• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

Modeling Teacher Attrition: Teacher Characteristics and Working Conditions

ProQuest Dissertations and Theses, 2011
Dissertation
Author: Matthew A Cannady
Abstract:
This paper describes the literature on teacher attrition as either focusing on the working conditions faced by beginning teachers or highlighting variations in teachers' characteristics as causes for early teacher attrition. This study uses responses to the School and Staffing Survey (SASS) along with the Teacher Follow-Up Survey (TFS) to compare these contrasting views of early teacher attrition. Two logistic regression models were constructed and their relative efficacy in explaining teacher attrition were compared using three statistical techniques; model fit characteristics (e.g. pseudo-R 2 , Akaike Information Criteria, Bayesian Information Criteria); a comparison of their classification effectiveness, and results from Davidson and MacKinnon's J test (1981). A final model was also constructed using the predictive elements of each of the previous models. Results suggest that the working conditions model better fits the observed data than the teacher characteristics model. The final model highlights the importance of teacher commitment and engagement in the profession in teachers' career decisions.

C ontents

Chapter 1: Perspectives on Early Leaving

................................ ................................ ................

1

Purpose of this Study

................................ ................................ ................................ .......................

5

Scope of this Study

................................ ................................ ................................ ..........................

8

Organization of this Dissertation

................................ ................................ ................................ ....

10

Chapter 2: Review of the Literature

................................ ................................ ......................

12

Search Criteria

................................ ................................ ................................ ...............................

14

Organization of the Review

................................ ................................ ................................ ...............

16

Section 1: Big Picture of Teacher Turnover

................................ ................................ .....................

18

Section 2: Scope of Teacher Turnover

................................ ................................ ............................

22

Section 3: Constellation 1: Teacher Characteristics

................................ ................................ .........

26

Recruitment

................................ ................................ ................................ ................................ ......

27

Certification Route

................................ ................................ ................................ ............................

33

Student Teaching

................................ ................................ ................................ ..............................

42

Disposition

................................ ................................ ................................ ................................ ........

44

Teachers’ Education

................................ ................................ ................................ ..........................

49

Conclusion

................................ ................................ ................................ ................................ .........

51

Section 4: Constellation 2: Working Conditions

................................ ................................ ..............

52

Salary

................................ ................................ ................................ ................................ .................

52

Student Composition

................................ ................................ ................................ ........................

57

School Culture

................................ ................................ ................................ ................................ ...

59

Administrative Support

................................ ................................ ................................ .....................

68

Job Satisfaction

................................ ................................ ................................ ................................ .

72

Conclusion

................................ ................................ ................................ ................................ .........

74

Section 5: Interaction of Teachers’ Characteristics with Working Conditions

................................ ..

75

Conclusion

................................ ................................ ................................ ................................ .....

78

Chapter 3: Research Design

................................ ................................ ................................ ..

79

Data

................................ ................................ ................................ ................................ ...............

80

Methodological Approaches

................................ ................................ ................................ ..........

85

Sampling Weights

................................ ................................ ................................ .............................

85

Analytical Methods

................................ ................................ ................................ ...........................

86

Research Design

................................ ................................ ................................ ................................

89

Working Conditions Regression Model

................................ ................................ ...........................

90

Salary

................................ ................................ ................................ ................................ .................

90

School Culture

................................ ................................ ................................ ................................ ...

91

Induction and Professional Development

................................ ................................ .........................

92

Student Composition

................................ ................................ ................................ ........................

93

Administrative support

................................ ................................ ................................ .....................

94

Teacher Characteristics Regression Model

................................ ................................ .....................

95

Recruitment

................................ ................................ ................................ ................................ ......

95

Certification

................................ ................................ ................................ ................................ ......

96

Student Teaching Experience

................................ ................................ ................................ ............

98

Teachers’ Education

................................ ................................ ................................ ........................

100

Regression Model Construction

................................ ................................ ................................ ...

101

Chapter 4: Results

................................ ................................ ................................ ...............

110

v

Sample

................................ ................................ ................................ ................................ .........

110

Descriptive Statistics

................................ ................................ ................................ ....................

113

Working Conditions Variable Descriptions

................................ ................................ .....................

115

Teacher Characteristics Variable Descriptions

................................ ................................ ................

118

Working Conditions Model

................................ ................................ ................................ ..........

126

Teacher Characteristics Model

................................ ................................ ................................ .....

133

Model Comparison

................................ ................................ ................................ ......................

137

Combined Model

................................ ................................ ................................ .........................

142

Chapter 5: Discussion

................................ ................................ ................................ ..........

148

Teacher Attrition and Working Conditions

................................ ................................ ...................

148

Teacher Attrition and Teacher Characteristics

................................ ................................ ..............

151

Comparing Working Conditions and Teacher Characteristics Logistic Regression Models

.............

154

Combining Working Conditions and Teacher Characteristics Logistic Regression Models

..............

157

Limitations

................................ ................................ ................................ ................................ ...

161

Implications

................................ ................................ ................................ ................................ .

164

References

................................ ................................ ................................ ..........................

169

Appendix A

................................ ................................ ................................ .........................

178

Appendix B

................................ ................................ ................................ .........................

179

Appendix C

................................ ................................ ................................ .........................

180

vi

Table of Tables

Table 3.1: Agreement Between the Teacher Status File and the Teacher Follow - Up Survey

............................

82

Table 3.2: Sample Comparison between the Teacher Status File and the Teacher Follow - Up Survey

..............

83

Table 3.3: Steps to Perform the J - Test

................................ ................................ ................................ ...............

107

Table 4.1 Research Questions

................................ ................................ ................................ ...........................

110

Table 4.2: Comparison of Weighted and Unweighted Estimates of Attrition

................................ ...................

112

Table 4.3 Demographic Descripti ve Statistics

................................ ................................ ................................ ...

114

Table 4.4 Descriptive Statistics: Working Conditions

................................ ................................ .......................

117

Table 4.5 Descriptive Statistics: Teacher Characteristics -

Certification

................................ ..........................

120

Table 4.6 Descripti ve Statistics: Teacher Characteristics -

Certification

................................ ..........................

121

Table 4.7 Descriptive Statistics: Teacher Characteristics -

Preparation

................................ ............................

122

Table 4.8 Descriptive Statistics: Teacher Characteristics -

Education

................................ ..............................

126

Table 4.9 Working Conditions Logistic Regression

................................ ................................ ..........................

132

Table 4.10 Teacher Characteristics Logistic Regression

................................ ................................ ...................

136

Table 4.11 Psuedo - R 2

and Information Criteria Comparison

................................ ................................ ............

138

Table 4.12 Steps to perform the J - Test

................................ ................................ ................................ ..............

141

Table 4.13 Comparison of Logistic Regression: J - test

................................ ................................ ......................

142

Table 4.14 Predicting Attrition with Logits from both models and their Interaction

................................ ........

143

Table 4.15 Combined Model: Stepwise Backward Entry

................................ ................................ ..................

145

vii

Table of Figures

Figure 4.1 Comparison of Logistic Regressions: Receiver Operation Characteristic Curve

................................ .....

140

Figure 4.2 ROC Curve: Combined Model, Teacher Characteristics Model, Working Conditions Model

................

146

1

Chapter 1: Perspectives on Early Leaving

When teachers leave the profession ,

they do so with more than their belongings. They leave with the investment that colleagues, administrators, and school districts have made in them along with the resources required to recruit them to the school and support them through the year. They al so walk out with the added professional experience garnered through the year(s) and the relationships developed with colleagues in the school. The true cost of exit is more than just these unrecoverable costs; when teachers leave, they leave other voids th at the school must expend even further resources to fill.

These departures from teaching generate costs, both financial (Texas Center for Educational Research, 2000) and instructional ( Johnson, Berg & Donaldson, 2005). In an attempt to estimate the fina ncial burden associated with teacher turnover, let alone the effect on instruction,

the National Commission on Teaching and America’s Future (NCTAF) studied five school districts across the country and found that the costs across school districts range from $4,366 to $17,872 per

teacher who leaves (NCTAF, 2003). NCTAF estimated 332,700 t eachers left their teaching assignments from the 2003 - 4 to the 2004 - 5 school year; their estimate of the cost of this turnover exceeded 7 billion dollars (NCTAF, 2007).

The problem of teacher attrition is especially evident with beginning teachers. Early attrition is particularly problematic, because it requires schools to replace a teacher they have just recently hired. Statistics on teacher attrition indicate 20% of beginning teachers leave the profession within their first three years (Henke, Chen, & Gi es, 2000) and nearly half leave within the first five years (Ingersoll, 2003). Moreover, national averages mask the even higher rates of early attrition among teachers in schools in urban and rural areas, which often are the least

2

equipped to deal with th e additional impact on the budget and traditionally serve students with the greatest educational needs (Hanushek, Kain & Rivkin, 2004).

Research on early attrition offers a variety of explanatory theories , but, as is demonstrate d in Chapter 2, th e s e

rese arch studies can largely be categorized into one of two constellations .

The term constellation is used to describe groupings of research studies

while recognizing the variety of separate elements that comprise each grouping. The shared perspective the st udies take on the explanation for early teacher attrition forms the constellations. For example, studies in t he constellation I call “working conditions,”

the one most frequen tly discussed in the literature,

takes the perspective

that teachers leave the pr ofession as a result of one , or a combination of several, factors in their working experience.

Although there are a variety of theor etical frameworks

within this constellation, one common theory derives from labor economics, which claims that teachers , li ke workers in other fields,

are rational beings and make choices about their career decisions based on their preferences for “wages, working conditions and other unobservable factors” (Strunk & Robinson,

2006 , p. 67). Another major theory within this same

constellation comes from sociology, specifically the study of organizations. Research deriving from this theory seeks to determine how leaders within schools, considered organizations, can respond to issues of teacher attrition by promoting the retention

of desired teachers within a school. The term working conditions in this study refers to all the experiences encountered by teachers while they are working in the classroom , including their salaries, induction experiences, and administrative support ; thi s differs from some other researchers who use the same phrase to refer to only the conditions at the school, such as quality of the facilities, behavior of the student body, or collegiality among the staff. Here, the phrase working conditions accounts for

those aspects of teachers’ jobs as well as a teachers’ salary, their

3

induction upon entering the profession, the behavior of the students in the school, the school culture and the support of the administration. Thus, r easons for early teacher attrition

f rom studies within this constellation include: inadequate salaries, low job satisfaction, unwelcoming

school cultures, poor administrative support, and problems with student behavior (e.g. Ingersoll, 2004; Luekens, Lyter, Fox & Changler, 2004) . Furthermor e, the lack of state and school district infrastructure to support teachers during their professional careers has also been noted as common factors associated with teachers’ exiting the field (Johnson & Project of the Next Generation of Teachers, 2004; Dar ling - Hammond, Berry, Haselkorn & Fideler , 1999; NCTAF, 2003).

R ecognition of the high rates of early teacher attrition has led to various intervention programs aimed at supporting beginning teachers

(Ingersoll & Kralik, 2004) . As Ingersoll and Kralik po int out, some schools have attempted to offer support to new teachers through their “induction” phase, including the provision of mentoring, peer observations, and additional support in the classroom (e.g., teacher aides or instructional coaches). There i s some evidence that it is the teachers entering the profession without an undergraduate degree in education

who

see the greatest increases in their retention rates when participating in induction programs (Duke, Karson & Wheeler, 2006). This suggests

it is the characteristics , including the amount of preparation, that teachers have upon entering the profession or the conditions that they face once there that most influences their decisions to remain a classroom teacher.

The second constellation of research studies looks at aspects within the teachers’ themselves to offer explanations of early teacher attrition. H ere some studies take the perspective

that teachers are ill prepared to manage the working conditions that they

encounter (Haberman, 20 05). This line of reasoning stands in contradiction to the previous argument about

4

working conditions being the primary factor leading to teacher attrition by positing that new teachers enter the profession with an awareness of impending difficulties and o ver half of them, despite experiencing these conditions first hand, still choose to stay in the profession (Haberman, 2005). Thu s, the researchers

in this second constellation

argue that it is not the working conditions that push teachers to leave, but rat her the new

teachers’ characteristics –

that is, their skills, experiences and attributes –

upon entering the classroom

that leads to their early exits . Some examples of relevant teacher characteristics

include: lack of commitment to teaching as a career ( Fleener & Dahm, 2007; Shen, 1997); lack of preparation to work with diverse populations (Haberman, 2005); limited amount, and/or low - quality, of teacher preparation in pedagogy (Boe, Cook & Sunderland, 2006) or mentored classroom experience (Reynolds, Ros s, & Rakow, 2002; Shen, 2003; Fleener & Dahm , 2007); and a lack of resiliency in the teachers themselves (Bernshausen & Cunningham 2001).

The theories that guide research within this constellation are widely varied. For example, there are several studies

that rely on economic theories, such as, labor economic theory and human capital theory. There are other studies within this constellation that utilize theories from psychology, such as Bandura’s work in self - efficacy, or social cognitive theory. Furthe rmore, there are some studies that are inspired by theories within sociology, socio - cultural career theory and social organizational theory. While the elements within this constellation define a broader theoretical space, their proximity is defined by the ir shared perspective that early teacher attrition is a product of teacher characteristics, rather than defined by external influences.

Despite the high rates of teacher attrition and the variety of explanations offered, there is

a dearth in literature e xploring the link of teachers’ characteristics

and working conditions with their career trajectories (Johnson, Berg & Donaldson, 2005). In fact, in their thorough review of

5

the literature on teacher retention, Johnson, Berg and Donaldson (2005) emphasize t he need for research that follows teachers over time through their transition from preservice to inservice teaching. These authors note that current research lacks adequate information regarding background, experience, and workplace conditions of teachers,

which would allow for an examination of how these factors relate to each other and to retention over time.

Purpose of this Study

The purpos e of this research, therefore, wa s to investigate the strength and merit of two constellations of research studies

which offer overlapping yet contrasting perspectives

o n

early teacher attrition

--

teachers’ characteristics and teachers’ working conditions. By comparing

and testing

these perspectives, policy makers can have a better understanding of where to direct r esources to promote teacher retention. Additionally, teacher educators can better understand the role of their work in supporting the retention of their teacher candidates into their careers . Specifically, this study address es

the following four

researc h questions regarding prediction of teacher attrition. Results of this line of inquiry will offer insight into a comparison of the perspectives

of teacher attrition noted above:

1.

To what extent do beginning teachers’ working conditions predict their attrit ion the year

following

the SASS administration ?

2.

To what extent do beginning teachers’ characteristics

predict their attrition the year

following

the SASS administration ?

3.

Which, if either, constellation of studies, working conditions or teacher characterist ics, better predicts attrition the year

following

the SASS administration ?

6

4.

How do the unique elements within the teacher characteristics and working conditions models combine and interact to predict early teacher attrition

the year

following

the SASS administration ?

The study s ought

to answer these questions using a national sample collected via the National Center for Educational Statistics’ School and Staffing Survey and Teacher Follow -

up Survey. The research was

executed by comparing the

perspectives from the two constellations of research on teacher attrition in the first five years in the profession. The

research studies forming

the first constellation

suggest

that working conditions lead to low job satisfaction, which in turn, leads to

the early attrition of teachers. The second constellation

holds that early attrition is a function of teacher characteristics

prior to entering the classroom, which accounts for the variation in teachers’ ability to withstand the difficulty of the job. Specifically, this study investigates t he extent to which working conditions and/ or teacher

characteristics

are able to predict a beginning teacher’s status (still teaching or left) for the year after the y are initially surveyed .

Data

The School and Staffing Survey (SASS) and the Teacher Follow - Up Survey (TFS) are designed and administered by the U.S. Department of Education’s National Center for Education Statistics (NCES). This survey system is the largest and most extensive survey of

primary, middle and secondary schools in the United States and, with minor changes, has been conducted every five years since the 1987 - 1988 school year. The SASS comprises five separate surveys to each of the following groups: school - district personnel, s chool staff, teachers, administrators, and library and media center personnel. This study focused on the SASS Public School Teacher Questionnaires. This particular survey asks participants about their general teaching assignment,

7

their preservice experien ces, perceptions of preservice preparation, their work conditions and attitudes about their working conditions. To measure the subsequent status of the teachers from the SASS, the Teacher Follow - Up Survey is administered to a subset of SASS participants on e year following the SASS administration. R esults from the Teacher Follow - Up surveys are used to determine the subsequent status of teachers who responded to the previous year’s SASS administration. This study examined the 2003 - 2004 SASS Public School Que stionnaires, which were linked to the 2004 -

2005 Teacher Follow - Up Survey.

The SASS Public School Teacher Survey collects data from approximately 38,000 public school teachers; however, t his available sample was

reduced to the sampl e of interest. T he sa mple of interest wa s defined as teachers who, upon taking the SASS survey, were in their first five

years of teaching and either stayed in teaching the following year (their 2 nd , 3 rd , 4 th ,

5 th

or 6 th

year) or left the profession the following year.

In

this study, t eachers

who transfer schools, movers, were consider ed

in the same category as teachers who stayed in their same school .

T he goal

of th is

research wa s to investigate

the difference between those who

continue teaching

and those who leave

teaching ;

individuals who transfer schools, while a part of te acher turnover, continue to teach and are therefore part of the same group as those who stayed in the same classroom . While recognizing that these

categories are a

simplification of teachers’ career trajectories, this approach

is useful in determining what encourages teachers to leave classroom

teaching and is not without precedent (e.g., Adams, 1996; Fleener & Dahm, 2007).

Using the SASS and T F S

sample of teachers in their first five years of teaching, two logistic regression models were

constructed. One of the regression models was informed by the first constellation of research

studies and use d

measures of working conditions as predictor

8

variables, while the other model, informed by the secon d constellation

of research studies , used

variations in teacher characteristics to predict attrition in the first five years of teaching. This study then compare d

the two models in several ways to determine which, if either, offers a statistically and sub stantively better prediction of teacher attrition. Finally, a combined model of the most predictive elements within each model was constructed to examine how these elements combine and interact to predict early teacher attrition.

Scope of this Study

Despit e the rigor of the research design and methodology, there are inherent limitations to the inferences that can be drawn from this work. For example, approaching the phenomenon of teacher attrition using multiple logistic regression models implicitly assume s (1) that the causal relationship is in one direction; (2) each factor contributes to the model independently (although interactions will be included in the model); and (3) the contribution of each factor is a fixed amount, meaning that the relationship b etween the factor and the outcome does not vary over time. This limits the

ability to consider potential feedback influences, such as a snowball effect of attrition where, for example, teachers who leave are influenced by the propensity of teachers around

them to leave. This research methodology also prevents understanding precisely the way these factors influence the teacher’s decision t o leave the classroom. That is, the

model s

capture the direction and magnitude of the effect, but this model does not c apture the mechanism of how these factors influence the observed effect.

There are also limitations due to the data collection procedure and the structure of the data set. For example, using responses from a single SASS and T FS

administration cycle pres ents a possible his tory effect. The data are collected in a single year, so teachers who have already left the profession in years prior to the survey administration are not surveyed and may

9

have different reasons for leaving than their former colleagues did in this particular year. For example, it is possible that economic conditions or educational policies in the survey year may have had a special influence on the decisions of teachers to leave the profession that does not exist in other years.

In the y ears that these data were collected (2003 - 04 and 2004 - 05) the nation was in a relatively stable economic environment.

These limitations, along with others, are discussed in greater detail in Chapter 5.

Despite the limitations on the inferences that can be drawn from this research, it remains an important endeavor. Results

of this work

will

allow researchers to understand how two different constellations of research about

teacher attrition, teacher

characteristics

and working conditions, compare ac ross a na tional data set in a recent SASS administration. This work will allow researchers to better understand the ways in which teachers’

characteristics

and working conditions can influence teachers’ career decisions. With this knowledge, intervention designer s and policy makers will have a better understanding of where to focus limited resources to have the greatest impact on the attrition of teachers, either in the preservice stage of teacher development or during the beginning years of their inservice career s. These interventions can lead to not only improved teacher induction practices and mentoring of new teachers, but also improved methods for preparing teachers to deal with the inservice stresses they are undoubtedly going to endure. Furthermore, this re search serve s

to inform future research examining critical points to provide support and interventions for either preservice or inservice teachers. Ideally, this research along with the work of others will help to transform teaching from “a profession tha t eats its young” to one that prepares and supports its newcomers (Osborne, 1992) .

10

Organization of this Dissertation

This dissertation

comprises five

chapters. Chapter 2 includes a review of the empirical and conceptual work that has framed the discussion of teacher attrition. Most specifically, the objective of the literature review is

to provide an explanation of how other

researchers

have examined

the influences on teachers’ career decisions and to understand how the present work fits into this larger discussion. A rticles included in the review were

garnered though electronic database searches as well as an existing stand alone literature review ( Johnson, Berg, & Donaldson

2005), a meta - analysis (Borman & Dowling, 2008)

and a literature review from another dissertation (Scheopner, 2009). Databases searched include the ERIC database, which includes journals from Resources in Education

and Current I ndex to Journals in Education , and EconLit, which houses the American Economic Association’s Journal of Economic Literature

and the Index of Economic Articles . The review focus es

on scholarly work that seeks to differentiate the characteristics of teacher s with different career trajectories within the United States at public and private K - 12 schools. This chapter contains five sections, the first two sections describe the big picture of teacher attrition, the definition and the scope of the problem. The n ext two sections

also describes the different perspectives of the two constellations of research studies as well as how the

perspectives

overlap.

Chapter 3 describe s the proposed methods of analysis for this study. Specifically, this chapter describes the

rationale supporting the uses of survey responses to answer these research questions and a clarification of the statistical modeling employed and why those models were chosen over other models. Finally, this chapter discusses the integrity of the research

design and the limitations of the results based on the research design and data collected.

11

In Chapter 4, I present the results of the analyses described in Chapter 3. Beginning with a detailed description of the sample in the analysis, the chapter then d escribes the results of the inferential statistical models. This chapter shows that the statistical models, informed by the two constellations of research, provide statistically significant predictions of the subsequent teaching status of beginning teache rs and that the working conditions model is a slightly better fit to the observed data than the teacher characteristics model. However, this chapter also shows that both models are able to statistically explain the residuals of the other model, implying th at neither offers a sufficient picture of why teachers leave. This chapter then shows how the two models can be combined to provide a more complete model of teacher attrition.

In Chapter 5, I discuss the implications of the findings described in Chapter 4 . It is in this chapter that the proper and improper inferences from these results are described and placed in context with previous research. This chapter offers suggestions for future work and calls for additional focus on the careers of beginning teac hers.

12

Chapter 2 : Review of the Literature

In 1983, the United States Department of Education released a report entitled A Nation at Risk

(Gardner, 1983) , in which the authors argued that the dismal quality of the knowledge amongst youth in the United S tates put the country at a disadvantage technologically and that this disadvantage would eventually result in a reduction of the nation’s security. In the same decade, several reports suggested impending teacher shortages due to two converging changes in t he demographics within schools:

Full document contains 191 pages
Abstract: This paper describes the literature on teacher attrition as either focusing on the working conditions faced by beginning teachers or highlighting variations in teachers' characteristics as causes for early teacher attrition. This study uses responses to the School and Staffing Survey (SASS) along with the Teacher Follow-Up Survey (TFS) to compare these contrasting views of early teacher attrition. Two logistic regression models were constructed and their relative efficacy in explaining teacher attrition were compared using three statistical techniques; model fit characteristics (e.g. pseudo-R 2 , Akaike Information Criteria, Bayesian Information Criteria); a comparison of their classification effectiveness, and results from Davidson and MacKinnon's J test (1981). A final model was also constructed using the predictive elements of each of the previous models. Results suggest that the working conditions model better fits the observed data than the teacher characteristics model. The final model highlights the importance of teacher commitment and engagement in the profession in teachers' career decisions.