DOCUMENT RESUME ED 283 853 TM 870 373 AUTHOR Simons, P. n. J. TITLE Individual Differences in the.Self-Regulation of Learning, Emerging from Thinking Aloud Protoco1s. PUB DATE Apr 87 NOTE 21p.; Paper presented at the Annual Meeting of the American Educational Research Association (Washington, DC, April 20-24, 1987). PUB TYPE Speeches/Conference Papers (150) -- Reports - Research/Technical (143) EDRS PRI.CE MF01/PC01 Plus Postage. DESCRIPTORS Correlation; *Individual Differences; Learning Processes; *Learning Strategies; Metacognition; *Protocol Analysis; Reading Comprehension; Secondary Education; *Self Control; Training Methods; Transfer of Training; Word Problems (Mathematics) IDENTIFIERS *Self Regulation ABSTRACT Underlying_the research reported in this paper is a theoretical framework which defines the self-regulation of learning as the number_and kinds of teaching tasks students perform themselves. Three studies are presented on individual differences in regulation-processes_emerging from thinking_aloud protocols. Protocols of good and weaker_performing subjects were compared and related to test scores (impulsivity, intelligence, etc.). Furthermore, students were trained to changu_their regulation-processes. Training_programs_consisted of a combination of awareness-training and regulation-training. The subjects were 14, 10 and 6 students from a normal secondary school and two special secondary schools (fcr. children with learning problems), respectively. The restats showed some relations between process-differences such as_the quality and quantity of te5ting7processes, the breadth of orientation, on-line regulation_and mind-orientation on the one hand and performance on the other hand. There were also influences of task difficulty on the process-data. Training appeared to be effective for only some of the students. Transfer effects failed to show up. (Author BAE) ***************************** *** ********************* * * Reproductions supplied by EDRS are the best that can be made * * from the original document. * ****************** ********* * INDIVIDUAL DIFFERENCES IN THE SELFREGUL A TION OF LEARNING, EMERGING FROM THINKING ALOUD PROTOCOLS toiRgii. SIMONS TILBURG UNIVERSITY POSTBOX 90153 5000 LE TILBURG THE NETHERLANDS U-$ OEFARTMENT OF EDUCATION "PERMISSION TO REPRODUCE THIS Mica of Education& Flegailreh wig invetevediat MATERIAL HAS BEEN GRANTED BY EDUCATIONAL RESOURCES INFORMATWN J- CENTER (ERIC) 3Ztrius Document nes Dern reproduC.ed II rrIMeiced from the pereen Or orgefillgIal originating 0 Miner clienges Dave been evade v, lateral reprOduCtion duality. Feints etyma pr opinions staled& thiNIPCii TO THE EDUCATIONAL RESOURCES rnent de net necessarily represent &Mal CERI position or policy, INFORMATION CENTER (ERIC): piper presented to the Annul Alerting of the Anrion Educational Resratich Assoolationt WaaNngton, 0611987. 2 BEST COPY AVAILABLE 1 Individual differences in the self-regulation of learning, emerging from thinking aloud protocols. 1) dr. P.R.J.Simons Tilburg University Department of psycholo--- Postbox 90153 5000 LE Tilburg The Netherlands Abstract After the presentation of the beginnings of a theory on the self-regulation of learning, three studies are reported on individual differences in regulation-processes, emerging from thinking aloud protocols. Protocols of good and weaker performing Ss were compared and related to testscores (impulsivity, intelli- gence (etc). Furthermore, students were trained to change their regulation-processes. Training programs consisted of a combina- tion of awareness-training and regulationtraining. The Ss were 14, 10 and 6 students from normal a secondary school and two special secondary schools (for children with learningproblems), respectively. The results showed some relations between process- differences such as the quality and quantity of testing-proces- ses, the breadth of orientation, on-line regulation and mind- orientation on the one hand and performance on the other hand. There were also influences of task difficulty on the process- data. Training appeared to be effective for some of the students only. Transfer effects failed to show up. Introductiorl Underlying the research to be reported in this paper, is a theoretical framework derived from theories of Boekaerts (1982), Brown (1980) , Gagné (1977), Hettema (1979), Klauer (1985) Kuhl , (1983) and Lawson (1984). According to this framework the self- regulation of learning is defined as the number and kinds of teaching tasks students perform themselves. Figure 1 presents the various teaching tasks. The first teaching task (preparing students) has subtasks borrowed from Gagné (1977) and Galperin (1964). The second task (facilitating learning) is formulated in accordance with propo- sals Boekaerts of and (1982) Klauer (1985). For the third teaching task we extrapolated from theories of Brown (1980), Hettema (1979), Lawson (1984) and our own research. The fourth task (feedback and judgement) comes from Gagné and the (1977) last one (keeping students motivated and concentrated) stems from research on teacher perceptions of their tasks. We believe that Kuhl's theory om mind- and activity-orientation forms an impor- tant tool in conceptualising this last teaching task. Apart from external distractions which draw the students' attention away from the learning task at hand, mind-oriented cognitions form distractors,too. 2 Prevvaring learning * Orientation on goals, strategies, time, etc. * Planning of learning (time, anticipation of problems, choice of strategies). * Gaining attention. * Promoting self-confidence. * Informing students on goals. * Stimulating recall of prerequisite learning U Facmilitating learning * aimed at remembering and compre- hension. * aimed integration with other information. * aimed at problem solving. Ilg-tilating aI learning Monitoring. * Testing and questioning. * Revision (re-orientation, reflecting, diagnosing, repairing). Evaluating learning processes IV Giv-ng feedback and jiidging performance V Keep)J_ng students concentra ed aiad motivated --------------------- Figure I: A categorization of Teaching Tasks Teachears should be alert to mind-orientations such as goal- fixatic>u, planning-fixation, failure-fixation and success- Moreover, fixaticari. they should try to lead students towards utivit;y-orientations in which the goal-state to be reached, the state, present= the difference between the goal and the present state exrid the plan that may be used to change the present state goal state irto tkie each get sufficient attention. One aid they aght 1.3.met to reach these states in students is goal-setting. Setif-regulation, in our opinion, thus pertains to the eaent 1to which one is able to be one's own teacher. This means, . arcorli_rilg to the scheme of teaching tasks presented: being able toprer>aure one's own learning, to take the necessary steps to (aiming at recall learn and comprehension, integration and pmblenx solving), to regulate learning, to provide for one's own feedba=k and judgements and to keep oneself concentrated and In our motivatIced. conception( see also Hettema, 1979: Lawson, t=l1ree levels or perspectives 19134), in respect to these tasks should I>e discerned (see figure 2): metaco nitive knowled e. and cenceot,fLons (for instance knowledge of study strategies, knowing Om to use certain strategies, or conceptions of self-regula- tin), etxecutive control_ or re-ulation -rocesses (for instance deddiagx on a plan, attention-maintenance, monitoring, or repair- mhanismms) and transformations or executive skills (for instance paraphreising, underlining, or reading). In agreement with Lawson ex7nA recent Russian theories (1984) (e.g. Zak, we assume 1980) 4 3 that metacognitive knowledge arises from_reflection, (being an executive control process itself) on executive control processes or transformations. (metacognitive) knowledge and conceptions executive control or regulation processes transforma ions Figure 2: Three perspectives and their interrelations The most extreme form of selfregulated learning occurs when students perform all of these teaching tasks themselves. Most of the time, however, teachers (or their substitutes, for instance books or computers) take care of at least part of these tasks. In essence, there always seems to be a division of tasks. Extension of the responsibility of students for their own learning may in some cases improve learning. Lodewijks for instance (1981), showed that students learning science concepts in a self-chosen sequence performed better than students learning these concepts in a predetermined sequence. Likewise, Van der Sanden (1986) showed that some students (especially the better ones) performed better on a practical construction task without instructions than with detailed and explicit advice from a teacher. According these to and other studies improvements of learning might be reached by giving students more opportunities to regulate their own learning. This, however, is problematical in practice Apart from the students who might profit from these . opportunities, there are also students who will perform (still) worse when teacher advice is absent (Lodewijks,1981, Van der Sanden, 1986). A differentiated system with opportunities for self-regulation for the better performing 'students and sound advice for the weaker students, however, encounters many practi- cal disadvantages and problems. As discussed by Larsson -;as (1983) paradoxes of teaching should al;-0 be taken into account. Some teachers would like to give students more freedom to learn, but do not believe that students are able to handle this freedom. Some students believe that only the teachers should make deci- sions on learning and seem to hand over all responsibility to the teachers. In our opinion there is only one way out of these and other paradoxes and circularities and that is_ by training students in self-regulation. One main goal of training programs should then be to convince students that they have a responsibi- lity for their own learning and that they can become able to regulate their learning. Before such training programs can be_ developed, however, we need more information about individual differences in self-regu- lation. As yet we do not know to what extent students have the 4 adequate conceptions and metacognitive de we know how good students are regulation proc. in 3iod programs,_in our view, should depart fr, a .soun& li-Itynosis of the entering behavior of the students Eau) Camf4one and "1,e-,;: Armbruster, 1984). Our research strategy- ha_b i`..o set up tr4't- 7414Dirk small7scale in depth studies, didual differen- searching f--.or ces in the self-regulation of desining learniinl, partial, prototypical training programs aiming to inuence !!hrzse indivi- dual differences and evaluating the reszl'cTi of tra17-7'2:ng in terms of process- as well as performance differes The main research questions werAz individual a) w214.t differences in self-regulation occur and', 'Jf these relate to performance differences? b) Is there an, training on process- ,-;f and performance-variables? Ia tudies, reported t1 below, the middle level of Figure 2 culzive control) is C accentuated. In designing training programs had WO the following starting-points: a) We based the training on the differences in processes observed during a pretest-session. We stressed bl metacognitive awareness by letting students reflect on their own way of learning and that of other students. c) We emphasized the importance of regulation processes by letting students practice with a set of questions one may pose oneself during learning (e.g. Do I understand this part? What went wrong? Is this in line with the learning goal?) and techniques and skills one may find useful in answering these questions (e.g. paraphrasing, reflec- tion, thinking of new examples, selftesting). d) Finally, non- cognitive variables like concentration, self-motivation, attribu- tionj and mind-orientations were also included when possible. STUDY 2. METHOD SUbAv.cts The subjects were 14 students from the second year of secondary school. Ages ranged from 13 to 15. There were 7 boys and 7 girls. Materials In this study three sets of learning tasks were used, thought to be representative of academic tasks used in lover classes of secondary schools (compare Doyle, 1984). The learning tasks consisted of two parallel texts of 900 words on probabili- ty, the .one introducing principles, problems and examples of chances with replacement, the other dealing with chances without replacement, two sets of 20 French words and their Dutch trans- lations and two parallel problem solving assignments, in which simple probability principles (introduced in a separate text) had to be applied. In these assignments students were to calculate the number of_correct answers to multiple choice tests on the basis of guessing. Following Olshavsky (1976), red dots were put in places in the texts where verbalization was thought crucial. Text-comprehension was tested by means of two multiple-choice- 6 5 tests of 13 items each, emphasizin etaric3irr .7 instead of retention; Vocabulary-learning was tested PVleans offf a randomi- zed list of the French words 'that stUde08 had to t=ranslate in Dutch. The quality of the problem _solvar;gOutcome was used as indication for the result on the,essignmerX%. For the training program two ca 4itorje were written depicting two totally different WaYs of = passive way lettningo 0 and an active way incorporating several metHiagno _=ic routines and heuristics. Also a short booklet $.14$ 40_ttes in la-which a set of questions was described one may ask ~elf dutist.ng learning (e.g. "What do I already know of this kaject-mat5z ter?", "Why can't I understand this part?" and " which Legts are 4-..tAficult for me?"). Moreover, for each question SuggestAm _were mput forward on possible ways to answer these questioksFitally. a set of practice materials (texts, , words and probleks) were coccrIstructed. Procedure On a one to one basis students and 4Iex 'mementers pro- ceeded as follows after regular sOhOo hairs: aftlter a short introduction each student was asked to lealnme of the texts, one set of words and to solve one set of problems They were asked to read and think aloud and to lotid whatever sas1 out thoughts occurred to them. te They were to14 th4t test items would pertain to comprehension and amxad a.pOdcation not to reproduction of facts. All the subjeet'sverbal zaJations were taped and their observable behavior was tz:he experi- recc:Idedi Dy menters. Afterwards the tests were complet0. Half of the students participated as a woUptin thne training program, which occupied three hours in all. Xn the fir.7st session the group discussed different ways of studyincon>oorating the two case-histories. During the second sez=ssions the third arlol students learned the set of questionS anti the PosSib:ole ways to answer them and practiced receivio these, from the feedheok : experimenter. One or two weeks after the training tO _sks were learned, again reading and thinking aloUd ip individual sessions. Finally, the paraliel-tests were complete Desicn The design was non-randomized a t-posttesmst control t're group design. The training was administeted tothe sUbf,ljects that were most in need of it according to the_viQcheadMasteer of tha school. The remaining students thus forMe4 a gmVP for -.-omparison only instead of a real control group. Data-analvsis Verbal protocols and experimenter's observatj.ons were combined. As first a step, units of altaaysiowere de-eetermined, following a procedure proposed by Wouters Jon_lag (1982), atmd de which stressed meaningful units rather tD40 late/Ices, The data _ were categorized using the categories presehtod in ap-(apendix 1. Definitions and examples of the main catecion'Aare pre -sented in appendix 2. Intercoder-agreement was sa parQet3t, '7 6 RESULT:-S_ 14..---ab1e 1 presents the average results on the three task r. for weaker good students z-.---and separately, for the main uategc..=ries on3.y. Table 2=1 Mean frequencies of processes in protocols .7 and of good Ss (during pretest-session) weaker Text Vocabulary Problem- solving Procsses good weak good weak good weak Exectv=ition 53.4 50.3 18.7 21.7 14.2 0.6* Oriermatation 1.4 2.5 3.7 1.3 10.8 4.6* Monit-=.0ring 25.7 25.3 8.6 4.7 9.0 3.5* Regtil_ation 6.7 8.1 9.1 19.1 12.6 4.6* Testi _ng 21.6 93* 24.0 27.8 3.4 2.6 iagn_mosing 6.7 5.7 0.4 0.3 0.6 0.8 Evalu zating 0.7 1.6 2.2 2.7 0.8 0.9 Total 116.3 118.7 69.3 49.3 51.4 25.6* q<,05 There wesoere no differences in frequencies on the main categries good and weaker between Ss in text-processing. However, wit=hin the catewegory "testing" a significant difference showed up. G---Zood tested more student frequently on understanding than wea:ker (19.0 vs 8.4, p< .05) stuclent Weaker students, however tes =tied . loore oft=iten on knowledge (11.1 vs 1.3, n.s.) Thus, not the . gmantit1W of testing differentiated good from weaker subjects, I but its quaLlTlity. vocabulary learning, however, the quantity of test:: ing differentiating tti vias variable. Better students th3L=ree tested often as weaker Ss times (27.8 vs 9.3 p< .05) Differeno-ces good . and weaker Ss between for the problem-solving task relall-ted to the total number of processes registered, and of the number renitori-ng, regulation and orientation-processes. The-= training program had a significant effect (p<.051 on te:the perforrnance of the students on the comprehension test admir=i- stored a zf ter studying the fiats text. The trained group increased average score with 2 points (out of 13) whereas the comparion , group re,inained on the same level as in the pretest. Perferrnar=ice and proca esses on the vocabulary-test and the problemsolving-tamtsk did neitUher increase in the training-group, nor in the cormpar-m-i- sorvoup inewe training effect on the comprehension test also shoiowed process-data. -ithe u.p &zm In the training group the number of on understanding increased testingo from 10.0 to 20.6 (pc05) a_and the nunelp(mr of testings on knowledge decreased from 9.4 to 3 (ee Figure 3) . In the comparison-group . no changes in showed up. Piocesse Mean Frequency 22 Cornivarison group, understanding 20 18 16 14 under-mtanding 12 Trained gr=oup 10 knowLdge 8 6 4 2 Corn arison group, knowledge 0 *- pretest posttest session session Figure 3: Mean fre uencies of testngs before and after training DISCUSSION This study relations 2-rteresting reveled between some z-ences in performance and control executive individual dif processes. In text-processing the tuning of self-diagnostics to the learning goal proved to be arta most important aspect. In of self testing differen- vocabulary learning, however, the armount tiated between good In the and performing students. weaker monitorrig, regulation and orientation problem-solving-task both processes more frequaltly with showed up better Ss. Thus, individual differences in procemses correlating to performance task-dependent. Training prow-ed to be were for the effective text-processing only, both for perff.ormance- and process-data. because of the Three cautionary notes should Iae made. First, small this study, participating ir k. number of Ss replication on a larger scale is necessary. Currem-rly we are doing so in a study with 64 Ss. Second, one wary of and ef fact- should b cause relations.. The quality of executi_.re control processes may cause text-proces- the effectivity of task-performance. In case of the sing data we have reasons to belie-Nre that this is wimt happened. Apart from the performance and processes, betweexx correlation there also was a change in the nu_imber of testings on understan- the training progrm) ding (induced by coinciding with a change in performance. For the ltwo tasks, however, the causal other be in the revere direction. Weak relation might also performance (for instance caused by low abilitiis) might cause the occurrence of particular processes, n.4=ticing negative results, or like distinction keeping on between good and planning. Therefore, a bad executive control-processes wouM.d be helpful. Kuhl's distinc- mind- and activity-orented tion between processes might form an improvement (see the next studies) Third, one might also argue . that relations and self-regulation between performance processes should not always be expected, Some ways of processing might be preferable to others without leadng to better performance, for instance because they are more effJ_clent, they keep or because processing going in difficult circumstances. 9 8 The results of the present study also suggest the importance of on-line processes for self-revulation as opposed to processes occurring at the start and at the end of learning (orientation and evaluation). Many training-programs stressed the importance of the latter processes. Nevertheless, in the_present investiga- tion these before and after processes did not a) occur very often, b)did not differentiate between good and weak students and c) did not increase in frequency through training. _The Opposite counts for on-line processes such as sclftesting during learning. The design of the traiming study was, due to practical arrangements, not optimal. In future studies a real control group consisting of untrained weak students, is to be preferred. Moreover, training should be extended in time, particularly as to the vocabulary and problem-solving parts. Also transfer- and long-term effects should be taken into account. STUDY 2 Introduction Resarch-questions for this study were about the same as for the first study. Now, however, transfer effects were accentuated. Furthermore, an attempt was made to discern good and bad regula- ting processes, using Kuhl's theory on mind- and activitY- orientations. The categorization-scheme was extended with mind- oriented processes (directed to failure or success experience or valuations of the task at hand, e.g. "This is too difficult for me" or "I hate these sums"), and task-irrelevant statements (distractions). Moreover, processing-measures were related to impulsivity, concentration-ability, verbal intelligence and motivation. Finally, students from a school of special education were the subjects of this study. METHOD sub'ects The Ss were 10 boys from a secondary school for special educa- tion. They were selected because of their weak concentration abilities. Ages ranged from 12 to 14 years. materials In this study word arithmetic problems formed the main learning materials. Because of the learning disabilities of the Ss, the tasks used in study 1 could not be used. Arithmetic word problems were chosen because of the difficulties they pose for this kind of pupils (according to the teachers) We wanted to restrict the training . to one type of task in order to prevent confusions between strategies for different tasks. In total 11 word problems like the following constituted the training material: " A train departs at 21.47 hours. Traveling time is 3 hours and 36 minutes. what time At will the train arrive?". Another set of 7 of these story problems formed the pretest and still 7 more were the posttest. Also, both at the pretest-session and at the posttest-session transfer-tasks were administered: 12 fraction-problems like and 4/..=6/9 2 problem-solving tasks. 10