HOW DO FACULTY TEACH
How do faculty go about their teaching? And how has that changed, if at all, in the past generation? Three out of four faculty members across all institutional types agree on three primary teaching goals: (1) mastery of a body of knowledge in the discipline; (2) developing critical thinking skills; and (3) preparing students for employment.9
The near ubiquitous use of the lecture as the primary mode of instruction reflects, not surprisingly, the strong subject-matter orientation of higher education. Eighty percent of a sample of undergraduate arts and sciences faculty, reported a recent study, lectured during all or most of their dasses.'° When queried about their propensity to lecture, faculty members noted the lecture's usefulness in facilitating student understanding by simplifying an ever-expanding volume of complex material.
The lecture method was not quite universal and its use varied by institutional type Bach with its characteristic expectations and culture. Subject-matter oriented university faculty, one study found, showed a greater tendency to lecture, use detailed notes, and assume student interest in the content. The more "studentcentered" faculty members at community colleges, in contrast, lectured less, favored individual assignments, more closely followed a textbook, and provided more structure to their classes.ll
Within the university sector, discipline and gender could predict patterned differences. Natural scientists, one study showed, were most likely to lecture exclusively; literature faculty favored discussion, and social scientists employed both techniques Among faculty at a midwestern university, another study reported, women faculty were significantly more person-oriented and student-centered, more concerned with the classroom's emotional climate and with involving students in the learning process.l3
The National Center for Education Statistics collected data on faculty
instructional approaches for the first time in its 1993 National Survey
of Postsecondary Faculty. Recent changes in these trends must await analysis
of this data.
PREPARATION FOR TEACHING
Virtually all faculty received advanced training in an academic discipline or a professional field; virtually none received any pedagogical training.l4 Nor did many come to their first full-time academic appointment with prior teaching experience except perhaps for a graduate teaching assistantship that did not involve full responsibility for a course. While some groups promoted reforms, most novices still began their teaching careers armed with memories of an influential teacher and little else.l5 It must all be learned "on the job."
Recent research focused on the first years of a college teaching career,
how faculty were introduced to teaching, and how they adapted to their
teaching assignments These studies depicted an intense period of stress.
Overwhelmed by the demands of preparing a host of courses for the first
time--often including courses outside their area of expertise, fearful
of student disapproval, and of appearing to need help--new faculty tended
to be highly structured and lecture oriented, to overprepare, and to teach
"defensively." That is, to strive to avoid error and adverse student reaction.l7
New faculty members rarely had access to senior colleagues, since most
institutions lacked formal mentoring programs and most novices would not
admit ignorance or incapacity to department chairs responsible for contract
renewals or promotion decisions. New faculty members found that teaching
made inordinate time demands, provided few intrinsic rewards at the novice's
skill level, and counted little towards promotion and tenure. 18 By the
third or fourth semester, many new faculty memberss established negative
teaching orientation and style that persisted throughout their careers.
They "wrote off" teaching.l9
The entry experience described here tended to characterize new full-time faculty members at four-year colleges and universities.20 Many new faculty members at community colleges, in contrast, came from secondary schools or acquired experience during lengthy part-time stints.2l Two-year college faculty usually taught more class sections but may have had fewer new course preparations, if only because community colleges offer a much more limited array of courses.
The entry experience may also vary by gender or the stage of life at
career entry. Academic women historically entered full-time teaching later
in life, following child-rearing. That pattern appeared to be changing
more among more recent female cohorts who entered full-time teaching prior
to or during child-rearing years. Family demands on these faculty members
may have created additional time pressures that complicated early career
Teaching, for new faculty, was a source of stress and even a "distraction"
from rewarded research, but it became a significant source of professional
satisfaction for mid-career and senior faculty. Students, one study found,
inspired teaching vitality for senior faculty at 11 diverse campuses.23
The evolution of faculty professional interests, another study reported,
paralleled Levinson's stages of adult development.24 Seasoned faculty,
in a stage of life where "generativity" concerns--preparation of the next
generation--were ascendant, came to value their teaching and their work
with students for satisfying that basic drive.25 Moreover, both teaching
and scholarly interests, some evidence suggested, may change from a specialized,
empirical focus, to an integrative, multi-disciplinary orientation.26
Institutional factors may affect the extent to which faculty members
increased their commitment to teaching. Department chairs and other academic
administrators, one study found, could 'broker' opportunities for renewal.
Chairs, for example, could assign temporary responsibilities to senior
faculty or change their work setting.27 Beyond periodic "refreshers," administrators
could facilitate and enhance collegial interaction around teaching by supporting
team teaching, course clustering, common examinations in multi-section
Administrative and collegial support played key roles, then, in sustaining
faculty commitment to teaching.29 Conversely, without this support, faculty
members facing increased course loads and student contact hours might suffer
"burnout." Burnout, several scholars found, often occurred among the most
active and engaged faculty members.30 But most faculty members took their
teaching ever more seriously and having survived the early stresses, made
it a major source of satisfaction.
How well do faculty teach? Evaluating the classroom effectiveness of college faculty developed from modest beginnings into a cottage industry in one generation. Nearly a dozen organizations distributed and scored instruments that record student perceptions of courses and instructors, including the Purdue cafeteria system, the Student Instructional Report (SIR) form of Educational Testing Service, and the Kansas State IDEA form. These evaluations assumed that students--the consumers of teaching--were in the best position to assess its quality.
Student evaluation of courses, instruction, and instructors is pervasive. Four out of five campuses use some student rating instrument to aid in personnel decisions, improvement of teaching, or student course selection.3l
Recently, some colleges complemented student evaluations with alumni ratings, faculty self-assessments, and peer assessments, including classroom visitation and review of curricular and teaching materials.32 This trend culminated in the use of teaching "portfolios," or documentation of teaching materials, activities, and eventually student products.33 Such systematic documentation could provide the basis for meaningful peer review of teaching just as publications provided the basis for peer review of faculty scholarship.
But student evaluations still constitute the dominant mode of assessment. As their use spread, pressing questions were raised about how solid a foundation the ratings provided. Were the ratings reliable and valid indicators of teaching effectiveness? Were they subject to situational distortions? How did they compare with peer ratings or faculty self-assessments? Such questions spawned a 20-year explosion of studies on student ratings of instruction, each usually reporting data from a single campus or from a single instrument. So voluminous were the findings, that a "meta-literature" reviewed and synthesized diverse, and frequently contradictory, findings.34
Analysis of this meta-literature was facilitated by the emergence of sophisticated quantitative techniques that permitted the reanalysis and integration of findings from many studies. Treating each individual study as a case or subject, these "meta-analyses" collected data on sample size and composition, variables examined, instruments employed, and the findings expressed as descriptive statistics--a zero-order correlation coefficient, for example. These efforts allow us to systematically relate patterns in the findings to characteristics of the studies themselves, thereby helping to locate the source of variation in the findings.
What does this comprehensive and integrated research base tell us about student rating instruments and their use? First, the ratings obtained by those "state of the art" nationally developed and disseminated instruments (Purdue, SIR/ETS; IDEA/Kansas State) were highly reliable; their results were consistent across administrations to the same group.35 The instruments were also highly valid in at least three respects:
* They accurately reflected student opinions assessed independently, i.e., they measured what they purported to measure.
* They were positively associated (r = 0.5) with student learning and academic achievement in the course. They measured something meaningfully related to good teaching.36
* They were highly correlated with colleague ratings, zero order correlations of about 0. 9.
What aspects of good teaching do these ratings instruments measure reliably and validly? Factor structures differed somewhat across diverse instruments and diverse studies, but Kenneth Feldman identified three basic dimensions that have proved subsequently to be virtually universal. 37
* Presentation skill, including items related to the ability to
"stimulate student interest," to "clarity," to "knowledge of subject matter."
* Rapport, including "sensitivity to individual student learning,"
"sensitivity to class progress," and "availability" or "openness."
* Course organization and classroom management, including syllabus,
assignments, grading practices, etc. 33
These were not the only elements of good teaching. They did not, for example, tap the "currency" of the instructor's course materials or the instructor's standing in the field. They did, nonetheless, consistently and reliably tap several acknowledged aspects of good teaching and could readily be supplemented by other sources.
What about bias? Didn't students rate those instructors highest when they get higher grades, for example? Faculty members widely perceived bias in student ratings but the literature suggested that there is much less bias than most faculty think. Indeed, when bias existed, the effects were relatively small, and they were consistent, that is, the patterns were well-recognized and can be accounted for.39 Most notable among these were the systematic differences in the ratings of courses in different disciplines. Table 5 arrays the disciplines from most to least highly rated. Courses in quantitative fields--such as mathematics, engineering, and the hard sciences--tended to receive lower ratings, while certain humanities fields-- art and music, for example--were rated highest. It is not clear to what extent these patterns reflected differences in the nature of the fields (the extent, for example, that learning was sequentially organized), in the nature of students majoring in different fields (different kinds of students rated differently), or in the nature of faculty teaching in these fields (scientists may have been poorer teachers).40
Beyond discipline, course characteristics may have modestly affected ratings though not always in the predicted direction. Smaller classes tended to be rated more highly than larger classes; but difficult courses tended to be rated slightly more highly than easier ones.4l Characteristics of the instructor, including gender and age, or of the student, including gender, age, grade level, GPA, anticipated grade, bore no statistically significant relationship to ratings.42
If the instruments are reliable and valid indicators of three important aspects of good teaching and are not subject to unreasonable bias, what about their use in personnel decisions and in instructional improvement? Sophisticated instruments, to be blunt, may not be matched by sophisticated use. About half of the faculty in a sample who used ratings in personnel decisions could not identify likely sources of bias in the results. Nor could these faculty members recognize standards for proper samples sufficient numbers of students per course, sufficient numbers of courses per instructor--or interpret common descriptive statistics.43 Instructional improvement, several studies showed, was only likely to occur when skilled consultants helped faculty interpret the ratings and develop better teaching strategies.44
Together, these studies suggest conditions for the appropriate use of ratings instruments. For personnel decisions, global or summary ratings--rather than ratings on individual components--provided the best basis for comparing faculty.45 Global ratings were relevant to all types of courses in all fields; specific dimensions such as rapport, for example, might be less relevant to large classes or to the natural sciences. Global ratings were more clearly related to student academic achievement, 0.5 zero order correlation versus 0.3 for individual dimensions, and were less subject to abuse by the untrained user trying to weigh various dimensions .4fi
Sample size was important--meaningful comparisons require sufficient
numbers of courses and of student raters. Systematic (average) disciplinary
differences must also be factored into comparisons. So must other indicators
of good teaching, syllabi or final examination results, for instance, that
tapped dimensions that ratings did not measure, including currency in the
field and student achievement.
Instructional improvement, in contrast, required a focus
on many descriptive and behavioral items. There was no need for concern
about sampling a single set of ratings. Ratings users, the literature suggests,
may need task-specific training, detailed instructions, and a skilled consultant
to guarantee appropriate use of ratings for personnel decisions and instructional
improvement.47 Student evaluations provided a reliable and valid vehicle
for assessing key aspects of good teaching as long as their use was informed
by a sense of their strengths and limitations.
TEACHING AND FACULTY RESEARCH
Stemming from von Humboldt's ideal for the l9th century German university,
the synergy of teaching and research became a touchstone in American academic
lore.48 Today, the oft-assumed correlation of effective teaching and faculty
involvement in research and scholarship is debated by state legislators
as they consider issues of faculty workload and productivity and in on-campus
forums about the academic reward system.49 Amid the whirl of that debate,
at least three hypotheses are discernable:
Research and teaching are positively correlated. Researchers, many partisans
argue, keep up with the latest developments in their field and therefore
bring greater subject matter knowledge into the classroom. Researchers,
some add, stimulated by their explorations, bring a greater enthusiasm
and excitement into the classroom. Research and teaching, others contend,
reflect a general ability factor so that good researchers teach 2 well,
and vice versa.
Research and teaching are negatively correlated. Time and effort expended
in research, advocates of this hypothesis assert, detract from time and
effort spent in teaching. Researchers and teachers, some state, are different.
Personality attributes that contribute to the performance of one detract
from the other.
Research and teaching are uncorrelated. These activities are totally independent, and proficiency in each is randomly distributed throughout the population. Scholars have evaluated these hypotheses by reviewing the many studies that relate
Fine and applied arts
Health and technology
Fine and applied arts
Letters and humanities
Physical and health education
Political science and government
Religion and theology
Social work and service
Trade and vocational- technical education
English language and literature
Health and technology
Letters and humanities
Physical and health education
Political science and government
Business and commercial technology
English language and literature
Social work and service
Mathematics and statistics
Computer and information science
Mathematics and statistics
Business and management
Computer and information science
Source: W.E.Cashin, 1990 ["Students Do Rate Different Academic Fields Differently," in M. Theall and J. Franklin, eds., Student Ratings of Instruction: Issues for Improving Practice, New Directions for Teaching and Learning, 43 (San Francisco: Jossey-Bass, 1990]
faculty research productivity to teaching effectiveness, measured by
student ratings. Research performance, however measured, was usually not
significantly related to global ratings of teaching effectiveness. But
the association was nearly always positive (about r = .12). 50 Moreover,
research activity was significantly related to individual components of
teaching effectiveness. Statistically significant, positive zero order
correlation coeffficients emerged in three of four tests of the relationship
of publications to the "intellectual competence" dimension of teaching
effectiveness. Nonsignificant correlation coefficients emerged in three
of four tests of the relationship of publications to the "rapport" teaching
A detailed meta-analysis reported the highest positive zero-order correlation
coefficients with publications for "instructor knowledge of subject matter"
(about 0.2), followed by "clarity," "stimulation of interest," and "value
of course material" (about 0.1).52 Items related to "enthusiasm," "sensitivity
to class progress," "availability to students," "feedback to students,"
"encourages questions" showed zero order correlations around zero. Faculty
research and publication appear positively, if modestly, associated with
instructor's subject matter knowledge but are independent of the socio-emotional
dimensions of teaching.
The research syntheses also addressed the purported conflict between
the time faculty invest in research and in teaching. The evidence was conclusive.
Time invested in research was positively associated with research productivity--assessed
by number of publications-but bore no association with teaching effectiveness
ratings.53 Time invested in teaching was negatively associated with research
productivity but was not associated with teaching effectiveness ratings.
Time invested in research may have thus detracted from time invested in
teaching, but the quality of instruction was not necessarily compromised.54
Conversely, the investment of extra time in teaching at the expense of
research may not have added substantially to the quality of instruction.
The reviews also discussed the influence of personality on the performance
of teaching and research responsibilities. Several studies reported positive
associations between global student ratings and some faculty personality
attributes, including "extraversion,""warmth," and "expressiveness."55
Personality correlates, a recent study suggested, may be course type specific.56
Two other studies reported that personality characteristics associated
positively with teaching effectiveness ratings, "supportiveness," "warmth,"
and "extraversion," were also often negatively associated with research
productivity.57 Teaching effectiveness may partly depend on personal attributes
not typically associated with the productive researcher. But there is no
evidence that a general ability or intelligence factor contributed to high
performance in both areas.
The empirically identified research-teaching relationship did not precisely
confirm any prevailing theory. The evidence resoundingly disconfirmed the
notion that investment and quality performance in one area was inimical
to the other. Indeed, the intellectual and scholarly aspects of teaching
may have benefited, if modestly, from faculty investment in research. Teaching
and research, the research also confirmed, may have drawn on somewhat different
strengths and attributes and that attributes associated with one may have
detracted from the other.
Our language is tentative since contextual factors may have mediated
the research-teaching relationship. Most studies were confined to a single
institution, mixed diverse disciplines, and grouped together faculty at
different career stages. A study that disaggregated the disciplines found
the strongest positive association between publications and student ratings
in the humanities, the weakest in the natural sciences and professional
fields. The social sciences fell in-between. The relationship may also
have been stronger for senior faculty and attenuated for new and junior
faculty for whom research and teaching made more urgently competing demands.
As for institutional context, one scholar found a more strongly positive
relationship at comprehensive and liberal arts colleges and an attenuated
relationship at research universities.59
FACULTY DEVELOPMENT AND CONTINUING EDUCATION
How, if at all, can faculty teaching be improved? The 1974 publication of Faculty Development in a Time of Retrenchment stimulated discussion of professional (in-service) development, or continuing professional education, for college and university faculty.60 Over the past decade, the recognition of increased student diversity, academic and social, has prompted faculty members and administrators to ask specifically about the relative efficacy of the leading approaches to instructional development. Much writing on faculty and instructional development remained impressionistic, or consisted of case study reports. But faculty developers and researchers, supported by the U.S. Department of Education and other government agencies, and by private foundations, are now engaged in more systematic empirical work.6l Much of that work focused on the community college, where the imperative for faculty and instructional development runs deepest.62
What have we learned about the possibilities for, and conditions of,
meaningful instructional development? First, faculty motivation to grow
was a crucial and complex issue. Typically, faculty participants in instructional
improvement activities were already recognized as good teachers but aspire
to be even better--the identifiable group of "volunteers."63 But for others,
the academic career provided windows of opportunity for professional growth
and development. New and junior faculty, we've already noted, often needed
advice and support to surmount initial career hurdles.64 Mid-career and
senior faculty often sought opportunities to contribute in new ways.65
These groups found discipline-based opportunities particularly attractive,
since they build on natural loyalties.66
Second, "quick fixes didn't work." Faculty members rated short-term
workshop experiences highly,67 but the evidence suggested that workshops
did not produce sustained reflection on or concrete changes in teaching
practices.fi8 Teaching patterns, scholars found, were formed early and
tacitly, and were relatively difficult to change unless two conditions
were met.69 First, faculty members must feel a real need for change. That
need often resulted from major or sudden external "threats" to survival,
such as fiscal crises or accelerated changes in composition of student
bodies.70 Perceiving a need for change may be facilitated, one study found,
if faculty members developed assessment data that indicates a problem requiring
their instructional intervention. 71
Next, faculty needed sustained support. Faculty members changed their
teaching methods, the late Joseph Katz demonstrated, only after a year
of intensive involvement in a program of peer classroom visitation and
student interviewing and only when a faculty partner and other colleagues
supported the changes.72 Providing faculty members with student feedback
early in a course, another study showed, led to course modification only
if an instructional consultant conveyed the feedback and then supported
the attempted changes.73 Colleagues, a decade of scholarship showed, could
play a powerful formative role in supporting faculty development.74
Seven propositions summarize our knowledge about college faculty as teachers:
1. College and university faculty saw themselves primarily as teachers and did, in fact, spend much of their time teaching. That has not changed appreciably in the last generation. Moreover, faculty did not want it any differently, except insofar as they needed to publish to obtain promotion and tenure.
2. Lecture was still the predominant mode of instruction, approximately 80 percent of class time, but there were clear differences by institutional type and discipline.
3. Faculty, save for secondary school teachers who go to community colleges, were illprepared, at best, for their initial teaching experience. The initial years of their first full-time academic appointment were therefore characterized by high stress generated by the teaching role, and by limited support. Faculty sustained the teaching orientation developed during this period throughout their careers.
4. Most faculty increased their interest in teaching during their careers. But many instructors, lacking the support and opportunity to sustain their teaching commitments, were subject to "burnout."
5. Teaching was typically assessed via student evaluations, but peer and self-assessment have recently received increased attention. Student ratings were highly reliable and valid; but their use in faculty personnel processes could be strengthened.
6. Three dimensions of good teaching reflected faculty classroom
practices: presentation skills (content mastery, clarity, and enthusiasm,
for example), rapport (sensitivity and openness to students and their progress),
and course organization and classroom management. Involvement in research
seemed to contribute to content mastery, less so to rapport and course
organization. But research commitments did not detract from teaching effectiveness,
and devoting more time to teaching might not translate into greater effectiveness.
7. Many faculty members desired to improve their teaching, but lacked the necessary knowledge and resources. Successful faculty development required institutional support at key points--just starting out, for example, and mid-career. Above all, teaching must be deisolated and discussed with colleagues, that is, seen as "community property."
A word of caution. Our knowledge about college faculty as instructors is primarily based on studies of full-time, four-year, liberal arts faculty, a shrinking segment of the profession. The American professorate is becoming an occupational group that is increasingly parttime (40 percent by headcount and rising), that teaches in the professional and applied fields, and that works in two-year institutions. The object of inquiry is changing before our eyes almost before we can get hold of it.
Forthcoming publication of the results of the National Survey of Postsecondary Faculty1993, conducted by the National Center for Education Statistics, will provide us with updated information about faculty career trajectories. The survey's large sample of part-time and community college faculty will enable us to trace the changing contours of the academic profession, and to use these changes to strengthen faculty and instructional development activities in the 21st century.