independently or with one other student. Talking to a co-worker or just look— ing out the window were examples. "Teacher-directed off-task" was defined when the student was off-task with the teacher present. Examples in- cluded such events as striking another child while in a teacher directed reading group or talking during the directed activity when the directions were to read silently. "Movement"was defined when the student was out of his or her seat without permission, directions, and/or paper or book in hand. Talking to neighbor when out of their seat and, in general, any non permissioned out-of- seat behavior was scored as movement. "Out-of—the room"was defined when the student was at some special activity out of the room or on an errand for the teacher. Other out-of-the-room non-permission instances were classified movement. The relationship of sets to subsets of student behavior can be represented as follows: 48 *— Universal Set of Student Behavior W Independent- Independent-study Study on-task Off-task Teacher-directed Off-task Teacher-direCted Movement On-task Out-of—the-room As with teacher behaviors, such a classification allowed for a detailed analysis of change due to a particular intervention. B. EXPERIMENTAL DESIGN 1. General Description and Lpgic The study employed a multi-element/multi-baseline design across subjects, settings, and behaviors. The design was developed from an Egpegrsubject replication point of view. The aim was to examine as many situations as possible. Each of the settings or situations was seen as a point on a generalization gradient. Its gradient position depending on the distance (temporally and/or substantively) from the intervention point (the setting in which the stimulus conditions, antecedent and/or consequent, are manipulated. With such a design, it was possible to examine in detail the change in any number of behaviors across situations. Multiple sub- jects allowed for the replication of any pattern of be- havioral change that might take place. 49 A graphic illustration of the design is pre— sented in Figure 2.1. The ordinate describes the various subjects and settings that are related to the multiple baseline portion of the design. The subjects include teachers and students; the settings are the morning read- ing period and the afternoon mathematics period. The two classrooms were not visualized as settings in terms of a point on a generalization gradient. They were, instead, conceptualized as experimental replications. The idea was to see if the two classrooms would produce the same patterns of behavior change (generalization) across their various settings. The abscissa depicts two settings that are re- lated to the multi-element portion of the design. By multi-element it is meant that there is a rapid alterna- tion of situations (Sidman, 1960; Ulman and Sulzer- Azaroff, 1975). In the present case the situations are represented as different days of the week and concomitant stimulus conditions where interventions do or do not occur (X's and O's respectively). The X's were called game days, i.e., days on which the good behavior game intervention was applied to the mathematics period. The O's were called non-game days, i.e., days on which the game was not applied. The dashes (—) represent weekend or planned vacations. The consecutive calendar days represent the exact number of days in the study exclusive of the prebaseline phase. 50 .mu0w>m:oa pco .mmcfluumm .muOOHQSu nuOuom :m«m00 ocwammmnnwu~=E\ucmem~wuuuasz .H.~ Tasmam mama umpcoanu m>«u=omw:00 xoxcxnlexcxssxcxoxnsxcxoxsIxOxoxnuxcxox uuuuuuuu xm“xoxulxoxoxuuxcxoxtnxoxoxuuxoxoxflsxoxoxunxcxoxInxoxoxnaxoxoxllxoxox _ . _ . _ _ xCxoxstxoxoxstxoxcxnixOxoxsIxoxoxnuxcxcx uuuuuuuuu xo xcxuuxoxoxtuxoxcxnuxcxoxllxoxo Ixoxoxllxoxoxnuxoxoxlnxcxoxcuxoxox COflUe—Quu‘ Mme—000.6 + OEMU MEMO QCfiHOmMm xoxoxuuxoxOXIIxcxcxuIxcxoxllxoxcxutxoxcx nnnnnnnnn xc—xoxlnxoxoxlvxoxoxllxcxoxInxoxowflxoxoxlnxoxoxulxoxoxnnxoxoxllxoxox _ . _ _ . . _ _ xcxoxnuxcxoxunxcxoxnIxoxoxuIxOxOxnuxcxox nnnnnnnnn xo xoxuuxoxOXIIxoxoxinxOXOXvaxoxox Ixoxoxnnxoxoxslxoxoxlnxoxoxnlxoxox sewucwuuc HO£OMOE + 05mm 0560 ocuawmmm esteemeqnvw Burpeau uoouzangv Bur ourpeau Buruxow soravmaqnsw uoouxaagv ii] IGQDEGL sauspnqs mooxsseta 51 As part A of Figure 2.2 indicates, the game was only applied to the afternoon mathematics period. All game days occurred on Monday, Wednesday, and Friday; and all non-game days occurred on Tuesday and Thursday. Part B of Figure 2.2 presents the situations where the second intervention, teacher attention, is applied. Notice that the game intervention was continued during the second intervention. In each of the situations, the multiple teacher and student behaviors described in the Behavioral Definitions section of this chapter were examined. 2 . Experimental Phases The experimental design involved four phases: Prebaseline, Baseline, Game, and Game plus Teacher Attention. The details of each phase are presented in the Experimental Procedures section of this chapter. The general structure and flow of activities is described here. a, Prebaseline. The purpose of this phase was to (1) train observers to use equipment and reliably re- cord, (2) have students and teachers adjusted to observer presents, and (3) work out the behavioral definitions so that they were as similar as possible between classes. The phase lasted three weeks. b. Baseline. The purpose of the phase was to determine the operant level of student and teacher be- 52 A. First Intervention: (X's) (O's) GAME DAY NON-GAME DAY (Monday, Wednesday, (Tuesday, Thursday) Friday) Morning Reading Period After- noon Mathe- + matics Period E. Second Intervention: GAME DAY NON-GAME DAY (Monday, Wednesday, (Tuesday, Thursday) Friday) Morning Reading * * Period After- noon * Mathe- * matics + Period Figure 2.2. Experimental settings examined in the multi- element/multi-baseline design. The plus (+) indicates the first intervention (the good behavior game) point and the asterisk (*) indicates the second intervention (teacher attention) points. Note that the game is still utilized in the second intervention. 53 haviors. Game training started during the last week of the phase. The phase covered four weeks. C. Game, The first intervention, the good-be- havior game, is introduced on Monday, Wednesday, and Friday of the afternoon mathematics period. The three other settings (game day reading, non game day mathematics, and non-game day reading) continued baseline conditions. Teacher attention training started the last week of the phase. The phase covered almost five weeks. d. Game plus Teacher Attention. The phase began officially the day after the teachers were shown the graphs of their consequence behavior and asked to apply to their classrooms what they learned about attention as a consequence for student behavior. Teacher administered attention was considered a supplementary intervention during the game period and the only intervention directly in operation during the other three periods. The extent to which attention could modify or maintain game initiated changes was of interest. The phase lasted seven weeks, one of which was spring vacation. 54 C. EXPERIMENTAL PROCEDURES 1. Subjects and Setting a. Setting. The research took place in two elementary classrooms of the second and fifth grades. The elementary school was located within a large univer- sity community in central Michigan. The community can be characterized as well educated and middle class. The population could in general be characterized as stable or non-transient. b. Teachers. Both of the teachers who partic- ipated in the research had over ten years teaching experi- ence and their master's degree. Both showed reluctance to participate; the reasons were related to their personal situations, time commitments, and the use to which the re- search results would be put (i.e., professional evalua- tion). After meeting with the experimenter and the principal they gave their consent. c. Students. Fourteen students were selected for observation by their respective teacher, seven from each class. Two of the second grade subjects were lost after baseline; one moved away and another went on a vacation that extended throughout the first intervention phase. The twelve remaining students all had average or above average I.Q.'s.that ranged from 100 to 130. All came from intact families. Each teacher's criterion for student selection was non-specific; it included students who were con- 55 sidered slow with class assignments but it also included those who were ahead on class assignments. In general, they selected students they thought would be interesting to observe, not necessarily ones who had problems re- lated to the behavioral measures being utilized in the present research. 2 . Observation and Recording Techniques a. Observers: training and reliability. Four observers were used throughout the study. They were undergraduate seniors with interests in educational pro- grams. Observer training occupied the three weeks of the prebaseline phase. The training involved (1) out-of- class recording practice, (2) in-class recording practice, and (3) the development of behavioral definitions. The out-of—class practice occupied the first week of the prebaseline phase. It involved role playing and observation of the role players as well as preliminary development and memorization of the behavioral categories. The in-class observational practice lasted the rest of the prebaseline phase (2 weeks). Practice at this time involved a procedure where two observers observed the same student or teacher and then debriefed themselves with the experimenter. During the debriefing they first determined the reliability of the recording session and then went over the intervals where disagreements existed and/or interesting events relating to observing/recording had occurred. 56 Reliability was determined by matching recording sheets interval by interval and then dividing the number of agreements by the number of agreements plus disagree- ments and next multiplied by 100. In this way reliability was expressed in terms of percent agreement. The reliability criterion required of all observers before the start of baseline observation was 85 percent over the last five reliability checks of four minutes or longer. Reliability checks during the study were carried out weekly on all observers. The experimenter checked an observer by connecting an earphone to the observers cassette recorder, waited for the first line recording cue and then recorded in the identical recording sheet location as did the observer. All checks were four minutes in length. b. Observation techniques. Two observers re- corded during the morning reading periods and two ob- servers recorded during the afternoon mathematics periods. Both teacher and students were recorded simultaneously, meaning that two observers were in the classroom at one time. Two observation samples were taken for each class- room period of morning reading and afternoon mathematics. The samples ranged from 12 to 18 minutes for each ob- servation sample. The result was from 24 to 36 minutes of observation per classroom period, per day. The gen- eral recording procedure is as follows: A sample of at 57 least 12 minutes was taken in the second grade classroom of both teacher and students and then the observers moved to the fifth grade classroom which was just beginning the period. After at least 12 minutes of re- cording both teacher and students in the fifth grade, the observers moved back to the second grade for another 12 minute sample. Finally, the observers moved to the fifth grade for the last sample of at least 12 minutes. This procedure was followed both morning and afternoon. The fluctuation in observational samples was due to the classes having late or early lunch, extra activities (e.g., art, theater, library), and other school programs. In general, classes were from 40 to 55 minutes in length. Observation within a sample was done on a time sample basis. Each of the seven students were observed sequentially with an observation/record interval of 3 seconds in length. Momentary observation was used with the rest of the three second interval delegated to re- cording. The procedure was as follows: on cue the observer looked at the first student just long enough to classify the individuals behavior and then recorded it. A second cue informed the observer to turn to the next individual and observe and record. This continued throughout the seven students and then the observer was told to pause. This cycle was repeated twice each minute for 12 minutes and then a one minute observer rest 58 pause occurred. Thus, each student was recorded twice each minute. The result was that each student was ob- served from 48 to 72 times per class, per day. Students were not aware that they were being recorded individually. MOmentary observations taken within 40 seconds of each other (in the present case 30 seconds) have been shown to approximate continuous recordings. Because of the interest in independent study be- havior (as mentioned in the Behavioral Definitions section) a minimum criterion of 15 independent study ob- servations had to be taken per class for the data to be considered as a representative sample. A minimum criterion of 72 observations had to be taken per class in order for the data to be considered as a representative sample of teacher behavior. Several pieces of apparatus were used by the ob- servers to assist them in the observation and recording of student and teacher behaviors. Recording was done on 21.8 by 35.8 cm recording forms designed specifically for the recording procedures employed (see Appendix B). The teacher and student observer observation/sample of student independent study behavior. The teacher observations were done concurrently with the students but a 10 second observe/record time sample was used. Again the observation was broken into main blocks of 12 minutes with 6 recordings per minute. 59 This resulted generally in from 144 to 216 observations per class, per day. Variations outside this range were caused by the teacher leaving the room, consulting with room guests, or equipment breakdown. Due to these exigent situations a record cues were controlled by minature recorders (Sony model TC-56) utilizing pre- programmed cassettes and earphones. The teacher wore a small wireless transmitter (Edcor model PM-lR) and the observer wore a small receiver (Edcor model PR-lR). The teacher observer was, thus, able to follow the teacher's verbal behavior in great detail. 3. Intervention Procedures Two intervention procedures were employed in the present research. The first was a classroom behavior game which utilized group contingencies. The second was a.teacher attention procedure which focused on altering the rate of teacher consequences. The teacher training related to these techniques will be discussed in the Teacher Training Procedures section which follows. The present section will examine the composition of these techniques and their use in the experimental classrooms. a. The group_contingency game procedure. The successful use of group contingencies in the classroom has been well documented (e.g., Barrish, Saunders, and WOlf, 1969; Harris and Sherman, 1973; Medland and Stachnik, 1972; Packard, 1970; Turner, Konarski, and 60 Johnson, 1976). In general, a group contingency involves a continuum of procedures which have been divided into three main types: independent, dependent, and inter- dependent group contingencies (Litow and Pumroy, 1975; O'Leary and O'Leary, 1975). An independent group contingency is established when the same response contingency is simultaneously in effect for all group members, but consequences are given to individuals meeting the response criterion. An example would be to require that in order to receive extra free time each individual of the group must com- plete 20 math problems during math period. Those who meet the criterion get the free-time those who do not will not receive extra free-time. Expressed as a con- tingency of reinforcement, the independent group con- tingency would be as follows: SD :R e SR I A Consequence given Same rules given to onl those who to all members meet t e response of the group. criterion. Some response required for all members of the group. A token economy is an independent group contingency. A dependent group contingency is established when the response contingency is in effect for some subset of 61 the group, but consequences are delivered to all the members of the group. For example Ascare and Axelrod (1973) modified the off-task behavior of a few children in various classrooms by having rewards for the class contingent on the reduction of inappropriate behavior. The concept of peer pressure is involved in this group contingency form. Represented as a contingency of-re- inforcement, the dependent group contingency would be described as follows: SD #‘rR :SR Consequence is given Response rules to all if subSet apply to a sub- meets response set of the criterion group The same re- sponse crite- rion is re- quired for the subset of in- dividuals An interdependent group contingency is defined when the same response contingency is simultaneously in effect for all group members at a group level, and con- sequences are delivered if group meets group response criterion. For example, Schmidt and Ulrick (1969) used classroom noise level as a criterion for earning a special activities time. The interdependent group contingency represented as a contingency of reinforcement would appear as fOllows: SD :R >SR A T Response rules Consequence is apply to the given to all if entire group group level crite- rion is met. The same response criterion is applied to all as a minimum or maxi- mum. In the interdependent contingency, one individual can keep the group from earning the specified consequence. Both the dependent and interdependent group contingencies involve the use of peer pressure to lessen the change of group lose of reinforcement due to a few individuals not following the behaviors that are reward contingent. The present study uses a variation of the inter- dependent group contingency that is often called a good- behavior game (Barrish et al., 1969). Here each class- room is divided into two groups each of which had the same contingency in effect. On-task behavior (see Appendix A) served as the response and a minimum criterion was set at 12 of 18 teacher made observation checks of the groups. If the whole group was on-task, it was given a point. A point could be gained by both groups. If one or both of the teams met the response criterion they won the game. The prize was extra free- time and a chance to wear the symbols which represented a winning team effort. (The present study used arm bands for the second grade and sun visors for the fifth.) 63 The teacher made observational checks were con- trolled by prerecorded cassette tapes programmed on a VIB' schedule. The teacher varied the tapes so that the room level signal could not be predetermined. ’The;' students were also informed that the teacher would not observe the groups exactly on the signal and that the signal only served as a reminder to her to observe the groups in the near future. A wrist counter was used to record observations. At the start of a game period (mathematics), the teacher introduced the game, put up a large game sign, asked students to repeat the rules of the game, put on their symbol if they had won the last game, and started the tape recorder. The announcement of winning groups was made at the end of the period. Appendix contains further details on the use of the game procedure. b. The teacher attention procedure. The system- atic use of teacher attention was one of the first be- havior intervention techniques to be applied to the classroom. It remains one of the most effective and widely used means of changing children's behavior (O'Leary and O'Leary, 1976). The effectiveness of atten- tion has been demonstrated across various subjects (e.g. , Madsen et al., 1968; Hall, Fox, Willard, Goldsmith, Emerson, Owen, Davis, and Porcia, 1971), settings (e.g., Schutte and Hopkins, 1970; McAllister, Stachowiak, Bear, 64 and Conderman, 1969), and Behaviors (e.g., Madsen et al., 1968; Kirby and Shilds, 1972). Attention is considered a consequence the teacher administers contingent on student behavior. The type of attention examined by the present study was divided into continuation and change attention delivered by the teacher. (A detailed presentation of teacher consequence behavior is given in the Behavioral Definitions section and in Appendix A.) The teachers were not asked to deliver a particular rate of attention to students as did Horton (1976) but to just use it as often as possible. The experimenter used prompts and praise during the attention phase to get the teacher to increase continua- tion attention and decrease change attention. Teacher attention training (see the Teacher Training Procedures section below) emphasized the analysis and kind of use to which attention could be put. 4 . Teacher Training Procedures Teacher training was divided into two parts. The first, game training was designed to be of short duration so that there would be a clear transition point between baseline and the first intervention. The second, atten- tion training, was longer in duration and involved a less clear transition point between phases. The details of the training are as follows: 65 a. Game training. The experimenter conducted game training lasted three one and one-half hour sessions that coincided with the termination of baseline. At this time an experimenter developed reinforcement menu was administered by the teachers to their respective classes (Appendix C). The first part of the training included how a game procedure works, the necessary considerations for setting up a game, an outline for the introduction of a game, and a sample of behaviors to which the game was to be applied. The details of this content is given in Appendix D. After the introduction to the game procedures, the teachers were asked to decide on the student composi- tion for each team (two teams per class), the rewards the winning team(s) was to get (winners symbol and type of extra free-time), and the name of their game. With the help of the experimenter the teachers determined the specific rules of the game for their class, organized their game introductions, and practiced with the equip- ment used in the game (tape recorder and wrist counter). Finally, the teachers were asked to role-play their introductions to the game. At no time during this train- ing was teacher attention or the use of social rewards mentioned. 66 b. Attention training. The heart of the atten- tion training covered five one and onerhalf hour sessions which coincided with the termination of the Game I phase. Three of these meetings occurred before spring break and two after. The content of the sessions revolved around the use of teacher administered consequences. The analysis of teacher consequences to student behaviors utilized the behavioral definitions used in teacher ob- servations (see Behavioral Definitions section of this chapter and Appendix A) and a graphic display of the teachers consequence behavior. The use of objectives, modeling, shaping, cues, and rules were covered within the context of a simple experimenter developed behavioral teaching model (see Appendix E). The teachers were asked to read Changing Children's Behavior by Krumboltz and Krumboltz. Finally, they were asked specifically to in- crease their rate of continuation consequences and try to lower their use of change consequences. A week after the above training, the experimenter met the teachers to inform them of the changes that had occurred. They were asked again to increase their con- tinuation consequences and decrease their change comments. ‘Following this meeting the experimenter continued un- systematically to praise and inform the teachers of their progress in increasing' continuation consequences and decreasing change consequences. 67 In terms of the stimulus control over the teachers use of consequences, the experimenter repeated on a number of occasions, "...that continuation consequences should be used as often and wherever possible." The information given them about their consequence behavior covered both reading and mathematics periods across game and non-game days. 5. Summapy and Schedule of Experimental Procedures Figure 2.3 summarizes the schedule of research events in terms of time periods (date, days of the week, number of school days), experimental procedures in effect (conditions), and the functions of teachers and observers during each experimental condition. 68 Date Days of Number the week of Conditions Teacher Function Observer Function School Days Jan. 3-21 M-F l5 Prebaseline l. Instruct class 1. Observational 2. Delineate class practice until schedule and reliability rules meets criterion 3. Instruct class 2. Assist in formulation of final student and teacher behavioral de- finitions Jan. 24- M,W,F 15 Baseline 1. Instruct class 1. Observe students 2. Attend in-service and teachers sessions to learn from 24 to 36 about good-be- minutes per T,Th 10 Baseline havior game (Feb. class period 21,22,23) and take reli- Note: 2nd grade teacher ability. starts game on Feb. 25 Feb. 28- M,W,F 14 Game (Math) 1. Instruct class 1. Observe students March 30 Baseline (Read- 2. Administer Game and teacher from ing) to students 24 to 36 minutes per class period and take reli- ability T,Th 9 Baseline l. Instruct class 1. Observe students (all classes) 2. Start attention and teachers training (March from 24 to 36 28, 29, 30) minutes per class and take reliability checks March 31- M,W,F 19 Game (Math) 1. Instruct class 1. Observe May 20 Attention 2. Administer game students and (all classes) 3. Give Attention teachers from 4. Finish attention 24 to 36 Training (April 12, minutes per 13) class and take reli- ability checks T,TH l3 Attention l. Instruct class 1. Observe (all classes) 2. Give attention students and teachers from 24 to 26 minutes per class and take reli- ability checks Figure 2.3. Research Schedule. CHAPTER III RESULTS The complexity of the behavioral definitions, the number of subjects observed, and the extended duration of the study resulted in a voluminous amount of data. The extent of the study can be realized when it is recognized that over a quarter of a million data points were acquired during the coursewxfthe observations. Data management, even with the computer, proved to be an extensive task. The reason for this being that computer software is not presently available for the study of the longitudinal research done from an N = 1 perspective. Although work has and is being done in this area, useful (validated) software is still in the future (Owen White, personal communication, 1977). The results that follow are concerned with asses- sing the patterns of behavioral change for students and teachers within and between experimental settings. The patterns of change presented are related to the three prob- lems of generalization research outlined in Chapter I: terminology, experimental design, and programming for gen- eralization. 69 70 A. OBSERVER RELIABILITY Over the course of 76 experimental days, observer reliability checks were performed 41 and 40 times for student and teacher observers respectively. The assess- ments were divided equally between the four observers and within each classroom. This averaged to a reliability check just under every fourth experimental day for each observer. The distribution of reliability checks across phases was rectangular. The reliability for student observations averaged 93.2% with checks ranging from 78.5 to 100%. The distribu- tion of reliability scores was negatively skewed with four of the 41 checks being 85% or less and 15 being equal to or greater than 95%. The reliability for the Baseline, Game, and Training phases of the study were 91.5%, 94.6%, and 94.7% respectively. The three percent discrepancy between Baseline and the other phases was due to a teacher rule in one of the classrooms, the problem was eliminated by the second day of baseline. The reliability of teacher observations averaged 94.2% with checks ranging from 80.8% to 100%. As with the students, the distribution of teacher reliability scores was negatively skewed. Of the 40 checks, three were 85% or less and 20 were 95% or better. The reliability for the Baseline, Game, and Training phases of the study were 92.3%, 94.6%, and 95.4% respectively. The increase 71 in reliability across phases appears to be due to practice and/or the rate of change in response categories used by the teacher. B. STUDENT BEHAVIOR Independent-study time gives a clear picture of how the teacher structured the student's classroom activities. It represents the time the student spent following (or not following) teacher direction but was not in direct contact with the teacher. Independent-study time is used as a measure because, first, the remaining time is group time (i.e., the time the student was in contact with the teacher, individually or in a small group); second, most of the student's time was spent in independent- study; and third, the student's on-task behavior (the target behavior for the game intervention) during group- time was almost always at or near the ceiling (i.e., 100%). Table 3.1 gives the mean percent of classroom-time, per setting, student, and grade spent in independent- study. The percentages are a little higher than "actual" independent-study time because when the teacher addressed the-class, the student was recorded as being involved in independent-study. Almost all teacher instruction was individual or small group, and thus, the percentages are not greatly discrepant. For both the second and fifth grade students, the percentage of time spent in independent- study was in the‘rmid- to high-nineties across mathematics '72 «n.4a .+q.ma ...oa ...ma nonhuman: h.vm +m.om a.~m v.mh DEMO v.5m ~.mm n.ah o.wa sequence .m.mm «+m.~m ao.om ao.oo~ couscouuc ~.mm +~.mm ~.ho n.5o ease o.mm ~.oa m.ha n.ma ocuuemem n.ma .+v.wo am.uo .m.mn ceaucoaut m.vm +v.va m.wm ~.nh Demo «.mm «.mm ~.~o «.mh acqumnem .~.~m i..Z...ma c~.cm co.cm :Oaucmuuc a.mm +~.vm m.na v.wa wane «.mm o.va w.va m.cm ocuummsn «c.5m a+o.am .v.mm .~.wh couucmuud n.mm +5.5m o.am ~.ah mseo b.vm ~.sm n.ho m.mm acquonsm «H.5m a+~.sm ah.oo to.~c seduceuu< a.om +w.mm ~.om ~.on 05mm >.om m.na ~.~> m.~h onuuomcm am.mm a+w.no .~.om .o.ea sequcouu< o.mo +m.mm c.~a a.v~ ease m.mm ~.vm m.vm ~.cm mcfiuomsm xea mice :02 use veto woo Dane so: fine memo Omani uofiuosmzuez mogusEozus: acupeem magnum: coocu0uu< nausea: ozukbmm Nam Hum cum am hm neatqns S EGVHD .ponnouucq on o» as: noducouue Heroes» zoaz3 cu sensed was nmcuuuom u « .ousua xOOu wasOOOOuQ oEem use code: cw mousse pas mosquumm u + .AEooummmHov mponm nose How mcwuuom one .mmmnm Housmseuedxo an.~m a+m.wm co.~a ca.mm couuceuu< m.om +v.~a ~.oc a.mn OEno mm n.5m >.ca a.vo v.oa onwaomsm c~.hm a+~.oa «N.mb «v.05 cO«u:wuu¢ m.no +e.nm ~.mh ~.va sane em ~.aa o.sa m..o o.ma mandamus am.va «+m.mm «o.hm ac.ma sewacouu< n.5m +c.mo c.5a ~.mm wEeU mm ~.ma v.ma a.mm m.~a ocuummmm .~.ma .+a.va ...mo .n.oo coaucmnnc o.mm +m.cm m.ch m.mh mEeU mm a.mm c.wm v.ma n.~a oraummsm .a.oa .+..aa .o.oe .a.sa coaucmusa c.mo +v.nm $.65 c.mh DEMO Hm 0.5m b.5o h.mo c.5h mcnuomsm and Osboucoz was oEoo >so OEeOucoz Nee osso sword S newueeozuez ovuussosunz weaves: acumen: fl coo:uwuu< mcwcuox fl. UZHBEWW W .H.m mamma .ucmpoun an upsunlucmpcommpsfi ca woman mswuasooummcao mo useouom some one Z HOURS 73 (P.M.) settings (game day and non-game day) and phases (Baseline, Game, and Teacher Attention). Both teachers had for the most part individualized mathematics instruc- tion. The percentage of independent-study time during reading (A.M.), across settings (game day and non-game day) and phases (Baseline, Game, and Teacher Attention), was lower and more variable in both classes than it was during mathematics. The lower percentage was attributed to the teachers' use of small group reading techniques. The variability within reading (i.e., between the two readings settings) was attributed to the somewhat scheduled meeting time with a particular group. The variability across phases was for the most part attributed to the changing of text books. Each time the teacher changed the text a group was using, they would spend a little more time with that group for a week or so. In general, both classes were structured the same in terms of the manner in which students spent their time. The greatest portion of the students time being given to large amounts of student-directed study-time. It is this portion of the students time to which the analysis turns. The student's independent-study time had three subcategories: on-task, off-task, and movement. Of the three on-task was of the greatest interest because the game intervention was directed at altering it and clear change patterns emerge from it. 74 Table 3.2 gives the mean percentage of on-task be- havior during independent study-time across settings and phases for both the second and fifth grades. Looking at the table it can be seen that the second grade has stable mean percentages across settings during the baseline phase. The mean percentage of on-task behavior for each student varied only for morning reading game day (Monday, Wednes- day, and Friday). In all cases on-task behavior for read- ing game day was lower than the other three settings which were within three percentage points for all students. Table 3.2 shows that the Baseline phase on-task behavior for the fifth grade was more variable than it was for the second grade. On-task in the mathematics settings was the least variable. Here, all students had their mean percentages for the two settings within five percentage points of each other. The variability across reading settings was much greater, from 5% to 24%. After the Baseline phase and with the instatement of game intervention, on-task behavior began to vary across subjects and settings for both grades. The patterns of change that resulted are relevant to the investigation of generalization and experimental design. The second grade on-task data of table 3.2 is pre- sented in figure 3.1. Part A of figure 3.1 gives the mean baseline on-task percentage and shows the variability across subjects for each setting. Part B of the figure shows the proportion of change in on-task behavior for the Game and 75 mm em mm am Z HGVHO «N.6N a+6.6m «n.6m a6.nm COAUCDuufl m.mm +m.mw ~.Nm 5.6m DEMO mum v.5. 6.6v m.6v ~.vo DaunDnmm cv.m~ +6.6v a6.mn c~.- :OaucDuud .6DneDuocm DD a o» on: scauceuus uwboeou rear: cu owners use nmcauuwm u c ~.v~ +m.6m m.- ~.v~ DEDU mum .DOeHQ n.vn m.mm ~.a~ o.v~ Drauonsm xOOD DDDODOOAQ Dena Dru sour: cu mousse use mocauuom a + c6.vm «+6.56 a~.N6 «5.65 :0«u:Duu< c~.5v a+v.~5 «n.06 ad.m6 :o«u:Duu¢ m.~v +v.55 v.66 6.66 DEMO 6mm w.~m +6.mm ~.mm 6.N6 DEso H.66 N.d6 v.6v m.v6 DeaaDwom 5.6m 6.5m «.mm n.~v DawaDmDm am.Nm a+o.vm cv.vw «N.mv :O«UCDDD< .m. c5.ev c+6.~5 a~.6m a5.6m :O«u:Duu< ~.~m +a.om ..om a.an «Eco awn“ a.~e +~.ae o.mv ~.a' mete ~.mn H.6v ~.vm 6.Nv DCAADDem 5.66 6.6m ~.~6 6.6m DcfidDwnm a6.~m u+v.~m am.m6 «v.~m :OquDuad ao.mv «+6.66 a6.6o .w.6m :Ouucouud 6.6m +1.66 6.5V ~.6m DEDU 6m 5.6m +6.6m 6.Nm v.vm DEDO 6.5M ~.vm v.mv ~.mm Daqummm v.m6 5.m6 v.n6 6.6m Daqumom am.6~ a+6.~o am.vm «6.65 saquDuuc «6.6v a+m.66 «n.6m «~.~m :OwucDuu< 6.6m +5.65 5.M5 6.~6 DEDU 5m ~.~m +m.6m n.~v 6.66 DEDU 6.~m «.mm ~.5v ~.~5 DaumDmmn v.mm ~.5m ~.mm 6.6v Dcqummm a6.mm a+m.mo «6.66 «v.—6 couucDuu< m.~v c+m.~6 am.~m a~.6m :OHuCDuuc ~.6m +m.v5 6.N6 m.m5 DEDC 6m v.6m +m.nm d.6v ~.vv DEDC o.ov n.~m ~.op a.mo masseuse a.nm a.am ..mm m.om wzsaomac 5mm DEnUuzoz N66 DEeU Mon Dieotcoz ,Nmm DEDU Decca W Nee D260 :02 Non D560 he: DEmU :0: Non Demo Dunne mofiusEDzunz mogusEDsuez mewpeoz maucnom q mowuoEDzunz nodumfiDsuez GawpoDm mcwvan cOOanuuc mousse: m. coccuDuut measuoz OZHEPmm 3 Uszfimm .AEooummmaov Donna boom MOM mswuumm 6cm .Dmnnm HmucDEfiHDQxD .uoenbsm 5b DsHunwosumuucDOcDmpsfi HON n0w>cson xmmulco mo usDouDd some Dee .m.m mamme iovans 76 . A m m .nmcfiuuDu can uDuezm snouus nDODnbsn Deena 6cm uOu uo«>sbob xnsuuco :« chaoosn Sosa Dmcezo ucDoqu "~.n Dusmeh m m a m: 1 mm A: mm m Aratnns Butane Defines a m 6min Asatnne saw-mun .M 3 sotunnqa sorrmmnnsfl sotasusqasu “Hafilflfllfl 83'! he nus-non m m m Keatnng- 5 6m: .4 3 mm: m 3.1. : oI o mm "a: o s... as w n... m~+ any a: a » 6m+ z s 3 3 a m5+ DcuHDnen coqucDuua genomes DEeo m uumm fl anon Iorneqag xssm-uo auaozaa 77 the Teacher Attention phases across settings relative to the baseline phase. Looking first at the Game phase, one can see that during "game time" (afternoon mathematics on Monday, Wednesday, and Friday), the on-task behavior for all but one subject was reduced. This one subject's on-task be- havior remained essentially unchanged from baseline. Four of the five subjects of the second grade class were on the same team. This team won the game 6 of the 12 times it was played during the Game phase. The other team with the one remaining subject won the game 10 of the 12 times. During the early stages of the Game phase for the second grade, the students employed the "correction" pro- cedure. This procedure involved having a teammate raise his/her hand, proceed to the rule breaker, ask him/her to return to task, and then return to his/her seat. During the later stages, the students employed the "voting" procedure to eliminate team members that they felt were precluding the team from winning. The subject on the team that won 10 of 12 times was voted out during part of the Game phase and, thus, did not participate in all of the rewards. The failure of the game procedure to facilitate on-task responding came as an unexpected event. It appeared to be related to the teacher's not instating the game as directed in training. The problem points came in the areas of game introduction and motivation, recording 78 accuracy, controlling the voting procedure, and the shaping of on-task behavior. The experimenter initially dealt with the teacher as one who was a passive partici- pant. This approach was modified towards the end of Game phase when it was realized that the teacher was actually resistant to implementing experimenter developed procedures. Even with the game procedure problems encountered in the second grade, a clear pattern of generalization emerges. Looking at the mathematics non-game day and the reading non-game day for the Game phase, one can see that all five of the subjects had a reduction in on-task respond— ing that was greater than that for the setting where the game intervention was instated. The reduction in on-task behavior ranged from 7.2% to 40% of baseline levels. The reading game day setting shows mixed outcomes for student on-task behavior. Two of the subjects increased sub- stantially but three decreased. The Teacher Attention phase shows an increase in the mean percentage of on-task behavior across all five subjects for the game-time setting (game procedure). This is a reversal of the direction of behavior change shown in the game-time setting for the Game phase. The reading non-game day setting shows that the responding of three subjects came close to their baseline levels and two subjects reverse their direction of behavior change enough to increase the mean percentage of on-task behavior over that of the Baseline phase. The levels of on-task for the 79 reading game day setting is still mixed across subjects. The reduction in on-task responding that occurred during the Game phase for the non-game day setting, continues during that setting for the Teacher Attention phase. The fifth grade on—task data of Table 3.2 is pre- sented in figure 3.2. Part A of the figure gives the mean baseline on-task percentage for each setting. Part B of the figure shows the proportion of change in on-task behavior for the Game and Teacher Attention phases across settings relative to the Baseline phase. Looking first at the Game phase, one can see that during game-time (afternoon mathematics on Monday, Wednes- day and Friday), the on-task behavior for all subjects increased over baseline. It also increased for all subjects in the reading non-game setting. The mathematics non-game day setting showed a decrease in on-task behavior for all subjects as it did for the second grade. The reading game day setting duplicated the second grade, the direction of change is mixed. The subjects of the fifth grade were distributed in teams for the game: three were on one team and four on another. Both teams won the game every time it was played during the Game phase. The students did not utilize the correction and voting procedures. The substantial increases in on-task behavior for the game-time setting during the Game phase follows the Part B 80 Aha umnyuou sotususqasu Arc amnyubn 6mm Abo1nuo unnumnnxu Teacher Attention mu1smu> 6m Ara.annyuou flnmnmnneu Asazamnynbu Dunn's Abc1un9 sore-Isuzu“ Kan mun) 6mm 6 1n 6 I“ 6 1n 6 6 5 m N N In H + + + l ' auTT -9833 MOI; 93331030 10 OSPEZOHI auaoxaa Asa unannou “Hflflflflfiflm Asa sues-non ownneu a S * .d Aha emu) t: 3 sorasusqasu s u . o. n m Aeainus banana 6 6 6 CbGD 6 6' 6 c: c: 6 6 o\ cs 5-~o an szD6 uo~>stm qu>s£D6 HOA>D£D6 xmm DEeoncoz Nso DSDO Noe Damoncoz Nee Disc aoAueEDzaez DoAusEDcuez chpcDm chpch coocuouu< chcuo: OZHBEmm mmouom DODszm Dnu Ou pDunADH DADS :OA53 HDsonDu DAD an ucDEEoo Dmsnno no soAuDscAusoo D chsAmusoo Moo: qu mAn>HDusA 6:00Dm 6A 60 “Dead: :DDE D29 . . 6 66.6 .A.Atdm.A ”6.A 66.6 .6. «6.6 .A. coAucDuu< 6.6 6. +n.A +6. A.n 6. 6.6 6. D500 66 5.6 A. N.6A 6. 6.6 6. 6.6 A. DcAADDsm .5.n .6.A "A.A H6.A «N.m «6. .6.6 .6. coAucDuuc N.6 6. +6.A +6. n.N N. 6.6 6. Demo 66 6.6 A. 6.6A 6. 5.6 A. 6.6 6. DcAADmom .6.m «A.A H6.A “6. aN.6 .N. .6.6 .6. coAucDuu¢ 6.6 6. +6. +6. 6.A A. 6.6 6. Demo «6 6.6 N. 6.6 6. 6.6 6. 6.6 A. DcAADmem «6.6 «A.A H6.A «A.A a6.6 .5. «5.6 «6. :OAucouu< 6.6 6. +6.A +A. 6. A. N.6 A. DEcO N6 6.6 A. 6.6A A. 6.6 A. N.6 N. DcAADnem «6.6 .6.A um.A ”A.A .6.6 .6. .6.6 .A. coAucDuu¢ 6.6 6. +N.A +6. 6.N 6. 6.6 A. DEDU A6 6.6 6. 6.6A 6. 5.6 6. 6.6 A. DcAADDsm Dmcmzu .ucoo Dmcsco .ucou Dmsezu .ucoo Dmcecu .6600 Debra N qu>enD6 ubwassDn qu>e£D6 qu>er6 q Nee DEecucoz use 0360 use DEDUusoz N66 DEDU fl. DDADDEDLDD: noAuseozuez chpeDN chUst o :OO:uDuu< chcaox 3 ozAFhmm .mmdAuan can mDmnsm AmusDEAHDme .m.m wand? Z SOVHU 85 The data displayed in table 3.3 and figure 3.1 for the second grade, indicate that the increase in continua- tion attention and the continued in changed attention on the average were correlated with the reversal of on-task during the mathematics game-time setting. But the decrease in on-task that occurred from the Baseline phase to the Game phase for afternoon mathematics on the non-game days was not eliminated for any subject. The data displayed in the fifth grade section of table 3.3 and figure 3.2 reveals that no clear pattern of change emerged during the Teacher Attention phase. The mean percentages remain generally the same even with the increase in continuation attention and the continued de- crease in change attention. As with the second grade, the addition of attention was not enough to overcome the negative contrast established in the mathematics non-game day setting, there was a further mean percent loss of on- task behavior for six of the seven subjects. The trend and variability of the contrast and in- duction phenomenon are of interest because contrast and induction have been considered transient phenomenon in a two-ply schedule and because it relates to generalization of behavior over time (maintenance). A longitudinal pre- sentation is given for four subjects, two from each class. They were selected because their performances represent . some of the considerations that have to be made in the analysis of generalization phenomenon. White's (1972) median trend analysis procedure is used. 86 Figures 3.3, 3.4, 3.5, 3.6 give the percentages of on-task behavior for subjects 2, 3, 7, 10 respectively. All four settings are indicated. Closed circles represent game-days for Mathematics (game-time) and reading, and open circles represent non game days for reading and mathematics. The dashed phase line indicates for reading the point at which changes due to the game intervention could have begun to occur. Although the contrast and induction patterns are evident for the mean percent of on-task behavior shown in figures 3.1 and 3.2, such patterns are not clearly evident in the examination of trend direction. For example, during the game time setting some subjects showed gradually in- creasing trend during the Game phase that reverses direc- tion during the Teacher Attention phase (e.g., SlO, figure 3.6). Others showed flat or decreasing trends during the Game phase that continued or began to decrease during the Teacher Attention phase (e.g. S7, figure 3.5). For the subjects (S7 and 810) and setting (game time) used in the above examples,the direction change does not seem too important because it is accompanied by large changes in "step". (A step indicates the difference between where the end of one trend line meets the phase change line and the beginning of the next trend line meets the phase change line. If the beginning of the second trend line meets the phase change line above the point at which the end of Percentage On-task Behavior 87 Subject 2 (2nd Grade) Mathematics Baseline Game Attention 100' Ollti I 80- I s A \/ / 201 w 4'0 5'0 6'0 7'0 8'0 9'0 10011'01éo H O 10 2O 30 Non-Game,Days 72:1? Game Days 0- 0M? 1 /\/ 0 1'0 2'0 36 4'0 56 6'0 7'0 8'0 9'0 1001i0 1520 Consecutive Calendar Days Reading It), _::2_ Figure 3.3: On-task behavior for Subject 2 across phases and settings. "S" indicates substitute teacher. "In" and "out" indicate ejection from and instatement back into the game. 88 Subject 3 (2nd Grade) Mathematics Baseline Game Attention. 100' _ I in 0 1'0 2'0 3'0 4'0 5'0 6 7'0 8'0 9'0 10011'0190 .... Non-Game Days ’33:? Game Days Percent On-Task Behavior I I I Readin 1001 g I . 80- ‘\ I 60: g ‘i 40.1%, r‘ I 20' I I n 0' 1O 2O 30 4O 50 60 7O 8O 90 100110 120 Consecutive Calendar Days Figure 3.4: On-task behavior for subject 3 across phases and settings. "S" indicates substitute teacher. "In" and "out" indicate ejection from and instatement back into the game. 89 Subject 7 (5th Grade) I -__.__L_—— A—T-_—-A.e.____ Mathematics Baseline AttentiOn 100' 80‘ 60' 4(3‘ '\ ‘\ 20‘] _§ 0 10 2'0 80I 4'0 56 6'0 7'0 8'0 9'0 10011'0 120 6 E __ Non-Game Day .2 __ Game Day 3 Reading T1oo~ c o .9-.-=,_/§4.—AZ—\ i1£30' . 8 o H D m 0 Figure 3.5: 1'0 2'0 3'0 .4'0 5'0 6'0 . 7'0 8'0 9'0 100 11'0 12o ConSecutive calendar Days On-task behavior for Subject 7 across phases and settings. "S" indicates substitute teacher. Percent On-Task Behavior 90 Subject 10 (5th Grade) Mathematics 100‘ Baseline Game ' Attention 80* 60‘ ‘\ 40' 20' 0 1'0 2'0 3'0 46 5'0 6'0 ,7'0 8'0 9'0 160 11'0 120 __ Non-Game Days H _--_, Game Days /, Reading 0 .1'0 2'6 30 4'0 5'0 6'0 70 80 90 100110120 Consecutive Calendar Days Figure 3.6: On-task behavior for Subject 10 across phases and settings. "S" indicates Sub- stitute teacher. 91 the first trend line meets the phase change line, it is called a step “up". The reverse is a step "down". The step indicates the immediate or initial effect of an intervention.) Taking trend direction and step into account, there is an even greater lack of pattern especially for non- game time settings. For example, subject 7 (figure 3.5) showed an initial step up for non-game day mathematics during the Game phase which indicates initial positive induction but over the course of the phase because the trend is decreasing the overall result is negative con- tast. Analysis is further complicated by the increasing trend in the non-game day mathematics setting during the Teacher Attention phase. Subject 10 (figure 3.6), on the other hand, seemed to maintain the negative contrast for the non-game day mathematics setting over both the Game and Teacher Attention phases. The slightly decreasing trend lines for both the phases indicate the maintenance of the negative contrast for subject 10. The induction and contrast that are present in the analysis of the mean percent on-task for the Game phase (figures 3.1 and 3.2) for the reading settings, covered all possibilities in terms of trend (e.g. figures 3.3 to 3.6). With the addition of teacher attention one can also see in these figures the differential changes across subjects. Some decreasing, some mixed, and some increasing trends. 92 Subjects 2 (figure 3.3)and 3 (figure 3.4) illustrate the effects of peer influence. Both of these subjects were removed from the game during the Teacher Attention phase by their peers and were later reinstated by a peer vote (see indicator arrows in figures). In both cases the reinstatement correlated with a high and stable rate of responding during most of the remaining game presentations. For subject 2 there was also a dramatic increase and stabilizing of on-task behavior for the non-game day reading setting. Subject 3 also has a correlated increase during the non-game day reading setting but it is not as dramatic nor as lasting as subject 2 had. The non-game day reading setting was the setting in which all of the fifth grade exhibited positive induction during the Game phase and in general increased the on-task behavior even further over baseline during the Teacher Attention phase (see figure 3.2). The second grade subjects in this setting exhibited decreased on-task behavior relative to the Base- line phase across all subjects. IBut during the Teacher Attention phase the decreased responding was almost eliminated in the setting (see figure 3.1). 93 C. TEACHER BEHAVIOR The mean percentage of class time each teacher spent engaged in instruction, consequence, and no-response behaviors across settings and experimental phases is pre- sented in table 3.4. It is a description of how the teachers spent their time. The instruction class of be- havior included the three previously defined classes of instructions, directions, and comments. Consequence be- havior includes the two subclasses of attention behavior: continuation and change. The no-response category is the time the teacher was not in direct contact with the students. The table shows that on the average teacher one spent 76% of her time on instruction across settings and phases. Of the remaining time, 13% was devoted to giving consequences and 11% to activities not directly involving the students. The pattern of time use was similar for teacher two. On the average across settings—and phases, 81% of teacher two's time was spent on instruction, 10% on consequences, and 9% on activities not directly in- volving students. The consequence category of behavior followed clear patterns of change across settings for both teachers during intervention phases. Figure 3.7 shows the percent- age of baseline consequences delivered across settings for the Game and Teacher Attention phases. During the Game 94 N.5 6.6A 6.N6 m.NA N.6 6.65 5.6 6.6 5.66 6.6 m.HA 5.N6 soAusmuug d U. 6.5 6.6H 6.65 6.6A 6.N 6.66 6.6 H.5 6.66 6.NH 6.6 6.A6 DEMO % a 6.NA 6.6H 6.A5 6.6 5.NA 5.65 A.HA 6.6 H.65 6.6 A.6 6.56 DcAADmDm z o A 2 O A 2 O ‘ A z o A mnoA>mrmm Non Damolsoz 566 DEMO woo DEDOICOZ 566 DEDO moAuDEDbunz mOAuDEDcumz chUDDm msAmem :oocHDDmd chcuoz ocADumm ADODAU numv HH HDnODDB H.6A H.6A 6.A5 6.6A 6.6 6.65 6.6 6.6A 6.66 6.6 6.6A 6.65 coAucDuu< a 6.6 6.6A N.m5 6.6A 5.6 6.65 6.6 5.5 5.N6 6.6A 5.6 6.65 DEDO W s 6.6 N.6A 6.55 o.AA H.6N 6.66 6.6 6.6A 6.65 6.6H 5.6H 6.65 DcAADmDm a 2 U H z 0 H z D H Z O H mAOH>m£Dm >66 Dsoulsoz Non DEMO >66 DEDOIGOZ moo DEDO mOAuDEDnumz DOAumsDbumz chmem msApDDm :OOGprmd 6CHCHOE ocADDDm Ameuu psmv H HDSODDB .NAD>ADODQDDH :2: com .:U= .eH: an ODDDOAOGA DAD mDAHOmDuDO DmsOQDDn on one .DOCDomecoo .cOAuosuumcH mo mDmmDAo HOA>D£DQ uDsonDu Dee hmDmmnd AmusDEAHDQxD pom mchuqu Dmouon DHOA>62DQ DchQDDHIoc can .DOCDDUDD Isoo :OAuoouumsA :A eDmmmcD uchm anODDu SODD DEAD DDDHO mo ucDOHDQ cDDS .6.m DAQDB 95 phase, teacher delivered consequences in the game setting fell to 29% of their baseline level for teacher one and 19% for teacher two. The morning reading non-game day also had a substantial reduction in consequence behavior with teachers one and two dropping to 55% and 74% of their baseline levels respectively. The morning reading game day setting had teacher one responding at 71% of her base- line level but teacher two was responding at 99% of base- line. For the fourth setting, afternoon mathematics non- game day, teacher one delivered consequences at baseline levels and teacher two was delivering them at 87% of base- line. The Teacher Attention phase results, displayed in figure 3.7, show that changes in consequence behavior for both teachers followed the same pattern. The two morning reading settings exhibited behavior change that resulted in both teachers having consequence behavior that was at or above baseline levels, thus, overcoming the reductions that occurred in the Game phase. The mathematics game- time setting showed an increase in consequence behavior for both teachers over the Game phase but still far less than baseline -- teachers one and two respectively dis- played 41% and 65% of their baseline behavior. The mathe- matics non-game day setting in which consequence behavior remained close to baseline levels during the Game phase, showed reductions to 78% and 68% of baseline levels for teachers one and two respectively. 96 .mmcfluumm can mommnm :oHucm>umu:H mmouom mocmEHOMHmQ mcHHmmmn Hflmzu ou m>Hu0Hmu Honommu comm HON H0H>mnwn mocmsvmmsoo mo wmmucmonmm some one «5.6 mmcfiuumm Butpeau m G 9 .A m .4 m. 1 To 3 8 Ken amps—non 5°?1PWEQ1PW Keg ewes-non Ava amps Burpeau SDIQPmaQQEW .A‘li ON 1 06 1 N nonomoe H umnomma m ov 1 cm 1 oo 1 O5 . E2511/(/’/C/1 Fj//////7| m am-..“ cm x ECL/J/C/C/(k’)/Z/C/’ cod. and- omfl. OMH. OVH1 IgI/Tz’L’L/Z/C/Cz’Q/Z/S/C/fx’37l :oHucwuud umaomms mOH>¢mmm muzmbommzou mmmum£om ooHumssHucoo : uoH>m£om omsmnu o 26 ON. 06. CV. cm. 00. 05. om. om. 661 CH. ON. om. ovL 66L 004 05. cm. cm. 006 sotnmmqimq Keg amps-non I'l. .mmcwuumm can mommnm mmouom HoH>mnmn omcmco 6cm coHumscHucoo mo psmunmm some one mmsfiuumm 66 em 96 mm 66 em 96 mm u o m m m e um“ u a m m m e u.+ . P P . . P o. . q 9,+ t 9 9?: m r 9 u mum nuu m m u no nuu mum m5 “a “a a m5 an “a a. a r. ”on“ a r. a.+ ... 6 26.. 6 26 ..A 1A.. m Honomme A I; H Moscow? I? cowucmuud umsomoa mEmo mOH>¢mmm MUZWDOmmZOU mmmo R4] 4— S4 As in the previous diagrammed arrangement, the options could be reversible or non-reversible with the reversible having compatable or incompatable contingencies. 139 The important point to be remembered in all of the arrangements outlined is that they represent important independent variables in the determination of behavior. If the pattern of behavior for two organisms are the same and the arrangements are different, one must look into the history of organisms for the answer. Similarly, if the pattern of behavior for two organisms is different in the same arrangement, 0ne must again look into the history of the organisms for the answer. These two statements are predicated on the knowledge that the researcher has checked and rechecked the adequacy of his or her experimental con- trol procedures and measurement apparatus. The historicalarrangements which could be called the phases of an experiment, an intervention change, or an organisms developmental elements can be represented as large sequential arrangements containing any of the types of arrangements so far outlined. The basic diagram would be: (1) (2) One element of past history (1) is shown with one element of present history (2). The arrow tells that the organism is progressing from the one to the next. Filling in the elements of history with two simple cycling discrimination 140 experiment sequences we get: (l) (2) ‘fi ' 'No ! No II ! _L 0 ,. R a R 0 R 0 R a [Sfaiksl [521112162 YesL. [53+R3P'S3 iIIS¢1+R«:I]"S4 Yes... The detailed specifications of S-deltas, home cage depriva- tion contingencies, etc. have been left out. Of interest If an organism with a history of continually vary- ing arrangements (1's) is repeatedly given one particular arrangement interspersed within the others (2'3) with the result that the behavior change for this contingency arrangement is the same, one is witnessing the controlling power of the elements of the contingencies of reinforce- ‘ment involved in that arrangement. It is related to the discovery of always-effective variables spoken of in an earlier section of this chapter. In this example the vari- ables involved (expressed here as arrangements of contin- gencies) were always-effective enough to reverse behavior change within a single subject. . Given that one is dealing with behaviors one does not care to reverse, the second arrangement of contin- gencies, (2), in the above discrimination experiment could proven always-effective by providing different organisms with 141 different prior histories (1's) and then seeing if the second arrangement (2) will produce the same pattern and rate of behavior in all the subjects. If it does the variables are proven to be always-effective across organisms. But what happens when the variables used in the arrangement of contingencies are not always-effective? What happens when the extrapolated experiment (the gen- erality-test) does not produce behavior change within the range of precision that was possible in the isolated system research or is needed to solve a social problem? The starting point for answering the question is an examina- tion of the arrangements of contingencies of reinforcement. It does not appear to matter if the problem is one of experimental procedures, actual differences in the prin- ciples of operation of the involved processes, or in the nature of the past history of the organism: the examination of contingency arrangements should start the researchers problem solving behavior in a useful direction. 2. Examples of Arrangements Throughout the presentation on the arrangements of contingencies of reinforcement the work of Guttman and Kalish (1956) has been contrasted and compared to the pre- sent research. At this point it would be instructive to diagram these two pieces of research and a few others that are of related interest. 142 The Guttman and Kalish (1956) experiment is dia- grammed in figure 5.1. In this illustration training is one segment of history and testing another. The behavior independent contingencies (B.I.C.) is an aspect of history that runs throughout the life of the organism, the organism is always recycled back to it. Looking at the training seg- ment of history one sees that the alternate set is not com- pletely specified. The physical specification is only a darkened chamber during the SA and if the response occurs during that time the reinforcing stimulus is not delivered. The history during testing indicates that no stimulus con- sequence is present. The outcome measure is indicated by the letter C; The Blough experiment (1967) is diagrammed in figure 5.2. Notice that the only difference in history is the continuation of the reinforcement contingency during the testing sequence. This represents for the testing phase the clearer specification of alternate sets. It is this slight difference in testing (present history that results in faster extinction and greater discrimination to the non- reinforced test stimuli. The outcome differences for Guttman and Kalish, and Blough are illustrated in figure 5.3. It indicates that with the Blough (1967) experiment there is little generalization beyond the organisms sensory limits to detect change. But Guttman and Kalish (1956) got a much wider generalization gradient. These subtle 143 .pousmmms mmz mHanum> ucmpsmmmp on» ones: mucwom 0:» moumowpcw :0: .coxou mm: musmmma acopcmmmp 0:» can AHm mm: msHseHum OGHUHOHCHoH o: .ommnm om may no mocomnm on» mm3 <6 man no :oHumoHHHoomm way was» oumowocH 60cmmcHu Icoo omega mchHmuu pcoomm on» :H mumxomun one .:0 who: mEchmmuo . ecu 0Hspo£om Hm>umucw anmHum> on» mucmmmumwu accomcHucoo 6chHmuu umuwm onu cH H> one .ucmEHummxo :oHumNHHmuwcom :oHumcwsHuomeucoc A666H6 cmHHMM pcm cmeuuso may now meocmmcHucoo Ho ucwEomcmuum was .H.6 9.862 mcqaunua 0600 one: ##1 — .coduu>qumoo. _ _ +6 +— _ —4 honousqusou . unaccomoch uoH>uson ”In? D 144 acquuea .peusmeea me3 eHceHue> ucepcemep ecu euecz mucaom ecu meueowpcu :0: .emecm mcuumeu ecu :H :exeu mes eusmeea ucepcemep ece .am ecu mo eocemeum ecu cw HHco ucc emecm mcucueuu ecu Scum pescwucoo me: msHsEHum mcwouow IcHeu e .emecm mewumeu ecu cH .emcommeu ecu mcuonHou msHsaHum acaouocheu o: me3 euecu uecu one am ecu mo eocemce ecu me3 <6 ecu mo coHueOHHHerm ecu uecu eueowpcw Hocemcuucoo mchHeHu pcooem ecu cw muexoeuc eca .co eueB mamwcemuo ecu eHcpecem He>ueucH eHceHue> ecu mucemeumeu Hocemcuucoo ochHeuu umuum ecu CH H> eca .uceEHuemxe A566H. cmson ecu uou meHocemcuucoo mo uceEemceuue ece .m.6 eus6um acucueua coco one: , ‘ u> _ _ + a + a _—h t— Accuue>que006 m o z + a —J _ z a aoceocducoo ucevceaeusu uoH>econ wink Proportion of responses to S Guttman éfxr & Kalish 145 Figure 5.3. 560 580 600 620 640 Wavelength (nm) A comparison of two ways to test for generalization. During testing in the Guttman and Kalish (1956) experiment, pigeons were not rewarded at all. During testing in the Blough (1967) experiment, pigeons Here rewarded for pecking during the S but not during the S . The result was a sharper gradient. For comparison She gradients are normalized about the S value. (From Rachlin, 1975). 146 differences in the generalization setting indicate that if expected rewards or the ABAB design arrangements do set up discriminations, the techniques needed to "break up" such discriminations may be profoundly simple. They may involve nothing more than some verbal statement that relates the reward condition with that of the non-reward condition. An interesting variation on the Guttman and Kalish (1956) technique is the one that produces a gradient called the peak shift (Hanson, 1959) where the highest level of responding occurs in the presence of an SA stimulus dis- placed away from the SD stimulus in the direction opposite the alternate but specified contingency, SA + R + I J. The difference in experimental procedure from Guttman and Kalish (1956), as illustrated in figure 5.4, takes place during training where the alternate contingency, SA + R + I ], is clearly specified. The outcomes of the peak-shift and the Guttman and Kalish (1956) experiments are presented together in figure 5.5. These dramatic differences illustrate what a slight shift in historical background, in terms of the relation between alternate sets, can do. The peak-shift is often referred to as demonstrating relational learning and the necessary condition for its occurrence is the clear specification of alternate sets in the organism's past history. It was the inability to specify the controlling variables in past history with respect to relational 147 U llilli IIIII U a + 6 .=0: >c peuoc we peusmees me: eHce IHue> ucepcemep ecu use ce>Hm me3 msHsEHum 6cHouoHcHeu o: .emecm mcHumeu ecu cH .emcommeu ecu mcHonHoM msHssHum mcHOHOHCHeH o: uecu eueercH muexoeuc ecu pee peucemeum we: msHssHum ucepeo neuce OHHHoemm e uecu meueOHUCH :oHqucoe mchHeuu pcooem ecu mo <6 eca .co ewe: mEchemuo ecu eHspecom He>ueucH eHceHue> ecu mucemeumeu aocemcHucoo OCHcHeuu umuHH ecu cH H> ece .uceEHuemxe uchm xeem H666Hv scene: ecu Mom meHocewcHucoo mo uceeemceuue ecB .v.6 eusmHm. mchHmub 06:0 0.0: .coHua>Hunev. aeceosHucou uceocoaevcH uoH>ocen ”SHE + -148 150 1 SD SA Q + 6 B . m 100 . Hanson ‘ e m c o c. U) 8 o [H o c o °H 4.) 3 . Q4 50 d o o u Q4 Guttman & Kalish .7 I I I I I I’ I 1* TI’II I I 520 540 560 580 600 620 640 Wavelength (nm) Figure 5.5. A comparison of two ways to train for gen- eralization. During training in the Guttman and Kalish experiment, the S was not used as an alternate set. During training in the Hanson experiment, the S was used. The re- sult was a shift in keyupeck_responding away from SA. Neither experiment employed reward during testing. For comparison the gradients are normalized about the SD value. (From Guttman and Kalish, 1956, and Hanson, 1959). 149 learning that lead to the confrontation between behaviorist and gestalt psychologists in the first half of this century (Goldiamond, 1975). Contrast and induction experiments (e.g., Reynolds, 1961) have a great deal in common with peak-shift experi- [ments during the training phase (past history). Both have clearly specified the antecedent stimulus in thier alter- nate sets although they are not as close in terms of sensory modality in contrast and induction experiments. The major difference is that the contrast and induction experiment has the alternate set on a reinforcement schedule during the training phase instead of the extinction schedule used in the peak-shift experiment. The arrangement of the con- trast and induction experiment (Reynolds, 1961) is illus- trated in figure 5.6. Notice that response rate is taken during training in the contrast and induction experiment. The testing phases of the two experiments make radical de- partures. During the testing phase of this experiment notice that the second contingency is changed. The extent and direction of its alteration is what defines the be- havior change as contrast or induction. For example if the second component of testing is altered to an extinction contingency then one can get negative induction (i.e., the rate of the non-altered component decreases as does the altered extinction component) or positive contrast (i.e., the rate of the non-altered component increases as the .umeuuCoo e>HuHmom CH muHsmeH uCeEemceuue ucemeum ecu CH Hexmv CoHu IUCHuxe 60 em: eca .Aeemeeuoee.uo cemeeuOCH uceeeuuoHCHeu .e.H6 eeueuHe mH emecm 6CHumeu ecu CH acce6CHucoo eCooem ecu 30c mH HCeoo HHH3 COHuoCe ICH uo umeuucoo uecuecB CH muceCHEueuee ecu 60 eCO .memecm 6CHumeu ece mCHCHeuu cuoc mCHHce .=U= ecu 6c eeueOHeCH me .Ceceu we: eusmeefi uceecem lee eca .memecm 6CHumeu pCe mCHCHeHu ecu CH pews eue: uCeEeououCHeu mo eHseecom caxmv COHuUCHuxe eCe AH>6 He>ueuCH eHceHue> .uceEHuemxe umeuucoo eCe COHuoseCH AH66H6 m.pHocem ecu How meHoCe6CHucoo Ho ucesemCeuue ecB mCHumeH 150 .6.6 musmHm emeu e50: ACoHue>Huerv HecemCHuCOU uCepceerCH uoH>ecem + 151 extinction component decreases). The variables deter- mining which occurs are related to the interaction of con- tingencies in this particular arrangement. Positive in- duction and negative contrast can occur when the altered component has an increase in reinforcement in the testing component. The four types of results are illustrated in figure 5.7. Such contrast and induction results have been found to be either permanent or transient (Mackintosh, 1974). It would be interesting to integrate the data on these generalization experiments by performing a multiple schedule contrast experiment but also put in the wide range of stimulus that are involved in the testing phase of the Guttman and Kalish, Blough or peak-shift experiments. It would be interesting to see how they could vary as the various forms of contrast or induction developed. Richards (1972) found that delay of reward also produced behavioral contrast but not as reliably across subjects as did a change in the reinforcement schedule to extinction. This indicates that neither a reduction in reinforcement frequency nor response rate during the alternate component are necessary for the production of contrast. More important it shows that some variables being dealt with or encountered in the interaction of con- tingencies situations are not always-effective. The interaction of schedules need not occur only between the components of the schedules being reinforced. Figure 5.7. 152 Schematic diagrams of the four types of be- havioral interactions. In part A, Baseline behavior has stabilized in two different settings with the same reinforcement schedule. During the Intervention phase, the reinforce- ment schedule associated with the second setting (dashed line) is changed to extinc- tion (no reinforcement); the result is a de- crease in response rate. The upper portion of part A indicates positive contrast (shaded area) or an increase in response rate in the first setting; the lower portion indicates negative induction (shaded area) or a decrease in response rate in the first setting. In both the upper and lower sections of part A, the schedule of reinforcement was never changed in the first setting (solid line). In part B the Baseline phase is identical with part A. During the Intervention phase, the reinforcement schedule associated with the first setting (solid line) is changed so that greater reinforcement is given; the result is an increase in rate of responding. The upper portion of part B indicates negative induction (shaded area) or a decrease in response rate in the second setting; the lower portion indicates positive induction (shaded area) or an increase in response rate in the second setting. In both the upper and lower portions of part B, the schedule of reinforcement was never changed in the second setting (dashed line). Responses per Minute Responses per Minute Baseline 153 Intervention I. 0.. 00.0... ‘5..- .. 0... g “0.. .O'Q.’ 5.0.. .0.“ . a ‘0." 0 o ' :9 'v.'. ' .. .... 0 .:..::. 1.3,: ‘80.. 0......- .. ~.::.\ -- . "2'": - 000000 “.... .. I". ...” .fi...."J ' \-----——--- Positive Contrast Negative Induction Baseline Sessions A Intervention O o o s 0.. =.; 5.1.... ... ' . ‘10:: ”a... ... 0‘. h: 0: .....:; I. 00:. .... '0. .51.... .5. . g . 0‘. \.l . fl 0 0‘. I... ...... ‘ .0 .0. . o o ' C J \M‘~:J H‘Fflfir ‘o:..:::’: ’:::...‘..O. -. .L. __- 2 ‘-.* -_- Negative Induction ’--——*.—-qp- I. ’. . .. ....‘ ‘~ .0. .0“. .0 ”-W:.,.- .3. o“ . o . ...:‘1'..a...u..':~";.3o0 :- c .3 O. o. I b 9", I: .. .f. “.:“' '6‘" -.\ on 0 ' S ' o O. ..5. ' l’o "‘.. ..l ‘0 .5. ‘0'. ~’..~ : ': Positive Induction Sess B ions 154 They can occur where the organism is not under an experi- mental contingency. Schedule-behavior is an example. Jacquet (1972) used multiple schedules in an arrangement similar to contrast experiments but access to water was on a behavior independent contingency (B.I.C.). The experi- ment is diagrammed in figure 5.8. During the training phase (past history) and the testing phase (present history), the animal had free access to water (the be- havior independent contingency). With the changing of the second food-reinforced set during testing, contrast occurred in the food reinforced behavior. But there was also a concomittant change in drinking behavior, indicating interactions outside the closely related food reinforced set. In other words the extent of behavioral change is dependent on what is available for the organism to respond to. In the complex applied environment, access is greatly expanded. And it could be that differential access across experiments is one reason for the promotion or non-pro- motion of generalization even when the number of settings are the same. . The present experiment expanded not only the number of behaviors available and recorded in a setting but the number of settings as well. The results as indicated in chapter III were contrast and induction relative to the altered contingency across settings for one of the be- haviors, i.e., on-task. The arrangements of contingencies 1£55 .H0a>mzmn cmoscca oaaomnom mm on couumwmu mfl mcwummu mswuno acau locwuxo ou mcwcamuu mcduao chEGOHOMGMOH Hm>uwucw mandaum> scum cmumuam ma oaacmnom mcwmmmum Hun may £053 madammu umnu mcaxofia umumz :« wwcmno 0:9 .uwumz mm can mumaawm mzfimn Mm and: mcfixowa H0063 am can mcfimmmum Han mucmmmummu Hm was .JUHmv mocomcwu 1:00 unaccommccw uoH>mnon m can .Asxmv coauocwuxm .AH>V Hm>umuzw manmfium> mum3 mmmmnm wcflumma can mewcwmua may cw 6mm: unmawouowcfiwu mo mmaavmnom was .ucmawummxm m.m muamwh mawxowa amosvcw «Haconom Amhmav umavomh on» How mmwocmmcwucoo mo unwfiwmcmuum one . mcuuaoa . uHm | nan» M+N¢f- mm+N~m.+n_ ‘ auoamfio U 0 mcuuuou axm - Ha na+mfmm mm+mfflm m:«=«mus munu mac: 0 u cucumBOU noduaou caudal-00 unusuouu .coaum>wummc. aocmmcwucoo acmachmUCH uo«>m£om » 156 for the students is diagrammed in figure 5.9. Although R components of the contingencies cannot be the SD and 8 clearly specified, one can see that a four set sequence of a three-ply concurrent cycle exists during each phase of the experiment with each cycle leading back to the non- school setting which is precluded by eight years or so of divergent history (not shown). Even with the eight years of divergent history not included in the analysis of arrangements, the patterns of outcome for the students are consistnet considering the control procedures used. Returning to figure 5.9, the four set sequence corresponds to the game and non-game days for mathematics and reading. The three-ply concurrent cycle is related to the on-task, off-task, and movement response classes observed for the student. The reciprocal interaction be- tween teachers and students would be specified by altering the sequence and cycle in which the change in contingencies took place. The complex interactions that occurred in this experiment indicate that the advancement of generalization research will take place when researchers begin to control environments so that they know what arrangements are in- volved and the extent of the behaviors that are accessible to the organism. Research with such knowledge should be able to establish what variables are needed to maintain or shape behavior in various environments given its initial 157 Figure 5.9. The arrangement of contingencies for the pre- sent experiment. The pluses (+) and asterisks (*) indicate the settings across phases where interventions were instated. The response classes measured are represented by R (on-task), R (off-task) and R (movement). Suc arrange- mgnts indicated that interaction resulted in both contrast and induction relative to the instated contingencies (see results and dis- cussion in Chapters 3 and 4). 158 Non-observed Settings (Home and School) Reading Game Day +1». :1 . s+ -|sD » n2 .. $1“) +» » «3 .9 Mathematics Game Day Mathematics Game Day +30... +t +80% a. E} 5+ ‘1 Yea Mathematics .Gama Day“ Envy} {sow} .) Reading . Non-Gama Day +Rl+s {50.1.3.5li Mathematics , Non-Gale Day 1““? L“: 3“} a. 4. TEACRER ATTENTION PRBSB _y % 5+ 159 induction. The knowledge should also allow the researcher to explore the parameters of initial induction. Controlling arrangements and accessibility to con- tingencies is a primary concern for the experimental analysis of behavior; but applied behavior analysis need not and cannot rely on it. The analysis and measurement of arrangements and accessibility may even be of greater concern. If one analyzes and measures such things, one can, as Baer (1978) put it, do generality-testing research. Such research will be useful because the analyses and measurements made will outline the similarities and dif- ferences between experimental and applied research; one will be able, as Sidman (1960) has said, to "recognize similarities in their principles of operation." The long run result is the systemization of knowledge and the clarification of research areas and problems; these as noted earlier in this chapter, were the primary criteria for an adequate classification system. A final note is that unrelated areas of psychology can be integrated using the arrangement of contingencies classification system. For example, the work of Seligman (1975) on helplessness can be classified in terms of the arrangements of contingencies as can all the transfer of training experiments that have been concerned with con- cepts like proactive and retroactive inhibition. In both of these cases it is the arrangements in past history and 160 the interaction of the components of the contingencies that lead to the results obtained. The reader is invited to apply the outlined classification system and the con- cept of the interaction of contingencies to these areas of research to witness the power of the scheme to systematize areas of research. REFERENCES REFERENCES Allen, G.J. Case study: implementation of behavior modification techniques in summer camp settings. Behavior Therapy, 1973, 4, 570-575. Allen, K.E., Hart, B.M., Buell, J.S., Harris, F.R. and Wolf, M.M. Effects of social reinforcement on isolate behavior of a nursery school child. Child Development, 1964, 35, 511-518. Altman, K. Effects of cooperative response acquisition on social behavior during free play. Journal of Experimental Child Psychology, 1971, 12, 387-395. Ayllon, T. and Roberts, M.D. Eliminating discipline problems by strengthening academic performance. Journal of Applied Behavior Analysis, 1974, 7, 71-76. Baer, D.M.. On the relationship between basic and applied research. In T.A. Brigham and A.C. Catonia (Eds.), The Handbook of Applied Behavior Analysis, New York: Irvington, 1978. Baer, D.M., Wolf, M.M. and Risley, T.R. Some current dimensions of applied behavior analysis. Journal of Applied Behavior Analysis, 1968, 1, 91-97. Bailey, J.S., Timbers, G.D., Phillips, E.L. and Wolf, M.M. Modification of articulation errors of pre- delinquents by their peers. Journal of Applied Behavior Analysis, 1971, 2, 265-281. Bandura, A. Influence of models' reinforcement contin- encies on the acquisition of imitative responses. Journal of Personality and Social Psychology: 1965, 1, 589-595. Bandura, A. Principles of Behavior Modification. New York: Holt, Rinehart and Winston, 1969. Barrish, H.H., Saunders, M. and Wolf, M.M. Good behavior game: effects of individual contingencies for group conseqeunces on disruptive behavior in a classroom. Journal of Applied Behavior Analysis, 1969, 2, 119-124. 161 162 Blough, D.S. Stimulus generalization as signal detection in pigeons. Science, 1967, 158, 940-941. Bolstad, O.D. and Johnson, S.M. Self-regulation in the modification of disruptive behavior. Journal of Applied Behavior Analysis, 1972, 5, 443-454. Broden, M., Bruce, C., Mitchell, M.A., Carter, V. and Hall, R.V. Effects of teacher attention on attending behavior of two boys at adjacent desks. Journal of Applied Behavior Analysis, 1970, 3, 199-203. Buell, J., Stoddard, P., Harris, F. and Bear, D.M. Collateral social development accompanying reinforce- ment of outdoor play in a preschool child. Journal of Applied Behavior Analysis, 1968, 1, 167-173. Burchard, J.D. and Tyler, V.O. The modification of delinquent behavior through operant conditioning. Behavior Research and Therapy, 1965, 2, 245-250. Campbell, D.T. and Stanley, J.C. Experimental and Quasi- Experimental Designs for Research. Chicago: Rand- McNally, 1966. Catania, A.C. Elicitation, reinforcement, and stimulus control. In R. Glaser (Ed.), The Nature of Rein- forcement. New York: Academic Press, 1971, Pp. 196- 220. Catania, A.C. The concept of the operant in the analysis of behavior. Behaviorism, 1973, 1, 103-116. Conway, J.B. and Bucher, B.D. Transfer and maintenance of behavior change in children: a review and suggestions. In E.J. Mash, L.A. Hamerlynck and L.C. Handy (Eds.), Behavior Modification and Families. New York: Brunner/ Mazel, 1977, pp. 119-159. Cooke, T.P. and Apolloni, T. DevelOping positive social- emotional behaviors: a study of training and generaliza- tion effects. Journal of Applied Behavior Analysis, 1976, 9, 65-78. Fallon, M.P. and Goetz, E.M. The creative teacher: the effects of descriptive social reinforcement upon the drawing behavior of three preschool children. School Applications of Learning Theory, 1975, 7, No. 2, 27-45. 163 Ferritor, D.E., Buckholdt, D., Hamblin, R.L. and Smith, L. The noneffects of contingent reinforcement for attenting behavior on work accomplished. Journal of Applied Behavior Analysis, 1972, 5, 7-17. Ferster, C.B. and Skinner, B.F. Schedules of Reinforce- ment. New York: Appleton-Century-Crofts, 1957. Findley, J.D. An experimental outline for building and exploring multi-operant behavior repertoires. Journal of the Experimental Analysis of Behavior, 1962, 5, 113- 166. Frederiksen, L.W., Jenkins, J.D., Foy, D.W. and Eisler, R.M. Social-skills training to modify abusive verbal outbursts in adults. Journal'of Applied Behavior Analysis, 1976, 9, 117-125. Garcia, E. The training and generalization of a conversa- tional speech form in nonverbal retardates. Journal of Applied Behavior Analysis, 1974, 7, 137-149. Gladstone, B.W. and Sherman, J.A. Developing generalized behavior-modification skills in high-school students working with retarded children. Journal of Applied Behavior Analysis, 1975, 8, 169-180. Goetz, E.M and Baer, D.M. Descriptive reinforcement of "creative blockbuilding" by young children. In E.A. Ramp and B.L. Hopkins (Eds.), A New Direction for Education: Behavior Analysis, 1971, Lawrence, Kansas: University of Kansas Press, 1971, 72-79. Goetz, E.M. and Baer, D.M. Social control of form diversity and the emergence of new forms. Journal of Applied Behavior Analysis, 1973, 6, 209-217. Goetz, E.M. and Salmonson, M.M. The effects of general and descriptive reinforcement on "creativity" in easel painting. In G.B. Senb (Ed.), Behavior Analysis in Education, 1972, Lawrence, Kansas: University of Kansas Press, 1972, 53-61. Goldiamond, I. Alternate sets as a framework for behavior formulations and research. Behaviorism, 1975, 3, 49-86. Griffiths, H. and Craighead, W.E. Generalization in operant speech therapy for misarticulation. Journal of Speech and Hearing Disorders, 1972, 37, 485-494. Guttman, N. and Kalish, H.I. Discrimination and stimulus generalization. Journal of Experimental Psychology, 1956, 51, 74-88. 164 Hall, R.V., Fox, R., Willard, D., Goldsmith, L., Emerson, M., Owen, M., Davis, F. and Porcia, E. The teacher as observer and experimenter in modification of disputing and talking-out behaviors. Journal of Applied Behavior Analysis, 1971, 4, 141-149. Hanson, H.M. Effects of discrimination training on stimulus generalization. Journal of Experimental Psycholqu, 1959, 58, 321-334. Harris, V.W. and Sherman, J.A. Use and analysis of the "Good Behavior Game" to reduce disruptive classroom behavior. Journal of Applied Behavior Analysis, 1973, 6, 405-417. Hauserman, N, Walen, S.R. and Behling, M. Reinforced racial integration in the first grade: a study of generalization. Journal of Applied Behavior Analysis, 1973, 6, 193-200. Hawkins, R.P., Peterson, R.P., Schweid, E. and Bijou, S.W. Behavior therapy in the home: amelioration of problem- parent child relations with the parent in a therapeutic role. Journal of Experimental Child Psychology; 1966, 4, 99-107. ' Hay, W.M., Hay, L.R. and Nelson, R.O. Direct and collateral changes in on-task and academic behavior resulting from on-task versus academic contingencies. Behavior Therapy, 1977, 8, 431-441. Herbert, E.W. and Baer, D.M. Training parents as behavior modifiers: self-recording of contingent attention. Journal of Applied Behavior Analysis, 1972, 5, 139-149. Hersen, M. and Barlow, D.H. Single Case Experimental Designs: SErategies for Studying Behavior Change. New York: Pergamon, 1976. Hingtgen, J.N. and Trost, F.C. Shaping cooperative responses in early childhood schizophrenics. In R. Ulrich, T. Stachnik, and J. Mabry (Eds.), Control of Human Behavior, Volume I. Glenview, Ill: Scott, Foresman and Co., 1966, pp. 110-113. Holman,;ln,Goetz, E.M. and Baer, D.M. The training of creativity as an operant and the examination of its generalization characteristics. In B.C. Etzel, J.M. LeBlanc, and D.M. Baer (Eds.), New Developments in Behavioral Research: Theory, Methods, and Applications. In Honor of Sidney W. Bijou. Hillsdale, N.J.: Lawrence Erlbaum Assoc., 197 . ‘ 165 Horton, G.O. Generalization of teacher behavior as a function of subject matter specific discrimination training. Journal of Applied Behavior Analysis, 1975, 8, 311-319. Jacquet, Y.F. Schedule-induced licking during multiple schedules. Journal of the Experimental Analysis of Behavior, 1972, 17, 413-423. Johnston, J.M. and Johnston, G.T. Modification of con- sonant speech-sound articulation in young children. Journal ongpplied Behavior Analysis, 1972, 5, 233-246. Jones, R.T. and Kazdin, A.E. Programming response main- tenance after withdrawing token reinforcement. Behavior Therapy, 1975, 6, 153-164. Kazdin, A.E. The effects of vicarious reinforcement on attentive behavior in the classroom. Journal of Applied Behavior Analysis, 1973, 6, 71-78. (a) Kazdin, A.E. Methodological and assessment considerations in evaluating reinforcement programs in applied settings. Journal of Applied Behavior Analysis, 1973, 6, 517-531. (b) Kazdin, A.E. Behavior Modification in Applied Settings. Homewood, 111.: Dorsey Press, 1975. (a) Kazdin, A.E. Recent advances in token economy research. In M. Hersen, R.M. Eisler and P.M. Miller (Eds.), Progress in Behavigr Modification, Volume 1. New York: Academic Press, 1975. (b) Kazdin, A.E. On resolving ambiguities of the multiple- baseline design: problems and recommendations. Behavior Therapy, 1975, 6, 601-608. (c) Kazdin, A.E. The Token Economy: A Review and Evaluation. New York: Plenum Pub., 1977. (a) Kazdin, A.E. Vicarious reinforcement and direction of behavior change in the classroom._ Behavior Therapy, 1977, 8, 57-63. (b) Kazdin, A.E. and Bootzin, R.R. The token economy: an evaluation review. Journal of Applied Behavior Analysis, 1972, 5, 343-372. Keeley, S.M., Shemberg, K.M. and Carbonell, J. Operant clinical intervention: behavior management or beyond. Behavior Therapy, 1976, 7, 292-305. 166 Kirby, F.D. and Shields, F. Modification of arithmetic response rate and attending behavior in a seventh-grade student. Journal of Applied Behavior Analysis, 1972, 5, 79-84. Koegel, R.L. and Rincover, A. Research on the difference between generalization and maintenance in extra therapy responding. Journal of Applied Behavior Analysis, 1977, 10, 1-12. Leitenberg, H. The use of single-case methodology in psychotherapy research. ~Journal of Abnormal Psychology, 1973, 82, 87-101. Litow, L. and Pumroy, D.K. A brief review of classroom group-oriented contingencies. Journal of Applied “Behavior Analysis, 1975, 8, 341-347. Lovaas, O.I., Freitag, G., Kinder, M.I., Rubenstein, B., Schaeffer, B. and Simmons, J.Q. Establishment of social reinforcers in schizophrenic children using food. Journal of Experimental Child Psychology, 1966, 4, 109-125. Lovaas, 0.1., Koegel, R., Simmons, J.Q. and Long, J.S. Some generalization and follow-up measures on autistic Achildren in behavior therapy. Journal of Applied Behavior Analysis, 1973, 6, 131-166. McAllister, L.W., Stachowiak, J.G., Baer, D.M. and Conderman, L. The application of operant conditioning techniques in a secondary school classroom. Journal of Applied Behavior Analysis, 1969, 2, 277-285. Mackintosh, N.J. The Psychology of Animal Learning. New York: Academic Press, 1974. Madsen, C.H., Becker, W.C. and Thomas, D.R. Rules, praise and ignoring: elements of elementary classroom control. Journal of Applied Behavior Analysis, 1968, 1, 139-150. Marholin, D., II and Siegel, L.J. Beyond the law of effect: programming for behavioral maintenance. In D. Marholin (Ed.), Child Behavior Therapy. New York: Gardner Press, 1977. Marholin, D., II, Siegel, L.J. and Phillips, D. Treat- ment and transfer: a search for empirical procedures. In M. Hersen, R.M. Eisler and P.M. Miller (Eds.), Progress in Behavipr Modification, Vol. 3. New York: Academic Press, 1976. 167 Marholin, D., II, Steinman, W.M., McInnis, E.T. and Heads, T.B. The effects of teacher's presence on the classroom behavior of conduct-problem children. Journal of Abnormal Child Psycholpgy, 1975, 3, 11-25. Martin, J.A. Generalizing the use of descriptive adjectives through modeling. Journal of Applied Be- havior Analysis, 1975, 8, 208-209. Mechner, F. A notational system for the description of behavioral procedures. Journal of the Experimental Analysis of Behavior, 1959, 2, 133-150. Meddock, T.C., Persons, J.A. and Hill, K.T. Effects of an adult's presence and praise on young children's performance. Journal of Experimental Child Psychology, 1971, 12, 197-211. Medland, M.B. and Stacknik, T.J. Good-behavior game: a replication and systematic analysis. Journal of Applied Behavior Analysis, 1972, 5, 45-51 Meichenbaum, D.H. The effects of instructions and rein- forcement on thinking and language behavior of schizophrenics. Behavior Research and Therapy: 1969, 7, 101-114. O'Dell, S. Training parents in behavior modification: a review. Psychological Bulletin, 1974, 81, 418-433. O'Leary, K.D. and Becker, W.C. Behavior modification of an adjustment class: a token reinforcement program. Exceptional Children, 1967, 9, 637-642. O'Leary, K.D., Becker, W.C., Evens, M.B. and Sandargas, R.A. A token reinforcement program in a public school: a replication and systematic analysis. Journal of Applied Behavior Analysis, 1969, 2, 3-13. O'Leary, S.G. and O'Leary, K.D. Behavior modification in the school. In H. Leitenberg (Ed.), Handbook of Behavior Modification and Behavior Therapy, 1975, Pp. 475-515. Osgood, C.E. The similarity paradox in human learning: a resolution. Psychological Review, 1949, 56, 132- 143. Packard, R.G. The control of "classroom attention": a group contingency for complex behavior. Journal of Applied Behavior Analysis, 1970, 3, 13-28. 168 Parsonson, B.S. and Baer, D.M. Training generalized improvisation. Paper presented at the third annual convention of the Midwestern Association of Behavior Analysis, Chicago, May, 1977. Parsonson, B.S., Baer, A.M. and Baer, D.M. The applica- tion of generalized correct social contingencies by institutional staff: an evaluation of the effective- ness and durability of a training program. Journal of Applied Behavior Analysis, 1974, 7, 427-437. Patterson, G.R. Behavioral intervention procedures in the classroom and in the home. In A.E. Bergin and S.L. Garfield (Eds.), Handbook of Psychotherapy and Behavior Change. New York: Wiley, 1971. Pp. 751-765. Patterson, G.R. Interventions for boys with conduct problems: multiple settings, treatments and criteria. Journal of Consulting and Clinical Psychology: 1974, 42, 471-481. Patterson, G.R., McNeal, S., Hawkins, E. and Phelps, R. Reprogramming the social environment. Journal of Child ngchology and Psychiatry, 1967, 8, 181-195. Peterson, R.F., Merwin, M.R., Moyer, T.S. and Whitehurst, AG.J. Generalized imitation: the effect of experimenter absence, differential reinforcement, and stimulus complexity. Journal of Experimental Child Psychology, 1971, 12, 114-128. Peterson, R.F. and Whitehurst, G.J. A variable influencing the performance of generalized imitation. Journal of Applied Behavior Analysis, 1971, 4, 1-9. Phillips, D. Case history of a behavior-modification pro- ject in a public school. In F.S. Keller and E. Ribes- Inesta (Eds.), Behavior Modification: Applications to Education. New York: Academic Press, 1974, Pp. 63-106. Pinkston, E.M., Reese, N.M. and LeBlanc, J.M. Independent control of a preschool childs agression and peer inter- actions by contingent teacher attention. Journal of Applied Behavior Analysis, 1973, 6, 115-124. Quay, H.C. and Hunt, W.A. Psychopathy, neuroticism, and verbal conditioning: a replication and extension. Journal of Consulting Psychology, 1965, 29, 283. Rachlin, H. Self-Control. Behaviorism, 1974, 2, 94-107. 169 Redd, W.H. and Wheeler, A.J. The relative effectiveness of monetary reinforcers and adult instructions in the control of children's choice behavior. Journal of Experimental Child Psychology; 1973, 16, 63-75. Reynolds, G.S. Behavioral contrast. Journal ofithe Experimental Analysis of Behavior, 1961, 4, 57-71. Reynolds, G.S. A Primer of Operant Conditioning, Second Edition. Glenview, 111.: Scott, Foresman, and Co., 1975. Richards, R.W. Reinforcement delay: some effects on behavioral contrast. Journal of the Experimental Analysis of Behavior, 1972, 17, 381-394. Rincover, A. and Koegel, R.L. Setting generality and stimulus control in autistic children. Journal of Applied Behavior Analysis, 1975, 8, 235-246. Schmidt, G.W. and Ulrich, R.E. Effects of group contin- gent events upon classroom noise. Journal of Applied Behavior Analysis, 1969, 2, 171-179. Schumaker, J. and Sherman, J. Training generative verb usage by imitation and reinforcement procedures. Journal of Applied Behavior Analysis, 1970, 3, 273- 287. Schwartz, B. and Gamzer, E. Pavlovian control of operant behavior. In W.K. Honig and J.B.R. Staddon (Eds.), The Handbpok of Operant Behavior. New York: Prentice Hall, 1977. ‘ Schwartz, M.L. and Hawkins, R.P. Application of delayed reinforcement procedures to the behavior of an elementary school child. Journal of Applied Behavior Analysis, 1970, 3, 85-96. Schutte, R.C. and HOpkins, B.L. The effects of teacher attention on following instructions in a kindergarten class. Journal of Applied Behavior Analysis, 1970, 3, 117-122. Seligman, M.B.P. Helplessness. San Francisco: W.H. Freeman and Co., 1975. Sidman, M. Tactics of Scientific Research. New York: Basic Books, 1960. 170 Siegel, L.J. and Steinman, W.M. The modification of peer- observer's classroom behavior as a function of his serving as.a reinforcing agent. In E. Ramp and G. Semb (Eds.), Behavior Analysis: Areas of Research and Application. Englewood Cliffs, N.J.: Prentics-Hall, Inc., 1975, Pp. 321-340. Skinner, B.F. The Behavior of Organisms. New York: Appleton-Century-Crofts, 1938. Skinner, B.F. Contingencies of Reinforcement. A Theoretical Analysis. New York: Appleton, 1969. Skinner, B.F. About Behaviorism. New York: Alfred A. Knopf, Inc., 1974. Skinner, B.F. The shaping of phylogenic behavior. Journal of the Experimental Analysis of Behavior, 1975, 24, 117-120. Skinner, B.F. Herrnstein and the evolution of behaviorism. American Psychologist, 1977, 32, 1006-1012. Solomon, R.W. and Wahler, R.G. Peer reinforcement control of classroom problem behavior. Journal of Applied Behavior Analysis, 1973, 6, 49-56. Staddon, J.B.R. and Summelhag, U.L. The "superstition" experiment: a reexamination of its implications for the principles of adaptive behavior. Psychologlcal Review, 1971, 78, 3-43. Staddon, J.B.R. On the notion of cause, with applications to behaviorism. Behaviorism, 1973, 1, 25-63. Steinman, W.M. Generalized imitation and the discrimina- tive hypothesis. Journal of Experimental Child Psychology, 1970, 10, 79-99. (a) Steinman, W.M. The social control of generalized imita- tion. Journal of Applied Behavior Analysis, 1970, 3, 159-167. (b) Stokes, T.F. and Baer, D.M. An implicit technology of generalization. Journal of Applied Behavior Analysis, 1977, 10, 349-367. Stokes, T.F., Baer, D.M. and Jackson, R.L. Programming the generalization of a greeting response in four re- tarded children. Journal of Applied Behavior Analysis, 1974, 7, 599-610. 171 Sundel, M. and Sundel, S.S. Behavior Modification in the Human Services: A Systematic Introduction to Concepts and Applications. New York: John Wiley and Sons, 1975. Surratt, P.R., Ulrick, R.E. and Hawkins, R.P. An ele- mentary student as a behavioral engineer. Journal of Applied Behavior Analysis, 1969, 2, 85-92. Tennov, D. Maintenance and generalization through unex- pected reward. Presented at the Midwestern Association for Behavior Analysis, Chicago, Illinois, May, 1976. Tracey, D.A., Briddell, D.W. and Wilson, G.T. Generaliza- tion of verbal conditioning to verbal and nonverbal behavior: group therapy with chronic psychiatric patients. Journal of Applied Behavior Analysis, 1974, 7, 391-402. Turner, P.F., Konarski, E.A. and Johnson, M.R. The "Good Behavior Game": a systematic and direct replication in two unruly traditional classrooms. Paper presented at the third annual convention of the Midwestern Association of Behavior Analysis, Chicago, May, 1976. Ulman, J.D. and Suzler-Azaroff, B. Multielement baseline design in educational research. In E. Ramp and G. Semb (Eds.), Behavior Analysis: Areas of Research and Application. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1975, Pp. 377-391. Wahler, R.G. Oppositional children: a quest for parental reinforcement control. Journal of Applied Behavior Analypis, 1969, 2, 159-170. Warren, S.F. A useful ecobehavioral perspective for applied behavior analysts. Paper presented at the Kansas Conference on Ecological Perspectives in Behavior Analysis, Lawrence, Kansas, October, 1976. Wheeler, A.J. and Sulzer, B. Operant training and gen- eralization of a verbal response form in a speech- deficient child. Journal of Applied Behavior Analysis, 1970, 3, 139-147. Whitehurst, G.J. Observational learning. In T.A. Brigham and A.C. Catania (Eds.), The Handbook of Applied Behavior Analypis. New York: Irvington,’1978. Whitman, T.L., Mercurio, J.R., Caponigri, V. Development of social responses in two severely retarded children. Journal of Applied Behavior Analysis, 1970, 3, 133-138. 172 Wilcox, B., Meddock, T.O. and Steinman, W.M. "Generalized imitation" on a nonimitative task: effects of modeling and task history. Journal of Experimental Child Psychology, 1973, 15, 381-393. Winett, R.A. and Roach, E.M. The effects of reinforcing academic performance on social behavior. Psychological Record, 1973, 23, 391-396. Winkler, R.C. Management of chronic psychiatric patients by a token reinforcement system. Journal of Applied Behavior Analysis, 1970, 3, 47-55. APPENDICES APPENDIX A Teacher and Student Behaviors II. TEACHER BEHAVIOR CONTINUATION (+)/CHANGE (-) A teacher's verbal and non-verbal contact with students will be viewed as either directed at changing or continuing the behavior of the student. Performing the continuation comment can be viewed operantly as: d _D _y» r+ d p {/8 val r»S /S 'fiRl‘l room situation 1 teacher comment student student behavior to continue response its double desired function by teacher The change comments can be viewed operantly as: d _ __ r- d s 1R1 is /s ___—>sz room situation 1 adjusted response student behavior teacher comment or action to change to response 2 A (+) will indicate a continuation comment, a (-) a change comment. Continuation and change classifica- tions will apply to On-task, Academic, General attention, and Physical contact categories; they will NOT apply to the neutral categories of Instruction, DIEEctions, No response, and Comments. CATEGORIES OF TEACHER BEHAVIORS Academic (A): Comments by the teacher concerned with the student(s)' academic behavior as defined by teacher directions and rules. ”I would like you to finish your reading assignment." (A-) "Your reading has really improved, John." (A+) "Great work, everyone." (A+) "I want you to all get back to your math!" (A-) Note: "You did great on your assignment." (A+) "You really paid attention to the assign- ment." (0+) "You were given directions, do them!" (A-) 173 174 Use of the words "assignment" or "directions" puts it in the academic category UNLESS REFERENCE IS MADE TO ATTENTION, ON-TASK (0): GENERAL ATTENTION (G): PHYSICAL CONTACT (P): Note 3 an which is (0). Comments by the teacher that refer to the class, a group, or an individual, and are concerned with the student(s)' attend- ing to teacher-specified activities. "Stop chewing gum, Mary." (0-) (against teacher-specified rule) "I like the way you are quiet, Mary." (0+) "Mary, I told you to pay attention to the assignment." (0-) (in this case it is 0 and not Academic because of referral to attention) "Shush, Jane." (0-) "I asked you boys to sit in your seats!" (0-) Comments by the teacher concerned with non-academic behavior. Such behaviors include personal aspects of the students, activities outside the classroom, or any activity not incompatable with academic activities (nose-picking, smoking - unless there is a rule which would make it off- task, etc.) "Fantastic socks, Harpo." (G+) "Your soccer game was very good yesterday, Jim." (G+) "Jumpy, you need to please take a bath tomorrow, OK?" (G-) Actions by the teacher related to an in- dividual. It may be used alone or usually concurrently with another category. "Great work, John'" (pats shoulder) (PA+) (teacher puts child forcibly into seat) (P-) "Sit in your seat!" (puts into seat) (PO-) arm on chair while instructing student does NOT count, but an arm ON his shoulder does count. INSTRUCTION (I): Comments by the teacher directed at shaping a student(s)' performance before or during a task activity. This also in- cludes non-verbal "checking" or students as they work. 175 "There are four new words in today's vocabulary...now say them with me..." (I) "Your 'N' is almost correct (takes hand)...but it is really more like this (moves hand to form criterion 'N') ...see?" (PI) "Notice the topic sentence, what does it tell you?...now answer #3...#4... that is correct." (I) Note: If followed by a "great!" or "fantastic", scored as A+. If simply a "correct" or "OK", just scored as "I". DIRECTIONS (D): A subclass of Instructions which in- volves comments by the teacher concerned with giving the students academic tasks to do. This includes "given" directions, not repeated every day, for things such as activities to do when work has been completed. "Please read chapter six, everyone." (D) "Read pages 30-69, James." (D) "You all may take out your art and finish it." (D) COMMENTS (C): A subclass of Instructions which in- volves statements by the teacher related to granting permission to do a task, general information questions not related to instruction, and statements directed at some non-academic event that will take place. "John, is the teacher in the hall?" (C) "Yes, you may begin your project now that you have finished your math." (C) "Lunch is changed to 11:30 today so we can attend the play at 12:30." (C) "Because the snow is so bad, we will stay in for recess." (C) -- non-academic NO RESPONSE (N): Times when the teacher is not inter- acting with the class. (teacher is out of the room) (N) (teacher is sitting alone at desk) (N) -- even if checking on students? or is this I? (teacher marking papers) 176 III. DIRECTION OF TEACHER ATTENTION Class (C): "Everyone, please open your book to page 37." ,(C) "Class, line up at the door." (C) Group (G) of two or more students: "This group is doing great!" (G) "You two had better stop talking." (G) Individual (I): IV. "Zelda, please do the next ten problems." (I) "What time is it, Burt?" (I) Target (number): scored as individual. Each target is given a number. If John is number 4, a comment to him is (4). scored aslgroup. If John is number 4 and Jim‘is number 5, a comment to both of them only is (4,5). HIERARCHY OF RECORDING CATEGORIES If a ten-second recording interval contains more than one category, the category to be recorded will depend on its place in the hierarchy of categories that was determined by the researcher's interests. In other words the category that yields to another lower on the hierarchy. The hierarchy is as follows Academic I On-task General Attention > Physical Contact 1 Instructions J No Response Physical contact was recorded with all the no-response category. It is recorded alone only if it is performed without comment by the teacher. 177 SCORING TEACHER BEHAVIORS 1. The direction of attention (C or G or I or target number) is marked in the first box as the examples show. The distinction between class, group, or individual, or target is dependent on the orientation of the teachers verbal statements and physical presents. If there is no specific orientation to a group or in- dividual, then the direction of the statement is to be classified as class. If the teacher clearly specifies a group by name or presents, then the direc- tion of the behavior is to be marked as group. For example, Mrs. '3 room the reading groups she sits with in the morning would be considered a group. For a target or individual to be classified as the direction of attention the teacher must have physical proximity and/or verbal specificity related to an in- dividual or target. I 2. If the direction of attention is to a group and a target is part of the group, then a target number is included in the middle box next to the group designa- tion. Examples G 4 3. If the direction of teacher attention is to a group and a non-group target in the same recording interval, then the direction of attention is marked with the target number. The target has precidence over all other direction classifications. The individual has pre- cidence over the group and the group over the class. 4. If two individuals are addressed by name and one is a target, mark the target number in the direction box. 4 5. The specificity or category of the teachers be- havior is marked in the third box. Example 1 I L0+l 6. If the interval contains an attention category and part of it is continuation and part of it change directed, then the change (minus) is given priority. 178 7. If a behavior category and direction begin and continue through several intervals then draw a continua- tion line. Example: G 4 I 4 A+ 8. If the direction pp category of teacher be- havior changes, then use only a partial continuation line. Examples: a. G 4 I Ar b. I I G 9. Remember when scoring teacher behavior that the interval to be recorded is the one announced on the tape. ThE—iHEerval ends with the presentation of the number on the tape. After recording the direction and specificity of the teachers' behavior the rest of the interval is devoted to observation. During the observa- tion try to make as much visual 35d auditory contact as you can with the teacher. STUDENT BEHAVIORS On-Task (+): Marked when the student is following the teacher's directions and/or permission. 1. Students can be marked on-task if working together but they must clearly be working together. For example, working on a math problem or figuring out a word in a text or library book is to be con- sidered on-task. 2. During a class or group reading time, a student can look at the person reading or at book. They can not play with things like rubberbands or draw on paper or book (these are considered group off- task). Off-Task (-): Marked when the student is not following the teachers' directions and/or permission. 1. Drawing in class is marked off-task. 2. Looking away from work is marked off-task. 3. Hitting, kicking, talking, etc. at desks is marked off-task. 179 Movement (M): Marked when the student is out of his/her seat without permission, directions, and/or paper or book in hand. 1. Out of seat is defined as not being able to touch the top of desk without bending or taking a step. A student may be out of his chair and still not be out of his/her seat. If the student stands in line to see teacher he/ she is to be marked (M). Marked (G) only if in close proximity to teacher and having the teacher's attention. Sharpening pencils is marked (M). Going to the bathroom is marked-(M). Returning from out of the room is marked (M). Movement to the sign-out sheet is marked (M). Exception: out-of-the-seat is marked on-task (+) when the student has paper or book in hand. If student stops to talk or interact with others about non assignment matters, he/she is to be marked (M). Out-of-Room (O): Marked when the student is out of the room. Group Work (G): Marked when the student is in a group with the teacher or is receiving individual attention/instruction from the teacher. The category is marked (G-) if the student's behavior is off-task in teacher's presence. NOTES ON OBSERVING FOR HIGH RELIABILITY: 1. Follow the definitions to the letter -- do not judge student behavior or infer what is going to happen. 2. Find an observational viewpoint that lets you see all targets without turning head back and forth. 3. Before you begin recording go over the observation pattern (visual sweep) and spot and place all targets. 4. If on the scan of the previous place of the students finds that he/she is not there, then look at teacher to see if contact is taking place there and if it is not then mark (M) and look during the next pause for the student. 180 If you must move to a better location. Relax and try to get an immediate "grove" to the re- cording. Try to feel at ease in the room -- this is best done by following the first five rules above. Memorize students and definitions! When in doubt about class assignments or rules, ask teacher. Can you think of any more? APPENDIX B Student and Teacher Recording Forms 5322': NO . OBSERVER DRE Teacher %_%.=|JU_L_J WEED—1 6 41 1 [fl 11 I I ll 14L H 1;]:31 IDI:] u .5." .r...‘.'..._.u-—-r——"1 182 Student CBS DATE LN LIIIII 'HVER a bi 535-1 30. I I '7 Is] I I I I I I30] Ins] Isfi IIIII III I I I I IIII II Ij III 1? [Ir I I I75] _IJ9E I I I I I I I {’1 PI LJ‘Ij LI II I IIII Ins] I I J I I I120] I I r III'IusJ I I I I I [1533 I I -1151: III I I 180 IS I 30I III ,er III II I II IIIJ I I I asJ I soI I 157 I 90] Ims] I I IILI I I I I I I I I I I I FIJI Iii - I III I I I . I IKI I II_ IILII IILI I J I IIII I—I I I J I III I Izzcfl IusI I I I III I VI Iis’o I I I I I T III I I I I I I ITTII 1.80 I 15 I301 Ins] I I I ”I I I I I60J I I I I I IJ I III I I I II III L_ r111 IJJI I III I I I I I J J II I75] III I II I‘IIIIII I III IJOJ I IIIJ II I. III —V '- Q L E I 'I 7 I [105:] L I120] I I I III IIII I I I I I I I4 I III_ II I 115;] Iis'oJ 15' I III I II [I It I 1: 5' , '7 D 1 II IzIrI 7 a I I I V ’ I II I I180_I IEI I I I VI. III I I APPENDIX C Reinforcement Menues th nd 5 Grade and 2 Grade 5th Grade NAME: HOW MUCH WOULD YOU LIKE FREE-TIME TO DO OR 38 INVOLVED IN THE FOLLOWING? lo. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Choose one answer for each. not at all a fair amount very much Having time to work on special individual pmjecuOOOOOOOOOOOO........0OCOOOOOOOOOOOOOOO Having time to work on special group projects. Holding class outdoors in good weather. . . . . . . . T‘lkingtoatri.nd O.....OOOOOOO0.00.0.0.0... Having a rap session with a teacher and/or Primo.OOO.....OOOOOOOOOOOOOOOOOOO0.0.0.0... H‘vmgaclu‘ PutYOOOOOOOOOOOOOOOO0.0.000... Writing a letter to a friend.................. Writing letters to penpals around the world... Buying your wayout of homework or exams through paying attention to work during class time.... Displaying your work on bulletin boards in hallw‘YSOOOOOOOOOOOO......OOOOOOOOOOOOOOOOOIOO Displaying your work in the principals office. Being a group leader for a lesson (teacher for a 1.3.on!)....OOOOOOOOOOOOOOO0.00000000000IOOO Playug gm, at you: chaiCQCOOOOOOIOOOOOOOOOI Extra gym time for individual sports.......... Displaying your picture in the hallway when your work is finished on time................. Displaying your picture in the principal's office when your work is finished on time..... Extra gym for group sports.................... Being in charge of taking attendence.......... Being awarded a ”Great Work” button for finishing your work on time................... Performing on a musical instrument. . . . . . . . . . . . 183 21. 22. 23. 24. 25. 26. 27. 184 (2) Looking at interesting buildings...... ...... .. Looking at beautiful scenery.................. Hiking or walking............................. Taking fieldtrips to see a. animals................................. b. science projects........................ c. historical places....................... d. others: ........... Singing a. alone...................... ..... ........ b. with others............................. Being right I a. by guessing what somebody is going to. b. in an argument.......................u.. c. about your work........ ...... ........... d. on a bet................................ Solving problems a. by yourself.......... ....... ............ b. with others............................. c. math problems........................... d. science problems........................ e. puzzles................................. f. your problems........................... 9. other‘s problems........................ not at all a fair very amount much 28. 29. 30. 31. 32. HOW WOULD YOU LIKE TO BE IN THE FOLLOWING SITUATIONSR 33. 34. 35. 36. 185 (3) Completing a difficult job a. by yourself............................. b. with others............................. Having people seek you out for your company... Having people seek you out for your advice.... Having people seek you out for your help- fulness....................................... Being praised a. about your appearance....a.............. b. about your work......................... c. about your hobbies...................... d. about your physical strength............ e. about your athletic ability............. f. about your clear thinking...............‘ g. about your friendliness................. h. about your being helpful to others...... 1. about your understanding of others...... You have just completed a difficult job. Your friends come by and praise you for a ”job well “n.0.00000.0.0.0.0.........OOOOOOOOOOOOOOOOIO You are at a party. Somebody walks across the room and smiles at you in a friendly way, and says, ”I'm glad to meet you. I've heard many good things about you from your friends!...... You have just led your team to victory. An old friend comes over and says,'That was great work! Let me treat you to a Big Mac.‘......-. You are walking along a mountain pathway. You notice beautiful lakes, streams, flowers, and trees. You think to your self how great it is to be able to see all of this, to have the chance to wander alone out in the country...” not at a fair very all amount much ‘ /,iu [/1 186 2nd Grade NAME: i. I Ilk'e... ‘I'o have. IVES Free-Time. No if. I like. e. le 4,. es 'I‘hanK WIIfi For Inar-d Y war-K. . . Gem! 00 weak Swept)! No 187 . I es 3.I/lkg+° be Augie-J \/ b e a fart nd. 1.; ; ” ’ I No 4 I Wow/E //%’e. an aware! IYes 7E0!- o’ofnj My (JO/”K. k No 188 5. I (Mei-0 read. yes NO 42,/”’4‘ 0/ 1— 6. I- Iike. CopIe ‘I'o Yes NO See. mg. picjure. SNOO‘ I’ItNRQI I 189 :7. I wau Id like. ex-Ir-a ar-I- ‘I‘n'me. Yes NO I Like ~ To Paint-r! aI-JIKQ. +0 .SI'I" bu1 YES NO 'Fr‘ltnds . PHI '0‘ 'FRICNdSI ado ' 190 9, 1 IIKQ To do ma-I-h YES I /0.r wauId IIKQ ‘I’o aIuJaLtS do mu‘ way-K. . Dahe‘. yes No) 191 \I. I wan‘I‘ ‘I‘ne. "I'exachex~ ‘I‘o IIKE. me. I12. I 8:.qu Iihe. more ‘I'ime. To 'I'aIK +o m~1 +ri€nd$~ yes I APPENDIX D Game Procedures NOTES ON PLAYING A CLASSROOM COOPERATION GAME I. Introduction: A cooperation game is designed to foster compat- ible behaviors in the classroom between students and students and teachers. What follows will go over how the game procedure works, some considerations that need to be made when playing a game, an outline for introducing a game, and a set of behaviors that should guide the forma- tion of the rules of the game. II. H y the gamp_procedure works: Compatible behaviors can be conceptualized as sets of behaviors that may be different but contributeixathe same end. The parts of any machine are good examples: they all are different but they all contribute to the same end -- the operation of the machine. If the parts con- tribute as designed, then the end is achieved. In the classroom the teacher contributes and so does the student. If the contributions are not in line with the goal, the goal will not be reached. The concept is simple; the hard part is trying to make it happen. In order to make it happen, one needs to understand (in terms of analysis and synthesis) the contingencies that exist in the classroom. A contingency of reinforcement can be represented as follows: 59---.R---..sR 192 193 The SD represents the various conditions in the classroom that set the occasion for some response, R, to occur. These conditions include a variety of things from the structure of the room, the way the students sit (tables, individual desks, study carrels). The number and activ- ities of other students, and the teacher's behavior. All of these and many others are in every classroom. The idea is to arrange them so that they are all compatible with and contribute to the same goal. The game procedure arranges the behavioral components of the classroom so that the goal can be reached. The response of the students, R, can be seen as the goal to be reached. Responses like attending-to-work can be seen as subcomponents of the important goals like read- ing and math. You can teach such behaviors like attending- to-work directly or indirectly. The game attempts to teach them direction so that one has confidence that major goals like reading and math can be achieved or if they are not, the reasons have a better chance of being determined. The reSponse, R, is followed by a consequence, SR, which makes the response more or less probable of occurring in the future under the same conditions (SD). Thus, the three terms, S2--»R--SR, represent a contingency of rein- forcement that will raise or lower the chance that a re- sponse will occur. If the chance is that the response will occur more frequently, then the consequence is called 194 punishment. The game procedure utilizes positive techni- ques. Examples of contingencies in the classroom: 1. The teacher gives a set of math problems to students and says, "please go right to work on these, I would like everyone to finish by recess." (The set of problems and the teachers instruc- tions which specify the form, duration, and latency of the response, are the anticedent con- ditions or SD of the contingency.) Next the children work hard and everyone finishes by recess. (The reSponse component of the con- tingency.) Just before they go to recess the teacher says, "Since you all finished before recess, like I asked, you can all have five minutes of extra recess." (The consequence, SD, if positive will raise the chance that that class of behavior will occur again in the same and similar situations.) Technically this is a class-level contingency because the same conditions, responses, and consequences apply to everyone in the class. 2. The teacher is working with a reading group and tells them to work quietly by themselves while she works with another group. (The anti- cedent conditions, SD, are workbooks, the members 195 of the group, and the teacher's instructions which specify the form of the response.) The children work quietly while the teacher is away. The teacher who occasionally looks over at the group and notes the quiet work behaviors that the group is emitting. Upon her return to the group she says, "I noticed that all of you were quiet while I was away and because of that you can, if you can still be quiet (remember that others are still working), you can work on your art projects." (The consequence for performing the response is doing art projects. If doing art projects is a reward, the probability of working quietly in a group will go up and will be reflected in the future by quiet group work.) This group-level contingency applies the same conditions, response, and consequence to every individual in the group. A group or class level contingency is used in a cooperation game procedure because it is easy to administer and, most importantly, because it helps foster compatible contingencies in the classroom. It does this by establish- ing one strong contingency that interrelates the members of the class so that they must work together to achieve some mutually desired objective. In any classroom many contingencies of reinforce- ment exist. Some are in conflict with and some are 196 compatible with the learning objectives of the teacher. For example an incompatible contingency can be represented as: Let the first contingency (1) represent the contingency arranged by the teacher. verbally it could be described as students given math problems in a group with the in- structions to work quietly and quickly so that they finish by lunch time (S? conditions). The students working quietly and quickly would bethe response (R1). The consequence could be extra free-time (8?). The second contingency could represent a contingency set up by a student's peers. The description could go something like the following: The student's are given math problems and the instructions to work quietly and quickly in their group (the same SE conditions as the first contingency). But instead of working the students talk and play (R2). The consequences could be peer attention (SE). The real question is which contingency is the strongest, which produces a consequence that will raise the chance that the response associated with it will occur more frequently in the future? From the amount of talking and playing that goes on in the classrooms of most any school, the question is not hard to answer. The cooperation 197 game is one of the primary procedures for eliminating such incompatible contingencies in the classroom in a positive manner. III. Some considerations lg setting pp §_game: The first consideration deals with finding out what is a reinforcement or positive consequence to all the students in a class. This consideration can be taken care of by administration of a reinforcement menu. This is just a list of things and activities the students might like to have or do. Observation of student activities can also be used. What do they do when given free-time? If they do or participate in an activity with a high frequency, then it is probably a rewarding activity. Finally, one can just directly ask the student what kinds of activities he/ she would like to participate in. The problem with this procedure is that they do not think of many of the activities that you as a teacher have at your disposal. The second consideration is for how one introduces the game to the class. This is important because it sets the tone of all that is to follow. It is a matter of "selling“ the game and the consequences that have been picked. A little show biz and step-this-way enthusiasm are the keys to success. It is good to know and remember that the game is a temporary thing on the road to building in- dependent learners -- even they will laugh. at it in the future. But for the present it is a needed tool. 198 The third consideration is for defining the be- haviors of interest. It is a very important consideration because it is closely tied with a clear and concise intro- duction, and determines if the teacher can objectively evaluate student behavior and if students can identify if they are following the behaviors established for the game. The best technique is to define a general class of behavior like attending-to-work and then go over all the examples and non-examples that apply to your classroom, and the various special situations that arise from.time to time. The fourth consideration is for variation of con- sequences and conditions of the game. This means: have something new or unknown always in the fire -- burn them with curiosity. This makes them think of the present by way of the future. This requires that the game-maker has to look and listen for what is new or the latest fad. New words or phrases to talk about the game help vary conditions. A surprise consequence will keep them on their toes. IV. Ag outline for the introduction g£_a_game: The teacher can follow the points below in out- lining the game to a class. In the higher grades it is often good to put the game in an instructional context. In other words relate it to their development as in- dividuals, preferably to something that you have done in social studies. The introduction would include the follow- ing: 199 l. The teacher explains that they are about to play a game that involves dividing the class into teams and that each team could win the game. 2. The game will be played on Monday, Wednesday, and Friday, and only in mathematics (specify exact time). 3. When a team or teams won the game, the teams or team would receive extra privileges and free- time. 4. Example that there are certain rules to follow -- like in all games -- in order to win. 5. Whenever all the team members follow the rules then and only then could they win the game. They are to work as a team. 6. The teacher will check on their following of the rules by use of a tape recorder and a wrist- counter (explain and show how it all works). 7. The teacher will check on the students follow- ing of the rules about __ times during the game period. And if the group was following the rules __ of the __ times, then the team(s) would win the game. 8. If the team had an individual that precluded the team from winning, the team could vote the individual out of the game and, thus, lose all privileges and free-time the group gained by winning the game. 200 9. If the team had an individual who they felt had really improved or helped them as a team win the game they could give that person a special award (suggest some things they have to pick from or a special symbol to give to the person). 10. If a member of the team is not following the rules, there is a procedure whereby the other team members can remind the individual to follow the rules. They do this by following the permission procedure of raising the hand and then proceding to the person to ask them what they are doing and if such a thing follows the rules of the game. In introducing the game, set everything out step by step. Write it all down on the board. Give them as many choices as possible within the game but it is sug- gested that the game not be an option in itself. A parent does not let the child choose if they want a vegetable or not but gives a choice of what kind of vegetable. Good nutrition like good behavior cannot be achieved without knowledge of what is needed to be healthy (be it physically or behaviorally). V. Game behaviors: The behaviors will be gone over carefully during our meeting prior to the start of the game. The main class of behavior should be attending-to-work. Some examples are as follows: 201 l. Sitting in your seat. 2. Eyes on work. 3. Asking permission. Some non-examples: 1. Talking during independent work. 2. Throwing things. 3. Hitting others. The important thing is to make the game such that they are given as much freedom to move around as possible. But make it contingent on thinking about what they are doing before they do it. Call the procedure gaining self-per- mission and they will see it in a positive light. For example, they can move about the room when they need to by first raising their hand and thinking about what they are going to do (or they put on a I-have-thought-in-advance hat) and then they can proceed to do the thing that they have considered and found necessary to do. If the signal to check on the students is given then you can tell that they are not just going places for the fun of it (if you find them breaking the rules with the hat on, that can mean not only the non-gaining of a point but the loss of a point). APPENDIX E Teaching/Learning Model used in Teacher Training Teaching/Learning Model _____7, __y Establish Instruc- 4———d I tional Environment I Cue to Respond j Educational * Objectivesfi I 1. Conditions ; 2. Behaviors I f 3. Standards Response Does R Contingency meet objective > of criteria Reinforcement ? Yes Consequence Ii Consequence —“_" I II Use Use correction positive _‘W procedures I— rewards y)\_ practice required Increasing the probability of response Yes Practice ./ 202 203 THEfiTEACHING/LEARNING MODEL: SOME CONSIDERATIONS I. Establishing Instructional Environments 1. What supports are necessary for the response to occur? a. Prompts b. Modeling c. Rules 2. Make the conditions of the terminal response as close to those in the natural environment as. possible. 3. Consider the speed with which prompts are faded. II. Cue to Respond 1. What modality is more appropriate (facilitating) for student response in question? 2. Is the cue part of the natural environment? If not gradually replace with the cues that do exist in the natural environment. III. Response 1. Is the response easy to identify (operationally defined)? 2. What response prerequisites does the student possess? (This is related to the degree of environmental structure that needs to be arranged.) IV. Consequence I: Correction Procedure 1. What modality will best facilitate correction? a. Physical guidance VI. Note: 204 b. Repeating the rules c. Remodeling and imitation of response com- ponents d. Combinations of a, b, and c. Consequence II: Positive Reinforcement 1. Reinstate the contingency directly or by question procedures. If response is new and critical for the learner, make the reinforcing stimulus as strong and obvious as possible. Consider the reinforcing stimulus modality and its relation to the students reinforcement history. a. verbal praise b. Physical contact (the pat or the hug). Practice 1. Consider if the response is critical to the learner. If it is the response probability must be high and this requires practice. Consider if the response is complex. If it is then practice will integrate the response com- ponents. There is a need to consider the simultaneous opera- tion of different response levels. In the establish- ment of a contingency of reinforcement in the class- room, the interaction of the response levels needs 205 to be considered. Three levels exist that need direct attention. These are the social, emotional, and academic levels.