SLOWING DOWN MAY BE A WAY OF SPEEDING UP

mary budd rowe --- american educator spring 1987 --- pp38-43

see: AETS journal Jan-Feb 1986

Rowe's original research was done in 1969 --- and here in an article dated 1987 -- she finds not much has changed.

What will you do to change that?

Effect OF WAIT TIME ON STUDENTS AND Teachers

To "grow;" a complex thought system requires a great deal of shared experience and conversation. It is in talking ahout what we have done and observed and in arguing abOut what we make of our experiences that ideas multiply, become refined, and finally produce new questions and further explorations. While listening to tape recordings of high school biology students discussing laboratory findings (Rowe and Hurd, 1966) and the conversations and "talk" in elementary school classes during a "hands.on" science program (Rowe, 1968; 1969a, b 1 made two observations common to both sets of data. The pace of interaction between teachers and students was very rapid for both elementary and high school classes, except for three recordings in each group where the pacing seemed slower and the level as well as the quantity of student participation was greater Wait time one, the interval between the end of a teacher question and the start of a student response, was three to five seconds on the average for the three special tapes in each set. For all the other recordings, wait time for pausing was less than one second and was too brief to measure reliably with a stopwatch.

Effects ON Students

To help document the astonishing speed at which teacher and student exchanges took place, I fed the sound from the tapes into a servo-chart plotter The servo-chart plotter made a graph of the speech patterns and pauses and revealed another pause location that might be important, wait time two. Walt time two, the accumulation of pauses between student utterances before the teacher speaks again, in most of the recordings averaged 0.9 seconds, but on the three special tapes it exceeded three seconds. Servo-chart plots showed substantial pauses in the body of student explanations. Quick reactions by teachers appeared to cut off student elaboration. At this juncture it was necessary to determine if these protracted pauses of three seconds or longer; wait times one and two, played a part in producing desirable student outcomes observed in the three tapes at each level or whether they were just interesting anomalies. To answer that question for the elementary group, teachers and staff members in the trial center and I began a series of studies that lasted a number of years and involved both small groups of students and whole classes. We manipulated wait times one and two separately and then together to observe what happens (Rowe, 1972,1973; 1974a;b,c,d,e; 1975)In addition; I monitored the consequences of protracted exposure to longer wait time schedules in order to examine both immediate and long-term effects. I found consequences for both students and teachers, highlights of which are listed below and all of which were subsequently verified by other researchers.

2. More inferences are supported by evidence and logical argument. Under one-second wait times, the incidence of qualilied inferences is extremely low, hut it becomes quite common at the three-second wait time threshold (Anderson, 1978, Arnold, Atwood, and Rogers, 1973)

3. The incidence of speculative thinking increases.

4. The number of questions asked by students increases as does the number of experiments they propose.

As a rule, students ask questions infrequently, and when they do, the questions are usually to clarify procedures and are rarely directed to other students. This situation changes rather dramatically under the three-second regimen.

5. Student-student exchanges increase; teacher-centered "show- and-tell" behavior decreases.

Under very short wait times, students compete for turns to perform for the teacher. There is little indication that they listen to each other Under the three-second regimen, however, they show more evidence of attending to each other as well as to the teacher, and as a result, the discourse begins to show more coherence. This outcome is particularly influenced by wait time two.

6. failures to respond decrease.

"I don't know" or no responses are often as high as 30 percent in classrooms with mean wait times one and two of one second, which is the most common pace. Increasing wait time one to three seconds is particularly important for this outcome. During training, teachers often ask, "What if the student just doesn't know? Wait time will just be an embarrassment." The practical answer to that is to provide an "I pass" option. A student who has that option and exercises it at the end of three seconds is 70 percent more likely to be back in the discussion spontaneously before the period is over than is the case under the normal one-second regimen.

7. Disciplinary moves decrease.

Students maintained on a rapid recitation pattern show signs of restlessness and inattentiveness sooner than do students on the longer wait time treatment plan. At first this seems counter- intuitive to teachers. It appears that fast-paced teacher questioning is a device for maintaining control of behavior in fact, it not only inhibits the kind of thinking teachers seek to encourage but it can also increase the need of discipline. At this point, it may not be apparent why increased wait time should be a factor for improved class room discipline. The explanation may lie in a remark by a fifth grader to his mother about his teacher who was experimenting with three-second wait times. "it's the first time in all my years in school that anybody cared what I really thought - not just what am supposed to say" Protracted wait time appears to influence motivation, and that in turn may be a factor in attention and cooperation.

8. The variety of students participating voluntarily in discussions increases. Also, the number of unsolicited, but appropriate, contributions by students increases. Under the short wait time pattern, a major portion of responses comes from a small number of students: Typically six or seven students capture more than hall of the recitation time. Under the three-second regimen, the number of students usually rated as poor performers who become active participants increases. interestingly, this change in verbal activity gradually influences teacher expectations for students because more students do more task-related talking. (Verbal competence appears to be a salient factor in teacher judgments concerning a student's capabilities.)

9. Student confidence, as reflected in fewer inflected responses, increases.

Under a short wait time schedule, student responses are often inflected as though a tacit question such as "Is that what you want?" were attached to their statements. In a series of investigations to assess growth of con- fidence and a shift of reliance away from unsupported declarations by a powerful source, I presented a laboratory apparatus and a controlling variables problem to individual students chosen from different science settings (Rowe, 1968, 1969b, 1971) To assess the strength of an evidence-inference linkage, when subjects discovered and stated a relationship as a result of working with the apparatus, I would say, "I disagree." I wanted to observe what they did as a consequence. Could they persist through three disagreements? Some students came from the experimental science program classes with the usual short wait time pattern; others came from classes that in addition to the experimental science program also had three-second wait time regimens. A third group of students came from classes still engaging in the city's standard science program. I found that three-fourths of the new science and long- wait-time-group persisted through three disagreements by returning to the system, demonstrating their findings, and arguing the logic of their explanations. The other groups did much less well. For those in the experimental program under a short-pause procedure, less than hall lasted through three disagreements. For the standard program (largely from a book) only 2 percent met the criterion - most could not even make a start on the problem presented to them (see Honea, 1981, for consonant results in an attitude/wait time study using social studies content) In a wait time investigation conducted with Pueblo Indian students, Winterton (1977) found that students who were previously described by teachers as nonverbal contributed spontaneously twice as often in the long wait time classes as did their counterparts in science classes operating on the short wait time regimen. 'Winerton also reported increased values on other verbal indicators identified by Rowe. 10. Achievement improves on written measures where the items are cognitively complex.

Effects on Teachers

Once teachers stabilize longer wait time patterns, certain characteristics of their discourse change. These changes are treated as outcome variables because they are influenced by the wait time factor.

1. Teachers' responses exhibit greater flexibility. This is indicated by the occurrence of fewer discourse errors and greater continuity in the development of ideas. Under the short wait time schedule, the discourse does not build into structural propositions. To put it another way, there are more discontinuities in the discourse between students and teachers. Instead of a well-prepared banquet of ideas, the sequence of discourse resembles a smorgasbord at which everyone goes along, commenting on what she or he picks up, but paying no attention to the doings of others. One can calculate a discontinuity index for classroom discourse in much the same way one does when evaluating a computer-assisted-instruction program (Rowe, 1978) The index is higher for short wait time regimens.

2. The number and kind of questions asked by teachers changes. There are fewer questions, but more of them entail asking for clarification or inviting elaboration or contrary positions. As teachers succeed in increasing their average wait times to three seconds or more, they become more adept at using student responses - possibly because they, too, are benefiting from the opportunity afforded by the increased time to listen to what students say Boeck and Hillenmeyer (I973) reported that wait time one following a complex question tended to be longer than after a low-level question. Rice (1977), Doerr (1984 ~ and Hassler, lagan. and Szab (1980) confirm the original finding that increased wait times result in a cognitively more advanced pattern of teacher questions and reactions.

3. Expectations for the performance of certain students seem to improve. Under the longer wait time schedule, some previously "invisible" people become visible. Expectations change gradually, often signaled by remarks such as "He never contributed like that before. Maybe he has a special 'thing' for this topic." This effect was particularly pronounced where minority students were concerned. They did more task-relevant talking and took a more active part in discussions (Rowe, 1969b, 1974e, 1975) than they had before. While protracted wait times were never intended for use in drill and practice, neither I nor other researchers (e.g., Jones, 1980, Arnold, Atwood, and Rogers, 1974) have found markedly different wait time one values to be related to the level of question. I reported rather that this value was more influenced by teacher expectations. 1 asked teachers, prior to wait time training, to list the top five and bottom five students in their classes. Teachers gave the top five an average of 1.2 seconds of wait time one and the bottom five slightly less than one second (Rowe, 1974a, b, c, d, e; 1978) Gore (1981) suggested that teachers gave more wait time to one sex than the other. However, his measurement of wait time did not conform to the definitions.

Training for Wait Time

In their eagerness to elicit responses from students, teachers often develop verbal patterns that make the achievement of wait time two unnecessarily difficult. Chief among the inhibitors is the habit of mimicry. repeating part or all of what a student says. A high mimicry rate cuts off extended wait times and reduces the quantity and quality of student responses. An anecdote illustrates the unintended consequences of a mimicry pattern. In a classroom where the teacher was changing his pattern in order to increase wait time two, one of the students asked, "Mrs. B., how come you are not repeating things any more?" Before she could reply, another student answered the question. "1 know She knows that we can tell from the tone of her voice which answers she likes and which she doesn't, and we can stop thinking." There are other verbal signals to consider avoiding or reducing in conjunction with wait time, e.g., 'Yes. . . -but and" . - though" constructions because they signal the student that an idea is about to be rejected without the consideration due it.

Various procedures have been tried to help teachers learn to increase wait times (e.g., Anshuta, 1975; Atwood and Stevens, 1976) So far, the procedure that gets the most people to achieve relatively stable criterion three-second wait times in classroom settings takes longer that we would like, six to twelve hours. More-over, it is a bit aversive because it involves transcribing ten-minute segments of tape recordings from three teach-reteach cycles using groups of four students. (When teachers work with small groups, wait times are as short as when they work with a whole class Rowe, 1973.) The procedure is further complicated by the fact that teachers have seen their servo-chart plots for each teach-reteach cycle.

With the teach-transcribe-reteach procedure 70 to 80 percent of people achieve three-second criterion wait times (Rowe, 1973, 1974, b, 1978; McClathe, 1978) One must be aware however, that in the third or fourth week after teachers start using longer wait times in their classes, they revert to the original fast pace unless they have a chance to talk about what they are experiencing. What appears to happen in this transition interval is that grounds for decision making are less clear cut than was the case under the fast schedule. For example, teachers cannot decide how long to let student-student interaction go or how they feel when the nature of student-teacher interaction changes. In short, there are role and norm transformations taking place, and until these get settled, some teachers feel uncomfortable. A little support during this transition, even some advance warning that it will happen, appears to be sufficient to reinstate the three-second wait time average and to get teachers through the transition period. Garigiiano (1973) followed a teach-reteach regimen in a wait time training experiment but dropped the transcribing procedure out of one group in favor of having teachers listen to their tapes and identity and measure both species of wait time. His best performing treatment group (transcribed) attained 2.8-second averages. He confirmed the student effects described by Rowe, provided that average wait times did not drop much below this value.

Swift and Gooding (1983) and DeThre (1984) found that written training protocols are virtually useless in helping teachers achieve three-second wait times. In Swift's study, teachers averaged 1.35 seconds for wait time one and 0.68 seconds for wait time two, values that differed little from the means of his untrained group. Similarly, DeTure reported averages of 1.47 and 0.87 seconds for wait time one and two respectively for people trained with written or oral protocols. Swift and Hawkins (1979) and Gooding, et al. (1982) introduced an electronic monitoring device, the basic concept for which was initially developed jointly with Rowe, as a substitute for the feedback function supplied by the servo-chart plotter. Their voice-actuated relay system flashed a green light when wait times were satisfactorily long and a red light when wait times were too short. Teachers could have immediate wait time feedback while they were interacting with students. This method did result in some improvements but did not help the group attain criterion wait times until the procedure was accompanied by supportive intervention. Swift, Swift, and Gooding (1983, 1984) report that when the wait time devices were removed, despite supportive intervention, teachers reverted to short wait times. It may be that the presence of the mechanical device, while somewhat helpful, prevented teachers from attending to the fundamental changes in student-teacher interaction that take place with longer wait times, namely, the decisions occasioned by subsequent shifts in roles and norms.

DeTure (1985) remarks at the conclusion of a review of training procedures that the quick fix for this variable may not be feasible. Transcribing tapes as part of the training procedure in teach-transcribe- reteach -cycles is time consuming but remains the procedure that enables more people to achieve a three-second average wait time and successfully transfer it to the classroom. Based on research, it is clear that wait time two is more important than wait time one in many of its effects. Ironically, some training programs and teacher competency rating schemes mistakenly focus only on wait time one (DeTure, 1985) All the training techniques may be useless if teachers believe they will lose control of the class under the longer wait time schedule. As the Soars so aptly observed (1983), teachers confuse management of ideas with management of discipline. They need to know that behavior management is actually easier with protracted wait times (Rowe, l974a)

Adaptation of Wait Time for Lecture Formats

Often in high school, and particularly in college, there is a need to convey complex content, and the lecture appears to be the most commonly chosen format. For the lecture situation, I developed a ten-two procedure for college and the eight-two for high school. Based on a theory about how short-term and long-term memory interact, 1 identified four types of mental lapses that take place on the part of listeners in science classes (Rowe, 1967a, 1980, 1983) Using the ten-two and eight-two formats, participating science faculty would lecture for eight to ten minutes then stop for two. In the strictly regulated two-minute intervals, students in sets of three shared their notes and helped each other clarity concepts. All unresolved questions were to be reserved for the last live minutes of the period.

Experimental groups following this regimen generally show improved performance over control groups on the more complex test items, more delayed retention, and more positive attitudes toward the subject and method. The quality of student questions also improves as does the usefulness of their notes.

Rewards

Another line of research that impacts on the wait time situation deals with teacher sanctioning. The effects of protracted wait times are enhanced if the teacher sanctioning pattern (either positive or negative statements by the teacher) is reduced. That is, a high positive or negative sanctioning pattern reduces some of the effects of protracted wait times, particularly the following: student confidence, speculation, and elaboration (Rowe, 1974b,d; also see McGraw, 1978; Soar and Soar 1983)

SPECIAL EDUCATION

Exposure to longer wait times is as useful to talented students as it is to lesser-ability youngsters. Gifted and talented high school students participating in a summer science program found the extended wait times particularly motivating for the same reason as did thc fifth grader mentioned earlier. Bright students see man'. connections between ideas but they never get to talk about them. With increased wait times, the changes in their production of ideas, in the variety of moves under the game model of the classroom, and in their expressions of relief at being able to go beneath the surface ideas are evident Servo-chart plots of their explanations show that explanations come in bursts separated by substantial pauses (often in excess of five seconds as does the speech of most students if they are not interrupted during the process by short wait time two intrusions. Thus the protracted wait times help both fast and slow learners, but for different reasons. Two recent studies, one with mildly handicapped subjects and one with severely handicapped, showed some desirable outcomes for a five-second interval as opposed to the usual one-second pace (Korinek, 1985; Lee, 1985) In these cases, fundamental processing just takes more time. Extended wait time one was particularly important in the study by lee. Shrum (1985) found that wait time two (post-response wait time) in second language classes is even shorter (.73 seconds) than the .90 seconds reported by Rowe, much too short for thoughtful cognitive processing. She reports that average wait times are longer fol-lowing questions in the native language than they are in the second language (see also Rochester, 1973)

Conclusion

Under a wide variety of instructional situations and levels ranging from first grade through university level, from classrooms to museum and business settings, the quality of discourse can be markedly improved by increasing to three seconds or longer the average wait - times used by teachers after a question and after a response. These pauses are ordinarily so brief, one second or less on the average, that an adequate exchange of ideas and the nurturing of new ideas cannot take place. Wait time, however, is just another technique if one does not understand why fostering more productive exchanges among us all is so important. Gwen Frostic, a poet and artist, tells us in her book Beyond Time: We must create a great change in human direction - an understanding of the interdependency by which the universe evolves Know - that knowing - is the underlying foundation for the life we must develop. ... We cannot leave it to the scientists -nor any form of government -each individual must fuse a philosophy with a plan of action. Wait time provides a context in which teachers and students may dialogue together in the service of that purpose.

6/25s/95