AEQ392763
AEQ61410.1177/0741713610392763Collins and PrattAdult Education Quarterly
The Teaching Perspectives
Inventory at 10 Years
and 100,000 Respondents:
Reliability and Validity
of a Teacher SelfReport Inventory
Adult Education Quarterly
61(4) 358­–375
© 2011 American Association for
Adult and Continuing Education
Reprints and permission: http://www.
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0741713610392763
http://aeq.sagepub.com
John B. Collins1 and Daniel D. Pratt1
Abstract
The Teaching Perspectives Inventory (TPI) measures teachers’ profiles on five
contrasting views of what it means “to teach.” The inventory can be used in
aiding self-reflection, developing statements of teaching philosophy, engendering
conversations about teaching, and recognizing legitimate variations on excellence
in teaching. Available at www.TeachingPerspectives.com, the TPI is a free, selfreport, self-scoring inventory that promotes a pluralistic understanding of teaching
and equips respondents with a more explicit vocabulary for reflecting on their
own teaching and that of others. Ten years of accumulated responses for more
than 100,000 respondents in more than 100 countries has provided a rich data
bank for analysis of the instrument’s reliability, validity, and utility in promoting
conversations about teaching that are respectful of disciplinary and professional
signature pedagogies as well as cultural and social variations on how teaching is
understood and valued.
Keywords
teaching perspectives, teaching philosophy, teaching styles
1
University of British Columbia, Vancouver, British Columbia, Canada
Corresponding Author:
John B. Collins, Department of Educational Studies, University of British Columbia,
Vancouver, British Columbia, Canada V6T 1Z4
Email: [email protected]
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
Collins and Pratt
359
Introduction
Increasingly, teachers at every level and in every context are being asked to articulate
and reflect on their approach to teaching. They do so for many reasons, some more
benign (e.g., as part of a workshop) and others more critical (e.g., as part of an evaluation). Few within the education community, whether educating youth or adults, argue
against this movement. Most simply presume it is a worthy and appropriate task,
perhaps assuming that it will provide better understanding and more equitable judgment of teaching.
At the same time, there is a move to adopt a single, dominant view of effective
teaching, usually one that is “learner centered” and based on a constructivist view of
knowledge and learning. In other words, teachers are asked to reflect on who they are
and how they teach but with an implied message that that reflection should conform to
some preconceived notion of a “good” teacher. In part, the argument for this move is
a reaction against teacher-centered instruction, which has dominated much of education, particularly adult and higher education, for much of the past century. Yet to argue
for a singularity of good in teaching while also asking people to reflect on their teaching implies a false promise of opportunity to be different from the dominant view of
teaching (Pratt, 2005a). It also contradicts a mounting body of evidence that effective
teaching depends on context (Pratt, Sadownik, & Jarvis-Selinger, 2011), discipline or
field of practice (Shulman, 2005), and culture (Tweed & Lehman, 2002). Clearly, one
size does not fit all.
Researching Perspectives on Teaching
Nearly three decades ago, Pratt and his graduate students began assembling evidence
through lengthy observations and interviews to learn how teachers in adult and higher
education conceptualize what it means “to teach.” They assembled transcripts and
observation notes for more than 250 teachers of adults on two continents and in dozens of locations across five different countries—the United States, Canada, Singapore,
China, and (then) Hong Kong. This work is summarized in their text, Five Perspectives
on Teaching in Adult and Higher Education (Pratt & Associates, 1998). Details of the
Five Perspectives on Teaching are available as an online supplement at http://aeq.
sagepub.com/supplemental. As the title suggests, they identified five distinctly different
views of what teachers of adult learners do and why. These they labeled Transmission,
Apprenticeship, Developmental, Nurturing, and Social Reform. In essence and substance, these perspectives are echoed in the work of Kember (1997), Apps (1996),
Brookfield (2006), and others, although sometimes by different names and descriptions. Furthermore, they soon learned that, in addition to variations in teaching perspectives, there were also differences in how commitment to teaching was expressed;
different teachers held different Beliefs about teaching, set themselves different
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
360
Adult Education Quarterly 61(4)
Intentions to accomplish, or undertook different Actions in their instructional settings.
From their work emerged a powerful heuristic for simplifying the myriad things
that can occur during any instructional event, a general model of teaching (Figure 1)
that abstracts a teaching session into five elements (teacher, learners, content, context, and
ideals) and three relationships (teacher↔learner, teacher↔content, and learner↔content).
Collectively, these eight features help researchers and practitioners organize and
classify narratives about how teachers differ in approach and justification of their
teaching.
The general model also provides a means by which adult educators can articulate
their approach to reflect meaningfully on their teaching and its possible improvement.
Importantly, the general model of teaching, and its constituent elements and relationships, presumes nothing about “effective teaching.” Nor does it suggest a causal
relationship between teaching and learning. Rather, it respects teaching as a personal activity that is socially mediated, culturally authorized, and historically situated. The model, therefore, respects adult and higher education practices wherever
they occur, by describing a set of elements and relationships that are neutral with
respect to the form and context of practice as well as the ends one wishes to achieve
through teaching.
Instrumenting Perspectives on Teaching
However, to arrive at such “thick descriptions” of how instructors conceptualized
their teaching required 2 to 3 hours of one-on-one interviews, observation, and subsequent analysis for each teacher. The obvious question was whether there might be
a more streamlined self-reporting option that could characterize teachers’ perspectives
faster and with less effort but with reasonable fidelity. Thus was born a decade-long
effort to develop a self-administering inventory (eventually online, self-scoring, and
with automated report-back) for respondents to typologize their individual profiles as
teachers. Pratt and his colleagues had collected hundreds of pages of transcripts,
observations, field notes, and other phenomenographic data during their interviews
with teachers in various academic, cultural, basic education, skills acquisition,
religious/spiritual, and leisure learning settings. Embedded in this data set were thousands of utterances reflecting teachers’ diversity of perspectives and commitments to
teaching. These transcripts and field notes became the initial source of “I statements”
reflecting each item’s focus on certain perspectives while differentiating it from others. Since the research team had first conjectured then ratified with each teacher what
their dominant perspectives were, it was a straightforward task to abstract characteristic utterances from teachers in each perspective and to cluster them into categories
believed to reflect Transmission, Apprenticeship, Developmental, Nurturing, and
Social Reform.
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
361
Collins and Pratt
Context
C
L
‘Ideals’
Z
T
Figure 1. A General Model of Teaching
From these data, we developed the following research questions for streamlining
that process:
Research Question 1: Can selected utterances be refined and restated such
that teachers’ endorsements of different statements reflect their dominant
teaching perspectives and distinguish them from nondominant or recessive
perspectives?
Research Question 2: Can such an inventory demonstrate acceptable standards
of reliability and validity?
From more than 200 of these original stems, items were reworded and refined to
characterize each perspective more expressly and to distinguish it from the remaining
perspectives. In successive refinements, these were pared to 155, then 120, and finally
to 75 items reflecting simultaneously balanced representations of Beliefs/Intentions/
Actions and learner–teacher, learner–content, and teacher–content relationships
(Chan, 1994). One-week test–retest sorting into categories by 10 experienced judges
demonstrated an 80% consistency. These 75 items yielded six clear factors (Apprenticeship
emerged as a “practice vs. modeling” distinction), which Chan subsequently reworked
as scale scores and tested against 471 adult night school teachers’ sociodemographic
characteristics and instructional practices.
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
362
Adult Education Quarterly 61(4)
After Chan’s analysis, the 75 items were revised once again and streamlined into a
45-item version with nine defining statements per perspective, each represented by
three Belief statements, three Intention items, and three Action declarations. This
45-item version was first tested in paper–pencil format on several hundred respondents (including almost the entire student cohort graduating with University of British
Columbia elementary, secondary, and adult teaching certificates in 2000) and examined for scale consistency (α reliability) and confirmatory factor structure. These
items, together with background information about sociodemographic attributes and
professional histories, were placed online as a self-administering, self-scoring instrument in late 2000. Streamlining of web page layout, response formats, and background
information questions continued into mid-2001. In its current format, each question is
a 5-point agree/disagree statement or a 5-point frequency report (never, rarely, sometimes, usually, always). Three sample items are shown in Table 1.
Implementation and Standardization
Following its launch, the Teaching Perspective Inventory (TPI) quickly gathered
momentum and an overall pattern of its respondents emerged. Part of its online
appeal was its instant feedback, which provided respondents with an immediate profile of relative strengths of their perspectives as well as an abbreviated summary
showing how their beliefs, intentions, and actions coalesced to differentiate among
their perspectives. Early reports of the inventory’s psychometric properties have been
noted elsewhere (Pratt & Collins, 2002) but the following sections outline much
larger scale results.
Going Live on the Web
In mid August of 2001, the current version of the TPI went online. The data bank’s
index case (August 17, 2001) is a Puerto Rican teacher/researcher whose subject
specialty was curriculum and instruction and whose learners were mature adults. By
the end of that same day, 3 more people had completed the inventory; within a week,
54; by the end of the first month, 342; and by the end of 2001, there had been 1,066
respondents. The response rate has steadily expanded such that in 2010 the daily
returns average just under 75 per day. The respondent count passed 50,000 in
September 2006 and exceeded 100,000 in April 2009. Although the largest fraction of
respondents are located in the United States (45%) and Canada (23%), there are currently more than 38,000 completed profiles from more than 120 other countries
around the world. Furthermore, although most people reported English as their first
language (78%), substantial numbers also reported Spanish (8.1%), French (1.3%),
Mandarin (1.6%), Cantonese (0.8%), and Hindi, Punjabi, and more than 100 “other”
(9.9%) languages.
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
363
Collins and Pratt
Table 1. Sample TPI Items
Teaching should build on what people already know.
I expect people to become committed to changing society.
I encourage people to challenge each other’s thinking.
SD D N A SA (Beliefs)
N R S U A (Intentions)
N R S U A (Actions)
Note: TPI = Teaching Perspective Inventory; Beliefs: SD, D, N, A, SA = strongly disagree, disagree,
neutral, agree, and strongly agree, respectively; Intentions and Actions: N, R, S, U, A = never, rarely,
sometimes, usually, and always, respectively.
Who Takes the TPI?
The TPI was designed principally with teachers in adult and higher education in mind,
but respondents to the inventory span the spectrum of educational roles. About half
report their primary role as teacher (47.5%), administrator/manager (8.9%), practitioner (11.6%), researcher (3.5%), student (17.7%), or “other” (10.9%). Their instructional responsibilities include learning groups ranging from young adults (44.9%) to
mature adults (17.1%), youth (17.7%), children (18.6%), older adults (1.3%), and
seniors (0.3%); thus, teachers of school-age children form about a third of the respondents. Overall, substantially more women have completed the TPI (65.7%) than men
(34.3%). The largest fraction of respondents report that their “usual learners” are
university undergraduates (22.2%); followed by graduate students (8.3%); community
college, vocational, or adult night school learners (14.4%); secondary school learners
(16.6%); elementary school learners (15.9%); and a variety of others (general public,
workshops, conferences; 8.9%). Most teach in settings that grant university or college
credit (34.9%), K-12 credit (22.4%), or certificate or diploma recognition (15.4%),
whereas 27.4% teach in noncredit learning situations. Respondents’ educational
qualifications span the spectrum from high school diplomas (12.4%), through bachelor’s (39.0%), master’s (26.4%), to doctoral degrees (11.9%). “Other” designations in
education (9.4%) often include specialty qualifications, current enrollment in masters
or doctoral programs, or dual qualifications (e.g., MD plus PhD). Although about 14%
(mostly students) are at the beginning of their teaching careers, most respondents have
substantial histories as educators (M = 7.9 years) or practitioners (M = 8.4 years) and
report that teaching responsibilities consume a majority of their professional responsibilities (55.0%). The large majority report an employment home base in the educational sector (69.5%), whereas others are located in health (13.0%), government
(4.5%), business (4.2%), volunteer (1.5%), and “other” (7.3%) sectors. Just more than
half report that they were prompted to complete the TPI as a course requirement
(52.7%). The next highest group took the TPI on recommendation from a colleague
or friend (19.0%). The remaining respondents report that it was a discovery while
searching the web (8.3%), a career search (6.8%), or some “other” (13.4%) reason for
taking the TPI.
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
364
Adult Education Quarterly 61(4)
Reporting Results to Respondents
The TPI website automatically scores each person’s answer pattern and provides an
instant graphic profile of their scores (Figure 2), which they can print, along with five
“Explanatory Paragraphs” describing the essence of each perspective as well as a
10-step interpretation guide.
Respondents’ results are also automatically e-mailed to the address they supply
while completing the questions. The profile sheet consists of five vertical bars–one for
each perspective—running from a theoretical minimum of 9 to a maximum of 45. At
the top of each bar is printed each respondent’s numerical totals as well as the three
commitment subscore components (Beliefs, Intentions, and Actions) that constitute
the total.
Teaching Perspectives
Although scores for each perspective could vary from 9 to 45, suggesting a midrange
average of 27, in practice, the scale means average around 34, suggesting an overall
upward bias of about 7 points on the response scales (SD, D, N, A, SA [strongly disagree, disagree, neutral, agree, strongly agree] or N, R, S, O, A [never, rarely, sometimes, often, always]). Individual means, with standard deviations in parentheses,
vary from scale to scale. Transmission averaged 33.1 (4.6), Apprenticeship 36.2 (4.2),
Developmental 34.5 (4.3), Nurturing 36.7 (5.0), and Social Reform 28.8 (6.0), indicating that items on the Nurturing or Apprenticeship scales are generally more appealing
(or socially desirable) than Social Reform or Transmission items. Table 2A details
these means, standard deviations, and other commonly accepted parameters of scale
performance.
Variations on Commitment
Each perspective comprises three manifestations of “commitment to teaching”:
Beliefs, Intentions, and Actions. In Pratt’s foundational conversations with teachers,
these expressions of commitment most often emerged early in the interviews (“Well,
what I really believe about teaching is . . .,” “What I’m actually trying to accomplish
with my learners is . . .,” or “What I usually do in the classroom is . . .”), followed later
by further disclosures indicating preferences for a Transmission, Apprenticeship,
Developmental, Nurturing, or Social Reform perspective. Each perspective score is a
nine-item sum of three Belief items, three Intention items, and three Action items
pertinent to that perspective. Thus, in addition to the five perspective scores, people’s
Belief scores can be aggregated across perspectives as can their Intention scores and
their Action scores. In theory, these Commitment scores can range from 15 to 75 and
would suggest a midrange average of about 45. In practice, they average between 55
and 56; Beliefs average 55.5 (6.2), Intentions 56.9 (6.9), and Actions 56.9 (6.7). An
inventory total across all 45 items could range between 45 and 225, but averages 169.3
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
365
Collins and Pratt
Teaching Perspectives Profile:Individual
Transmission
Tr: 33
Apprenticeship
Ap: 37
Developmental
Dv: 26
Respondent: Sample
Nurturing
Nu: 31
Social Reform
SR: 24
B:12, I:9, A:12
B:11, I:13, A:13
B:11, I:8, A:7
B:11, I:10, A:10
B:9, I:8, A:7
45
45
45
45
45
44
44
44
44
44
43
43
43
43
43
42
42
42
42
42
41
41
41
41
41
40
40
40
40
40
39
39
39
39
39
38
38
38
38
38
37
• 37 •
• 36 •
37
37
37
36
36
36
36
#8
TPI ID Number: 101203015147
Your scores at or above this line (35) are your DOMINANT perspective(s).
35
34
• 33 •
• 32 •
• 31 •
• 30 •
• 29 •
• 28 •
• 27 •
• 26 •
• 25 •
• 35 •
• 34 •
• 33 •
• 32 •
• 31 •
• 30 •
• 29 •
• 28 •
• 27 •
• 26 •
• 25 •
35
35
35
34
34
34
33
33
33
32
32
32
31
• 31 •
• 30 •
• 29 •
• 28 •
• 27 •
• 26 •
• 25 •
31
30
29
28
27
• 26 •
• 25 •
30
29
28
27
26
25
Your scores at or below this line (25) are your RECESSIVE perspective(s).
• 24 •
• 23 •
• 22 •
• 21 •
• 20 •
• 19 •
• 18 •
• 17 •
• 16 •
• 15 •
• 14 •
• 13 •
• 12 •
• 11 •
• 10 •
•9•
Transmission
• 24 •
• 23 •
• 22 •
• 21 •
• 20 •
• 19 •
• 18 •
• 17 •
• 16 •
• 15 •
• 14 •
• 13 •
• 12 •
• 11 •
• 10 •
•9•
Apprenticeship
• 24 •
• 23 •
• 22 •
• 21 •
• 20 •
• 19 •
• 18 •
• 17 •
• 16 •
• 15 •
• 14 •
• 13 •
• 12 •
• 11 •
• 10 •
•9•
• 24 •
• 23 •
• 22 •
• 21 •
• 20 •
• 19 •
• 18 •
• 17 •
• 16 •
• 15 •
• 14 •
• 13 •
• 12 •
• 11 •
• 10 •
•9•
Developmental
© Pratt & Collins 2001, Web Version: 2.0, August 2001
Nurturing
• 24 •
• 23 •
• 22 •
• 21 •
• 20 •
• 19 •
• 18 •
• 17 •
• 16 •
• 15 •
• 14 •
• 13 •
• 12 •
• 11 •
• 10 •
•9•
Social Reform
A Scoring Profile for the Teaching Perspectives Inventory
Figure 2. Sample profile
(16.6). Table 2A shows these Commitment parameters along with the whole-scale
total. Evidence that these three Commitment scores can be summed to yield perspectives scores is confirmed by their reasonably high overall collinearity (α = .77).
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
366
Adult Education Quarterly 61(4)
Dominant, Backup, and Recessive Perspectives
In most instances, people’s TPI profile (see Figure 1) charts show a “‘stepped” configuration with one perspective or sometimes two perspectives standing out notably
higher than the remainder. These prominent perspectives are those that teachers often
describe as “Where I’m most ‘at-home’” or “How I most often see myself.” These are
labeled one’s Dominant perspective. Similarly, one score is often notably lower than the
remaining four and is termed Recessive. On each respondent’s profile, two horizontal
bars indicate the boundaries at or beyond which a perspective is either Dominant or
Recessive. Numerically, these are computed as plus or minus one standard deviation
around the mean of each respondent’s five perspective scores; hence the spread separating dominant from recessive is individualized for each person. Between the two bars are
one’s backup perspectives—skills and strategies that can be called on when needed but
that are not always at the forefront of one’s instructional tool kit.
Among the 116,621 teachers in this sample, nurturing is the most common dominant perspective (50%), followed by Apprenticeship (38%), Developmental (18%),
Transmission (14%), and Social Reform (3%). Nearly a quarter of the respondents
have two (and rarely three) dominant perspectives, so these percentages total to more
than 100% (see Table 2A).
Psychometric Stability of the TPI
The TPI exhibits overall the psychometric precision and stability that most researchers would interpret as at least “satisfactory” if not “good.” Traditional measures of
reliability, validity (both convergent and discriminant), and sensitivity to expected
criterion group differences suggest that the TPI operationalizes and measures those
perspectives initially identified qualitatively among Pratt’s large sample of teachers
of adults.
Reliability
Interitem convergence. Scale coherence and stability are required hallmarks of all
multi-item inventories. Cronbach’s α reliabilities for respondents range from .70
(Developmental) to .83 (Social Reform) and average .76 across the five scales (see
Table 2B). All meet or exceed Nunnally and Bernstein’s (1994) α = .70 benchmark.
Interscale correlations range from .15 between Transmission and Nurturing to .58
between the Apprenticeship and Developmental scales and averaged .41 among the
five scales combined. Together, these results indicate reasonable convergent (withinscale) and discriminant (between-scale) inventory properties (Campbell & Fiske,
1959). Furthermore, none of the 45 items correlates nearly as highly on any adjacent
scale as with the scale of which it is part. Similarly, α reliabilities for the TPI’s commitment components are .72 for Beliefs, .78 for Intentions, and .80 for Actions,
averaging .77.
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
367
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
36.2
4.2
34-39
9-45
9
38.3
.62
.75
.73
.23
.67
.66
.46
8.9
.04
−.06
.08
.05
.03
−.03
.69
.75
.72
.22
.57
.90
.46
9.9
−.02
−.06
.03
.02
−.06
−.09
Ap
33.1
4.6
30-36
9-45
9
13.6
Tr
.05
.11
.04
.12
−.03
.04
.46
6.5
.21
.58
.77
.66
.62
.70
34.5
4.3
32-38
9-45
9
17.8
Dv
.21
−.21
−.06
−.03
−.15
.05
.49
5.4
.34
.54
.94
.69
.65
.82
36.6
5.0
34-40
9-45
9
50.5
Nu
.04
−.05
.01
.05
−.12
−.21
.49
5.0
.35
.62
.88
.71
.87
.83
28.8
6.0
25-33
9-45
9
2.7
SR
Note: TPI = Teaching Perspectives Inventory; Tr = Transmission; Ap = Apprenticeship; Dv = Developmental;
Nu = Nurturing; SR = Social Reform; Bel = Beliefs; Int = Intentions; Act = Actions;
SIQR = semi-interquartile range; CFV = common factor variance.
A: Parameters
Mean
SD
SIQR
Minimum-maximum
n (Items)
% Dominant
B: Reliability
Test–retest(1-2)
Test–retest(2-3)
α Reliability
C: Convergent validity Item intercorrelation
Mean factor loading
Scale/factor intercorrelation
D: Discriminant validity Highest cross-scale correlation
% CFV accounted for
E: Criterion group validity (bias) Gender
Education
Role: (student/teacher)
Experience (years)
Instruction level
First language
Perspectives
.05
−.05
.06
.06
−.09
−.18
—
—
.14
—
—
.63
.72
.72
56.9
6.7
52-59
15-75
15
—
Bel
.07
−.11
.01
.02
−.08
−.03
—
—
.19
—
—
.65
.79
.78
56.9
6.9
53-61
15-75
15
—
Int
Commitment
Table 2. Psychometric Properties of the Five TPI Perspectives Scales and Three Commitment Subscales
.12
−.08
.01
.04
−.09
−.01
—
—
.21
—
—
.68
.74
.80
55.5
6.2
53-61
15-75
15
—
Act
.10
−.06
.05
.05
−.10
−.08
—
35.6%
.15
—
—
.69
.79
.89
169.3
16.6
159-180
49-224
45
—
Total
116,621
116,621
76,027
115,470
116,621
116,621
116,621
116,621
116,621
116,621
67,028
116,621
116,621
116,621
116,621
—
116,621
516
63
116,621
Number
368
Adult Education Quarterly 61(4)
Test–retest reliability. Many respondents take the TPI more than once, sometimes
later the same day, sometimes a few weeks apart, and occasionally at 1- or 2-year
intervals “to see if much has changed,” as they often explain. Table 2B shows a selection of these first-to-second TPI results for more than 500 respondents indicating test–
retest reliabilities averaging .67 and ranging from .62 for Developmental to .71 for
Social Reform, indicating passable stability over time. A further subset of second-tothird administrations for a sample of 63 people showed even greater stability, averaging .73 and ranging from .65 for Nurturing to .87 for Social Reform.
Validity
Many forms of validity exist; some of the more prominent are internal, external, face,
concurrent, predictive, criterion group, construct, and consequential (Messick, 1989).
A variety of legitimate approaches exist to examine each form of validity. Since the
TPI is the first operationalization of this cluster of teaching perspectives, there are no
preexisting instruments to serve as external benchmarks. Hence, validity analyses
must rely largely on internal or consequential approaches.
Face validity. Early in the item refinement stages of the TPI’s development, a group
of about 75 senior graduate students who were familiar with the conceptual underpinnings distinguishing the five perspectives were assigned the task of sorting stacks of
potential item statements printed one to a card into five clusters representing one or
another of the five perspectives (Pratt, 1998). That they were able to accomplish this
with about 95% accuracy suggested that, even at that early stage, the items were reasonable representations (and operationalizations) of the perspectives underlying them.
In subsequent workshops or face-to-face meetings with teachers who have taken the
TPI, it is a common occurrence for people to exclaim, “Oh yes, this is me” or for colleagues who know each others’ teaching styles to share their profiles and to comment,
“Indeed this does capture you” or (on more than a few occasions) to laugh and say,
“Wow! This has got you dead to rights.” Thus, there is gathering evidence—both
anecdotal and systematic—for the face validity of the TPI’s items and scale results.
Internal validity. Factorial approaches are convenient because (a) they examine the
foundations on which the perspectives were derived and (b) they illuminate the contribution that each item makes to its respective perspective. In brief, factorial studies
confirm that (1) five factors is the optimal number, (2) each rotated factor accounts for
roughly the same fraction of variance (5.0% to 9.9%), (3) each item is correctly
assigned to its proper scale, (4) factor scores correlate highly with scale scores (r ~ .83),
and (5) each item is more saturated with its perspective identity than with any of the
commitment aspects (Beliefs, Intentions, or Actions).
The perspectives were originally intuitively and inductively derived from interviews with 253 practicing teachers, and the “germ” of each TPI item originated in their
own statements (“a” above). The notion of commitment (Beliefs, Intentions, and
Actions) often emerged first in the interviews, followed by further clarification as to
what their intentions were and how they justified their teaching strategies. Nevertheless,
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
Collins and Pratt
369
the factorial emergence and structure of the TPI’s current 45 items is organized around
perspective differentiation rather than commitment aspects. This indicates that their
own conceptual understanding of what it means “to teach” is structured primarily
around distinctions among perspectives and secondarily around expressions of their
commitment, even though those two components had emerged in reverse order in their
narratives.
As always, some items contribute more to the total variance accounted for and to
the definitions of the respective factors (“b” above) than others do. No items had communalities less than .30 indicating that all 45 items contributed meaningfully to the
defining of one or another of the perspectives. Principal component analysis yielded
nine factors with eigenvalues greater than 1.00 and accounting for 50.7% of the common factor variance. Extractions of nine, eight, seven, six, five, and four factors were
examined, and several rotational strategies (both oblique and orthogonal) were tested
over the course of inventory development. The most satisfactory was a quartimax rotation incorporating all 45 items into one or another of five factors accounting for 39.7%
of the common factor variance. Items loading on more than one factor were virtually
nonexistent. Nevertheless TPI scores are scale scores, not factor scores. Evidence that
the scales are faithful to their underlying factors is clear in the correlations (see Table
2C) between scale and factor scores, averaging .83. Specifically, the correlations of
each perspective’s scale with factor scores are Transmission .90, Apprenticeship .66,
Developmental .77, Nurturing .94, and Social Reform .88. Finally, the fact that each
TPI item correlates much higher with its parent scale (rs in the range of .6 to .8) than
with any adjacent scale (rs in the range of .1 to .3) confirms that scale items converge
on their respective underlying concepts but diverge from the latent continua of the
other perspectives since the rotated factors are both orthogonal and roughly equal
sized (see Table 2D).
Predictive approaches confirm that teacher experience differentially predicts the
Developmental perspective (r = .13), whereas practitioner experience predicts
Apprenticeship (r = .14), and gender predicts Nurturing (r = .20). Greater fractions of
instructional (as opposed to administrative) responsibility predicts stronger endorsement of all TPI items (r = .14). The more “seasoned” these instructors were (i.e., they
had a minimum of 5 years experience and excluded students, researchers, and administrators), the higher were their scores on all perspectives (r = .10), except Developmental
(r = .01) and Social Reform (r = .01). As well, there are expected perspective differences among teachers of various grades or levels of student maturity. Nurturing scores
were higher for elementary and secondary teachers than for university-level instructors; conversely, Developmental scores were highest for graduate-level university
instructors. Transmission scores were highest for secondary and vocational-technical
instructors and lowest for teachers of preschoolers. Grade-level differences were
weakly significant for all five perspectives (see Table 2E).
Criterion group differences. Tests for differences among recognized criterion groups
can be helpful in two ways: (a) by increasing face validity and (b) by illustrating an
absence of bias. In regards to face validity, logically expected differences help confirm
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
370
Adult Education Quarterly 61(4)
that scales are operating as expected (e.g., Nurturing is more evident among women
than among men). Regarding absence of bias, as long as scale-to-scale differences are
generally small, scales can be assumed to be unbiased across different groups of
responders (e.g., the scales perform similarly for people whose first language is not
English and for native English speakers). Gender differences were significant but
small (ranging across the five perspectives from r = −.03 to +.20). Correlations with
teachers’ educational levels were also generally small (r = −.11 to +.13), as were English as first language (r = −.21 to +.05), length of experience as a teacher (r = −.03 to
+.12) or as a practitioner (r = −.04 to +.14), and fraction of professional time devoted
to instruction (r = +.08 to +.11). Women had slightly higher Nurturing scores; more
experienced teachers had slightly higher Developmental scores; more experienced
practitioners had slightly higher Apprenticeship scores; respondents with higher education levels had somewhat lower Nurturing scores but higher Developmental scores,
and people whose first language was other than English had slightly higher Social
Reform totals. Differences between people who reported their primary role as “teacher”
and those who reported “student” were usually small, but significant. Teachers were
consistently one half to three quarters of a point higher on Transmission, Apprenticeship, Developmental, and Social Reform, but students were three quarters of a point
higher on Nurturing. Still, none of these differences were so strong as to suggest the
presence of scale bias (see Table 2E).
Because the TPI is brokered via many universities and organizations, such as
Arizona State, Clemson, Columbia, Cornell, Dublin, Harvard, Hawaii, Minnesota,
York, the Dental Hygiene Educators of Canada, and Education Resources Information
Center (ERIC), there is a steady influx of data. Yet between 2008 and 2009, when
respondent numbers increased from 60,000 to more than 100,000, the norms have not
changed significantly, suggesting there is stability in what has been reported above.
Consequential validity. Messick (1989) argues that there is no such thing as “validity
in the abstract”; a measuring instrument is valid only if it accurately and helpfully informs
decisions or actions. Thus, what educational operations are facilitated by information
deriving from the TPI? In workshops and debriefing sessions, teachers “see themselves” in their profiles and explanatory interpretations. They also recognize colleagues whose profiles are different from their own. From such activities they gain
understanding and respect for a plurality of approaches to teaching, as well as an
expanded vocabulary of concepts and terminology that frequently find its way into
personal statements of teaching philosophy. The fact that different people have different profiles confirms that there is more than one acceptable way to think about teaching and that people’s profiles and interpretational materials are powerful tools for
faculty development exercises and group goal clarification (Pratt, 2002).
There is yet another important aspect of the TPI that speaks to its consequential
validity and the necessity to acknowledge and respect a plurality of the “good” in
teaching. It has to do with the very nature of teaching. Teaching is, of course, a dynamic
and complex activity. But it is also an intellectual, relational, moral, and cultural activity. It is intellectual because it deals with claims to truth; it is relational because it
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
Collins and Pratt
371
places people in interdependent roles and responsibilities; it is moral because it makes
judgments of propriety, value, and worth; and it is cultural because relationships and
propriety are culturally and historically constituted. Therefore, by its nature, teaching
is not an individual act but, rather, a social act that is situated in time and traditions that
have intellectual, relational, moral, and cultural dimensions. Each of these is differentially respected and revealed through the constructs that make up the TPI.
The TPI in Practice
The TPI shows utility well beyond a simple exercise in self-examination and values
clarification among adults who hold varieties of teaching responsibilities. In systematic studies, it has shown itself to be a powerful tool in faculty development, teaching
assessment, teaching improvement, peer reviews of teaching, and a variety of research
investigations of adults teaching adults (and sometimes children) in learning settings
around the world.
Faculty Development
Self- and peer observations are enhanced by having frameworks that remind observers
that good teaching is much more than “doing it the way I would do it” (Courneya,
Pratt, & Collins, 2008). At the University of California, Davis School of Medicine,
the TPI is a central before and after descriptor of perspective change in yearlong programs of faculty development (Srinivasan et al., 2007). Similarly, the TPI is an
instructional tool at the University of Toronto School of Medicine and the University
of British Columbia Faculty Certificate Program on Teaching and Learning in Higher
Education (Hubball, Collins, & Pratt, 2005). At a more systemic level, the TPI is a
faculty development requirement at Strayer University where (to date) nearly 1,000
faculty members from across their 42 multistate campuses have completed the TPI.
At Republic Polytechnic, a new problem-based tertiary technology training institution in Singapore, all incoming teaching staff are required to complete the TPI during
their initial weeklong facilitation-training workshops. A year or two later, these new
instructors assemble a portfolio of supportive materials to present at a certifying interview that includes a second, more recent, update of their TPI profile. Second profiles
are significant predictors of who will gain certification on their first attempt and distinguishes them clearly (η2 = .31) from those whose certification will be delayed for
a second or third year or who will ultimately leave the institution. Some 1,950 TPI
records exist for these instructors.
Research Projects
The TPI has been the instrument of choice for dozens of master’s theses, doctoral
dissertations, and research projects in the United States and Canada and around the
world. These works have investigated a range of issues, learners, and contexts, such
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
372
Adult Education Quarterly 61(4)
as: student teachers’ “journeys” toward becoming teachers (Jarvis-Selinger, 2002),
conceptual equivalency of teaching perspectives across languages (Lu, 2006; Ruan,
2004), an exploration of the influence of context in educating medical technicians
(Tiffin, 2008), discontinuities between occupational therapy instructors’ beliefs,
intentions, and actions (Kehres, 2008), teaching perspectives and supervisory practices of cooperating teachers (Clarke & Jarvis-Selinger, 2005), and contrasts of beliefs
and intentions of teachers in online and face-to-face instructional settings (Panko,
2004). As well, the TPI is finding application in many countries and in a variety of
languages and cultures. To date, it has been translated and tested in Spanish, French,
Portuguese, German, Chinese, Korean, and Japanese, although only the English and
Spanish versions are currently online.
Peer Reviews of Teaching
Awareness of one’s perspective on teaching is useful, especially when teaching is
under review. Peer reviewers are often selected because they fit within the culture and
norms of teaching within an instructional unit. Thus, it is not uncommon for reviewers
to look for some reflection of themselves as the measure of good teaching. In doing so,
reviewers may project their own perspective on the teaching they are to review. This
can put teachers at risk, particularly if they are being reviewed by someone whose
approach to teaching is markedly different from their own. When a person’s TPI profile
is used as part of a prereview discussion, conceptual differences about teaching can be
clarified before judgments are made about the quality of someone’s teaching. Whether
in formal, nonformal, or informal educational contexts, this is of no small consequence
for anyone whose teaching is being reviewed for continuing employment.
Improving Teaching
Being aware of one’s perspective may help, but it is not a sufficient indicator of an
effective teacher. Each perspective can represent effective teaching and each perspective
can represent poor teaching. No perspective is inherently better than any other perspective. Perspectives are a blend of personal philosophy (beliefs and intentions) and
situational circumstances. Yet some perspectives are a better fit to some teaching
contexts than others. For example, Apprenticeship perspectives fit very well in workplace learning. That does not mean that everyone teaching in workplace settings
should adopt an Apprenticeship perspective. There are many highly effective teachers
in the workplace who hold views other than Apprenticeship. But our data suggest that
more often than not, those who teach in workplace settings hold an Apprenticeship
Perspective as their dominant view of teaching. Furthermore, there are strong “signature pedagogies” within the professions that make for a better fit between forms of
training or education, eventual workplaces, and particular teaching perspectives
(Shulman, 2005; Jarvis-Selinger, Pratt & Collins, 2007). However, what matters most
is not which perspective fits in what setting. Rather, the question is whether we can
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
Collins and Pratt
373
help people improve their teaching without having to change their perspective on
teaching.
Although we know a good deal about effective teaching of adults, we also know
that no teacher embodies all the findings that characterize highly effective teaching.
They do so in some cases, because some findings fit their personal, situational, and
cultural circumstances better than others (Pratt, 2005b, 2009, 2010; Pratt & Collins,
2000, 2002). Bain’s (2004) work with teachers from North America, for example,
shows that highly effective teachers know their content area very well and know the
essential questions, debates, and issues that characterize their discipline or field of
practice. They do not try to “cover” a lot of material. They try, instead, to help students
address the big questions: the larger picture governing a body of knowledge and/or
practice. Bain (2004) also found that highly effective teachers know how to engage
their students in those questions, issues, and debates. Those two findings give direction for all who want to improve their teaching but they are particularly helpful for
those who hold a dominant Transmission perspective. Teaching within each perspective can be improved by focusing on findings from research that are consistent with
that perspective.
Conclusions and Recommendations
Finally, in the words of a reviewer of this article, the “heavy lifting” related to the
philosophical implications of the five perspectives and the “know thyself” desirability
of teachers looking inward has been done in the Five Perspectives book (Pratt &
Associates, 1998) and several other articles. By contrast, this article presents the
critical technicalities of the TPI and some of its most general findings, establishing
with a fair degree of certitude the foundation for a larger discussion of what it means
“to teach.”
Consequently, however useful the TPI may be in helping teachers to clarify their
perspectives, it should not be viewed as a diagnostic tool for purposes of screening or
remediation. Such use would, presume that one (or more) of the perspectives is inherently better than another. We do not hold that view. Rather, the TPI should be used as
a discussion tool to promote reflection, discussion, clarification, and, most important,
respect for the intellectual, relational, moral, and cultural aspects that are essential to
understanding what it means “to teach.”
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interests with respect to the authorship and/or
publication of this article.
Funding
The author(s) received no financial support for the research and/or authorship of this article.
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
374
Adult Education Quarterly 61(4)
References
Apps, J. (1996). Teaching from the heart: Professional practices in adult education and human
resource development series. Malabar, FL: Krieger.
Bain, K. (2004). What the Best College Teachers Do. Cambridge: Harvard University Press.
Brookfield, S. D. (2006). The skillful teacher (2nd ed.). San Francisco, CA: Jossey-Bass.
Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait, multi-method matrix. Psychological Bulletin, 56, 81-105.
Chan, C. H. (1994). Operationalization and prediction of conceptions of teaching in adult education (Unpublished doctoral dissertation). University of British Columbia, Vancouver,
British Columbia, Canada.
Clarke, A., & Jarvis-Selinger, S. A. (2005). What the teaching perspectives of cooperating teachers tell us about their advisory practices. Teaching and Teacher Education, 21,
65-78.
Courneya, C. A., Pratt, D. D., & Collins, J. B. (2008). Through what perspective do we judge
the teaching of peers? Teaching and Teacher Education, 24, 69-79.
Hubball, H. T., Pratt, D. D., Collins, J. B. . & Pratt, D. D. (2005). Enhancing Reflective Teaching
Practices: Implications for Faculty Development. The Canadian Journal of Higher Education,
35(3), 57-81.
Jarvis-Selinger, S. A. (2002). Journeys from student to teaching: Charting the course of professional development (Unpublished doctoral dissertation). University of British Columbia,
Vancouver, British Columbia, Canada.
Jarvis-Selinger, S. A., Pratt, D. D., & Collins, J. B. (2007). Do academic origins influence perspectives on teaching? Teacher Education Quarterly, 34(3), 67-81.
Kehres, E. (2008). Faculty beliefs, intentions, and actions in occupational therapy education
(Unpublished doctoral dissertation). Ohio University, Athens.
Kember, D. (1997). A reconceptialization of the research into university academics. Conceptions of teaching. Learning and Instruction, 7, 255-275.
Lu, X. (2006). Conceptual equivalence of teaching perspectives in Chinese and English (Unpublished master’s thesis). University of British Columbia, Vancouver, British Columbia,
Canada.
Messick, S. (1989). Validity. In R. L. Linn (Ed.), Educational measurement (3rd ed., pp. 13-103).
New York, NY: Macmillan.
Nunnally, J. C., & Bernstein, I. (1994). Psychometric theory. New York, NY: McGraw-Hill.
Panko, M. (2004). The impact of teaching beliefs on the practice of e-moderators (Unpublished
doctoral dissertation). University of New England, Armidale, New South Wales, Australia.
Pratt, D. D. (2002). Good teaching: One size fits all? In J. Ross-Gordon (Ed.), An up-date on
teaching theory (No. 93, pp. 5-16). San Francisco, CA: Jossey-Bass.
Pratt, D. D. (2005a). Personal philosophies of teaching: A false promise? ACADEME, 91, 32-36.
Pratt, D. D. (2005b, April). Teaching international students (Grand rounds presentation).
St. Michael’s Hospital, University of Toronto Medical School, Toronto, Ontario, Canada.
Pratt, D. D. (2009, April). Why can’t everyone be like me: Variations on a theme of excellence
(Thomas Q. Morris Invited Lecture). Columbia University Medical School, New York, NY.
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015
Collins and Pratt
375
Pratt, D. D. (2010, February). The power of context and community in clinical teaching: How
do Australians compare? (Opening plenary for Education Week). Faculty of Medicine,
University of Melbourne, Melbourne, Victoria, Australia.
Pratt, D. D., & Associates. (1998). Five perspectives on teaching in adult and higher education.
Malabar, FL: Krieger.
Pratt, D. D., & Collins, J. B. (2000, February). Cultural variations on a theme of “good teaching” (by Ritsumeiken/UBC students and faculty). University of British Columbia, Vancouver,
British Columbia, Canada.
Pratt, D. D., & Collins, J. B. (2002, November). International perspectives on teaching. internationalising education in the Asia-Pacific region. Paper presented at the 30th Annual ANZCIES Conference, University of New England, Armidale, New South Wales, Australia.
Pratt, D. D., Sadownick, L., & Jarvis-Selinger, S. (2011). Pedagogical BIASes and clinical
teaching in medicine. In L. M. English (Ed.), Health and adult learning (pp. 273-296).
Toronto, Ontario, Canada: University of Toronto Press.
Ruan, X. (2004). Conceptual equivalency in translating Teaching Perspectives Inventory (TPI):
A Chinese perspective (Unpublished master’s thesis). University of British Columbia,
Vancouver, British Columbia, Canada.
Shulman, L. S. (2005). Signature pedagogies in the professions. Daedalus, Summer, 52-59.
Srinivasan, M., Pratt, D. D., Collins, J. B., Bowe, C., Stevenson, F., Pinney, S., & Wilkes, M.
(2007). Developing the master educator: Cross-disciplinary teaching scholars program for
human and veterinary medical faculty. Academic Psychiatry, 31, 452-464.
Tiffin, S. (2008). Effective clinical teaching for medical technologists in Canada: Five case
studies (Unpublished master’s thesis). University of British Columbia, Vancouver, British
Columbia, Canada.
Tweed, R. G., & Lehman, D. R. (2002). Learning considered within a cultural context: Confucian and Socratic approaches. American Psychologist, 57(2), 89-99.
Bios
John B. Collins is Adjunct Professor in the Department of Educational Studies at the University
of British Columbia. He specializes in program evaluations of educational training and initiatives, especially for mid-career adults in medicine, nursing, pharmacy, law, education, and other
social and health services. In 2008 and 2009 he was researcher-in-residence at Singapore’s
Republic Polytechnic, that nation’s first exclusively Problem-Based Learning institution.
Daniel D. Pratt is Professor of Adult & Higher Education at the University of British Columbia
and holds a cross-apointment in the School of Medicine, where he is Senior Scholar and past
Director of Clinical Educator Fellowships in Medical Education. In 2008 he received Canada’s
most prestigious university teaching award- the 3M National Teaching Fellowship.
Downloaded from aeq.sagepub.com at VIRGINIA COMMONWEALTH UNIV on February 11, 2015