'11

with plan defined as "any hierarchical process in the organism that can

control the order in which a sequence of operations is to be performed"

(Miller et al., 1960, p. 16, italics theirs).

The two scenarios discussed above illustrate what is meant by

"teacher as decision maker" in this paper (cf. Coladarci, 1959;

McDonald, 1965; Packer & Packer, 1959; Strasser, 1967). They also show

the critical connection betwcel. teaching skills and the basic teaching

skill: decision making.

Methods for Research on Teachers' Decisions

The various conceptualizations of the teacher as decision maker have

not led to empirical verifications (Coladarci, 1959; McDonald, 1965;

Strasser, 1967). The purpose of this section is to review methodology

for examining teachers' extemporaneous decisions (e.g., Stolurow, 1965;

Turner, 1971). Perhaps this tack will stimulate research on teachers'

decision-making processes.

Examining. Decisions from Descriptions

of Teacher-Student Interaction

To examine teachers' decisions from Jescriptions of teacher-student

interaction, basic teaching skills ("alternative acts") are identified

(cf. Smith et al., 1962), or sequences of basic skills described (cf.

Figures 2-4; Bellack et al., 1966; Taba & Elzey, 1964; Taba, Levine, &

Elzey, 1964; Smith et al., 1966).

Alternative basic teaching skills available to teacher (cf. alter-

natives in a decision matrix) can be described in a number o1 different

ways. One is the logical approach, exemplified by Smith et al. (1962).

N STANFORD CENTER Lb S DEPARTMENT
OF HEALTH,
FOR RESEARCH AND DEVELOPMENT EDUCATION & WELFARE
OFFICE OF EDUCATION
THIS DOCUMENT
PrN IN TEACHING DUCED EXACTLY
HAS BEEN REPRO
AS RECEIVED FROM
N- THE PERSON OR

O IONS STATED
ORGANIZATION ORIG
INATING IT POINTS
OF `dIEW OR OPIN
DC NOf NECESSARILY
REPRESENT OFFICIAL
OFF/CE OF EDU
CATION POSITION OR POLICv
CI
LIJ

Research and Development Memorandum No. 104

THE BASIC TEACHING SKILL: DECISION MAKING

Richard J. Shavelson

School of Education
Stanford University
Stanford, California

February 1973

Published by the Stanford Center for Research

and Development in Teaching, supported in part
as a research and development center by funds
from the National Institute of Education, U. S.
Department of Health, Education, and Welfare.
The opinions expressed in this publication do
not necessarily reflect the position, policy,
or endorsement of the National Institute of
Education. (Contract No. OEC-6-10-078,
Component 1B.)
 Introductory Statement

The Center's mission is to improve teechi-g in merican schools.

Too many teachers still employ a didactic ,ty.te aimed at filling passive
students with facts. The teacher's environment often prevents him from
changing his style, and may indeed drive him out of the profession.
And the children of the poor typically suffer from the worst teaching.

The Center uses the resources of the behavioral sciences in pur-

suing its objectives. Drawing primarily upon psychology and sociology,
but also upon other behavioral science disciplines, the Center has formu-
lated programs of research, development, demonstration, and dissemination
in three areas. Program 1, Teaching Effectiveness, is now developing a
Model Teacher Training System that can be used to train both beginning
and experienced teachers in effective teaching skills. Program 2, The
Environment for Teaching, is developing models of school organization
and ways of evaluating teachers that will encourage teachers to become
more professional and more committed. Program 3, Teaching Students from
Low-Iticome Areas, is developing materials and procedures for motivating
both students and teachers in low-income schools.

This memorandum reports research conducted by the Assessment System

component of the Program on Teaching Effectiveness. One problem for the
Assessment System is to refine measurement techniques linking teaching
acts to student achievement and aC,itude. In this paper, the teacher is
conceptualized as a decision maker, and a teaching act is viewed as the
result of a decision to use one or another teaching skill. To examine
only the frequency with which 'Ate skill is employed is to ignore impor-
tant information about when the teacher chooses to use it. It may be
that information about this latter decision accounts for student outcomes
better than frequencyof-use information, or at least in conjunction with
it. Various methods of rxamining a teacher's decisions are reviewed.

ii
 Abstract

Any teaching act is the result of a decision, either conscious or

unconscious. Previous_ research on basic teaching skills examined alter-
native teaching acts (e.g., explaining, questioning, reinforcing) without
examining how teachers choose between one or another act at a given point
in time. This paper argues that the basic teaching skill is decision
making. What distinguishes the exceptional teacher from his or her
colleagues is not the ability to ask, say, a higher-order question, but
the ability to decide when to ask such a question. This decision-making
process is examined using decision theory. Viewed from tIle decision
theory perspective, a teache- has a number of alternative acts from which
to choose. The choice may depend, for example. upon the teacher's sub-
jective estimation of a student's understanding of some material and the
usefulness of various alternatives in increasing that understanding.
Research on teaching skills is reviewed from this perspective to show
how current methodology can be applied to the study of teachers' decision
making. This conceptualization of the teacher as a decision maker in-
corporates previous research on basic teaching skills. Skills such as
questioning and explaining represent the teacher's repertoire of alter-
native acts from which to choose, while skills such as listening and
hypothesis generation influence the quality of information from which
the teacher estimates the student's understanding and the utility of
alternative acts. One implication is that teacher training should
include a decision-making component that integrates the other basic
skills.

iii
 THE BASIC TEACHING SKILL: DECIS ION MAKING

Richard J. Shavels on

Teachers make a great many decision s in the course of a day's teach-

ing. In fact any teaching act is the result of a decision--sometimes

conscious but more often not--that t he teacher makes after the complex

cognitive processing of available information.

Basic Teaching Skills and Decision Making

This reasoning is clari fied by Snow's (1968) model of teacher-student

interaction, shown in Figu re 1. (Some other model, such as Strasser's

[1967), might work as we

The di gram identifies cognitive events that are presum-

ably invol ed in heuristic teaching behavior. One can assume,
for examp e, that at some given instant in an ongoing group
discussi n a teacher attends to significant cues regarding the
course f discussion, makes inferences about that state of
confus on in some problem faced by the students, decides on a
kind f question or comment designed to open up a new aspect
of t e problem, and skillfully inserts the question or comment
int o the stream of discussion. It can be further suggested
th at both the current course of classroom events and the
t eacher's earlier acquisition of skills
will have been influ-
enced by the _eacher's aptitudes for teaching (and for learning
to teach), by his substantive knowledge and repertoire of tech-
nical and personal skills, and by his affective or temperamental
state at any given moment (Snow, 1968, pp. 78-79; see also
Snow, 1972).

The conceptualization of and research on basic teaching skills (e.g.,

Mc Donald & Allen, 1967; Rosenshine, 1970; Rosenshine & Furst, 1971), then,

have focused on the "exercise of skilled performance" almost to the exclu-

Parts of this paper were presented at the Symposium on Basic Teaching

Skills in Chicago, April 8, 1972. A version of it will appear in the
Journal of Teacher Education (in press).
 2

Professional Aptitudes for Inferences about

decisions teaching and learning state of learner
on next act Substantive knowledge
Technical and personal Hypcthesis generation
skills of teaching
Affective state
Information
Exercise of extraction
skilled
Cue attendance
performance

Class Given instant in time

begins
Fig. 1. Temporal course of teacher-learner interaction (eased on
Snow, 1968).

sion of other cognitive components of teaching behavior (some eNceptions are

Lundgren, 1972; Lundgren & Shavelson, 1972; Salomon & Sieber, 1970).

The connection between the current conceptualization of basic teaching

skills and teacher decisions is illustrated by two scenarios. The first is

as follows. As part of the research program at the Stanford Center for

Research and Development in Teaching in 1967, interns (teacher trainees) in

the Stanford Secondary Teacher Education Program were trained in a micro-

teaching laboratory to use various "technical skills of teaching" (e.g..

questions, reinforcement, probing, silence). Subsequently, the interns were

videotaped in their classrooms. These videotapes are particularly instruc-

tive. In one tape, an intern begins a discussion with a question. Several

students respond by raising their hands. The intern calls on one student;

he answers. Apparently dissatisfied with the answer, the intern asks this

student a second question. The student does not respond. The student is

asked a third question; apparently panicked, he mumbles something. Event-

ually the intern gives up on the student and redirects his questions to

other students. If one were to count the frequency with which the intern

used the technical skills of questioning and probing, he would receive a

high score. If one could ascertain what the student learned, it probably

would show little relation to the intern's "intent" or "plan." The

critical f,,ctor was the intern's decision or decisions to continue ques-

tioning the student instead of using some other basic skills such as

"explaining" or "refocusing."

This situation might be described as "individual decision making

under uncertainty" (cf. Luce & Raiffa, 1967; Raiffa, 1969) and can be

treated :ormally with concepts from decision theory. Our purpose is

heuristic; decision theory presents a frame of reference for viewing a

teacher's decisions.

Suppose the intern has a set of alternative basic teaching, skills

(e.g., questioning, explaining, refocusing) from which to choose his next

teaching act. This set of skills (alternatives) can be represented sym-

bolically A1, A2, ..., Ai, ..., Am. Suppose further that the intern's

preference for one skill instead of another depends on the student's

present learning state (called "state of nature"). The student's state

of nature might be "learned," "may have learned," or "unlearned" with

reference to some subject matter. The student's learning state may be

represented symbolically as s s s3. Corresponding to each learning

2'
 4

state is the probability that the learner is actually in a particular

state. This probability is assigned by the intern; it is subjective.

The intern's choice, then, can be represented by an m x n matrix (Table 1)

TABLE 1

Teacher's Decision Matrix

STUDENT LEARNING STATES

S ** S S
1 2 3

Pr(S )*** Pr(S ) Pr(S )

1 2 3
,

A 1.1 U u13
1 11 u12

A u u23
cf) 2 21 u22
-;

A u31 u32 u33

J' 3

A.
1

A u u u
m ml m2 m3

*A refers to the alternative teaching act (basic skill)

available to the teacher (intern).

**S refers to a particular learning state of the student;

i.e., "unlearned," "may be learned," or "learned."

***Pr(S) refers to the probability that a particular stu-

dent is in that learning state. The probability is estimated
subjectively by the teacher (intern).
 5

with m rows (basic skills) and n columns (student states), For every

cell in the matrix, there is an outcome. And, if the intern rank-orders

every cell according to some preferred outcome, u.. (assuming consistency;

ij

see Luce & Raiffa, 1967, Chapter 2), the intern should choose the row

(skill) that is optimal in some sense. One strategy is the following.

Any alternative basic skill (A ) can be represented as the sum:

1

uilPr(si) + ui2Pr(s2) + ui3Pr(s3). According to decision theory, the

1
intern should choose the basic teaching skill with the largest row sum.

The second scenario illustrating the relationship between basic

teaching skills and decision making focuses "in the sequential nature of

decision making. It is drawn from the work of Taba and Elzey (1964).

Using their observation system, student-teacher interaction ccui be repre-

sented by the flow chart in Figure 2.

When the teacher attempts to raise the level of thought

very early in the discussion, this typically results in
the children's returning to a lower level and in their
inability to sustain discussions at the higher levels of
thought [at this point in the decision tree, the teacher
inappropriately assigns a high utility (u) value to the
cell which cont s the lifting question and/or a high
probability value (Pr) to the "learned" state of the
student]. On the other hand, a strategy representing an
Effective pacing of shifting the thought onto higher
levels seems to follow a characteristic course [at each
choice point on the decision tree, assign the following
u and Pr values]. The level of seeking information is
sustained for a considerable time during the first por-
tion, of discussion [assign high u values to such basic
skills as lower-order questions, probing, explaining).
Grouping [e.g., forming categories for instances of a
concept] is requested after a large amount of information
has been accumulated [now assign a high u value to a

This type of analysis can be extended to a sequence o' decisions.

The intern's decision to use a particular skill is represented as a node
on a decision tree. The decision tree, then, represents the sequence of
choices made by the intern.
 IChild gives ! Teacher seeks [Child provides leacher attempts to Child provides Teacher
I
specific _,,, extension of requested lift level of thought reason as gives I

information- -"I information >1 information I from "information" -H requested by approval'

1

1
level to "reason" level teacher
I
L 1

Cr,

Fig. 2. Flow chart of teacher-child interaction, illustrating the

sequential nature of decision making (based on Taba & Elzey, 1964).
 7

higher-order question]. The result is that in a fairly

brief period, children transcend to labeling and then to
providing reasons for labeling and to inferences (Taba &
Elzey, 1964, pp. 532-33).

This sequence of decisions made by the teachcr can be characterized

by the TOTE unit (Miller, Galanter, & Pribram, 1960). TOTE stands for a

Test-Operate-Test-Exit sequence, illustrated in Figure 3. The arrows in

Figure 3 represent the transfer of control from one component of the

TOTE unit to the next. The TOTE unit can be applied to Taba and Elzey's

description as follows (illustrated in Figure 4). The teacher begins by

testing (T) whether the student has sufficient information available to

group instances of a concept; e.g., information available versus infor-

mation needed. If the test result is incongruity, i.e., the information

available is less than the information needed, the teacher moves to the

TEST i
EXIT
1
(Ccngruity)

(Incong.fuity)

+
OPERATE

Fig. 3. The TOTE unit (after Miller, Galanter, &

Pribram, 1960, p. 26).
 8

TEST
Does student have sufficient
Does Positive EXIT
information to group instances ?)
1

V'
OPERATE
!Execute basic skill designed
to provide necessary
information
L

Fig. 4. Description of a sequence of

decisions using the TOTE unit.

operate (0) component. Here the u values assigned to his decision matrix

remain the same, he readjusts the subjective probabilities with r:spect

to the student's learning state, determines the optimal basic sk,..11, and
2
then exercises that skill. Again, the teacher tests (T) whether the

student has sufficient information. If the test result is congruity,

i.e., the information available equals or exceeds the information needed,

\ .1
4
the teacher :xit; (E) this ph se, and moves to the next. The progress
\
from phase to phase can be ch iracterized as the teacher's clan (cf. the

decision tree or Talm and Elzey's [1964] plan in the scemtrio above),

2
At least four operations are carried out at the (0) component.
A better repre;entation of this ,omponent is a hierarchy of TOTE units
(see Miller et :J.., 1960, pp. 31-39). The hierarchy, however, introduces
complexities beyond what is useful for the present di!.cussion.
 with plan defined as "any hierarchical process in the organism that can

control the order in which a sequence of operations is to be performed"

(Miller et al., 1960, p. 16, italics theirs).

The two scenarios discussed above illustrate what is meant by

"teacher as decision maker" in this paper (cf. Coladarci, 1959;

McDonald, 1965; Packer & Packer, 1959; Strasser, 1967). They also show

the critical connection betwLe:. teaching skills and the basic teaching

skill: decision making.

Methods for Research on Teachers' Decisions

The various conceptualizations of the teacher as decision maker have

not led to empirical verifications (Coladarci, 1959; McDonald, 1965;

Strasser, 1967). The purpose of this section is to review methodology

for examining teachers' extemporaneous decisions (e.g., Stolurow, 1965;

Turner, 1971). Perhaps this tack will stimulate research on teachers'

decision-making processes.

Examining Decisions from Descriptions

of Teacher-Student Interaction

To examine teachers' decisions from Jescriptions of teacher-student

interaction, basic teaching skills ("alternative acts") are identified

(cf. Smith et al., 1962), or sequences of basic skills described (cf.

Figures 2-4; Bellack et al., 1966; Taba & Elzey, 1964; Taba, Levine, &

Elzey, 1964; Smith et al., 1966).

Alternative basic teaching skills available to teacher (cf. alter-

natives in a decision matrix) can be described in a number of different

ways. One is the logical approach, exemplified by Smith et al. (1962).

k
 10

Their unit of analysis--the episode--begins with and is classitied by an

expression or "entry" that triggers a verbal interchange about a topic,

and ends with the conclusion of discussion on the topic. The classifi-

cation of an entry--e.g., "defining" (a statement about how words or

other symbols are used to refer to objects)--depends on the response that

it logically demands as contrasted with the response that is actually

given. "Such a response is a schema. It is a form to which responses

to the members of a given class of entries would conform . . . were the

responses logically correct" (pp. 34-35). These responses, then, might

represent a list of alternative acts available to a teacher. The de-

cision problem is how to choose one response from all logically possible

responses.

A second approach is flow charting., as exemplified by the work of

Taba and Elzey (1964: also Taba, Levine, & Elzey, 1964). The relation

of this approach to teachers' decisions has already been discussed. One

important application of this technique is being investigated by Moore

(personal communication, 1972; 1973). She has constructed a game in

which a player (teacher) teaches a "student." At present, the game

simulates four types of students (man more types can be constructed to

test various hypotheses): (a) high intelligence--high subject matter

knowledge, (b) high intelligence--low subject matter knowledge, (c) low

intelligence--high subject matter knowledge, and (d) low intelligence- -

low subject matter knowledge. These types are defined operationally by

the student's position in the hierarchical structure of the learning

task (subject matter knowledge) and by his ability to move upward in the
hierarchy (intelligence). The player is given 84 moves with which to

teach the student. The moves, at present, are restricted to questions-

(higher- and lower-order) and explanations. To begin the game, the

player arranges his moves into a plan, as defined earlier. Once the

plan is established, the player interacts with one of the four students

selected at random. The interaction involves a player's move and a

student's binary response of "Yes, I can answer the question correctly,"

or "No, I cannot answer the question correctly." In general, any

teacher move that deals with information below the student's position

in the hierarchy receives a "Yes" response. A teacher move at the same

level as the student in the hierarchy or higher may get a "Yes" response,

depending on the teacher's past moves and the student's intelligence and

subject matter knowledge. The game continues until the student has

learned the terminal objective.

Moore's same can be described in decision theory terms. The moves

are the alternatives from which the player must choose his next acts.
3
The student's states are learned or unlearned. For each move, the

player's decision can be analyzed as in Table 1. And a flow chart (or

decision tree) can be built that shows the sequence of decisions made

by the player.

A third approach is sequential analysis of teacher or teacher-

student acts. The interaction data are examined for stable patterns of

acts (see, for example, Bellack et al., 1966). In some cases, these

3
Actually, the student can be in one of three states: learned,
transition from unlearned to learned, or unlearned. This is consistent
with current all-or-none mathematical models of learning (for a review,
see Kintscli, 1970).
 12

patterns tell about the teachers' (a) alternative basic teaching skills

(acts), (b) judgments about the various learning states of the student

or students (probabilities), and (c) preferences for various teaching

act/student state outcomes (utilities).

The application of sequential analysis to the investigation of

teachers' decisions is illustrated by S. Nicholson (personal communi-

cation). He used a statistical technique developed from information

theory (cf. Attneave, 1959) to analyze sequences of higher- and lower-

order questions. Preliminary results indicate a consistent pattern

of questions for most teachers at the beginning of a lesson. Initially,

the teacher asks lower-order questions. Then, a transition is made to

higher-order questions (cf. Taba & Elzey, 1964). The stage at which

this transition is made varies from one teacher to another.

If these results were cast into a decision framework, one possi)le

result would be Table 2. Two alternative aces are available to the

teacher of which he must choose cne, and students are characterized

simply as being in a learned or unlearned state. Since Nicholson's

results apply to the initiation of instruction, assume that the prob-

ability of the students being in the unlearned state is 1.00. (Vari-

ability between patterns of teacher questions might derive from differ-

ances in probabilities assigned to learning states.) Finally, rank-

order the various outcomes (cells) in Table 2 according to their desir-

ability (utility). The rankings indicate that outcomes in the learned

state are preferred to outcomes in the unlearned state, etc. (Vari-

ability between patterns of teacher questions also might be due to

differences in priorities.) To choose one of the two types of questions,

 13

TABLE 2

Hypothetical Decision Matrix for Nicholson's Data on

Sequences of Teachers' Questioning: Beginning
of Teacher-Student Interactions

Student's Learning State

Alternative basic skills Unlearned Learned Row Sum

Lower-order question 2 3 2.00

Higher-order questions 1 4 1.00

Probabilities associated
with student states 1.00 0.00

determine the expected utility of each: u1.J'14-u121)2

.101 (where u is the

ranking and j is the probability estimate of the student's learning

state). Table 2 shows that the expected utility (row sum) of lower-

order questions is greater than for higher-order questions; the teacher

therefore would be expected to choose a lower-order question.

At present, Nicholson has been unable to identify stable patterns

of questions for an individual teacher beyond the initial stage of in-

struction. This finding points to the complexity of the sequence of

decisions made by a teacher as instruction progresses. The absence of

stable patterns might be due, in part, to the teacher's (a) use of

other basic skills not represented in Nicholson's analysis of the

interaction data, (b) identification of many different student learning

states as instruction progresses, (c) inaccuracy in estimating student

learning states, (d) inconsistency in assigning prio.tities to outcomes.

 14

In conclusion, the various interpretations of Nicholson's work are

almost entirely speculative. It has been discussed here to show how a

,,:evential analysis of teacher acts for stable patterns might lead to

important hypotheses about teachers' decisions.

One problem with sequential analysis is that the inference from

the data to the teacher's decisions is often difficult to justify. The

last approach to be discussed here, stimulated recall (Bloom, 1953;

Jackson, 1968), offers a solution to this problem when combined with

sequential analysis. With this method a recording is made of classroom

interaction. It is played back to the teacher and is stopped at criti-

cal points in the interaction. The teacher is asked to report, say,

alternative possible acts, judgments about the students' learning states,

and the desirability of various outcomes. No study using this method

of examining teachers' decisions has been found.

Examining Decisions on the Basis of

Teaching Strategies

Teachers' decisions can be examined in the light of teaching stra-

tegies essentially in two ways. The first is post-hoc grouping. Student

achievement or teacher-student interaction data are used to group teach-

ers; differences between groups are examined. This form can be used in

several ways to investigate teachers' decisions. One way is to compare

groups of teachers on the basis of particular skills or combinations of

skills that they use. Then the correspondence between the teacher's

choice of skill and student achievement can be examined. For example,

Bellack et al. (1966) grouped teachers according to whether their stu-

dents had significantly higher-than-predicted or lower-than-predicted

 15

mean scores on a posttest on international trade. After the teachers

were grouped, a search was made for systematic differences in the two

groups' pedagogical moves and teaching cycles. No consistent differences

were found. A variant of this approach is disc-iminant analysis. With

this technique, a linear combination of skills, :Ain sequences, or

other variables is used to classify teachers into their original pest-

hoc groups. Once made, the classification into the original groups

would yield important information on similarities ard differences in

decision making between groups.

The second way of examining teachers' decisions in the light of

teaching strategies is a priori grouping. In this type of experiment,

teachers are assigned randomly to a sequence of decisions. Studies in

which teachers are assigned to one or another teaching method may be

classified here. But the problem with this type of study Is that teach-

ers \ary so much within groups that inter-group Lomparisons are difficult

to interpret (see Dubin & Taveggia [1968] for a review of sane of these

studies).

Methods that Simulate Teachers' Decisions

The rationale for studying machine simulations or models of a

master teacher, rather than the master teacher himself, is given by

Stolurow (1965, p. 225): "The idea of modeling the master teacher has

not worked. . . .
Since there probably are fewer ways to teach effec-

tively than to teach ineffectively, it is more likely that ineffective

teaching behaviors would be identified in observational studies of

teaching behavior."
 16

Stolurow studies teaching through computer simulation. (One of his

models is called SOCRATES: System for Organizing Content to Review and

Teach Educational Subjects.) His method is descriptive in the sense that

it makes "explicit the elements and relationships needed to account for

the phenalenon in which we are interested (e.g., a student's performance

on a learning task)" (p. 229). It is not descriptive of the human

teacher; the human teacher may operate in different ways to achieve the

same goal as the machine. Nevertheless, the programs (algorithms) used

in the machine simulation suggest the types of decisions that need or

ought to be made during extemporaneous instruction (cf. Stolurow's

[1965] tutorial process).

The objectives for the Tutorial phase of SOCRATES are to implement

the program (cf. the Operation phase of the TOTE unit, which might be a

hierarchy of other TOTE units) and to monitor the student's progress

(cf. the Test phase of the TOTE unit). The program contains two com-

ponents: the units of subject matter to be learned, and the decision

rules applied to the student's performance. The decision rules per-

form the following functions among others: (a) "specify the performance

that is sufficient to provide knowledge or results" (p. 232); (b) "speci-

fy that the response must be correct before the student is allowed to go

on to the next item or concept" (p. 233, cf. Tabs & Elzey's [1964] infor-

mation-gathering stage, or the Test of a TOTE unit); (c) specify that,

at some minimum (unacceptable) level of performance, the program is to

be changed (or the student is to be given more practice). "Each rule

that specifies a contingency is an aspect of the complex process called

 17

teaching; a complete set of rules defines a particular teaching program"

(cf. Landa, 1969, p. 233). "One interesting implication . . . is that

it is necessary to use a different program for each student entry when

only one outcome is accepted for all students" (p. 235; cf. aptitude-

treatment interaction in Cronbach 6 Snow, 1969).

Experimental studies involving SOCRATES might investigate the effects

of alternative programs (treatments) on students with different entry

behavior (aptitudes). Such an approach, then, would examine the effec-

tiveness of certain decision rules in matching treatments to individual

differences in students in order to optimize learning. Stolurow (1965)

reviews some studies using teaching machines (broadly defined) with

alternative "programs" that investigated the interaction of aptitude

with treatment (see Cronbach & Snow, 1969; Berliner & Cohen, 1972; or

Bracht, 1970, for more comprehensive reviews).

A second type of study is exemplified by Nuthall's (1968) extension

of work by Smith et al. (1962, 1966). Smith et al. (1966) defined a

venture as a single identifiable unit of subject matter; eight dif-

ferent types of ventures were identified. Nuthall chose to study the

conceptual venture: "A section of discourse in which the meaning, use,

implications of a class term are described or discussed" (p. 562). For

any type of venture, a particular unit of information was referred to

as a move. Nuthall defined teaching strategy as "the ordered set of

conceptual moves which occurred in the venture" (p. 563). Different

strategies, then, may be viewed as the result of differences in choices

among teaching skills. He identified four logical teaching strategies

 18

and developed programmed texts to simulate each type of strategy. In

this way, students could be assigned at random to teaching strategies.

Teaching strategies were identical within each of the four major types

but differed in specifiable ways between strategies. Nuthall concluded

that "differences in teaching strategy could be meaningfully related to

differences in student learning. But the generalizability of the results

is limited by the lack of a system of classifying meaningful concepts in

some pedagogi,:ally significant way" (p. 583).

Conclusion and Implications

Any teaching act is the result of a decision, whether conscious or

unconscious, that the teacher makes after the complex cognitive pro-

cessing of available information. This reasoning led to the hypothesis

that the basic teaching skill is decision making.

This conceptualization incorporates previous research on teaching

skills. Such skills as questioning, explaining, reinforcing, and

probing represent the teacher's repertoire of alternative acts from

which he must choose at any instant in time. Such skills as listening

and hypothesis generation influence the teacher to infer (a) the

probability that the student, let us say, does or does not understand

the concept the teacher is presenting, and (b) the relative utility of

each of the alternative teaching acts.

One implication of this conceptualization is that research on

teaching should examine teachers' decisions. Decisions may be the

principal focus of research, or they may be studied indirectly when

the main concern is the exercise of skilled performance. Another im-

 19

plication is that teachers' training should include a decision-making

component. The teacher should be taught not only how to ask, say,

higher-order questions, but also how to decide when to ask a higher-

order question.

It
 20

References

Attneave, F. Applications of information theory to psychology. New

York: Holt, Rinehart & Winston, 1959.

Bellack, A. A., K]iebard, H. M., Hyman, R. T., & Smith, F. L. The

language of the classroom. New York: Teachers College Press, 1966.

Berliner, D. C., & Cahen, L. S. Trait-treatment interactions and

learning. In F. N. Kerlinger (Ed.), ileview of research in
education. Itasca, Ill.: F. E. Peacock, Publishers, in press.

Bloom, B. S. Thought processes in lectures and disctssions. Journal of

General Education, 1953, 7, 160-69.

Bracht, G. H. Experimental factors related to aptitude-treatment inter-

actions. Review of Educational Research, 1970, 40, 627-45.

Coladarci, A. P. The teacher as hypothesis-maker. California Journal

for Instructional Improvement, 1959, 2, 3-6.

Cronbach, L. J., & Snow, R. E. Individual differences in learning

ability as a function of instructional variables. Stanford
University, 1969. Final Report, Contract Na. OEC-4-6-061269-1217,
U. S. Office of Education. (ED 029 001)

Dubin, R., & Taveggia, T. C. The teaching-learning paradox: A com-

parative analysis of college teaching methods. (Center for the
Advanced Study of Educational Administration, Monograph No. 18)
University of Oregon, 1968.

Gage, N. L. Teacher effectiveness and teacher education: The search

for a scientific basis. Palo Alto, Calif.: Pacific Books, 1972.

Jackson, P. W. Life in classrooms. New York: Holt, Rinehart &

Winston, 1968.

Kintsch, W. Learning, memory, and conceptual processes. New York:

Wiley, 1970.

Landa, L. N. The construction of algorithmic and heuristic models csf

thinking activity and some problems in programmed learning. In
W. R. Dunn & C. Holroyd (Eds.), Aspects of educational technology.
London: Methuen & Co., 1V69.

Luce, R. D., & Raiffa, H. Games and decisions. New York: Wiley, 1967.

Lundgren, R. E. The effects of listening training on teacher listening

and discussion skills. (Stanford Center for Research and Develop-
ment in Teaching, Technical Report No. 25) Stanford University,
1972. (ED 062 310)
 21

Lundgren, R. E., & Shavelson, R. J. Effects of listening training on

teachers' listening and discussion skills. Unpublished paper,
Stanford University, 1972.

McDonald, F. J., & Allen, D. W. Training effects of feedback and model-

ing procedures on teach'lg performance. (Stanford Center for
Research and Development in Teaching, Technical Report No. 3)
Stanford University, 1967. (ED 017 985)

McDonald, F. J. Educational psychologx. Belmont, Calif.: Wadsworth,

1965.

Miller, G., Galanter, E., & Pribram, K. H. Plans and the structure of
behavior. New York: Holt, Rinehart & Winston, 1960.

Moore, C. A. Styles of teacher behavior under simulated teaching

conditions. Unpublished doctoral dissertation, Stanford University,
1973.

Nuthall, G. A. An experimental comparison of alternative strategies for

teaching concepts. American Educational Research Journal, 1968,
5, 561-84.

Packer, C. K., & Packer, T. Cybernetics, information theory, and the

educative process. Teachers College Record, 1959, 61, 134-42.

Raiffa, H. Decision analysis: Introductory lectures on choices under

uncertainty. Reading, Milss.: Addison-Wesley, 1970.

Rosenshine, B. Interpretive study of teacher behaviors related to stu-

dent achievement. USOE Project No. 9B-010. Philadelphia: Temple
University, 1970.

Rosenshine, B., & Furst, N. Research on teacher performance criteria.

In B. O. Smith (Ed.), Research in teacher education: A symposium.
Englewood Cliffs, N. J.: Prentice-Hall, 1971.

Salomon, G. Interaction of communication-medium and two procedures of

training for subjective response uncertainty of teachers. Unpub-
lished doctoral dissertation, Stanford University, 1968.

Salomon, G., & Sieber, J. E. Relevant subjective response uncertainty

as a function of stimulus-task interaction. American Educational
Research Journal, 1970, 7, 337-49.

Shavelson, R. J. Dialogue paradigm: A paradigm for research on teach-

ing. Stanford Center for Research and Development in Teaching,
Stanford University, 1969. (mimeo)
 22
1
Smith, B. 0., et al. A study of the logic of teaching. Urbana:
University of Illinois Press, 1962.

Smith, B. 0., et al. A study of the strategies of teaching. Urbana:

University of Illinois Press, 1966.

Snow, R. F. A model teacher training system: An overview. (Stanford

Center or Research and Development in Teaching, R&D Memorandum
No. 92) Stanford University, 1972. (ED 066 437)

Snow, R. E. Heuristic teaching. (Stanford Center for Resea:Z.- and

Development in Teaching, Third Annual Report) Stanford University,
1968. Pp. 78-84.

Stolurow, L. M. Model the master teacher or master the teaching model.

In J. D. Krumboltz (Ed.), Learning and the educational process.
Chicago: Rand M'Nally, 1965.

Strasser, B. A conceptual model of instruction. Tile Journal of Teacher

Education, 1967, 18, 63-74

Taba, H., & Elzey, F. F. Teaching strategies and thought processes.

Teachers College Record, 1964, 65, 524-34.

Taba, H., Levine, S., & Elzey, F. F. Thinking in elementary school

children. San Francisco State 7ollege, San Francisco, 1964.
U. S. Office of Education Cooperative Research Project No. 1574.

Turner, R. L. Conceptual foundations of research in teacher education.

In B. O. Smith (Ed.), Research in teacher education: A symposium.
Englewood Cliffs, N. J.: Prentice-Hall, 1971.

It