Special Education Teacher Retention and Attrition:
A Critical Analysis of the Literature
Prepared for the Center on Personnel Studies in Special Education
Bonnie S. Billingsley
Virginia Tech, Blacksburg
(COPSSE Document No. RS-2)
Center on Personnel Studies in
UNIVERSITY OF FLORIDA
CENTER ON PERSONNEL STUDIES IN SPECIAL EDUCATION
UNIVERSITY OF FLORIDA
JOHNS HOPKINS UNIVERSITY
UNIVERSITY OF COLORADO BOULDER
INSTRUCTIONAL RESEARCH GROUP, LONG BEACH, CA
COPSSE research is focused on the preparation of special education professionals and its
impact on beginning teacher quality and student outcomes. Our research is intended to inform
scholars and policymakers about advantages and disadvantages of preparation alternatives and
the effective use of public funds in addressing personnel shortages.
In addition to our authors and reviewers, many individuals and organizations have contributed
substantially to our efforts, including Drs. Erling Boe of the University of Pennsylvania and
Elaine Carlson of WESTAT. We also have benefited greatly from collaboration with the National
Clearinghouse for the Professions in Special Education, the Policymakers Partnership, and their
parent organizations, the Council for Exceptional Children and the National Association of State
Directors of Special Education.
The Center on Personnel Studies in Special Education, H325Q000002, is a cooperative
agreement between the University of Florida and the Office of Special Education Programs of
the U. S. Department of Education. The contents of this document do not necessarily reflect the
views or policies of the Department of Education, nor does mention of other organizations imply
endorsement by them.
Billingsley, B. S. (2003). Special education teacher retention and attrition: A critical analysis of
the literature (COPSSE Document No. RS-2). Gainesville, FL: University of Florida,
Center on Personnel Studies in Special Education.
U. S. Office of Special
Additional Copies may be obtained from:
P.O. Box 117050
University of Florida
Gainesville, FL 32611
There are no copyright restrictions on this document; however
please credit the source and support of the federal funds when
copying all or part of this document.
Literature Review 8
Literature Review Results 12
A Critique of Definitions, Conceptual Models, and Methodologies Used to
Study Attrition 29
Priorities for Research 34
Summary and Implications 37
APPENDIX. Special Education Attrition and Retention Studies 43
Figure 1. Schematic Representation of Special Education Teacher Retention,
Transfer, and Attrition 7
Figure 2. A Conceptual Model of the Influences of Teachers' Career
Figure 3. Conceptual Model for Understanding Teacher Attrition and
Figure 4. Path Diagram for Intent To Stay 19
The lack of qualified special education teachers threatens the quality of education that students
with disabilities receive. Attrition plays a part in the teacher shortage problem, and efforts to
improve retention must be informed by an understanding of the factors that contribute to
attrition. Specifically, this paper provides a thematic synthesis of studies investigating factors
that contribute to special education teacher attrition and retention. Five major themes are
addressed: (1) teacher characteristics, (2) personal factors, (3) teacher qualifications, (4) work
environments, and (5) teachers' affective reactions to work. Following this thematic review, a
critique of definitional, conceptual, and methodological approaches used to study special
education attrition is provided as well as priorities for future research.
One of the most important challenges in the field of special education is developing a qualified
work force and creating work environments that sustain special educators' involvement and
commitment. For decades the supply and demand of special educators has been of concern to
policy makers (Morsink, 1982; Smith-Davis & Billingsley, 1993; Smith-Davis, Burke, & Noel,
1984). The field has an insufficient supply of new teachers as well as qualified teachers who are
not currently teaching to fill vacancies. For example, administrators report that a shortage of
qualified applicants is the greatest barrier to obtaining special education teachers (Carlson,
2001). Carlson found that local school administrators reported job openings for 69,249 special
educators (positions for which personnel were recruited in 1999-2000). In addition, Carlson
(2001) found that as of October, 1999, over 50,000 special education teachers were newly hired,
and, at that same time, over 12,000 positions were left vacant or filled by a substitute because a
suitable candidate could not be found.
These recent data are consistent with the trend over the last ten years. There has been a clear
upward trend in the number of teachers needed over time to fill vacancies and replace
unqualified personnel. Roughly 10% of special educators are not fully certified for their
positions, according to Bright Futures for Exceptional Learners: An Action Agenda to Achieve
Quality Conditionsfor Teaching and Learning (Council for Exceptional Children [CEC], 2000).
However, these statistics may not fully illustrate the extent of the teacher shortage problem in
special education. For example, school districts may reduce services to students with disabilities
or raise class size limits (Billingsley, 1993; Darling-Hammond & Sclan, 1996).
The shortage problem has serious implications for students with disabilities if they are being
taught by people with less than adequate preparation. Darling-Hammond and Sclan (1996)
discuss a range of consequences of the shortage problem, including inadequate educational
experiences for students, reduced student achievement levels, and insufficient competence of
graduates in the work place.
Although the causes of the special educator shortage problem are complex, the retention of
teachers is a critical part of solving the problem. As Ingersoll (2001) observes, the shortage
problem will not be solved by recruiting thousands of new people into teaching if many leave
after a few short years.
It is critical to know how many teachers are leaving and what they do upon leaving. Recent
evidence suggests that special education, math, and science are the fields with the highest
turnover and that special education teachers are more likely to depart than other teacher groups
(Ingersoll, 2001). There are limited data comparing special and general education teacher
attrition, but national data give some indication of rates. Boe, Barkanic, and Leow (1999) state
that there has been a "fairly high level of school attrition in the public school teaching force" (p.
6) with about 7% of teachers moving to other positions and 6% exiting teaching altogether for a
total of 13%. Special and general education teachers leave at about the same rates; however,
special educators are significantly more likely than general educators to transfer to other teaching
assignments (Boe, Cook, Bobbitt, & Weber, 1998).
^^^^ ^^^ ^^^ ^^^^ ^^^ ^^^ ^^^ ^^^^ ^^^ ^^^ ^^^ -^^^ ^^^ ^^^ ^^^ ^^^^ ^^^ ^^^ ^^^ ^^^^ ^^^ ^^^ ^-
The field of special education loses many teachers to general education (Billingsley & Cross,
1991b). A higher percentage of special educators transfer to general education than the reverse
(Boe et al., 1998). Schnorr (1995) reports that of teachers who plan to leave special education,
12% want to transfer to general education. For the most part, general education teachers who
hold both general and special education certification are not interested in transferring to special
education (Billingsley & Cross, 1991a).
Clearly, teacher attrition is a major contributor to the teacher shortage problem in special
education. Understanding the rates at which teachers leave is important, because most teachers
are hired to replace those who leave rather than to meet the "needs of expanding enrollment, or
smaller class sizes, or new programs" (Grissmer & Kirby, 1987, p. 2). (A comprehensive
discussion of attrition rates and how teacher attrition contributes to the supply and demand of
special educators is provided in McLeskey, Tyler, & Saunders (2002).
Efforts to reduce attrition should be based on an understanding of factors that contribute to
special educators' decisions to leave the field. Billingsley (1993) and Brownell and Smith
(1992) reviewed research through the early 1990s, mostly state data and small-scale studies. The
goal of the present paper is to extend their initial findings by synthesizing research since 1992.
Specifically, this paper considers: (1) thematic synthesis of studies investigating factors that
contribute to special education teacher attrition and retention; (2) a critique of definitional,
conceptual, and, methodological approaches used to study special education attrition; and (3)
priorities for future research.
Special education teaching force
(3) Exit attrition
Returns to same special
(2) Transfer attrition
Transfers to another
special education teaching
(Teacher Position Transfer)
Transfers to general
(Teaching Field Transfer)
Figure 1. Schematic representation of special education teacher
retention, transfer, and attrition (Billingsley, 1993).
Non-teaching positions in
education (e.g., counseling,
For this paper, electronic databases (i.e., ERIC, Psychological Abstracts) were searched using
terms such as special education teacher attrition, retention, turnover, and transfer. Studies prior
to 1992 were not included, because they have been previously reviewed in either Billingsley
(1993) or Brownell & Smith (1992). The Appendix summarizes 16 research-based articles in
scholarly journals that meet the criteria outlined above. Several research reports and
presentations are also included in this review, because the findings from several major funded
projects do not appear in journals. Reports consisting primarily of tables with minimal narrative
discussion were excluded. Moreover, research papers directed primarily at personnel supply and
demand, attrition rates, job satisfaction, stress, burnout, and general education attrition were used
only to provide a context for the findings in this paper. Dissertations were also excluded.
Definitions of Attrition and Retention
Researchers studying attrition and retention have used varied definitions. Billingsley (1993)
provides a four category schematic representation of special education teacher retention, transfer,
and attrition. (See Figure 1.) In this first category, "absolute" retention (Boe, 1990), teachers
remain in the same teaching assignment and the same school as the previous year. The second
category, "transfers to another special education teaching position," includes those who stay in
special education teaching but transfer to another position (either in the same or a different
district). The third category, "transfers to general education teaching," is of particular concern,
because this group reflects a loss to the special education teaching force (Billingsley, 1993).
The fourth group, "exit attrition," includes those who leave teaching altogether. This "exit"
group includes those who retire, return to school, stay home with young children, or take non-
teaching positions in education (e.g., counseling, administration). Boe, Bobbitt, and Cook
(1997) state "the most troublesome component of turnover is exit attrition, because it represents
a reduction in the teaching force, requiring a compensating inflow of replacement teachers" (p.
377). Researchers often combine more than one of the categories identified in Figure 1 in a
given study. For example, Boe, Bobbitt, Cook, Whitener, and Weber (1997) and Miller,
Brownell, and Smith (1999) studied teachers who moved to other districts as well as those who
left teaching altogether.
The above schema relate to what teachers actually do (i.e., stay, transfer, exit). It is important to
note that only about half of the special education attrition studies address attrition behavior.
Several researchers collected follow-up data from special educators who left their positions
(Billingsley, Bodkins, & Hendricks, 1993; Billingsley, Pyecha, Smith-Davis, Murray, &
Hendricks, 1995; Miller et al., 1999; Morvant, Gersten, Gillman, Keating, & Blake, 1995). Boe
and colleagues used the national Teacher Follow-up Survey data for their study of leavers (e.g.,
Boe, Bobbitt, & Cook, 1997; Boe, Bobbitt, Cook, Whitener, et al., 1997).
Determining the extent to which teachers exit and rearrange themselves in the work force
requires careful follow-up study that is difficult, time-consuming, and costly, an obvious
drawback to studying teacher attrition. This may be one reason that the majority of special
education researchers do not actually study special education teachers' career behaviors; instead,
they examine existing populations of current teachers to determine their intent to leave as a
proxy for attrition (e.g., Billingsley & Cross, 1992; Cross & Billingsley, 1994; Gersten, Keating,
Yovanoff, & Harniss, 2001; Littrell, Billingsley, & Cross, 1994; Singh & Billingsley, 1996;
Westling & Whitten, 1996; Whitaker, 2000). The use of intent as a variable is discussed later in
the section critiquing definitions, models, and methodologies.
Societal Economic Institutional
Professional Work Work Commitment
Qualifications Conditions Rewards
Educational oDistrict environment oIntrinsic
oSchool environment oExtrinsic
Entry path/ Teaching
Certification *Work assignment oAncillary field
Initial commitment Employability
Demographic Family Cognitive/Affective
Figure 2. A
of the influences
*View of Efficacy
Perceived Career Options
![ Teachers. |
..le r.ain IIilefo 'm rplatc
,LEA. SEA, Federal PolICy
Macromsytm Personal Factors
Perceptions of learnear
Economnv Perceived Career Options
Life Cycle Events
Figure 3. Conceptual model for understanding teacher
attrition and retention (Brownell & Smith, 1993).
LITERATURE REVIEW RESULTS
Two conceptual models provide a broad understanding of factors that influence special education
attrition and retention (see Figure 2 and Figure 3). Both models provide a description of the
wide range of factors that influence special educators' career decisions. Figure 2 provides
Billingsley's (1993) schematic representation of the range of influences on teachers' career
decisions, including External Factors, Employment Factors, and Personal Factors. External
Factors (economic, societal, institutional), which are external to the teacher and employing
district, have an indirect effect on teachers' career decisions. The center of this model focuses on
Employment Factors (professional qualifications, work conditions and rewards, and
commitments to school, district, teaching field and teaching profession). Billingsley (1993)
hypothesizes that when "professional qualifications and work conditions are not as favorable,
teachers are likely to experience fewer rewards and, thus, reduced commitment. Whether
teachers actually leave depends on a host of personal, social, and economic factors" (p. 147).
Personal Factors include variables outside of the employment arena that may directly or
indirectly influence career decisions, such as life circumstances and priorities.
The second model (see Figure 3) proposed by Brownell and Smith (1993) is an adaptation of
Bronfenbrenner's model that incorporates four nested, interrelated systems. These systems
include the microsystem (teacher's immediate setting and interactions that occur as a result of
student and teacher characteristics), the mesosystem (interrelations among several variables in the
work place, e.g., collegiality and administrative support), the exosystem (formal and informal
social structures, e.g., socioeconomic level of community, nature of district), and the
macrosystem (cultural beliefs and ideologies of the dominant culture as well as economic
conditions that impact schools and teachers' career decisions). Brownell and Smith propose a
framework for designing and interpreting attrition/retention research, but not necessarily as a
causal model to be tested. They expect that variable relationships may be complex and
reciprocal and that some variables may correlate more highly with attrition than others.
Both frameworks identify variables that may be related to attrition and propose relationships
between the variables. This section of the present paper provides a thematic synthesis of
findings, including: (1) Teacher Characteristics and Personal Factors, (2) Teacher
Qualifications, (3) Work Environment Factors, and (4) Affective Reactions to Work. During the
last ten years, the study of work-related factors has been central in special education attrition and
retention research; therefore, this review focuses heavily on these factors. Given that External
Factors were not directly addressed in recent studies, this area is not reviewed [see Billingsley
(1993) and Brownell & Smith,(1992) for a discussion of these factors.]
Teacher Characteristics and Personal Factors
The relationship of teacher characteristics to attrition has been studied fairly extensively in the
last two decades in general education attrition research (Darling-Hammond & Sclan, 1996;
Grissmer & Kirby, 1987), but has received less attention in special education studies. Although
some special education researchers have investigated the relationship between attrition and
demographic variables (i.e., age, gender, and race), few conclusions can be drawn about teacher
characteristics and attrition.
Age. Age is the only demographic variable that is consistently linked to attrition in the special
education literature. Researchers consistently show that younger special educators are more
likely to leave (or express intent to leave) than older special educators (Boe, Bobbitt, Cook,
Whitener, et al., 1997; Cross & Billingsley, 1994; Morvant et.al., 1995; Singer, 1993a). Singer
(1993a) found that young special education teachers leave at rates nearly twice that of mature
teachers. Earlier, Grissmer and Kirby (1987) showed that teacher attrition patterns for both
general and special educators followed a U-shaped curve. Attrition is high among younger
teachers, low for teachers during the mid-career period, and high again as teachers retire.
However, Boe, Bobbitt, Cook, Whitener, et al. (1997) report that age functions differently for
leavers (those who exit public school teaching) than movers (those who changed positions). They
show that leavers show the characteristic U-function with age, whereas the percentage of movers
declines systematically with increasing age. Boe and colleagues found this relationship holds for
both general and special educators. Miller et al. (1999) reported that younger special educators
were more likely to transfer than older teachers; however, this finding did not hold for leavers.
Teachers with less experience are more likely to leave (Miller et al., 1999) and also indicate
intent to leave more often than their more experienced counterparts (Cross & Billingsley, 1994;
Gersten et al., 2001; Morvant et al., 1995; Singh & Billingsley, 1996). This is to be expected,
because age and teaching experience are highly correlated. However, many people now begin
teaching when they are older, often as a second career, so age should be controlled while
Age impacts supply and demand as well as teacher retention. Because attrition rates are sensitive
to teacher characteristics, "teacher career persistence may change as the composition of the
teaching force changes" (Murnane, Singer, & Willett, 1988, p. 22). As Singer (1993a) points
out, there is a "possibility that future attrition rates computed across all special educators may
drop as older teachers comprise a larger fraction of the new teaching force and as special
educators hired in response to EHA reach the stable years of mid-career" (p. 274-275).
The reasons for higher attrition among younger teachers have been discussed fairly extensively
in the literature. While some new teachers find the job of teaching satisfying, others encounter
frustrations and initial difficulties that discourage them from continuing in their positions
(Grissmer & Kirby, 1987; Heyns, 1988; Singer, 1993a). Grissmer and Kirby (1987) also point
out that younger teachers have fewer debt obligations and are less invested in a specific
occupation or location. Experienced teachers who leave also face retraining costs, as well as the
loss of tenure and an experienced teacher's salary (Singer, 1993a). Age-related factors also
include lifestyle cycle stages (e.g., child-rearing, retirement) as well as personal needs and
Gender. The relationship between gender and attrition has been included in only a few special
education studies, and the findings are mixed. Boe, Bobbitt, Cook, Whitener, et al. (1997) did
not find a relationship between gender and attrition for a national sample of general or special
educators. Moreover no relationship between gender and turnover was found in state studies of
attrition behavior (Miller et al., 1999) and intent (Cross & Billingsley, 1994). However, in a
study of urban special educators, Morvant et al. (1995) found that male teachers are more likely
to indicate intention to leave. Singer (1993a) found that young female special educators leave
the classroom at a higher rate and return at the same rate as their male counterparts.
Inconsistent findings may be due to differences in the methods and samples used as well as
changes in the workforce over time. For example, Singer's data (1993a) was from a database
covering 1972 to 1983, while the more recent findings (Boe, Bobbitt, Cook, Whitener, et al.,
1997; Miller et al., 1999) reflect teachers over a decade later. As Darling-Hammond and Sclan
(1996) point out, in the previous era, younger women were more likely to leave than men or
older women. Women's labor force participation patterns now more closely resemble men's.
Race. No differences in attrition behavior were found between teachers of different races in a
recent national study of special educators (Boe, Bobbitt, Cook, Whitener, et al., 1997), as well as
studies in Florida (Miller et al., 1999), North Carolina and Michigan (Singer, 1993a). In Virginia
Cross and Billingsley (1994) found that Whites were more likely to stay but their study focused
on intent, not actual behavior.
In one study of special educators working in an urban system, a higher proportion of European
American teachers left than African American teachers (Billingsley et al., 1995). This finding is
consistent with Dworkin's (1980) findings that White faculty, particularly women, were more
likely to want to quit urban school positions than Black or Hispanic faculty.
Personal Factors. Personal finances and perceived opportunities may influence whether
teachers stay or leave. Special educators who were primary "breadwinners" were more likely to
stay than those who were not (Westling & Whitten, 1996). However, Billingsley and Cross
(1992) did not find any differences between "breadwinner" status and intent to stay or leave.
Special educators perceiving the likelihood of finding non-teaching positions plan to teach for
shorter periods than those perceiving fewer non-teaching opportunities (Cross & Billingsley,
1994; Singh & Billingsley, 1996). Cross and Billingsley suggest that teachers who had higher
levels of education, less experience, and belonged to a minority group were more likely to intend
to leave because of better career alternatives outside of education.
In several qualitative studies, teachers indicate that personal reasons unrelated to work
contributed to their decisions to leave (Billingsley et al., 1993; Billingsley et al., 1995; Brownell,
Smith, McNellis, & Miller, 1997; Morvant et al., 1995). In a study of 99 teachers who exited an
urban school system, Billingsley et al. (1995) found that 37% of the special educators compared
to 53% of general educators, left primarily for personal reasons, e.g., family or person move,
pregnancy/child-rearing, health, retirements. Personal reasons were usually given as first reasons
for leaving and appear to be "pivotal" in decisions to leave.
Teacher qualifications have received less attention in the special education attrition literature
than any other area. Most of the special education attrition studies include relatively easy-to-
obtain measures that are sometimes assumed to be basic indices of quality (e.g., certification
status, degrees earned, performance on tests, and experience). Because it is difficult to find
consensus on what teacher "quality" means (Blanton et al., 2002), the selection of any measure
will likely be controversial. Variables, e.g., the nature of preservice experiences, student
teaching, and teacher skill or efficacy have rarely been addressed in the majority of special
education attrition reports.
Certification. There is some evidence that links certification status to special education teacher
attrition. In a study of over 1,000 Florida special educators, Miller et al. (1999) reported a higher
level of attrition among uncertified teachers than certified teachers. In their logit analyses, Miller
and colleagues found that certification is a predictor for exit, but not transfer attrition. In Boe,
Bobbitt, Cook, Barkanic, and Maislin (1999), being uncertified was associated with a higher
level of transfer. Boe, Bobbitt, Cook, Whitener, et al. (1997) reported that higher levels of
turnover were associated with teachers who are not fully certified in their main assignment when
general and special educator samples were combined, but not for general and special education
samples separately. (This may be due to an inadequate sample size of special educators).
Certification has also been linked with intent to leave. Carlson and Billingsley (2001) reported
that uncertified special educators were more likely to indicate plans to leave as soon as possible
than their certified counterparts. Because two large-scale studies (Miller et al., 1999; Carlson &
Billingsley, 2001) found a relationship between lack of certification and higher levels of attrition,
special education teachers on provisional or emergency certificates should be considered at high
risk of leaving and in particular need of support.
Academic Ability, Degrees Earned, and Teacher Preparation. Few studies address the
relationship of attrition to academic ability, degrees earned, or the quality of teacher preparation,
so few conclusions can be drawn. Most likely, the strongest link is between attrition and
performance on standardized tests. For example, Singer (1993a) found that teachers with higher
National Teacher Exam scores were twice as likely to leave than those with lower scores.
Similarly, Frank and Keith (1984) found that special educators who were more academically able
(as measured by the Scholastic Aptitude Test) were more likely to leave teaching than those with
lower academic performances. Although these measures are not indicators of teaching
competence, it is of concern that teachers with higher tested ability are leaving the field.
None of the attrition studies relate level of academic degrees to leaving, moving, or exiting.
However, in two studies of intent to leave, those with more training were more likely to indicate
they intended to leave (Cross & Billingsley, 1994; Westling & Whitten, 1996). Cross and
Billingsley state that those teachers with higher degrees perceive greater employability in non-
teaching positions and therefore are more likely to leave.
Some studies include variables such as perceived preparedness (George, George, Gersten, &
Grosenick, 1995; Miller et al., 1999), ranking of own abilities (Westling & Whitten, 1996),
ratings of competence (George et al., 1995), and self-efficacy (Brownell, Smith, McNellis, &
Lenk, 1994-1995; Miller et al., 1999). However, neither perceived preparedness nor self-
rankings have been related to attrition and retention. In a study of Florida teachers, Miller et al.
(1999) did not find that self-efficacy was significantly related to teacher attrition. Although no
relationship was found between attrition and teacher quality, it is possible that the measures used
in some of these studies may not be adequate to detect differences.
Few conclusions can be made about the relationships between teacher preparation, quality, and
attrition. Although we do not have data to support a relationship between teacher quality and
attrition in special education, Darling-Hammond (1999) argues convincingly that if teachers are
well-prepared in both content and pedagogy, "it makes an enormous difference not only to their
effectiveness in the classroom, but also whether they're likely to enter and stay in teaching" (p.
16). She further states that better preparation increases career longevity. She states that it is
"more expensive to under-prepare people, and then let them spin out again, than it is to prepare
people more effectively and keep them in the profession" (p. 17). She further states, that the
quality of preparation and support is "as integral to the task as the development of incentives to
boost up the supply of people coming in" (p. 18).
District, school, and classroom environments define many aspects of a teacher's work life,
including the salary and benefits received, nature of the community and student population,
physical facilities, role expectations, as well as the professional communities in which teachers
live out their everyday work lives. In a national study of teachers, Ingersoll (2001) states that:
The data show that, in particular, low salaries, inadequate support from the school
administration, student discipline problems, and limited faculty input into school decision-
making all contribute to higher rates of turnover, after controlling for the characteristics of
both teachers and schools. (p. 7)
Overall, the special education attrition and retention research shows that work environments are
important to teachers' job satisfaction and subsequent career decisions. Researchers in attrition
and retention define work environments in a range of different ways, use both broad and
narrowly defined variables, define similarly named variables differently, and use a range of
analytic approaches to investigate the relationships between work-related variables and attrition.
This section addresses the relationship of attrition to specific work environment variables, e.g.,
salary, school climate, administrative support, colleague support, support through induction and
mentoring, professional development, teacher roles, paper work, and students and case load
Salary. Several special education studies suggest that salary is related to turnover. In three
studies, researchers looked at the salaries earned by teachers who actually left and those who
stayed. Boe, Bobbitt, Cook, Whitener, et al. (1997) report that for a national sample of special
and general educators, moving and leaving decreased as salary increased. Both Miller et al.
(1999) and Singer (1993a) also found that special educators with higher-paying jobs were more
likely to stay than those with lower-paying jobs. Billingsley et al. (1995) reported that 10% of
those who left an urban setting gave salary as one of the primary reasons for leaving their
In summary, researchers found that salary is clearly associated with attrition behavior in special
education. Henke, Choy, Chen, Geis, and Alt (1997) suggest that compensation is an important
consideration for current teachers weighing the "tangible and intangible costs and benefits of
remaining in the teaching field or in a particular district or school" (VI-1). Given the consistent
findings in three studies of attrition behavior among special educators, salary should be a
strategy that school systems consider to increase retention. However, there are equity
implications. As Henke et al. (1997) point out, districts and schools that cannot offer
competitive salaries are likely to be at a serious disadvantage when it comes to hiring and
retaining teachers. It is interesting to note that over half of the nation's large districts use
financial incentives, such as cash bonuses or placement on a higher step of the salary schedule,
as a recruitment strategy (Carlson & Billingsley, 2001). These bonuses help attract teachers, but
the effects of these bonuses on the retention of teachers beyond the negotiated period remains a
School Climate. One of the broadest work environment variables included in the special
education attrition literature is school climate. Two large-scale studies (A High-Quality Teacher
for Every Classroom, Study of Personnel Needs in Special Education [SPeNSE], 2002; Miller et
al., 1999) suggest that teachers who rate school climate positively are more likely to stay than
those who have less positive views of school climate. Although both found climate important to
retention, they measure climate differently. Miller and colleagues define school climate in a
three-item scale (e.g., "the morale of the staff in my current school is good" (p. 207). In the
SPeNSE study, entitled A High-Quality Teacher for Every Classroom, school climate is
measured by a scale that includes a range of items such as:
school administrative behavior is supportive and encouraging
necessary materials are available when you need them
most of your colleagues share your beliefs and values about what the central mission of
the school should be
there is a great deal of cooperation among staff members
this school is a safe place for students
you feel included in the school.
In summary, school climate is clearly linked to attrition and the Miller et al. (1999) study looked
at actual leavers. These two studies suggest that when teachers have overall positive feelings
about the climate of their schools, they are more likely to stay in teaching. Many attrition
researchers attempt to separate various work-related influences; this is difficult because these
influences are inextricably linked. The climate variable is important, in essence, researchers are
asking, "Overall, is your school/district a good place to work?"
Administrative Support. Research suggests that teachers are more likely to leave teaching or
indicate intent to leave without adequate support from administrators and colleagues. In a
national study, Boe et al. (1999) reported that teachers who stay in their positions are almost four
times more likely to strongly perceive administrators' behavior as supportive and encouraging
than leavers. Miller et al. (1999) also found that perceived support from building administrators
was significantly related to attrition behavior. Research on intent supports these findings.
George et al. (1995) found that when teachers of students with emotional disorders perceive
supervisory support as "adequate" or "more than adequate," there is a greater likelihood that they
plan to remain in the field. Special and general educators who report higher levels of principal
support are more likely to be less stressed, more satisfied with their jobs, and more committed to
their employing school divisions than those receiving less support (Billingsley & Cross, 1992).
Westling and Whitten (1996) found that teachers who plan to stay are more likely than leavers to
indicate that they receive support from school administrators for inclusion, program
enhancement, students taught, and problem solving. In a study of incentives to teach in special
education, Schnorr (1995) reports that the top-rated incentive was a supportive principal (88%).
The relationship between central office administration and attrition are mixed. In a study of
intent to leave, Gersten et al. (2001) found that central office administrators exert an indirect
influence on attrition through professional development opportunities and stress (Gersten et al.,
2001; see Figure 4). However, Miller et al. (1999) did not find a relationship between principal
support and attrition in their study of attrition behavior. It is likely that the different dependent
measures may have influenced these differences or the different analyses employed.
In a study of urban teachers, special educators indicate dissatisfaction with central office
administrators as frequently or more frequently than with principals (Billingsley et al., 1995).
Billingsley et al. (1995) states that 25% of those who left teaching in an urban setting identify
dissatisfaction with support from central administration and 20% indicate that dissatisfaction
with principal support influenced their decision to leave. In contrast, general educators were less
likely to report dissatisfaction with support from central administrators (10%) and principals
(12%). The finding that central office administrators play a critical role in the lives of special
educators is not surprising, given that district administrators usually have a key role in
determining special education policies, regulating IDEA requirements, and identifying and
placing students with disabilities.
Recent path analyses provide a better understanding of how administrative support influences
intent to leave, through other important mediating variables, e.g., job satisfaction stress, and
commitment. More specifically, in all three path studies, a higher level of support from
principals is directly or indirectly associated with more professional development opportunities
(Gersten et al., 2001), fewer role problems, greater job satisfaction, reduced stress, and higher
levels of commitment (Cross & Billingsley, 1994; Gersten, 2001; Singh & Billingsley, 1996).
(Gersten et al. included both principals and teachers in their support variable.)
Defining support is difficult, because it must be comprehensive in nature and varied in type
(Gold, 1996). In this sense, support is a global construct that has many dimensions. Littrell and
colleagues (1994) found that emotional support (e.g., showing appreciation, taking an interest in
teachers' work, maintaining open communication) is perceived as most important to special
educators. They also found that emotional and instructional support (providing needed materials,
space, and resources) correlate positively with both job satisfaction and school commitment.
The fact that Littrell and colleagues did not find a significant relationship between administrative
support and intent to leave is consistent with the path analysis studies described above.
Administrative support likely influences attrition through other key mediating variables, e.g.,
role problems, stress, job satisfaction, commitment, and professional development.
Site = Silver City
R2 for intent to stay = .24
X2 = 30.0 df 18
Adjusted goodness of fit = .93
Root mean square error of approximation
Figure 4. Path diagram for intent to stay
(Gersten et al., 1999).
Colleague Support. Although administrative support is prominent in the special education
attrition literature, less attention has been given to the relationship of colleague support and
attrition, and the findings from these studies are mixed. Miller et al. (1999) in a large-scale study
found that a lower level of colleague support was associated with leaving behavior and a higher
level with staying. Although George et al. (1995) found that about one-fourth of teachers of
students with behavioral disorders indicated that support from classroom teachers was "totally
inadequate," collegial support did not discriminate between those who intended to stay and those
who intended to leave. In a study of 99 teachers who left an urban setting, 4 leavers indicated
that problems with colleagues contributed to their decisions to leave (Billingsley et al., 1995). In
an open-ended study of 42 teachers who left their positions, none of the respondents identified
colleague factors as contributing to their decisions to leave (Billingsley et al., 1993). Reasons
for these differences may have to do with how teachers responded (e.g., open-ended survey
versus questionnaire items), measuring intent versus leaving behavior, and large differences in
sample sizes for these studies.
Although most of the special education attrition studies to date focus on the role of the
administrator in supporting teachers, Singh and Billingsley (1998) suggest that the principal
enhances commitment through fostering a collegial environment. They suggest that principals
who share goals, values, and professional growth foster supportive and collegial learning
communities. Gersten et al. (2001) suggest that it makes more sense to examine building-level
support as the "cumulative impact" of the building principal, assistant principal, and fellow
teachers at the school than to examine support from the building principal separately.
Although administrators clearly have important roles in supporting teachers, it is limiting to think
of support as something that one person provides and another receives. Important to creating a
positive school climate is reciprocity of support among special and general educators,
administrators, parents, paraprofessionals, and other service providers.
Support through Induction and Mentoring. Focusing on the support needs of early
teachers is important because teachers are at risk of leaving during these years. A large body of
literature in general education suggests that the optimism that beginners bring to their work is
often replaced with disappointment, discouragement, and disillusionment (Gold, 1996). Novice
teachers struggle with a range of problems, e.g., discipline, parent difficulties, insufficient
support, apathy from colleagues, and problems with student behavior (Gold, 1996; Veenman,
1984). Special educators, like general educators, must engage in educational planning,
understand the curriculum, and become familiar with school routines. Special educators have
numerous additional responsibilities and concerns related to working with students with
significant learning and behavioral problems. A few qualitative studies have documented the
concerns experienced by beginning special educators, e.g., managing paper work, making
accommodations for instruction and testing, developing and monitoring IEPs, scheduling
students, and collaborating with teachers, paraprofessionals, parents, and related services
personnel (Billingsley & Tomchin, 1992; Boyer & Gillespie, 2000; Kilgore & Griffin, 1998;
MacDonald, 2001; Magliaro & Wildman, 1990).
It is critical that teachers obtain support during the early stages of their careers when they are
most likely to leave. Although early career teachers are at risk of leaving, only two special
education attrition studies specifically report on the relationship between induction experiences
and attrition. Whitaker (2000) investigated what beginning special educators perceive as
effective mentoring programs and examined the impact of such programs on their plans to
remain in special education. Although the effect size is small, perceived effectiveness of
mentoring is significantly correlated with teachers' plans to remain in special education and with
special educators' job satisfaction. Billingsley (2002a) did not find the level and helpfulness of
induction support provided to beginning teachers to be significantly related to their plans to stay;
however, those with higher levels of induction support were more likely than those with lower
levels of support to see their roles as manageable, believe that they can get through to the most
difficult students, and believe they are successful in providing education to students with IEPs.
Billingsley and Whitaker used different types of measures and populations, which may account
for conflicting results on the relationship between induction support and career intent. For
example, Billingsley (2002a) investigated intent to stay over an entire career span, while
Whitaker (2000) looked at plans to remain in or leave special education for the following school
year and in the next five years.
An important contribution of Whitaker's (2000) study is the identification of specific aspects of
effective mentoring, which includes selecting a special education mentor (as opposed to a non-
special educator), even if that special educator works in a different school. Assistance provided
in the area of emotional support and the mechanics of the job are particularly important.
Billingsley (2002a) and Whitaker (2000) also found that more informal contacts were perceived
as more effective than formal mentor programs.
Carefully designed induction programs can help teachers cope with these challenging tasks
(Gold, 1996; Rosenberg, Griffin, Kilgore, & Carpenter, 1997; Zepeda & Ponticell, 1997).
Novice teachers who are given reasonable assignments, adequate feedback, and personal support
are more likely to acquire the skills needed for a satisfying teaching career and to develop greater
commitment to teaching (Yee, 1990).
Reducing attrition should not be the primary goal of induction programs; indeed, carefully
designed induction programs would be needed even if a teacher surplus existed. Induction
programs must be designed with the primary purpose of helping teachers become more effective
in supportive and carefully designed environments. If this is the primary goal and teachers
develop competence and satisfaction in their work, attrition will likely be reduced. A model for
providing individualized support for special educators proposed by Rosenberg and colleagues
(1997) is an excellent resource for those developing induction programs. Their model considers
both the "full series of dynamic interacting factors that are specific to the individual and specific
to the teaching environment and tasks that are to be accomplished" (p. 301).
Professional Development. Professional development may be thought of as one dimension
within the broad concept of support. Several researchers found a relationship between
professional growth opportunities and attrition (Brownell et al., 1994-1995; Gersten et al., 2001;
Morvant et al., 1995). Miller et al. (1999) did not find that satisfaction with professional
opportunities related to the attrition behavior of special educators. In a study of teacher attrition
in three urban systems, Gersten et al. (2001) found that professional development has an indirect
effect on teachers' intent to leave and a direct influence on teachers' commitment to the
profession. In the Gersten et al. study, professional development opportunities referred to the
degree to which special educators perceive that they have opportunities to grow and advance
professionally. "This scale measures satisfaction with items such as opportunities to learn new
techniques and strategies and opportunities for professional advancement and promotion (pp.
555-556). Teachers who perceive greater professional development opportunities also
experienced less role dissonance. Findings from the Gersten et al. study suggest the importance
of the support at both the district and school levels. It is interesting to note that over half of those
surveyed in this urban study did not feel there were many opportunities to learn new techniques
and strategies in their district (Morvant et al., 1995).
Brownell et al. (1994-1995) report in their interview study that stayers are more likely to assume
at least some responsibility for their own professional development and to initiate actions to
continue their own learning. They also report that only stayers discuss the importance of
university training to their professional development. Brownell and colleagues suggest that the
degree of satisfaction with professional development opportunities may be influenced by the
content, timing, and quality of the opportunities as well as incentives for participating.
Special educators devote considerable time to continuing professional development, averaging
59 hours in 1999-2000 as reported by A High-Quality Teacher for Every Classroom, (SPeNSE,
2002). Although school districts support staff development, they do not reliably incorporate
best practices, e.g., engaging teachers in the learning process, allowing time to plan now to
implement new skills. Teachers want to continue to learn and grow throughout their careers
(Rosenholtz, 1989), and it is ironic that schools are not better organized to promote teachers'
professional development (Rosenberg et al., 1997). Quiocho and Rios (2000) discuss the need
for responsive professional development opportunities for teachers and suggest that they occur in
"qualitatively different ways than are experienced in teacher credential programs. They must be
directed toward professional nurturing, be systematic, and change over time as professional
needs change." (p. 522)
Teacher Roles. Problems with role overload and design have been strongly linked to special
education attrition, as much or more than any other work-related factor. Regardless of whether
quantitative or qualitative methods are used, research results provide convincing evidence that
role problems significantly interfere with special educators' ability to be effective with their
students and job satisfaction.
The specific role-related problems are not single or simple problems. As Billingsley et al. (1995)
state, "Multiple problems interact and create what teachers sometimes view as stressful,
overwhelming work situations" (p. 7.21). As one special educator pointed out, her paper work
increases every time she receives another student (Billingsley et al., 1995). Corcoran, Walker,
and White (1988) also suggest that the lack of resources increases the teacher's work load. And
a heavy work load makes it very difficult to use the available resources. Many of the work-
related problems identified in these attrition and retention studies are similar to the concerns
expressed by special educators in the recent report on working conditions, Bright Futures for
Exceptional Learners: An Action Agenda to Achieve Quality Conditions for Teaching and
Learning (CEC, 2000).
Role problems. A number of researchers (Billingsley & Cross, 1992; Cross & Billingsley,
1994; Gersten et al., 2001; Singh & Billingsley, 1996) report a relationship between different
types of role problems (e.g., role problems, role overload, role conflict and ambiguity, role
dissonance) and intent to leave teaching. Work load manageability has also been included as a
measure in some attrition studies. Teachers who plan to leave special education teaching as soon
as possible are significantly more likely to rate their work load as "not at all manageable"
compared to those who indicate intent to stay (Carlson & Billingsley, 2001). Morvant et al.
(1995) found that only half of the special educators in their study agreed that their work load was
manageable, 68% felt they had too little time to do their work, and almost one-third found
conflicting goals, expectations, and directives a frequent source of stress (pp. 4-29). However,
Miller et al. (1999) did not find that role conflict or manageability had an effect on leaving
behavior. Miller and colleagues studied teachers in Florida where the school psychologist and
counselors handle much of the testing and paper work burden.
Research results suggest that role problems create stress and decrease job satisfaction (Cross &
Billingsley, 1994; Gersten et al., 2001). An important finding of these studies is that teachers
who perceive greater principal support also perceive fewer role problems than those receiving
less support; the Gersten et al. variable included support from both principals and teachers.
Qualitative studies also support the relationship between role problems and attrition. In open-
ended interviews, Billingsley et al. (1993) report that special educators gave job-design factors
(e.g., lack of time, resources, paper work, excessive meetings) as a reason for leaving more often
than any other factor.
Other job-design factors have been identified with teachers' plans to stay. Westling and Whitten
(1996) identified a number of role-related items associated with teachers' plans to stay: teacher
responsibilities are clearly defined, teacher is never directed to complete tasks not considered to
be their responsibility, teacher has adequate time to complete paper work, teacher has adequate
time to plan instruction and prepare materials, the job has clear program goals, teacher agrees
with program goals for job. To improve retention, more attention must be paid to role design.
As Gersten et al. (2001) asks: "Does the job, with all it entails, make sense? Is it feasible? Is it
one that well-trained, interested, special educational professionals can manage in order to
accomplish their main objective-enhancing students' academic, social, and vocational
competence?" (p. 551).
Paper work. Paper work is a major contributor to role overload and conflict. Recent studies
consistently identify paper work as a problem that contributes to teacher attrition (e.g.,
Billingsley et al., 1993; Billingsley et al., 1995; Brownell et al., 1994-1995; George et al., 1995;
Morvant et al., 1995; Schnorr, 1995; Westling & Whitten, 1996). In particular, findings from the
largest study investigating paper work, Paperwork in Special Education (SPeNSE, 2002) suggest
that paper work problems are significantly related to special educators' intent to leave teaching,
after many other work condition variables are controlled. According to Westling and Whitten
(1996), those planning to leave rated the item "has adequate time to complete paper work"
significantly lower than those planning to stay. In a study of Alaska teachers, Schnorr (1995)
found that 71% of respondents indicated that paper work is a major deterrent to special education
teaching. Qualitative results support these findings. In open-ended interviews, Billingsley et al.
(1995) found that 60% of special educators who plan to leave teaching in an urban district gave
paper work as a major contributor to their decision. Billingsley et al. (1995) emphasized in their
three-year study, "Whenever teachers were given an open-ended opportunity to express
concerns, paper work was sure to emerge as one of their greatest frustrations" (p. 7.14).
The extent of paper work is also significantly related to the overall manageability of special
educators' jobs. The typical special education teacher reports spending five hours per week
completing forms and doing administrative paper work. This is as much time as they spend
preparing for lessons. Over half of the special educators report that routine duties and paper
work interfere with their teaching to a "great extent." General educators were significantly less
likely to indicate that routine duties and paper work interfered with their teaching in Paperwork
in Special Education (SPeNSE, 2002). Billingsley et al. (1995) also reported that general
educators are less likely than special educators to view paper work as a problem: 35% of the
special educators and 12% of the general educators cited paper work as one of their most
Understanding what teachers mean by excessive paper work is highlighted in some of the
qualitative attrition studies. Special educators describe paper work as overwhelming,
unnecessary, redundant, and intimidating (Billingsley et al., 1995). Some interviewees said that
they did not have the time to complete required paper work, that there was too much pressure to
complete paper work, and that paper work requirements were inconsistent or unnecessary
(Billingsley et al., 1995). Morvant et al. (1995) provided an in-depth discussion of the paper
work problem for special educators. As one leaver described:
You don't only have to test 'em. You have to write up your results. But, before you ever do
it, you have to get all these permission forms signed and all the referrals and the request for
services-and the paperwork...gets worse every year. And then test, write up the results, get
all the paperwork ready for the first conference, notify all the other people that have to sit in
on that. And then you have your professional conference, and then you have to have another
one where the parent comes. And it just goes on and on. And you have paperwork for every
one of these conferences. (Morvant et al., 1995, p. 3-14)
Although paper work is clearly a problem for many special educators, not all leavers view paper
work as contributing to their decisions to leave. Why some teachers see paper work as a major
obstacle and others do not cannot be answered in this review. It is likely that different state,
district, and/or school practices influence the paper work responsibilities that teachers are given.
In some systems, school psychologists may have a larger share of the responsibility for testing
and identification, reducing responsibilities for teachers. In another district, one of several
teachers in a school may serve as a school coordinator and have a disproportionate share of the
paper work burden. Still, other teachers may have found effective ways of keeping their paper
work burdens to a minimum.
Service-delivery models and shifts in responsibility. No conclusions can be drawn
about the relationship between teaching in a particular service-delivery model (e.g., resource,
self-contained, inclusion) and attrition. George et al. (1995) found that self-contained teachers of
students with emotional disorders plan to leave the field significantly more often than resource
room teachers. However, as districts have moved toward greater inclusion, some special
educators struggle with changing roles and lack support for these new roles. For example, in
interviews with leavers, Morvant et al. (1995) state, "Teachers indicated a desire to spend more
time providing direct instructional services to students and less time coordinating with classroom
teachers and serving essentially as 'case managers' of students' schedules and programs" (p. 3-
26). Some of these leavers related the problems they had in implementing inclusive programs.
Teachers indicated the loss of paraprofessionals (who were reassigned to general education
classrooms), the increased complexity of scheduling students, and the expectation for meaningful
teacher-to-teacher collaboration without planning time. As one teacher states:
It's an idealistic situation, and I can't see it working. If I have children from three different
classrooms and two different grades, they can come together and form a group. And I can
work with them in a short time period and accomplish something. But now, with [the new
model] I am supposed to go to these individual rooms. Now where in an hour can I go to
three different rooms and accomplish anything? (p. 3-13)
In an investigation of teacher burnout, Embich (2001) concludes that teachers who work
primarily in general education classrooms are at more of a risk of burnout than teachers who
work in more traditional settings (e.g., resource, self-contained classrooms). Embich states that
responsibilities of those who team teach have expanded and include a wider range of services
(e.g., teaching, work in regular classrooms, collaboration) than those working in resource or self-
contained models. These team teachers are often involved in working where they are not wanted
and in areas in which they have had little preparation.
As Morvant et al. (1995) suggest, the move toward inclusion may contribute to role dissonance
or conflict for some special educators. If special educators have beliefs that differ from the
philosophy of the school, they may seek other positions. Special educators who find it difficult
to implement an inclusive program because of inadequate support systems or resistance from
general educators may also find their work unfulfilling and look elsewhere. Administrators
need to be particularly aware of the support needs of teachers as they make changes in their roles
Students and Case Load Issues. In a large-scale study of teachers between 1972 and 1983,
Singer (1993a) found that special educators in secondary schools stay an average of 1.6 years
less than their colleagues in elementary schools. Additionally, teacher attrition by disability area
varied. Teachers of students with learning disabilities, physical/multiple disabilities, and mental
retardation were the least likely to leave teaching. Those working with students with emotional
problems were somewhat more likely to leave. Teachers working with students with speech,
hearing, or vision impairments were the most likely to leave, perhaps because of more
opportunities outside of education (Singer, 1993a). As some states move toward non-categorical
and inclusive programs, such comparisons across disability may make less sense today. A recent
national study of special educators revealed that 80% of teachers work with students with two or
more exceptionalities, and 32% of teachers work with students with four or more different
primary disabilities reported in A High-Quality Teacherfor Every Classroom (SPeNSE, 2002).
No empirical studies have shown a relationship between the number of students on teachers' case
loads and attrition. In two large-scale studies, no relationship is reported between the number of
students taught and intent to leave (Carlson & Billingsley, 2001). George et al. (1995) also
found no relationship between size of case load and teachers' intent to leave in a study of
teachers who taught students with emotional disorders.
Although there have not been any clear quantitative findings relating attrition to numbers of
students taught, in qualitative studies, teachers consistently report problems with case load and
give case load issues as reasons for leaving (Billingsley et al., 1993; Billingsley et al., 1995;
Brownell et al., 1994-1995; Morvant et al., 1995; Schnorr, 1995). In a study of urban special
educators who left teaching because of dissatisfaction, (Billingsley et al., 1995), 33% indicated
"class size/case load too large" and 25% indicated "inappropriate placement of students with
disabilities" (25%). Teachers indicate that it is not simply the number of students, but rather the
diversity of case loads that is problematic for teachers as well. In a recent national study of
special educators, Carlson and Billingsley (2001) reported that teachers who planned to leave as
soon as possible were significantly more likely to teach students with four or more different
primary disabilities (42%) compared to special educators planning to stay.
Related to case load issues are the problems that teachers encounter with students. Researchers
link a range of student-related problems to attrition, such as discipline problems (Billingsley et
al., 1993; Billingsley et al., 1995; Brownell et al., 1994-1995; Brownell et al., 1997), attitudes of
students (George et al., 1995), lack of student progress (Billingsley et al., 1995; Brownell et al.,
1994-1995), safety issues (Billingsley et al., 1995; Brownell et al., 1997), and diversity of
student needs (Billingsley et al., 1995; Brownell et al., 1997). In a review of general education
teachers who depart because of low job satisfaction, Ingersoll (2001) found that a quarter or
more of the dissatisfied teachers who leave indicate that student discipline problems and lack of
student motivation are two primary reasons. However, few conclusions can be drawn about the
extent to which student issues contribute to the attrition problem in special education. Miller et
al. (1999) did not find a link between satisfaction with student relationships and attrition.
Billingsley et al. (1995) found that student issues were a less important factor in attrition than
other types of problems, e.g., inadequate administrative support, case load, and role problems.
Westling and Whitten (1996) stated that, "Teachers who planned to leave were not doing so
because of the students they were teaching or the type or severity of their disabilities" (p. 330).
Morvant et al. (1995) suggested that 83% of teachers felt satisfied with their accomplishments
with students, 85% felt they were making a significant difference in their students' lives, and
96% enjoyed their students.
Affective Responses to Work
Excessive and prolonged work problems lead to negative affective reactions, such as increased
stress, lower job satisfaction, and reduced organizational and professional commitment. The
combination of multiple, interacting work-related problems (e.g., too many students, too much
paper work, too little support, and the lack of needed resources) clearly weakens the teacher's
ability to be effective and therefore reduces their opportunities for the positive intrinsic rewards
that are important to teachers (Billingsley et al., 1995).
Stress. Stress was one of the most powerful predictors of attrition behavior in a large-scale
Florida study (Miller et al., 1999). Other studies found that stress is related to intent to leave
(Billingsley & Cross, 1992; Cross & Billingsley, 1994; Gersten et al., 2001; Morvant et al.,
1995; Schnorr, 1995; Singh & Billingsley, 1996). Morvant et al. (1995) stated that almost 80%
of those who planned to leave indicated that they felt under a great deal of stress on a weekly or
daily basis, compared to just over half of stayers. They also report that leavers indicated
significantly more frequent stress than stayers due to: (1) the range of students' needs and
abilities; (2) bureaucratic requirements; and (3) conflicting expectations, goals, and directives.
Researchers have studied the effects of stress and burnout among special educators for over two
decades (Banks & Necco, 1987; Cooley & Yovanoff, 1996; Embich, 2001; Greer & Greer, 1992;
Weber & Toffler, 1989; Weiskopf, 1980; Wisniewski & Gargiulo, 1997; Zabel & Zabel, 1982;
Zabel & Zabel, 2001). Most researchers identify stress as a major contributing factor in burnout.
Burnout refers to exhaustion, powerlessness, and depersonalization in response to chronic work
stress (Maslach, 1982). Not all burned-out employees leave. Of particular concern is the quality
of education delivered to students by teachers experiencing burnout (Weber & Toffler, 1989).
The relationship between the stress/burnout literature and the attrition literature needs
integration. Some of the factors that influence burnout are also associated with attrition (e.g.,
age, support systems).
Moreover, many of the strategies designed to reduce burnout have also been suggested to
improve retention (e.g., administrative and collegial support). Cooley and Yovanoff (1996)
found that a stress management workshop and a peer collaboration program improved teachers'
job satisfaction, burnout, and organizational commitment. They taught stress management
techniques (e.g., analyzing physiological responses, changing destructive thinking, and learning
about techniques for changing a specific situation) to special educators and related service
providers. Participants also learned about a collegial dialogue process to help participants
improve their ability to collaboratively identify and solve problems. Both teachers and service
providers who participated in both treatment groups outperformed control groups on job
satisfaction, burnout control, and organizational commitment.
Job Satisfaction. Increasing teachers' job satisfaction is one of the most important ways to
reduce attrition, because job satisfaction and attrition are strongly linked in studies of career
intentions (e.g., Billingsley & Cross, 1992; Billingsley et al, 1995; Brownell et al., 1997; Cross
& Billingsley, 1994; Gersten et al., 2001; Singh & Billingsley, 1996; Westling & Whitten, 1996;
Whitaker, 2000). Gersten et al. (2001) found that "satisfaction and personal assessment of
rewards" reflects greater differences between those intending to stay and leave than other factors.
Morvant et al. (1995) reported that stayers rated their perceived effectiveness in making a
significant difference in the lives of their students significantly higher than leavers. Miller et al.
(1999) did not find that satisfaction was significantly related to attrition behavior, which may be
accounted for by different measures used, differences in the samples, or study of leavers vs.
those who intended to leave.
The path models described earlier (Cross & Billingsley, 1994; Gersten et al., 2001; Singh, &
Billingsley, 1996) show how different work conditions influence job satisfaction. Paying
attention to creating supportive relationships with teachers and principals, reducing stress,
clarifying roles, and providing professional support should help teachers derive more satisfaction
from their work.
Commitment. Several special education studies suggest that teachers with higher levels of
professional and organizational commitment are more likely to stay (Miller et al., 1999) or intend
to stay in teaching (Billingsley & Cross, 1992; Gersten et al., 2001; Littrell et al., 1994).
Commitment was defined by Mowday, Porter, and Steers (1982) as an attitude that includes: (a)
strong belief in and acceptance of an organization's/profession's goals and values, (b)
willingness to exert significant effort on behalf of the organization/profession, and (c) strong
desire to maintain membership in organization/profession. Billingsley (1993) suggests several
types of commitment, including commitment to school, district, teaching field, and teaching
profession. Teachers may be committed to special education teaching, not necessarily the
employing district or assigned school.
In special education, higher levels of commitment are associated with leadership support
(Billingsley & Cross, 1992; Cross & Billingsley, 1994; Gersten et al., 2001; Littrell et al., 1994);
fewer role problems, e.g., conflict, overload, dissonance (Billingsley & Cross, 1992; Cross &
Billingsley, 1994; Gersten et al., 2001; Singh & Billingsley, 1996); lower levels of stress (Cross
& Billingsley, 1994; Gersten et al., 2001; Singh & Billingsley, 1996); more teaching experience
(Cross & Billingsley, 1994); and higher levels of job satisfaction (Billingsley & Cross, 1992;
Cross & Billingsley, 1994; Gersten et al., 2001; Littrell et al., 1994). The degree to which initial
commitment contributes to subsequent career decisions (Brownell et al., 1994-1995; Billingsley,
1993; Chapman & Green, 1986) deserves further consideration.
Summary of Research Findings in Attrition and Retention
A decade of research shows that teacher and work factors are critical to special educators' job
satisfaction and their subsequent career decisions. Attrition researchers have identified several
key teacher characteristics and personal factors that influence teachers' decisions to leave special
education: (1) there are consistent reports that younger and inexperienced special educators are
more likely to leave than their older, more experienced counterparts; (2) there is support from
two major studies that uncertified teachers are more likely to leave than certified teachers; (3)
special educators with higher test scores are more likely to leave; and (4) teachers' personal
circumstances and priorities influence attrition and retention.
The majority of attrition studies focus on the effects of district and school working conditions,
work assignment factors, and teachers' affective reactions to their work. Work Environment
factors associated with staying include: (1) higher salaries; (2) positive school climate; (3)
adequate support systems, particularly principal and central office support; (4) opportunities for
professional development; (5) reasonable role demand; and (6) manageable case loads.
Problematic district and school factors-especially low salaries, poor climate, lack of
administrative support, and role overload and dissonance-lead to negative affective reactions,
e.g., high levels of stress, low levels of job satisfaction, and low levels of commitment. These
negative reactions can often lead to withdrawal and eventually attrition.
A CRITIQUE OF DEFINITIONS, CONCEPTUAL MODELS, AND
METHODOLOGIES USED TO STUDY ATTRITION
Compared to the exploratory literature prior to 1992, the current research base is much stronger
and provides direction to those interested in improving teacher retention. However, the current
frameworks used to study attrition and the range of definitions, samples, measures, and analysis
strategies used make it difficult to answer important questions. This section reviews and
critiques definitional, conceptual, and methodological considerations in attrition studies.
Definitions of Attrition
The schema in Figure 1 illustrate the different types of teacher attrition. Special educators
transfer to other special or general education positions (within the same or different districts) or
leave teaching altogether. The Appendix highlights the definitions of attrition that are used in
each special education study. Only a few researchers differentiate between different types of
leavers in their studies (i.e., Boe, Bobbitt, & Cook, 1997; Boe, Bobbitt, Cook, Whitener, et al.,
1997; Miller et al., 1999). Moreover, only about half of the studies include special educators
who actually left their positions. The remainder of researchers use "intent to leave" as a proxy
Note that the use of different attrition definitions depends on the focus of the study. As Grissmer
and Kirby (1987) state, there is no single definition of attrition, and it is the policy or research
context that frames the definition. This section highlights the need to study actual leavers, the
importance of studying different types of leavers, and the intent variable.
Future studies of attrition need to focus on attrition behavior (teachers who actually leave their
positions). Although half of the studies identified in the Appendix include those who left special
education (i.e., Billingsley et al., 1993; Billingsley et al., 1995; Brownell et al., 1994-1995;
Brownell et al., 1997; Boe, Bobbitt, & Cook, 1997; Boe, Bobbitt, Cook, Whitener, et al. 1997;
Morvant et al., 1995; Miller et al., 1999; Singer, 1993a), several of these studies were large-scale
studies (i.e., Boe, Bobbitt, & Cook, 1997; Boe, Bobbitt, Cook, Whitener, et al., 1997; Miller et
al., 1999; Singer, 1993a).
Of these large-scale studies of actual leavers, one of the strongest and most comprehensive is
Miller et al. (1999). These researchers surveyed a large, representative teacher sample in
Florida, included a range of teacher and work-related variables, and differentiated between types
of attrition (i.e., transfer and exit attrition). Singer's study (1993a) highlighted a number of
important findings related to demographic variables, teacher quality, and specific teaching
assignment variables. However, the database used did not allow for the study of different work
variables, and various exit categories could not be differentiated (see Figure 1). One of the
problems in using state databases is that "teachers who leave a state disappear from state
databases, giving the appearance of having left the field" (p. 377). Unless researchers use
follow-up surveys (e.g., Boe, Bobbitt, & Cook, 1997; Boe, Bobbitt, Cook, Whitener, et al., 1997;
Miller et al., 1999), distinctions between different leavers are not possible.
Boe and colleagues (e.g., Boe, Bobbitt, & Cook, 1997; Boe, Bobbitt, Cook, Whitener, et al.,
1997) have conducted the only national studies of teacher attrition behavior comparing special
and general educators. Boe and colleagues provide important information about the relationships
of certain important variables (e.g., salary, age) and attrition behavior. Moreover, they studied
both moving and exit behavior. However, the relatively small number of special educators in the
sample and the constraints of the database limit their study.
In contrast to looking at teachers' behavior, the use of intent allows researchers to assess the
relationship of teachers' career plans to a range of district and teacher variables without the
expensive and time-consuming task of finding those who left. Therefore, it is not surprising that
approximately half of the studies included in the Appendix use intent-to-leave as a proxy for
This intent variable is controversial, with some questioning whether the intent variable is related
to attrition behavior (Boe, Bobbitt, Cook, Whitener, et al., 1997; LeCompte & Dworkin, 1991).
Boe, Barkanic, et al. (1999) report that "plans to stay in the same school next year are associated
with actual staying, and plans to leave the school are associated with voluntary moving,
voluntary leaving, and involuntary leaving" (p. 12). Recently Boe (personal communication,
October, 2001) shared his observations that, for three waves of the Teacher Follow-up Survey
combined (1988-1989, 1991-1992, 1994-1995), based on national estimates of public school
teachers, 93% of teachers who said they would stay in the same school the next year actually did;
44% of teachers who said they would move to a different school actually did; and 52% of
teachers who said they would leave actually did. Future studies should consider the time frame
for looking at whether those who intend to leave actually do leave. For example, Gersten et al.
(2001) suggested that those who have criticized the intent variable consider a time frame that is
too short, one year or less. Gersten et al. (2001) report that of 33 teachers who planned to leave
within the 15-month period, 69% had actually left special education teaching.
Researchers also need to carefully consider how they ask teachers about their career plans. For
example, researchers include different time frames for leaving as well as what leaving means,
e.g., whether one plans to leave teaching, a district, or the teaching field altogether. Westling
and Whitten (1996) provide multiple-choice items to teachers asking about intent to remain in
same position/similar position in the next school year and in five years. Cross and Billingsley
(1994) asked a group of special and general educators at various stages of their careers, "Please
check which of the following comes closest to describing how long you plan to teach," followed
by five choices until retirement. Westling and Whitten (1996) address a relatively short time
frame, whereas Cross and Billingsley look at how teachers see their futures over a longer span of
time. Researchers will clearly get different kinds of information depending on how these
questions are asked. Both types of questions and measures are appropriate, depending on the
specific populations, intents, and purposes of the study. It would be logical to assume that short-
term plans will have a stronger relationship to attrition than long-term plans.
In summary, both studies of leavers and those who intend to leave provide critical knowledge
about factors related to teachers' career decisions. It is important to consider Grissmer and
Kirby's (1987) reminder that it is important to analyze and measure subgroups of teachers who
leave. The special education research base lacks information about actual leavers, which has only
been included as a variable in a few major studies of attrition. Boe, Bobbitt, Cook, Whitener et
al. (1997) and Miller et al. (1999) found differences between those who exit teaching and those
who transfer to other positions. Whenever possible, future studies should include various types
There is a positive relationship between career intentions and later decisions, although more
needs to be known about the strength of this relationship. Studying intent is important if we
consider Gold's (1996) suggestion that we develop a broader definition of retention:
One, which encompasses not only teachers' decisions to leave teaching or to stay but also the
concept of engagement or involvement in teaching. This definition suggests a corresponding
commitment to teaching that needs to be a focus of retention, not simply retaining all teachers
on the job (p. 548).
Studies of special education attrition over the last ten years have been more comprehensive than
the earlier exploratory studies, due in part to stronger conceptualizations of the factors associated
with attrition (e.g., Billingsley, 1993; Brownell & Smith, 1992; Gersten et al., 2001) and the
support offered through federal funding. Recent reports include important variables that have
not been investigated in earlier special education studies such as school climate (Carlson &
Billingsley, 2001; Miller et al., 1999), mentoring (Whitaker, 2000), manageability of work (A
High-Quality Teacher for Every Classroom, SPeNSE, 2002; Morvant et al., 1995), and self-
efficacy (Miller et al., 1999). Some of the more recent studies have also included comparisons
between general and special educators (Billingsley et al., 1995; A High-Quality Teacher for
Every Classroom, SPeNSE, 2002; Boe, Bobbitt, Cook, 1997; Boe, Bobbit, Cook, & Whitener et
al., 1997). Although the knowledge base is growing, greater attention needs to be given to the
framing of these studies, particularly models for examining turnover behavior.
An important question is the extent to which existing special education conceptual models of
attrition (e.g., Billingsley, 1993; Brownell & Smith, 1993; Gersten et al., 2001) help to frame and
interpret attrition research findings. Many existing research studies include many variables, but
few theories. Unlike most special education research, attrition research in general education
tends to be more focused and theory-driven. Gold (1996) summarizes the basics of attrition
models used in general education, e.g., social learning theory, organizational theory, career
choice theories, economic cost-benefit analysis, and human capital theory. Future studies need
to be carefully conceptualized and focused.
In general, studies conducted in the last ten years include larger samples (Carlson, 2001; Gersten
et al., 2001; Miller et al., 1999; Singer, 1993a) from more geographically diverse areas than
earlier studies. The studies listed in the Appendix include samples from a number of states (e.g.,
Alaska, Florida, North Carolina, Michigan, South Carolina, Virginia); several urban cities
(Billingsley et al., 1995; Brownell et al., 1994-1995; Gersten et al., 2001; Morvant et al., 1995);
rural settings (Westling & Whitten, 1996); and nationally representative groups of teachers (Boe,
Bobbitt, & Cook, 1997; Boe, Bobbitt, Cook, Whitener, et al., 1997; Carlson, 2001).
Interpreting special education attrition findings is problematic, because samples include
variations in the data collection time frame, and different attrition definitions lead to different
types of samples. Some samples address only a particular group of teachers, such as teachers of
students with emotional disorders (George et al., 1995) or beginning special educators
(Whitaker, 2000). Other samples are from specific states (e.g., Florida, Virginia) where state
characteristics and district practices may vary. One-third of the studies included small samples
of fewer than 100 teachers. Moreover, most studies only gathered data at one point in time. An
exception to this was the longitudinal study by Singer (1993a), which examined the largest group
of special education leavers (over 6,600 teachers from two states between 1972 and 1983).
Singer examined the characteristics of special educators who returned to teaching in a second
study (Singer, 1993b).
Quantitative Studies in Attrition Research
Most attrition studies in the Appendix involved the use of questionnaires and survey methods to
explore the range of variables associated with attrition. Researchers analyzed these data from
several approaches. One widely used approach in both general and special education studies is
to investigate bivariate relationships to determine if a particular variable (e.g., age, gender,
salary) is associated with special education attrition (e.g., George et al., 1995; Grissmer & Kirby,
1987; Morvant et al., 1995). Some of the researchers who investigated bivariate relationships
focused primarily on the relationships of a particular class of variables and turnover, such as
demographic variables (Grissmer & Kirby, 1987); others focus on a greater number of work and
demographic variables (Morvant et al., 1995; Westling & Whitten, 1996). Other researchers use
multivariate methods to investigate attrition and retention. Miller et al. (1999) and Westling &
Whitten (1996) use logit models to identify significant predictors of attrition. Miller et al.
(1999) argue that the multinomial logit analyses identify the variables that have the most direct
effect on attrition, thus allowing for more parsimonious models of attrition. Other researchers
use path models to test causal relationships among various work-related variables (e.g., support,
professional growth, role demands, commitment) believed to be important to job satisfaction,
commitment, and ultimately teacher retention (Cross & Billingsley, 1994; Gersten et al., 2001;
Singh & Billingsley, 1996). Researchers should continue to use multivariate methods, which
can demonstrate the dynamic interactions between important variables and career decisions
while controlling for effects of other variables.
Large-scale quantitative studies of attrition are also needed to investigate the relationships among
different factors included in the conceptual models as well as subgroups of teachers. For
example, little is known about attrition in rural versus urban systems or attrition among teachers
who work with low incidence disabilities versus high incidence disabilities.
Only a few researchers use qualitative methods in attrition studies. These studies include open-
ended surveys of leavers (Billingsley et al., 1993), interviews with teachers who have left
(Brownell et al., 1994-1995; Brownell, Smith, et al., 1997; Morvant et al., 1995), and those who
intend to leave (Billingsley et al., 1995). These survey and interview studies provide a basic
understanding of factors that influence career decisions, but do little to illustrate the lives of
special educators and critical transition points that lead to withdrawal and eventually attrition.
Although some contextual information is provided in these qualitative studies, future studies
should address teachers' perspectives, observations of their work lives, and analyses of journals
and other documents to provide a full understanding of important contributors to job satisfaction,
commitment, stress, and career decisions.
PRIORITIES FOR RESEARCH
The previous discussion addresses issues important to the design of future attrition research in
special education (e.g., attrition definitions, samples, methodologies). This section identifies
knowledge gaps in the literature and proposes several priorities for research.
Although teacher questionnaires are used in most of these studies, few researchers gave teachers
the opportunity to frame issues from their perspectives. Only a few researchers asked special
educators why they left or solicited their views on their work lives (Billingsley et al., 1993;
Billingsley et al., 1995; Brownell et al., 1994-1995; Brownell et al., 1997; Morvant et al., 1995).
Although these studies requested information in an open-ended manner, the data were collected
at only one point in time, usually soon after teachers left their positions. Very little attention has
been paid to factors within a school, descriptions of what these problems mean to teachers on a
day-to-day basis, and how problems and issues contribute to decisions to leave over time.
Moreover, an in-depth analysis of stayers would provide a better understanding of resilience and
why some special educators remain involved and committed to working with students with
disabilities for many years.
Teacher Preparation and Quality
The relationship between teacher quality and retention has received little attention, and few
conclusions can be drawn. Singer (1993a) provided evidence that teachers with higher test
scores are more likely to leave. However, little is known about how educational background,
preparation, or classroom practice is related to career longevity. Longitudinal studies of special
educators from their entry into teacher preparation programs through first years of teaching are
needed. A closer look is needed at the role that teacher preparation plays in the development of
special educators' career dispositions (e.g., involvement, initial commitment) and decision to
stay or leave.
Also, little is known about strategies that faculty in teacher preparation programs might employ
to increase special educators' survival in teaching. What steps might institutions of higher
education do to help future administrators and teachers advocate for effective working conditions
in special education? Although two studies (Carlson & Billingsley, 2001; Miller et al., 1999)
suggested that uncertified teachers are more likely to leave than certified teachers, we know little
about why some of these uncertified teachers leave after a short time and others pursue
certification and remain in special education teaching.
The Early Career Period
Future research must address programs and strategies to reduce attrition among early career
teachers, given they are at most risk of leaving. A neglected aspect in the attrition literature is
beginning teachers' perspectives, qualifications, and work factors that influence their decisions
to stay and leave. As Pugach (1992) observed, a "major question that has not been addressed in
the attrition/retention literature is the socialization of what goes in between choosing to become a
special education teacher and choosing to leave" (p. 134). One way to address this gap in the
literature is to study teachers during their preparation programs and follow these teachers
through the early career period. A range of approaches might be used, including journals and
logs, interviews over time, observations, and recordings of critical incidents.
A better understanding of special educators' entry into teaching would provide critical
information on how to best support their transition into teaching. There is increasing interest in
supporting beginning teachers (e.g., Boyer & Gillespie, 2000; Rosenberg et al., 1997; Whitaker,
2000); however, few studies examined the specific needs of beginning special educators, and
even fewer reports are available on the effectiveness of induction programs. As Gold (1996)
states, "There is an urgent need for data regarding the effectiveness of different types and sources
of support for new teachers" (p. 560).
Support has been shown to be critical to teacher retention, particularly administrative support.
One way of providing support to teachers is through collaborative environments. However, the
relationship of collaboration to attrition has received only scant attention in the attrition
literature. Given the different cultures in general and special education (Pugach, 1992) and the
isolation that many special educators experience (CEC, 2000), collaborative environments have
the potential to help cultivate better understanding between general and special educators, and
foster a sense of belonging for special educators. Moreover, research suggests that collaborative
environments have the potential to benefit teachers by preventing burnout, heightening teachers'
sense of efficacy, and improving teachers' knowledge base (Brownell, Yeager, Rennells, &
Riley, 1997). Future research should consider the nature and extent of collaboration and its
effects on special educators' affective reactions to work and career plans. Moreover, because
administrative support is strongly related to attrition among teachers, we need to know more
about what supportive administrators do and how they promote positive school climates and
working conditions in special education.
Reducing Role Overload and Dissonance
Certain working conditions are important for effective teaching, e.g., reasonable role
expectations; time to teach and collaborate; support of colleagues, administrators, and parents.
Unfortunately, many special educators report problems with role overload and stress (CEC,
2000). The attrition research in special education clearly documents that perceived work
conditions are related to teachers' career plans and attrition. One question that remains
unanswered is the extent to which school, district, state, and federal requirements contribute to
the overload that teachers report. Some qualitative data suggest that it is not the paper work
itself, but the combination of meetings, forms, testing, scoring, written reports, scheduling, and
paper work that creates the problem. In-depth qualitative studies are needed to promote an
understanding of issues that researchers have primarily described at a superficial level, e.g., role
problems, paper work concerns. Some focused questions include:
What in particular contributes to role overload and dissonance? What contextual factors
make role problems better or worse?
What needs to happen in a district/school so that teachers' roles can be restructured to
devote more time to student-centered tasks? Are there states or districts in which special
educators report overall fewer role-related problems than others? If so, what are these
districts/states doing differently?
What are reasonable case loads given the demands of different service-delivery models,
e.g., resource, self-contained, inclusion? What supports and structures are needed to
allow teachers sufficient opportunities to collaborate?
Comparisons of particular groups of teachers will help provide important information about the
differential effects of teacher characteristics, teacher qualifications, and work conditions on
attrition-related variables. It is important to investigate those who leave (both transfer and exit)
and those who stay, as several researchers have done (Boe, Bobbitt, & Cook, 1997; Boe, Bobbitt,
Cook, Whitener, et al., 1997; Miller et al., 1999). Given the need to recruit and retain teachers
from different races in special education, it is critical that race be included as a variable in special
education studies. Future studies should address the relationship of race to different types of
districts and the match between teachers' race and that of their students.
We still know little about the attrition in different school districts and geographic regions.
It is interesting to note that although three major research projects address attrition in major
urban areas (Billingsley et al., 1995; Levine, Campeau, Doorlag, & Cegelka, 1994; Morvant et
al., 1995), the knowledge gained from these projects provides little about the unique challenges
of teachers working in urban districts. Most of the findings in these urban reports are consistent
with the findings from non-urban areas. The attrition of teachers from rural areas has also
received minimal attention.
We have learned that special educators teaching students with four or more disabilities are more
likely to indicate intent to leave than those who teach a less diverse case load (A High-Quality
Teacher for Every Classroom, SPeNSE, 2002). We need to know more about the needs of
teachers who work with different student populations, particularly teachers working with
students from low-incidence categories.
An ideal large-scale study of attrition would enable comparisons, such as: (1) factors that
influence different types of attrition (exit, transfer, stay); (2) differences among geographic
regions and types of school districts (e.g., rural, suburban, urban); (3) teachers working in more
inclusive models versus traditional models; (4) teachers working with high-incidence versus low-
incidence student groups; (5) differences in high and low attrition districts; and (6) teachers who
work with children from racial and cultural backgrounds different from their own. Moreover,
large-scale studies should always include a report on teacher characteristics, e.g., age, gender,
race, basic preparation factors.
SUMMARY AND IMPLICATIONS
In summary, this review finds that a wide range of factors influence attrition, including teachers'
personal circumstances and priorities. Most of the attrition studies focus on problematic work
environment variables and their relationships to attrition. This review suggests that work
environmental factors, (e.g., low salaries, poor climate, lack of administrative support, role
problems) can lead to negative affective reactions, e.g., high levels of stress as well as low levels
of job satisfaction and commitment. These negative reactions can lead to withdrawal and
eventually attrition. In addition, teacher characteristics and qualification variables that are linked
to attrition include the following: (1) special educators who are younger and inexperienced are
at higher risk of leaving that their older and more experienced counterparts, (2) those who are
uncertified are more likely to leave than those who are certified, and (3) those with higher test
scores are more likely to leave than those with lower scores.
Policy makers and administrators interested in reducing attrition must facilitate the development
of better work environments for special educators (for implications of attrition research, see
Billingsley, 2002b). Addressing issues such as teacher role overload and the need for critical
supports (e.g., administrative support, professional development) must be addressed to ensure
that teachers can be effective in their work. Focusing on one or two aspects of teachers' work
lives will probably be insufficient to substantially reduce attrition. For example, providing
beginning teachers with formal induction programs is not likely to be effective in the long run
unless their work assignments are also reasonable. A holistic look at creating positive work
environments should not only reduce attrition behavior, but should also help sustain special
educators' involvement and commitment in their work.
Banks, S. R., & Necco, E. (1987). Alternative certification, educational training and job
longevity. Action in Teacher Education, 9(1), 67-73.
Billingsley, B. (2002a). Beginning special educators: Characteristics, qualifications, and
experiences. SPeNSEFactsheet. Retrieved March, 2002, from www.spense.org
Billingsley, B. (2002b). Improving special education teacher retention: Implications from a
decade of research. The Journal of Special Education Leadership, 15(2), 20-26.
Billingsley, B. S. (1993). Teacher retention and attrition in special and general education: A
critical review of the literature. The Journal of SpecialEducation, 27(2), 137-174.
Billingsley, B., Pyecha, J., Smith-Davis, J., Murray, K., & Hendricks, M. B. (1995). Improving
the retention of special education teachers: Final report. Research Triangle Institute
(Prepared for Office of Special Education Programs, Office of Special Education and
Rehabilitative Services, U. S. Department of Education under Cooperative Agreement
H023Q10001). (ERIC Document Reproduction Service No. ED379860)
Billingsley, B. S., Bodkins, D., & Hendricks, M. B. (1993). Why special educators leave
teaching: Implications for administrators. Case in Point, 7(2), 23-38.
Billingsley, B. S., & Cross, L. H. (1991a). General education teachers' interest in special
education training: Deterrents, incentives, and training needs. Teacher Education and
Special Education, 14(3), 162-168.
Billingsley, B. S., & Cross, L. H. (1991b). Teachers' decisions to transfer from special to general
education. The Journal of Special Education, 24(4), 496-511.
Billingsley, B. S., & Cross, L. H. (1992). Predictors of commitment, job satisfaction, and intent
to stay in teaching: A comparison of general and special educators. The Journal of
Special Education, 25(4), 453-471.
Billingsley, B. S., & Tomchin, E. M. (1992). Four beginning LD teachers: What their
experiences suggest for trainers and employers. Learning Disabilities Research and
Practice, 7(2), 104-112.
Blanton, L., Sindelar, P., Correa, V., Hardman, M., McDonnell, J., & Kuhel, K. (2002).
Research summary: Conceptions of beginning teacher quality: Models for conducting
research. Gainesville, FL: University of Florida, Center on Personnel Studies in Special
Boe, E. E. (1990, November). Comprehensive retention and attrition model (CRAM). Paper
presented at the Research Forum on Differing Approaches to Defining and Measuring
Personnel Supply and Demand, Washington, DC.
Boe, E. E., Barkanic, G., & Leow, C. S. (1999). Retention and attrition of teachers at the school
level: National trends and predictors. Philadelphia: University of Pennsylvania,
Graduate School of Education, Center for Research and Evaluation in Social Policy.
(ERIC Document Reproduction Service No. ED436485)
Boe, E. E., Bobbitt, S. A., & Cook, L. H. (1997). Whither didst thou go? Retention,
reassignment, migration, and attrition of special and general education teachers in
national perspective. The Journal of Special Education, 30(4), 371-389.
Boe, E. E., Bobbitt, S. A., Cook, L. H., Barkanic, G., & Maislin, G. (1999). Teacher turnover in
eight cognate areas: National trends and predictors. Philadelphia: University of
Pennsylvania, Graduate School of Education, Center for Research and Evaluation in
Boe, E. E., Bobbitt, S. A., Cook, L. H., Whitener, S. D., & Weber, A. L. (1997). Why didst thou
go? Predictors of retention, transfer, and attrition of special and general education
teachers from a national perspective. The Journal of SpecialEducation, 30(4), 390-411.
Boe, E. E., Cook, L. H., Bobbitt, S. A., & Weber, A. L. (1998). Retention and attrition of
teachers at the district level: National trends in special and general education. Reston,
VA: National Clearinghouse for Professions in Special Education.
Boyer, L. & Gillespie, P. (2000). Keeping the committed: The importance of induction and
support programs for new special educators. Teaching Exceptional Children, 33(1), 10-
Brownell, M. T., & Smith, S. W. (1992). Attrition/retention of special education teachers:
Critique of current research and recommendations for retention efforts. Teacher
Education and Special Education, 15(4), 229-248.
Brownell, M. T., & Smith, S. W. (1993). Understanding special education teacher attrition: A
conceptual model and implications for teacher educators. Teacher Education and Special
Education, 16(3), 370-382.
Brownell, M. T., Smith, S. W., McNellis, J., & Lenk, L. (1994-1995). Career decisions in special
education: Current and former teachers' personal views. Exceptionality, 5(2), 83-102.
Brownell, M. T., Smith, S. W., McNellis, J., & Miller, M. D. (1997). Attrition in special
education: Why teachers leave the classroom and where they go. Exceptionality, 7(3),
Brownell, M. T., Yeager, E., Rennells, M. S., & Riley, T. (1997). Teachers working together:
What teacher educators and researchers should know. Teacher Education and Special
Education, 20(4), 340-359.
Carlson, E. (2001, February). Preliminary results from the study of personnel needs in special
education (SPeNSE): Local administrator's efforts to fill job openings for special
education teachers. Paper presented at the OSEP Personnel Preparation Conference,
Carlson, E., & Billingsley, B. (2001, July). Working conditions in special education: Current
research and implications for the field. Paper presented at the OSEP Project Directors
Conference, Washington, DC.
Chapman, D. W., & Green, M. S. (1986). Teacher Retention: A Further Examination. Journal
ofEducational Research, 79, 273-279.
Cooley, E., & Yovanoff, P. (1996). Supporting Professionals-at-Risk: Evaluating Interventions
to Reduce Burnout and Improve Retention of Special Educators. Exceptional Children,
Corcoran, T. B., Walker, L. J., & White, J. L. (1988). Working in Urban Schools. Washington,
DC: Institute for Educational Leadership.
Council for Exceptional Children [CEC]. (2000). Bright futures for exceptional learners: An
action to achieve quality conditions for teaching and learning. Reston, VA: Author.
Cross, L. H., & Billingsley, B. (1994). Testing a model of special educators' intent to stay in
teaching. Exceptional Children, 60(5), 411-421.
Darling-Hammond, L. (1999, September). Paper presented at The National Commission on
Mathematics and Science Teaching for the 21st Century. Retrieved June 18, 2001, from
Darling-Hammond, L., & Sclan, E. M. (1996). Who teaches and why: Dilemmas of building a
profession for twenty-first century schools. In J. Sikula, T. J. Buttery, & E. Guyton
(Eds.), Handbook of research on teacher education (2nd ed., pp. 67-101). New York:
Simon & Schuster.
Dworkin, A. G. (1980). The changing demography of public school teachers: Some implications
for faculty turnover in urban areas. Sociology ofEducation, 53(April), 65-73.
Embich, J. L. (2001). The relationship of secondary special education teachers' roles and factors
that lead to professional burnout. Teacher Education and Special Education, 24(1), 58-
Frank, M., & Keith, T. Z. (1984). Academic abilities of persons entering and remaining in
special education. Exceptional Children, 51(2), 76-77.
George, N. L., George, M. P., Gersten, R., & Grosenick, J. R. (1995). To leave or to stay? An
exploratory study of teachers of students with emotional and behavioral disorders.
Remedial and Special Education, 16(4), 227-236.
Gersten, R., Keating, T., Yovanoff, P., & Harniss, M. K. (2001). Working in special education:
Factors that enhance special educators' intent to stay. Exceptional Children, 67(4), 549-
Gold, Y. (1996). Beginning teacher support: Attrition, mentoring, and induction. In J. Sikula, T.
J. Buttery, & E. Guyton (Eds.), Handbook of research on teacher education (2nd ed., pp.
548-594). New York: Simon & Schuster.
Greer, J. G., & Greer, B. B. (1992). Stopping burnout before it starts: Prevention measures at the
preservice level. Teacher Education and Special Education, 15(3), 168-174.
Grissmer, D. W., & Kirby, S. N. (1987). Teacher attrition: The uphill climb to staff the nation's
schools. Santa Monica, CA: The RAND Corporation.
Henke, R. R., Choy, S. P., Chen, X., Geis, S., & Alt, N. (1997). Teachers in the 1990's: Profile
of a profession (NCES 97-460). Berkeley, CA: Office of Educational Research and
Heyns, B. (1988). Educational defectors: A first look at teacher attrition in the NS-72.
Educational Researcher, 17(3), 24-32.
Ingersoll, R. M. (2001). Teacher turnover, teacher shortages and the organization of schools
(Document R-01-1). Seattle, WA: University of Washington, Center for the Study of
Teaching and Policy.
Kilgore, K. L., & Griffin, C. C. (1998). Beginning special educators: Problems of practice and
the influence of school context, Teacher Education and Special Education, 21(3) 155-
LeCompte, D., & Dworkin, A. G. (1991). Giving up on school: Student dropouts and teacher
burnouts. Newbury Park, CA: Corwin.
Levine, R., Campeau, P., Doorlag, D., & Cegelka, P. (1994, November). Attrition of special
education and regular education teachers in California: Analyses of CBEDS data. Paper
presented at the annual meeting of the Teacher Education Division of the Council for
Exceptional Children, San Diego, CA.
Littrell, P., Billingsley, B., & Cross, L. (1994). The effects of principal support on special and
general educators' stress, job satisfaction, school commitment, health, and intent to stay
in teaching. Remedial and Special Education, 15(5), 297-310.
MacDonald, V. (2001). Making time: A teacher's report on her first year of teaching children
with emotional disabilities. The Journal of Special Education, 35(2),84-91.
Magliaro, S. G., & Wildman, T. M. (1990, April). Meeting the needs of beginning special
educators. Paper presented at the American Educational Research Association Annual
Meeting, Boston, MA.
Maslach, C. (1982). Burnout: The cost of caring. New York: Prentice-Hall Press.
McLeskey, J., Tyler, N., & Saunders, S. (2002). The supply and demand of special education
teachers: The nature of the chronic shortage of special education teachers. Gainesville,
FL: University of Florida, Center on Personnel Studies in Special Education [COPSSE].
Miller, M. D., Brownell, M., & Smith, S. W. (1999). Factors that predict teachers staying in,
leaving, or transferring from the special education classroom. Exceptional Children,
Morsink, Catherine V. (1982, February). Changes in the role of special educators: Public
perceptions and demands, Exceptional Education Quarterly, 15-25.
Morvant, M., Gersten, R., Gillman, J., Keating, T., & Blake, G. (1995). Attrition/retention of
urban special education teachers: Multi-faceted research and strategic action planning.
Final performance report, Volume 1. (ERIC Document Reproduction Service No.
Mowday, R. T., Porter, L. M., & Steers, R. M. (1982). Employee-organization linkables: The
psychology of commitment, absenteeism, and turnover. New York: Academic Press.
Murnane, R. J., Singer, J. D., & Willett, J. B. (1988). The career paths of teachers: Implications
for teacher supply and methodological issues for research. Educational Researcher,
Pugach, M. C. (1992). Uncharted territory: Research on the socialization of special education
teachers. Teacher Education and Special Education, 15(2), 133-147.
Quiocho, A., & Rios, F. (2000). The power of their presence: Minority group teachers and
schooling. Review ofEducational Research, 70(4), 485-528.
Rosenberg, M. S., Griffin, C. C., Kilgore, K. L. & Carpenter, S. L. (1997). Beginning teachers in
special education: A model for providing individualized support. Teacher Education
and Special Education, 20(4), 301-321.
Rosenholtz, S. J. (1989). Workplace conditions that affect teacher quality and commitment:
Implications for teacher inductive programs. The Elementary School Journal, 89(4), 421-
Schnorr, J. M. (1995). Teacher retention: A CSPD analysis and planning model. Teacher
Education and Special Education, 18(1), 22-38.
Singer, J. D. (1993a). Are special educators' career paths special? Results from a 13-year
longitudinal study. Exceptional Children, 59(3), 262-279.
Singer, J. D. (1993b). Once is not enough: Former special educators who return to teaching.
Exceptional Children, 60(1), 58-72.
Singh, K., & Billingsley, B. (1996). Intent to stay in teaching: Teachers of students with
emotional disorders versus other special educators. Remedial and Special Education,
Singh, K., & Billingsley, B. (1998). Professional support and its effects on teachers'
commitment. The Journal of Educational Research, 91(4), 229-239.
Smith-Davis, G., Burke, P. J., & Noel, M. (1984). Personnel to educate the handicapped in
America: Supply and demand from a programmatic viewpoint. College Park, MD:
University of Maryland, Institute for the Study of Exceptional Children and Youth.
Smith-Davis, J., & Billingsley, B. (1993). The suppy/demand puzzle. Teacher Education and
Special Education, 16(3), 205-220.
SPeNSE. A high-quality teacher for every classroom (n.d.). SPeNSE factsheet. Retrieved
January 9, 2002, from www.spense.org
Veenman, S. (1984). Perceived problems of beginning teachers. Review of Educational
Research, 54(2), 143-178.
Weber, D. B., & Toffler, J. D. (1989). Burnout among teachers of students with moderate,
severe, or profound mental retardation. Teacher Education and Special Education, 12(3),
Weiskopf, P. E. (1980). Burnout among teachers of exceptional children. Exceptional Children,
Westling, D. L., & Whitten, T. M. (1996). Rural special education teachers' plans to continue or
leave their teaching positions. Exceptional Children, 62(4), 319-335.
Whitaker, S. D. (2000). Mentoring beginning special education teachers and the relationship to
attrition. Exceptional Children, 66(4), 546-566.
Wisniewski, L., & Gargiulo, R. M. (1997). Occupational stress and burnout among special
educators: A review of the literature. The Journal of Special Education, 31(3), 325-346.
Yee, S. M. (1990). Careers in the classroom: When teaching is more than a job. New York:
Teachers College Press.
Zabel, R. H., & Zabel, M. K. (2001). Revisiting burnout among special education teachers: Do
age, experience, and preparation still matter? Teacher Education and Special Education,
Zabel, R. H., & Zabel, M. K. (1982). Factors in burnout among teachers of exceptional children.
Exceptional Children, 49, 261-263.
Zepeda, S. J., & Ponticella, J. A. (1997). First-year teachers at risk: A study of induction at three
high schools. The High School Journal, 8-21.
Appendix. Special Education Attrition and Retention Studies
Definition of Methodology/
Author(s) Purpose Attrition/Retention Sample Analysis
Billingsley & To determine the 286 former Mailed
Cross (1992) variables that influence Intent to leave special questionnaire;
commitment, job educators Regression
satisfaction, and intent analysis
to stay in teaching
Billingsley, To investigate special Teachers who left 42 former Open-ended
Bodkins, & education teachers' special education special mailed
Hendricks reasons for leaving teaching (included educators in questionnaire;
(1993) teaching and related transfer to general Virginia Qualitative
work experiences education and exit analyses
Billingsley, Study 1: To understand Study 1: Those who Study 1: 99 Study 1: Mailed
Pyecha, from special educator- left their primary special questionnaires,
Smith-Davis, teachers' perspectives: teaching positions in a educators who descriptive
Murray, & why they left their large, urban district left their analyses
Hendricks positions in an urban teaching
(1995) school district and what positions in a
work they did the large, urban
following year district
Study 2: To gain a "a Study 2: 81 Study 2:
better understanding of (committed and current Standardized,
committed and '
the influence of career satisfied teachers who special open-ended
plans, commitment, and intend to remain educators; 27 guide used to
job satisfaction for teaching in district teachers in conduct
special educators in an "Leavers" (those who each of the interviews with
1Leavers" (those who
urban setting are dissatisfied, not following teachers.
committed, and plan to groups:
leave district) stayers, Qualitative and
"Undecided" (those leavers, quantitative
who are unsure about undecideds analyses used to
who are unsure about
career plans) reduce, display,
Boe, Bobbit, &
To analyze from a
four components of
turnover in the teaching
force in public schools
and teachers' activities
To investigate from a
national perspective, a
wide variety of teacher,
school, and district
characteristics that are
either known to be or
suspected to be
associated with teacher
turnover and retention
of both public school
general and special
To determine personal,
educational, and work
place variables and their
influence a special
decision to leave or stay
in the classroom
migration, and attrition
of special and general
Leavers (left public
school teaching entirely
or are teaching in
Movers (transferred to
a public school in the
same or different
both leaver and mover
Stayers Those who
stayed in the same
Leavers and stayers
from a large
district in the
Survey and the
the Schools and
up Survey 1989)
Brownell, Identify why special 93 randomly Individual
Smith, educators leave the Leavers (those who selected phone
McNellis, & special education switched to general Florida interviews using
Miller (1997) classroom and education, moved to teachers (did protocol
make distinctions administration, or left not include
between disgruntled and teaching altogether) speech Qualitative and
non disgruntled leavers pathologists) quantitative data
Carlson (2001, Identify variables 867 preschool Structured
2002) related to intention to Intent to leave teachers, telephone
(Several leave 2,031 teachers interviews
research briefs of visual and
and (Part of larger SPeNSE hearing Bivariate
presentations) Study of Personnel impairments, analyses and
Needs in Special 856 teachers regressions
Education) of emotional
Cross & To determine the extent 412 special Path analysis
Billingsley to which work-related Intent to leave educators and used to examine
(1994) variables, as well as 130 teachers the direct and
teaching assignments, of students indirect effects
and personal with of principal
characteristics explain emotional support, role
expressed intentions to disorders problems, stress,
stay in teaching from Virginia job satisfaction
(response rate and commitment
was 82% for on intent to stay
both groups) in teaching
George, To compare variables 96 current Mailed
George, associated with intent to Intent to leave (current special questionnaire
Gersten, stay and leave among teachers of educators (51 and follow-up
Grosenick teachers of behaviorally behaviorally disordered who plan to interviews:
(1995) disorders (BD) students students) stay and 45 at
risk for Bivariate
Gersten, To use a path analysis to 887 special Survey in
Keating, examine specific job- Intent to leave educators in 3 statement;
Yoranoff, & design variables and large urban LISREL
Harniss, their effects on teachers school structural
(2001) and intent to stay in or districts modeling
leave the field analysis
Littrell, To identify general and 385 special Mailed
Billingsley, & special educators' Intent to leave educators of questionnaire,
Cross (1994) perceptions of principal students with ANOVA,
support received LD, MR, and regression
ED and 313 analyses
To determine the effects general
of perceived support on educators
teacher stress, job from Virginia
health, and intent to stay
Miller, Determine those work 1,208 Florida Survey
Brownell, & place variables that Teachers who left special instrument;
Smith (1999) were significant special education education Multinomial
predictors of teachers' teaching or transferred teachers logit model used
decisions to leave or to a similarjob responded to distinguish
transfer from the special (compared to stayers) (80% stayers, leavers,
education classroom response rate) and transfers.
Morvant, Study 1: Study 1: Special Study 1: 17 Study 1:
Gersten, educators who left one teachers Interviews
Gillman, Study 2: urban district Study 2: 868 Study 2: Mailed
Keating, & Study 2: Intent to leave special Questionnaire
Blake (1995)* educators
To determine how many
years special education
teachers continue to
teach and in what years
they tend to leave;
To determine if the risk
of leaving differs by
year of hire, personal
To examine specific
and how they influence
teaching, and intent to
continue in the
profession and to
determine if factors are
the same among
teachers of BD and
other special educators.
Teachers who left
Intent to leave
extant data tapes
Westling & Determine which Intent to leave 158 special Mailed
Whitten (1996) conditions or attitudes education questionnaire
differed between teachers from
teachers who indicated 14 Bivariate
they planned to remain predominantly statistics and
and those who did not rural counties logistical
plan to remain and in one regression
Whitaker To determine the impact Intent to leave 156 first year Mailed
(2000) of mentoring programs special questionnaire on
on first year teachers' educators in mentoring, plans
plans to remain in South to remain in
special education Carolina special