With semiconductor manufacturing efficiency improvement efforts primarily focused upon equipment technologies, KBI recognizes that overall labor effectiveness (OLE) improvement efforts will soon be required to support that industries performance improvement goals and maintain pace with laws governing the industry . Illinois State University published this research paper which clearly defines the strategies incorporated with the KBI Performance Series line of learning products. (KBI has published a WBT product to summarize the research and draw parallels to benefits the semiconductor manufacturing community may realize. Visit our website http://www.knowledgebrokerinc.com)
Teaching Brain-Behavior Relations Through Stimulus Equivalence
1. JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2010, 43, 19–33 NUMBER 1 (SPRING 2010)
TEACHING BRAIN–BEHAVIOR RELATIONS ECONOMICALLY WITH
STIMULUS EQUIVALENCE TECHNOLOGY
DANIEL M. FIENUP, DANIEL P. COVEY, AND THOMAS S. CRITCHFIELD
ILLINOIS STATE UNIVERSITY
Instructional interventions based on stimulus equivalence provide learners with the opportunity
to acquire skills that are not directly taught, thereby improving the efficiency of instructional
efforts. The present report describes a study in which equivalence-based instruction was used to
teach college students facts regarding brain anatomy and function. The instruction involved
creating two classes of stimuli that students understood as being related. Because the two classes
shared a common member, they spontaneously merged, thereby increasing the yield of emergent
relations. Overall, students mastered more than twice as many facts as were explicitly taught, thus
demonstrating the potential of equivalence-based instruction to reduce the amount of student
investment that is required to master advanced academic topics.
Key words: college students, neuroanatomy, programmed instruction, stimulus equivalence
________________________________________
Instructional interventions based on stimulus recognized as ‘‘belonging together’’ or ‘‘mean-
equivalence and related principles of stimulus ing the same thing’’). In the language of
class formation (hereafter called equivalence- stimulus equivalence, the emergent ability to
based instruction or EBI) focus on teaching relate stimuli that were not previously paired,
generatively, that is, giving students the con- but that share a common associate, is called
ceptual building blocks that allow them to transitivity (Sidman) or transitive inference.
reliably ‘‘go beyond the information given’’ Transitive inferences are expected to emerge
(Bruner, 1957; p. 41). The defining feature of reliably only if certain foundational facts are
EBI is that students learn overlapping condi- learned and only if these reflect overlapping
tional discriminations that promote the forma- conditional discriminations (Sidman).
tion of additional, unpracticed abilities (e.g., Principles of stimulus class formation have
Critchfield & Fienup, 2008; Green & Saun- been applied mainly toward enhancing the
ders, 1998; Sidman, 1994; Stromer, Mackay, & repertoires of young children and persons with
Stoddard, 1992). For example, having learned developmental disabilities. Relevant academic
to associate the spoken word ‘‘cat’’ with both a instruction (for a seminal review, see Stromer et
photograph of a cat and the printed word cat, a al., 1992) has, therefore, focused on basic
child may be able without further teaching to academic repertoires such as translating between
relate the cat picture to the printed word cat, fractions and decimals (Lynch & Cuvo, 1995),
even though the child has never experienced identifying letters (Connell & Witt, 2004; Lane
these together (e.g., see Sidman & Cresson, & Critchfield, 1998a) or words (de Rose, de
1973). This suggests that the picture and the Souza, & Hannah, 1996), and translating
spoken and printed words have become func- between English and Spanish words (Joyce &
tionally interchangeable (i.e., the items are Joyce, 1993). In recent years, applications also
have emerged that focus on basic social and
We thank Ruth Anne Rehfeldt and Rocio Rosales for
helpful comments on a draft of the manuscript. ´ ´
communicative abilities (e.g., Perez-Gonzalez,
Address correspondence to Daniel M. Fienup, who is ´
Garcıa-Asenjo, Williams, & Carnerero, 2007;
now at the Psychology Department, Queens College, 6530 Rosales & Rehfeldt, 2007).
Kissena Blvd., Flushing, New York 11367 (e-mail:
daniel.fienup@qc.cuny.edu). The published literature includes few instanc-
doi: 10.1901/jaba.2010.43-19 es in which sophisticated academic material was
19
2. 20 DANIEL M. FIENUP et al.
taught to advanced learners using techniques by Lane and Critchfield (1998a) in which
based on research on stimulus classes (Fields et children with Down syndrome were taught
al., 2009). We are aware of only two relevant relations between written vowels and the
reports, in which college students learned algebra spoken label ‘‘vowel’’ (parallel instruction also
skills through a combination of spoken instruc- was used to teach relations involving conso-
tion and stimulus class formation (Ninness et al., nants). For example, the children were explic-
2005, 2006). The dearth of applications with itly taught to match ‘‘vowel’’ to the letter A and
sophisticated learners might not appear to be A to the letter O. Without further training,
much of an oversight if it is assumed that such they could also match O and ‘‘vowel’’ (transi-
learners require little academic assistance. Yet, as tivity). In the same way the children mastered a
learner capability increases, so does the complex- class consisting of ‘‘vowel’’ and the letters E and
ity of academic subjects. Consider, for example, U. Because two vowel classes shared a common
the challenge of instructing college students member (A), they merged to form one larger
about the biological bases of behavior (see class (A-O-E-U-‘‘vowel’’), thereby allowing the
Wilson et al., 2000). Even an elementary children to match additional vowels that had
textbook on this topic contains many hundreds never been paired directly during training (e.g.,
of unfamiliar terms and concepts related to E and A).
neuroanatomy and the biochemical functioning The present study used EBI to establish
of brain regions and individual neurons (e.g., relations among brain regions, their anatomical
Carlson, 2005). Each anatomical feature must be locations, and psychological functions and
understood in relation to its various physical and psychological problems associated with them.
functional characteristics. Couple this with the For efficiency, a few carefully selected relations
typical college course schedule, which includes were taught with the goal of promoting the
only a few hours per week of formal instruction, emergence of several untaught relations that
and it is easy to see why even in advanced would increase the number of relations learned
academic programs it is important to get the above the number of relations explicitly taught
most out of limited instructional time (e.g., (e.g., Fienup & Dixon, 2006; Sidman et al.,
Chew, 2008). The present study illustrates how 1985). Four classes of five learning stimuli each
this may be achieved through stimulus class (defined in Figure 1 and denoted here as A, B,
formation. C, D, and E) were employed to which the
One key feature of the present study was an participants were not exposed. Each represented
attempt to promote generative responding a lobe of the brain, and two small classes were
through a phenomenon called class merger, in established for each (A-B-C and A-D-E), both
of which incorporated emergent potential. It
which two stimulus equivalence classes sharing a
was expected that students would spontaneously
common member spontaneously merge to form
treat the B and C stimuli as ‘‘belonging with’’
one larger class (Fienup & Dixon, 2006;
the D and E stimuli (i.e., new relations would
Sidman, Kirk, & Willson-Morris, 1985). Class
emerge) because the two classes shared a
mergers increase the number of potential
member (A).
relations among facts that have never been
directly paired (i.e., more emergent relations;
see Lane & Critchfield, 1998b). Nevertheless, METHOD
this tactic of arranging for class mergers to Participants
develop more relations than those taught rarely Eight college undergraduates volunteered
has been employed to academic advantage. To after reading a flier posted on a recruitment
illustrate the potential benefits, consider a study bulletin board and participated after providing
3. TEACHING ABOUT THE BRAIN 21
Figure 1. Stimuli used in the study. See text for details of how stimuli were displayed.
informed consent. In exchange for participat- Setting and Materials
ing, they received vouchers that could be Equipment. The study took place in a
exchanged for bonus credit in psychology classroom equipped with 30 computer work-
courses. Volunteers were retained in the study stations, with multiple students working simul-
if they scored below 70% on all of the pretests. taneously. Instructional testing and training
This criterion was used to avoid ceiling effects procedures were automated using a custom-
that could preclude evidence that the experi- written computer program created with Visual
mental procedures promoted learning. No data Basic 2005 (Dixon & MacLin, 2003). Each
are reported for 4 individuals whose pretest student worked on an IBM-compatible desktop
scores indicated existing mastery of the material computer (with a 15-in. flat panel monitor,
to be taught. The 4 remaining students ranged keyboard, and mouse) that ran on the Microsoft
in age from 18 to 22 years (M 5 20.3, SD 5 Windows XP operating system.
1.7). Participants appeared to be roughly Learning and feedback stimuli. Figure 2
typical of undergraduates at the university at reproduces key features of the computer display
which the research was conducted based on self- that students viewed during the lessons. Learn-
reported college grade point averages (range, ing stimuli were presented in black Arial font
2.4 to 3.7) and ACT college-entrance exami- (24 to 28 point) within white boxes (approx-
nation scores (range, 21 to 27). Although imately 7.6 cm by 7.6 cm). For all training and
college students often serve as participants of testing, one box was presented at the top of the
convenience in behavioral research (e.g., Ecott screen as a sample stimulus and four boxes
& Critchfield, 2004), in the present study below as comparison stimuli. In the top right
college students were the population of practi- corner of the participant’s screen was a visual
cal interest. feedback box. During training the box dis-
4. 22 DANIEL M. FIENUP et al.
Figure 2. Participant’s display, with example stimuli, as it appeared during training phases.
played the mastery criterion and success towards or the relations among them that would be
mastery. During testing the box displayed the taught.
number of trials on the test and the trial Overview of lessons. The students completed
number the participant was currently complet- two lessons, each focusing on a unique aspect of
ing. the subject matter to be taught and each
During training phases, auditory accuracy consisting of a pretest, training, and a posttest.
feedback was provided through stereo head- The pretest and posttest of each lesson were
phones. Correct responses were followed by an identical. The computer program recorded the
ascending sound, and incorrect responses were amount of time that students spent engaged
followed by a descending sound (called a chime with the training phase of each lesson (omitting
and chord, respectively, in the Windows XP nonlesson activities such as informed consent,
operating system). No accuracy feedback was instructions provided by the researcher, and
provided during testing. transitions between tasks).
The lessons consisted of a series of trials in a
General Procedure match-to-sample format (Green & Saunders,
General instructions, provided before the 1998; Stromer et al., 1992). Each trial
study began, described the computerized lessons presented a sample stimulus (the question),
as in development for eventual classroom use. which appeared near the top of the student’s
The students were told that while working on screen, and up to four comparison stimuli
the lessons they would complete pretests, (possible answers), which appeared in the
training sessions, and posttests; that a score of bottom row of boxes (see Figure 2). One
at least 90% was required on each posttest to comparison stimulus was the correct response,
continue to the next part of the experiment; and and the remaining stimuli were incorrect. The
that they would be excused from the study after sample and comparison stimuli were presented
2 hr or completion of all of the lessons, simultaneously, and both stimuli remained on
whichever came first. Instructions did not, the screen until the student made a response to
however, describe the stimuli used in the lessons a comparison stimulus. During training por-
5. TEACHING ABOUT THE BRAIN 23
tions of the study, clicking on any comparison trials (e.g., either A1RB1 followed by A2RB2
box immediately produced auditory and visual or vice versa). This resulted in no more than
accuracy feedback (described above) followed two consecutive trials involving the same sample
by the next trial (no intertrial interval). During stimulus and ensured that the students gained
testing portions of the study, clicking on any approximately the same amount of experience
comparison stimulus immediately produced with each of the stimulus relations being taught.
visual progress feedback. Overview of research design. The three
To complete a lesson, a student had to stimulus sets were A-B-C (Lesson 1), A-D-E
demonstrate competence both during training (Lesson 2), and B-C-D-E (relations between
and on the posttest. During training, mastery Lessons 1 and 2). Figure 3 displays the order in
was defined as making 12 consecutive correct which testing and training were staggered. This
responses during a given learning unit (e.g., design is similar to that employed by Fienup
ARB). The cumulative probability of selecting and Dixon (2006) and was implemented on an
the correct one on 12 consecutive trials was .512 individual basis. The experimental design
(.0002) or .2512 (.00000006) given two or four followed the general logic of multiple baseline
comparison stimuli, respectively, on each trial. and multiple probe experiments, in which
On this basis, the mastery criterion was deemed abilities are measured at several times to
adequate for distinguishing between genuine determine whether changes correspond to the
mastery and spurious runs of correct responses. introduction of an intervention (Johnston &
Once a student achieved mastery of the training Pennypacker, 1980). The study began by
portion of a lesson, an on-screen message pretesting the Lesson 1 and 2 relations followed
declared, ‘‘You have passed! Click this button by training on Lesson 1 relations. Following the
to continue.’’ Clicking on this button began the mastery of Lesson 1, Lesson 2 and the B-C-D-E
next scheduled part of the procedure. relations were tested. This served as a control to
For posttests, mastery was defined as scoring help attribute the change in Lesson 1 scores to
at least 90% correct. A lower score initiated the instruction provided. Then, Lesson 2
remediation in which the training portion of instruction was implemented to replicate the
the lesson was repeated, after completion of effects of EBI on generative responding.
which the posttest was readministered. After the Note on symmetrical relations. The training
posttest mastery criterion had been met, the and testing phases of the study omitted
student proceeded to the next scheduled portion symmetrical variants of explicitly taught rela-
of the study. tions (e.g., ARB was taught and tested, but not
Randomization of trials. During both training BRA). We assumed that, for the skilled
and testing, the positions of the comparison learners on whom this study focused, compe-
stimuli were randomized such that each possible tence on the trained relations implied compe-
comparison stimulus was equally likely to be tence on symmetrical variants (e.g., Fields &
assigned to each possible screen location. The Reeve, 1997). This approach also allowed us to
trials that comprised the testing phases of each limit the number of trials in the training and
lesson (pretest and posttest) occurred in a testing phases. When examining potential
unique, randomized order for each student; transitive associations among stimuli that had
however, each student experienced the same not previously been paired, however, we tested
number of trials and exposure to different the two symmetrical forms of relevant relations
relations. During training involving learning (e.g., B as sample with C as comparison, and C
units with two trial types (e.g., A1RB1, as sample with B as comparison) and treated
A2RB2), the order was randomized every two each form as a separate relation. This approach
6. 24 DANIEL M. FIENUP et al.
Figure 3. Schematic showing the sequence of training and testing phases of the study.
was based on the common assumption that for trials, with all B stimuli presented as compar-
both of these to emerge, the learner must have isons.
mastered the symmetrical version of each ARC training. The students learned four
trained relation (e.g., Sidman, 1994). ARC relations. For this training the sample
stimuli were brain lobe names (e.g., frontal lobe)
Lesson 1 (Lobes and Functions): A-B-C Relations and comparison stimuli were additional brain
In this lesson students learned how the functions (e.g., Function 2: Involved in higher
following stimuli (textual descriptions) relate: cognitive functions). ARC training was similar
names of brain lobes (A stimuli), a psycholog- to ARB training. Training occurred in three
ical function associated with the respective brain blocks of trials that had to be mastered
lobes (B stimuli), and a second function separately. The first involved A1RC1 and
associated with the respective brain lobes (C A2RC2 relations, the second involved
stimuli). Figure 1 displays the stimuli, Table 1 A3RC3 and A4RC4 relations, and the third
shows the specific trials presented in each phase involved all four ARC relations.
of training, and Figure 4 (top left) shows the A-B-C pretest and posttest. The following
basic structure of training and testing, with transitive relations were tested: four BRC
trained relations depicted via black arrows and relations and four CRB relations. The test
expected emergent transitive associations de- included four trials of each relation type, for a
picted via gray arrows. total of 32 trials. All four possible comparison
ARB training. The students learned four stimuli appeared on each trial (e.g., for BRC
ARB relations. The sample stimuli were brain relations the comparison stimuli always were
lobe names (e.g., frontal lobe) and comparison C1, C2, C3, and C4).
stimuli were brain functions (e.g., Function 1:
Involved in movement). Training occurred in Lesson 2 (Lobes, Locations, and Disorders):
three blocks of trials, each of which had to be A-D-E Relations
mastered separately. The first training block In this lesson the students learned how the
involved A1RB1 and A2RB2 trials, with B1 following stimuli (pictures and textual descrip-
and B2 serving as comparisons along with two tions) relate: brain lobe name (A stimuli), an
blank boxes. The second block involved anatomical location of the respective brain lobe
A3RB3 and A4RB4 relations, with B3 and represented by a colored drawing (D stimuli),
B4 serving as comparisons along with two and an effect (disorder) that can arise due to
blank boxes. In the third block all four ARB damage to the respective brain lobe (E stimuli).
relations were intermingled in a block of Figure 1 shows the stimuli that were involved,
7. TEACHING ABOUT THE BRAIN 25
Table 1
Training Sequence
Condition Phase Sample/correct comparison All comparisons
ARB training 1 A1RB1 and A2RB2 B1 and B2
2 A3RB3 and A4RB4 B3 and B4
3 A1RB1 and A2RB2 B1, B2, B3, B4
A3RB3 and A4RB4
ARC training 1 A1RC1 and A2RC2 C1 and C2
2 A3RC3 and A4RC4 C3 and C4
3 A1RC1 and A2RC2 C1, C2, C3, C4
A3RC3 and A4RC4
ARD training 1 A1RD1 and A2RD2 D1 and D2
2 A3RD3 and A4RD4 D3 and D4
3 A1RD1 and A2RD2 D1, D2, D3, D4
A3RD3 and A4RD4
ARE training 1 A1RE1 and A2RE2 E1 and E2
2 A3RE3 and A4RE4 E3 and E4
3 A1RE1 and A2RE2 E1, E2, E3, E4
A3RE3 and A4RE4
Note. This table displays the trials that were presented in each phase of the study. For each type of relation (e.g., ARB,
ARC) participants began MTS training with stimuli from Sets 1 and 2. Following mastery, participants completed
training with Sets 3 and 4 followed by a phase with all four sets of stimuli presented (see Figure 1 for description of the
stimuli).
Figure 4. Top: structure of training and testing in the two lessons. Relations that were directly taught are shown as
black arrows; transitive associations that were expected to emerge without direct training are shown as gray arrows.
Bottom: structure of the test for class merger. Transitive associations that were expected to emerge without direct training
are shown as gray arrows. Note that in each phase of the study, the relevant relations were included for four stimulus sets,
each representing a different lobe of the brain.
8. 26 DANIEL M. FIENUP et al.
and Figure 4 (top right) shows the basic merger test were BRD (and DRB), BRE (and
structure of training and testing, which paral- ERB), CRD (and DRC), and CRE (and
leled that of Lesson 1. ERC). All four possible comparison stimuli
ARD training. The students learned four appeared on each trial (e.g., for BRD relations
ARD relations. The sample stimuli were brain the comparison stimuli always were D1, D2,
lobe names (e.g., frontal lobe), and comparison D3, and D4). Each relation was presented four
stimuli were pictures of the brain with the times, resulting in 128 total trials (8 relations 3
respective area highlighted. ARD training was 4 lobe classes 3 4 trials each).
similar to ARB training. Training occurred in Students completed the class merger test on
three blocks of trials that had to be mastered two occasions: prior to A-D-E training, before the
separately. The first involved A1RD1 and lessons provided any basis for relating the A-B-C
A2RD2 relations, the second involved and A-D-E stimuli, and following A-D-E training
A3RD3 and A4RD4 relations, and the third and the posttests for A-B-C and A-D-E relations.
block involved all four ARD relations.
ARE training. The students learned four RESULTS
ARE relations. For this training the sample
stimuli were brain lobe names (e.g., frontal lobe) Results of pretests and posttests are summa-
and comparison stimuli were statements about rized in Figure 5. Test outcomes are shown in
what could occur given damage to the respective terms of percent correct. Results of the trainings
brain region (e.g., Damage causes impulsiveness). are summarized in Table 2 in terms of the
ARE training was similar to ARB training. number of trials and amount of time required
Training occurred in three blocks of trials that in a given lesson to meet the mastery criteria.
had to be mastered separately. The first involved Because mastery was required prior to any
A1RE1 and A2RE2 relations, the second posttest, the reader may assume, without
involved A3RE3 and A4RE4 relations, and inspecting the details of Table 2, that each
the third block involved all four ARE relations. student demonstrated mastery on each learning
A-D-E pretest and posttest. The following unit of each training phase.
transitive relations were tested: four DRE relations Prior to any training, all of the students scored
and four ERD relations. The test included four well below mastery for what would be the focus
trials of each relation type, for a total of 32 trials. All of Lesson 1 (A-B-C) or Lesson 2 (A-D-E). Data
four comparison stimuli appeared on each test trial on the left side of the phase line of Figure 5
(e.g., for DRE relations the comparison stimuli summarizes pretest results. Mean correct student
always were E1, E2, E3, and E4). responding on the A-D-E pretest was 50% and
slightly higher on the A-B-C pretest. Subse-
Class Merger Test: B-C-D-E quently, as Table 2 shows, all students complet-
Lesson 1 established A-B-C relations, and ed Lesson 1 A-B-C training in 121 to 198 trials,
Lesson 2 established A-D-E relations. If these including any remediation (see below); this
two stimulus classes merged because of a shared required 10 to 13.5 min of engagement. On
member (A), then without explicit training the the ensuing test of emergent relations (data on
students would also be competent with novel B- the right side of the phase line of Figure 5)
C-D-E relations. Figure 4 (bottom) shows the involving the B and C stimuli, Students 1 and 3
transitive associations that were expected to demonstrated mastery. Students 2 and 4 met the
emerge without direct training. Note that mastery criterion after failing a first attempt at
previous tests evaluated relations between the the A-B-C posttest and repeating A-B-C train-
B and C stimuli and between the D and E ing. At this time, no improvements were evident
stimuli. Thus, the novel relations on the class in A-D-E relations (Figure 5, middle), and no
9. TEACHING ABOUT THE BRAIN 27
Figure 5. Performance on tests for taught (top and middle) and emergent relations (bottom).
10. 28 DANIEL M. FIENUP et al.
Table 2
Requirements to Meet Mastery Criteria
A1RB1, A1RC1, A3RC3,
Student A2RB2 A3RB3, A4RB4 All ARB A2RC2 A4RC4 All ARC Total
Lesson 1 (A-B-C): number of trials
1 12 22 91 12 26 35 198
2 13 (12) 21 (22) 12 (12) 16 (12) 12 (12) 24 (12) 180
3 12 12 60 12 13 12 121
4 19 (26) 27 (12) 13 (12) 13 (12) 12 (17) 12 (12) 187
Lesson 1 (A-B-C): time of engagement (s)
1 34 50 431 37 55 196 803
2 50 (38) 69 (88) 43 (47) 48 (34) 31 (38) 95 (39) 620
3 92 39 318 60 55 44 608
4 53 (64) 66 (37) 64 (73) 65 (48) 39 (51) 60 (80) 700
A1RD1, A3RD3, A1RE1, A3RE3,
Student A2RD2 A4RD4 All ARD A2RE2 A4RE4 All ARE Total
Lesson 2 (A-D-E): number of trials
1 21 15 14 17 14 12 93
2 12 12 12 12 12 12 72
3 13 13 12 13 12 12 75
4 17 13 12 12 12 21 87
Lesson 2 (A-D-E): time of engagement (s)
1 32 25 31 43 34 33 198
2 31 34 22 26 36 30 179
3 28 37 21 43 27 34 190
4 21 26 27 31 42 66 213
Note. Numbers in parentheses represent remedial training that was required after a student failed a posttest. In such
cases, the Total column includes both iterations of the training phase.
student showed substantial competence with B- each of four stimulus sets (each related to a
C-D-E relations (Figure 5, bottom). brain lobe). The study’s critical contribution
Next, all of the students completed Lesson 2 regards the number of relations that that
training (Table 2) in 72 to 93 trials (about 2 to emerged without explicit training. Training
3.5 min of engagement). On the ensuing test of promoted the emergence of four symmetrical
emergent relations involving the A-D-E stimuli, relations (BRA, CRA, DRA, and ERA) that
all students demonstrated mastery; thus, unlike were not explicitly tested but theoretically must
in Lesson 1, no remedial training was needed. be in place for transitive relations to emerge (see
At this time all students also demonstrated that Sidman, 1994) plus six transitive relations
they had retained the benefits of Lesson 1 (BRC, BRD, BRE, CRD, CRE, and
(Figure 5, rightmost columns, top), and that DRE) and their symmetrical variants (CRB,
the stimulus classes of Lessons 1 and 2 had DRB, ERB, DRC, ERC, and ERD). If
merged to create additional B-C-D-E emergent these symmetrical variants of transitive relations
relations (Figure 5, rightmost columns, bot- are treated as independent relations, as is the
tom). Accuracy was near 100% in all cases. norm in basic research and in Sidman’s (1994)
stimulus equivalence theory, then the teaching
of 16 relations (four for each of four brain-lobe
DISCUSSION classes) supported the emergence of 64 addi-
Results in Practical Context tional untaught relations (those just listed for
To summarize the training, each of the each of four brain-lobe classes) or 80 total
computerized lessons taught two relations for relations. Thus, students learned about five
11. TEACHING ABOUT THE BRAIN 29
times as many relations as were expressly taught. relations that emerged among the stimuli of
In a more conservative accounting, in which the two separate lessons depended on the
each symmetrical pair of relations is considered training of both lessons. Together, these
as one relation, the teaching of 16 relations findings support a maxim that Skinner (1968)
yielded 24 emergent relations (40 total rela- advanced long ago but bears repeating: In
tions), in which case students learned about two building complex repertoires, it is important
and a half times as many relations as were not to skip steps.
expressly taught. By either perspective, students Although the students showed clear academic
were able to ‘‘go beyond the information given’’ gains, including emergent ones, it is important
(Bruner, 1957, p. 41) in precisely the ways that to note that that this occurred in match-to-
stimulus equivalence theory predicts. EBI thus sample trials that emphasized selection-based
delivered the generative responding that consti- responding. By contrast, many academic assign-
tutes its major promise to education (Stromer et ments in higher education require topography-
al., 1992). Given the technical nature of brain based responding (e.g., this is a critical
function and anatomy as an academic subject difference between multiple-choice and essay
(Wilson et al., 2000), this was accomplished assignments), as do many professional tasks
with surprisingly little student investment (e.g., beyond higher education (e.g., writing a
about 13 to 17 min of engagement with the two research report to submit for publication).
lessons, excluding instructions, tests, and time Because topography-based responding was not
in transition between activities). assessed in the present study (or in the
To place the effectiveness of the lessons into a previously mentioned studies by Ninness et
practical context, consider that the mastery al., 2005, 2006), it is not known whether the
criterion defined student success during training repertoires established in selection-based proce-
phases as 100% correct, which would count as a dures would contribute to student success on
letter grade of A on any academic grading scale. topography-based tasks. This defines an impor-
Most likely this success was facilitated by generic tant direction for future EBI research.
features of good behavioral instruction, such as Conceptual Issues
frequent student responding, frequent feedback,
Class merger. The present study joins only a
and individualized progress through the lessons few others (e.g., Lane & Critchfield, 1998a) in
(e.g., Keller, 1968; Skinner, 1968). On tests, suggesting that spontaneous class merger,
which included only relations that had not been promoted by classes that share a common
directly taught in the lessons, students had to member, may be academically beneficial. When
score at least 90% correct (which would earn an two classes unite, emergent relations among
A on most college grading scales) to move on to their respective members become possible. Our
the next task. That they did so reliably study does not provide unequivocal evidence of
underscores the novel (generative responding) class merger, however, because strictly speaking
contribution of EBI to behavioral instruction. the term merger implies that two classes have
The multiple probes of the present experi- existed independently prior to their union
mental design were helpful in highlighting two (Sidman, 1994). In the present study, the
outcomes of practical interest to the instruc- stimuli of Lessons 1 and 2 included a common
tional designer. First, the relations that emerged member from the outset, so Lesson 2 might be
among the stimuli of a given lesson were described more conservatively as promoting the
dependent specifically on the training of that expansion of the classes that were created during
lesson (i.e., Lesson 2 abilities did not improve Lesson 1. To demonstrate class merger unam-
until after Lesson 2 training). Second, the biguously would require the creation of unre-
12. 30 DANIEL M. FIENUP et al.
lated classes during initial lessons, followed by instruction based on equivalence relations, and
an additional lesson during which the shared it is worth clarifying what this assertion does
associate was taught. The underlying distinction and does not imply. Methodologically speaking,
is important conceptually but possibly irrelevant students responded to the academic stimuli in a
to the instructional designer whose primary goal fashion that was consistent with Sidman’s
is to engineer new repertoires as efficiently as (1994) theoretical account of equivalence. Most
possible. notably, the students transitively matched
The role of verbal rules. Although EBI has dissimilar stimuli, that is, they treated the
been used to build new skills economically in stimuli as interchangeable to the extent that
people with learning difficulties and disorders, the instructional procedures allowed. This is not
few studies have explored the utility of this kind to say, however, that they would treat the same
of instruction for building high-level academic stimuli as interchangeable under all circum-
skills in typically developing individuals. Previ- stances. For example, few reading-capable
ous studies (Ninness et al., 2005, 2006) focused individuals would, in the abstract, judge that
on mathematics skills; the present study extends frontal lobe and damage causes impulsiveness
the generality of EBI to brain–behavior rela- ‘‘mean the same thing,’’ and many situations
tions. Unlike the studies by Ninness et al., may be imagined in which these phrases
however, this one was designed to evaluate EBI would be called ‘‘different.’’ It is important to
independently of instructor-generated verbal note that stimulus equivalence theory recogniz-
rules governing stimulus relations that are es the potential for contextual control in which
typical employed in a classroom environment. stimuli are equivalent in some circumstances
Specifically, Ninness et al. taught algebra skills but not others (Sidman). For example, dober-
via a treatment package that included explicit man may be said to ‘‘go with’’ poodle and
instructor descriptions of how various stimuli bichon in a discussion about the species canus
were related as well as relevant match-to-sample but not in a discussion about specific dog
training (involving overlapping conditional breeds. Similarly, no theoretical contradiction
discriminations, performance feedback, and exists in asserting that stimuli like frontal lobe
mastery-level training). The lessons were effec- and damage causes impulsiveness functioned as
tive, but the studies that evaluated them left equivalent within the confines of the present
unclear the relative contribution to student lessons but are not unconditionally inter-
mastery of instructor-generated explanations changeable.
and student practice with the component Communicating about academic stimulus rela-
relations of stimulus classes. The current study tions. The preceding discussion highlights two
is valuable in demonstrating that it is not always important points. The first point is that it is
necessary to provide college students with rules difficult to communicate about stimulus equiv-
about stimulus relations in order to use EBI alence to a nonspecialist audience, which
successfully in teaching advanced material. Of includes many readers of this journal. Lay
course, given that the students were verbally euphemisms (e.g., ‘‘means the same thing’’)
capable, they could have generated their own often are invoked in an attempt to ease the
rules. If so (our experiment was not designed to exposition because, on first blush, these seem to
evaluate this possibility), then this may be capture the general idea of stimulus equivalence.
viewed as another way in which EBI promotes Yet such expressions may obscure important
‘‘going beyond the information given.’’ nuances of stimulus relations. Nothing in the
Is this really stimulus equivalence? We have present lessons indicated to students that frontal
described the present lessons as exemplifying lobe ‘‘means the same thing’’ as damage causes
13. TEACHING ABOUT THE BRAIN 31
impulsiveness; and, as noted above, stimulus be mastered in even an introductory course this
equivalence theory does not make exactly this topic (e.g., see Carlson, 2005), and it could be
claim. Thus, lay vocabulary may carry unwant- argued that resulting emergent relations did not
ed conceptual baggage. As EBI interventions are really take students very far ‘‘beyond the
developed for use outside of research settings, information given’’ (Bruner, 1957, p. 41).
the need to communicate about them to Moreover, the ‘‘information given’’ was, as is
nonspecialists will increase, and it is an open often the case in introductory lessons, not very
question as to whether the expositional benefits precise. Brain regions are interconnected and
of lay expressions outweigh the potential for most psychological functions in which they
conceptual confusion that they introduce. participate are distributed across several regions
The second point worth stressing emphasizes (Carlson, 2005). Thus, the statements that
a different way in which the learning stimuli served as this study’s B, C, and E stimuli
may not ‘‘mean the same thing.’’ To illustrate (Figure 1) are oversimplified. The present
from the present study, note that damage causes lessons are best appreciated, therefore, as an
impulsiveness is not a synonym of frontal lobe example of how to establish selected aspects of
but rather a property of it. As relational frame an entry-level higher education repertoire
theory (Hayes, Barnes-Holmes, & Roche, regarding brain structure and function. Sub-
2001) teaches, many types of relations (e.g.,
ject-matter expertise would require considerable
opposites, superordinate-subordinate, etc.) can
additional instruction. To concede this point
spawn emergent abilities just as equivalence
does not contradict our assertion that the
relations do. Thus, many types of stimulus
subject matter was more advanced than is
relations could serve as the basis for instruc-
normally seen in EBI investigations (consider,
tional programming (e.g., Ninness et al., 2005,
e.g., the challenges that might be encountered
2006). In the present article, our reliance on the
language of stimulus equivalence was driven in teaching the same material to very young
mainly by practical rather than theoretical children or persons with developmental disabil-
concerns: This language mapped conveniently ities).
onto the instructional material, and it is It should be noted as well that the learners in
somewhat familiar to the journal’s audience. the present investigation were research volun-
There is no question, however, that the present teers whose learning was not linked to an
study could be interpreted within the frame- academically sanctioned course of instruction.
work of relational frame theory, and given The study was intended to mimic college
the diversity of stimulus relations that aca- instruction in important ways (the learners were
demic curricula probably incorporate, a well- similar to students who typically would en-
elaborated program of instruction could profit counter the topic of instruction, and the lessons
from drawing from the resources of both were administered to several students simulta-
relational frame theory and stimulus equiva- neously in a computer-equipped classroom).
lence theory. Nevertheless, like all other published studies of
EBI to date, the present one qualifies as a
Limitations and Future Directions feasibility evaluation; that is, it demonstrated
The present lessons were neither a complete learning-economy benefits under well-con-
program of instruction nor likely to promote trolled conditions. Whether similar lessons can
sophisticated understanding of the targeted have a significant impact on student progress
concepts. In the former case, the lessons taught through a course of instruction in a natural
only a few facts about brain anatomy and setting remains to be determined. From an
function from among the multitude that must evidence-based practice perspective (Chorpita,
14. 32 DANIEL M. FIENUP et al.
2003), the empirical evaluation of an interven- Fields, L., Travis, R., Roy, D., Yadlovker, E., de Aguiar-
Rocha, L., & Sturmey, P. (2009). Equivalence class
tion considers two broad features: its efficacy formation: A method for teaching statistical interac-
(beneficial effects under controlled conditions) tions. Journal of Applied Behavior Analysis, 42,
and its transportability (beneficial effects when 575–593.
Fienup, D. M., & Dixon, M. R. (2006). Acquisition and
extended to specific practice settings). The
maintenance of visual-visual and visual-olfactory
present study addresses efficacy. With respect equivalence classes. European Journal of Behavior
to transportability, field studies, in which EBI is Analysis, 7, 87–98.
evaluated as a part of an ongoing instructional Green, G., & Saunders, R. R. (1998). Stimulus
equivalence. In K. A. Lattal & M. Perone (Eds.),
system, are needed to determine whether Handbook of research methods in human operant
academic benefits that accrue in laboratory-like behavior (pp. 229–262). New York: Plenum.
environments also can be engineered reliably Hayes, S. C., Barnes-Holmes, D., & Roche, B. (Eds.).
(2001). Relational frame theory: A post-Skinnerian
under less ideal everyday circumstances. account of human language and cognition. New York:
Kluwer Academic/Plenum.
REFERENCES Johnston, J. M., & Pennypacker, H. S. (1980). Strategies
and tactics of human behavioral research. Hillsdale, NJ:
Bruner, J. S. (1957). Going beyond the information given. Erlbaum.
In J. S. Bruner, E. Brunswik, L. Festinger, F. Heider, Joyce, B. G., & Joyce, J. H. (1993). Using stimulus
K. F. Muenzinger, C. E. Osgood, et al. (Eds.), equivalence procedures to teach relationships between
Contemporary approaches to cognition (pp. 41–69). English and Spanish words. Education & Treatment of
Cambridge, MA: Harvard University Press. Children, 16, 48–65.
Carlson, N. R. (2005). Foundations of physiological Keller, F. S. (1968). Good-bye, teacher. Journal of Applied
psychology (6th ed.). Boston: Pearson Allyn & Bacon. Behavior Analysis, 1, 79–89.
Chew, S. L. (2008). Study more! Study harder! Students’ Lane, S. D., & Critchfield, T. S. (1998a). Classification of
and teachers’ faulty beliefs about how people learn. In vowels and consonants by individuals with moderate
S. A. Meyers & J. R. Stowell (Eds.), Essays from mental retardation: Development of arbitrary rela-
excellence in teaching. (Vol. 7, pp. 22–25). Retrieved tions via match-to-sample training with compounds.
from http://teachpsych.org/resources/e-books/eit2007/ Journal of Applied Behavior Analysis, 31, 21–41.
eit2007.php Lane, S. D., & Critchfield, T. S. (1998b). Increasing the
Chorpita, B. F. (2003). The frontier of evidence-based generativity of identity-based procedures for estab-
practice. In A. E. Kazdin & J. R. Weisz (Eds.), lishing arbitrary conditional relations. The Psycholog-
Evidence-based psychotherapies for children and adoles- ical Record, 48, 457–520.
cents (pp. 42–59). New York: Guilford. Lynch, D. C., & Cuvo, A. J. (1995). Stimulus equivalence
Connell, J. E., & Witt, J. C. (2004). Applications of instruction of fraction-decimal relations. Journal of
computer-based instruction: Using specialized soft- Applied Behavior Analysis, 28, 115–126.
ware to aid letter-name and letter-sound recognition. Ninness, C., Barnes-Holmes, D., Rumph, R., McCullen,
Journal of Applied Behavior Analysis, 37, 67–71. G., Ford, A. M., Payne, R., et al. (2006). Transfor-
Critchfield, T. S., & Fienup, D. M. (2008). Stimulus mation of mathematical and stimulus functions.
equivalence. In S. F. Davis & W. F. Buskist (Eds.), Journal of Applied Behavior Analysis, 39, 299–321.
21st century psychology (pp. 360–372). Thousand Ninness, C., Rumph, R., McCullen, G., Harrison, C.,
Oaks, CA: Sage. Ford, A. M., & Ninness, S. K. (2005). Functional
de Rose, J. C., de Souza, D. G., & Hanna, E. S. (1996). analytic approach to computer-interactive mathemat-
Teaching reading and spelling: Exclusion and stim- ics. Journal of Applied Behavior Analysis, 38, 1–22.
ulus equivalence. Journal of Applied Behavior Analysis, ´ ´ ´
Perez-Gonzalez, L. A., Garcıa-Asenjo, L., Williams, G., &
29, 451–469. Carnerero, J. J. (2007). Emergence of intraverbal
Dixon, M. R., & McLin, O. H. (2003). Visual Basic antonyms in children with pervasive developmental
for behavioral psychologists. Reno, NV: Context disorder. Journal of Applied Behavior Analysis, 40,
Press. 697–701.
Ecott, C. L., & Critchfield, T. S. (2004). Noncontingent Rosales, R., & Rehfeldt, R. A. (2007). Contriving
reinforcement, alternative reinforcement, and the transitive conditioned establishing operations to
matching law: A laboratory demonstration. Journal establish derived manding skills in adults with severe
of Applied Behavior Analysis, 37, 249–265. developmental disabilities. Journal of Applied Behavior
Fields, L., & Reeve, K. F. (1997). Equivalence class Analysis, 40, 105–121.
formation using stimulus-pairing and yes-no respond- Sidman, M. (1994). Equivalence relations and behavior: A
ing. The Psychological Record, 47, 661–686. research story. Boston: Authors Cooperative.
15. TEACHING ABOUT THE BRAIN 33
Sidman, M., & Cresson, O. (1973). Reading and Wilson, C., Marcus, D. K., Hamilton, S. B., Knox, T. A.,
crossmodal transfer of stimulus equivalence in severe Reardon, R., Durso, F. T., et al. (2000). Teaching
retardation. American Journal of Mental Deficiency, neuroanatomy and neurophysiology. In M. E. Ware
77, 515–523. & D. E. Johnson (Eds.), Handbook of demonstrations
Sidman, M., Kirk, B., & Willson-Morris, M. (1985). Six- and activities in the teaching of psychology: Physiological-
member stimulus classes generated by conditional- comparative, perception, learning, cognitive, and devel-
discrimination procedures. Journal of the Experimental opmental (2nd ed.) (Vol. 2, pp. 9–25). Mahwah, NJ:
Analysis of Behavior, 43, 21–42. Erlbaum.
Skinner, B. F. (1968). The technology of teaching. New
York: Appleton-Century-Crofts.
Stromer, R., Mackay, H., & Stoddard, L. (1992).
Classroom applications of stimulus equivalence Received March 19, 2008
technology. Journal of Behavioral Education, 2, Final acceptance September 18, 2008
225–256. Action Editor, Chris Ninness