What Do Teaching Qualifications Mean in Urban Schools? A Mixed-Methods Study of Teacher Preparation and Qualification

Abstract

The inequitable distribution of teachers in high-needs areas and the failure of teacher education programs have recently become focal points in the discussion of how to provide a quality education to all students. To address this concern, reformers have responded by mandating specific qualifications for teachers in all schools. These mandates have been established, however, without a real understanding of what these qualifications mean. This article adopts a mixed-methods approach to understanding what qualifications measure for novice teachers in urban districts in terms of teacher efficacy and one-year retention. Analysis of data from the Schools and Staffing Survey (2009-2010) and qualitative interviews reveal that qualifications do predict teacher efficacy, to an extent, yet they do not predict teacher retention. More research needs to be done to identify measurable qualifications that can actually predict what will happen in the first year of teaching.

Keywords

teacher efficacy highly qualified teacher teacher preparation urban teaching teacher retention

Introduction

In the current climate of educational accountability, the inequitable distribution of teachers and the “failure” of teacher education programs have become focal points in the discussion of how to provide a quality education to all students (Duncan, 2009). This discussion, furthermore, continues to be especially salient for schools serving poor minority students in urban areas who desperately need high quality teachers. The teachers in these schools in particular, according to rhetoric, are unprepared, ineffective, and transitory (Darling-Hammond & Green, 1990; Ingersoll, 2001; Jacob, 2007; Thernstrom & Thernstrom, 2003). The research community has been unable to determine not only how best to prepare teachers for the tasks of teaching and staying in these high-needs urban schools but also how to measure and evaluate incoming teacher qualification (Levine, 2006). Therefore, politicians and school districts continue to evaluate incoming teachers based on those readily available preparation-related credentials highlighted by the federal government such as certification, test scores, and the amount of coursework completed. These credentials, due to the immense diversity of certification pathways, have lost a good deal of their practical meaning (Constantine et al., 2009).

The lack of knowledge regarding how to train teachers for high poverty/high minority urban areas and the lack of distinction in the teacher credentialing process have created a policy problem that is especially detrimental to the urban districts that contain a majority of the high poverty/high minority schools in the United States (Knight & Wiseman, 2005; Wilson, Floden, & Ferrini-Mundy, 2001). In addition to the many challenges associated with educating students living in poverty, the teachers in these schools are generally less experienced and have much higher rates of attrition (Ingersoll, 2004). These schools also tend to hire later than wealthy suburban districts, have trouble attracting (and keeping) teachers, and have the fewest resources available to thoroughly evaluate incoming teachers (Ingersoll, 2004). This article examines the use of two readily available measures of incoming teacher qualification—amount of teacher education coursework and the highly qualified teaching credential as methods for predicting the teaching confidence and retention of incoming and novice teachers in high poverty/high minority urban schools. Although it is difficult to identify the precise methods districts use to make hiring decisions, it is clear that large urban school districts rarely have the resources to evaluate candidates beyond incoming qualifications. Data from the Teacher Policy Research center lends support to this assumption, as they provide documentation that school districts and principals are most likely to consider general information (such as degree, certification status, and major; Boyd et al., 2010). Politicians need, first, to know more than whether a teacher holds specific credentials—but also what those credentials mean—before any reforms can be made to teacher education policy.

In this article, I adopt a mixed-methods approach to addressing the question: To what extent do readily available, surface measures of incoming teacher qualification predict teaching confidence and teacher retention in high poverty/high minority urban schools? In doing so, I begin to create a crucial chain of evidence situated in the context of high poverty/high minority schools linking teacher preparation, first, to teacher efficacy (individual teaching confidence) and then to teacher retention (see Figure 1; Cochran-Smith, 2005). In creating this chain of evidence, it is important to connect what happens before teaching to what happens during teaching and to a longitudinal outcome. It is this type of empirical evidence that the field of teacher education research is lacking and, therefore, is unable to assist schools and districts to make informed decisions when hiring new teachers who can only document inputs at the time of hiring. Current research on teacher education rarely addresses what happens after teachers begin teaching. Research on preservice teachers is mostly confined to discussion of the experiences that they have with coursework and during the internship (Earley, 2005; Hollins & Guzman, 2005; Sleeter, 2001). There is definitely a need to increase research on teacher quality using outputs (such as student test scores; Cochran-Smith, 2005; Goe, 2007), but 1st-year teachers can only be evaluated using inputs at the time of hiring; therefore, it is crucial to understand the relationship between inputs and outputs.

Figure 1.

A chain of evidence

Theoretical Background

Teachers in High Poverty/ High Minority Schools

Teachers in schools serving poor minority students in large urban areas face a myriad of challenges—challenges that are almost all related to the fact that the students are often living in poverty. These students are also the most likely to be taught by teachers with the least teaching experience. Because these schools are most likely to employ novice teachers, and they educate a population in need, it is important to understand how well their incoming qualifications predict what happens in their 1st years of teaching. The students in these schools come from neighborhoods burdened with gang violence, have the highest rates of dropping out and getting pregnant, and have the lowest scores on standardized achievement tests (Gregory, Skiba, & Noguera, 2010; Lankford, Loeb, & Wykoff, 2002; Milanowski et al., 2009; Noguera & Wells, 2011). In addition to (or perhaps the reason for) the gap in student achievement, the schools that these students currently attend also suffer from a significant disadvantage in terms of elements such as funding and parental involvement, and these students have consistently been the most likely to encounter teachers who are new to teaching and enter the classroom with the least preparation (Buddin & Zamarro, 2009; Darling-Hammond & Green, 1990; Desimone & Long, 2010; Horng, 2009; Lankford et al., 2002; Levin, 2009; Talbert, 1990). To close the achievement gap and with the knowledge that those students on the lower end of the achievement gap are more likely to be taught by inexperienced teachers, the federal government mandates that all students be taught by teachers who hold accepted qualifications. Unfortunately, without an understanding of what those qualifications predict, it is difficult to determine the impact of the mandate.

One goal articulated under the reigning federal education policy, No Child Left Behind (NCLB), was to place a “highly qualified teacher” (HQT) in every classroom of every school by the end of the 2005-2006 school year (U.S. Department of Education, Office of Postsecondary Education, 2004). According to NCLB, “in order to be considered highly qualified under NCLB, teachers must hold a bachelor’s degree, have full state certification, and demonstrate competency in the core academic subjects they teach” (U.S. Department of Education, Office of Postsecondary Education, 2006, p. 1). Increasing the number of these HQTs has become one of the major avenues through which policy makers hope to improve schools located in large, urban areas. However, schools primarily serving disadvantaged students, especially ethnic minorities and those in poor urban and rural areas, are less likely to staff highly qualified teachers and have consistently been the furthest from accomplishing the HQT goal (Allgood & Rice, 2002; Clotfelter, Ladd, & Vigdor, 2010; Peske & Haycock, 2006). Although previous research has demonstrated mixed results regarding the HQT provision and student achievement (Boyd, Grossman, Lankford, Loeb, & Wykoff, 2006; Marszalek, Odom, LaNasa, & Adler, 2010), these results are inconclusive and this discussion has yet to include either teacher cognitions or teacher retention as outcome measures.

One of the major reasons regarding the need for an accurate way to measure the quality of teacher preparation regards the diversity of teacher credentialing pathways in the United States—diversity that reaches beyond the standard debate over traditional and alternative certification. In one major study concerning the debate between alternative and traditional certification, researchers found that there was very little difference regarding student outcomes between the two primary certification pathways (Constantine et al., 2009). They further explained that the most likely reason for this finding was the extensive diversity of program requirements within each of the pathways. In another study (Achinstein, Ogawa, & Speiglman, 2004), the researchers discovered that a unique interplay between state policy, teacher background (including personal background and preparation), and local context was responsible for sorting teachers into different tracks divided by “social” disparities among students and teachers (p. 592). The disparities found between groups of teachers included measures of program quality as well as more individualized measures of preservice achievement. Although these studies include information that highlights our lack of understanding regarding the measurement of preparation quality, no study has specifically questioned these two most readily available measures of assessing our incoming teachers: HQT status and coursework completion.

Admittedly, researchers do need to investigate the quality of teachers in, and teacher preparation for, high poverty/high minority urban schools. In reality, however, simply monitoring teacher credentials can be a problematic solution. The federal definition of a “qualified teacher” is based almost entirely on the assumption that teachers holding certain credentials will be more likely to ensure student success. In other words, they assume that a qualified teacher necessarily means a well-prepared teacher. In this study, I challenge the current assumptions concerning the inherent relationship between preparation and practice (and between qualified and well prepared) and investigate the way that the urban context impacts this relationship.

The assumptions that are discussed in this article have remained unchallenged in part because researchers have yet to find an alternative against which to measure incoming qualifications. Few articles, furthermore, have drawn attention to the importance of context in determining whether a teacher is well prepared. Drawing on the importance of school context, research on teacher preparation has yet to fully address the critical question: “prepared for what?” (Talbert-Johnson, 2006). Without access to student test scores or principal evaluations, I have chosen two measures that can help the research community better understand how prepared incoming teachers are for the at-risk urban environment. The first measure, teacher efficacy, documents an individual’s perceptions of how well prepared he or she is for teaching. The second, teacher retention, indicates whether a teacher remains in a given school after 1 year of teaching—and can indicate how resilient that 1st-year teacher was. Although there are concerns with each of these measures—namely, that teacher efficacy is wholly dependent on an individual’s perception and that teacher retention can be influenced by any number of elements beyond preparation for teaching—this research is still a crucial step in understanding how well teachers are prepared for the urban environment.

Teacher Efficacy

Teacher efficacy measures a teacher’s cognitions (perceptions) related to his or her capacity as a teacher—In other words, teacher efficacy measures a teacher’s perception of preparedness (Tschannen-Moran, Hoy, & Hoy, 1998). In this study, I utilize teacher efficacy as a method of beginning to understand how well prepared teachers are. Notably, researchers have been able to establish empirical relationships between high teacher efficacy and positive teaching behaviors, a willingness to remain in teaching and student achievement—Therefore, it is an important metric to utilize (Armor et al., 1976; Ashton & Webb, 1986; Shidler, 2009; Skaalvik & Skaalvik, 2010; Tschannen-Moran & Hoy, 2001; Tschannen-Moran et al., 1998).

Teacher efficacy consists of two distinct, but related, sets of beliefs: personal teacher efficacy (PTE) and general teacher efficacy (GTE; Tschannen-Moran et al., 1998). There are some crucial differences between the two concepts that require some attention (Figure 2). PTE measures how prepared a teacher feels to engage in and construct learning for his or her students. GTE, on the other hand, represents a type of outcome expectancy—a teacher’s confidence in his or her ability to encourage tangible success from his or her teaching regardless of the situation or context (Tschannen-Moran et al., 1998).

Figure 2.

Differences between personal teacher efficacy (PTE) and general teacher efficacy (GTE)

In studying these cognitions, researchers have linked various measures of teacher efficacy with stress levels, motivation to implement innovation and remain in the teaching profession, and student achievement (Armor et al., 1976; Ashton & Webb, 1986; Shidler, 2009; Skaalvik & Skaalvik, 2010; Tschannen-Moran & Hoy, 2001; Tschannen-Moran et al., 1998). Teacher efficacies can affect teacher behavior in a number of different ways. Gibson and Dembo (1984), for example, reported that teachers with higher levels of teacher efficacy were less likely to give up on a failing student, more likely to divide students into small groups for instruction, and less likely to criticize incorrect responses. Concerning student outcomes, Ashton and Webb (1986) reported that GTE and PTE greatly increased the amount of variance explained (24% and 46%, respectively) in a regression of student math scores. Research reveals that higher GTE teachers in particular tend to have students who outperform the students of lower GTE teachers on standardized tests (Bruce, Esmonde, Ross, Dookie, & Beatty, 2010; Goddard, Hoy, & Woolfolk Hoy, 2004; Ross, 1992). Research further shows that the presence of teachers with higher teacher efficacy predicts students with higher academic efficacy (or positive self-concept; Anderson, Greene, & Loewen, 1988; Tschannen-Moran et al., 1998). These studies highlight the importance of fostering teacher efficacy, not just for the teacher’s well-being but also as a factor influencing how long a teacher is willing to remain in a school and for the academic success of students.

Questions regarding perceived efficacy are specifically necessary for research on novice teachers (who are still actively engaged in learning) because, at its core, teacher efficacy is situated in theories of learning (social and social cognitive learning theories; Tschannen-Moran et al., 1998). Novice teachers have often recently completed preservice learning and are involved in the beginning stages of inservice learning. Novice teachers also generally experience a significant decrease in teacher efficacy during the 1st year of teaching. Although levels of teacher efficacy usually rise and stabilize, they never return to the levels found in preservice teachers (Hoy, 2000). Understanding these complex learning processes is crucial to improving and analyzing how teacher qualification is measured because they point to how prepared a teacher feels based on those qualifications. The goal of this study, therefore, is to understand the relationship between these crucial cognitions and teaching credentials of novice teachers.

Teacher Retention

To extend literature on what the readily available measures of teacher qualifications realistically mean in the United States, I analyze the relationship between these credentials, teacher efficacy, and teacher retention for novice urban teachers. Based on the research outlined in this section, it is clear that one important aspect of “prepared for what” is be “prepared to remain in high poverty/high minority urban schools.” Teacher retention, whether a teacher stays in his or her current school (or teaching altogether), is an extremely important concept in understanding the health of schools based on the positive impact on students of measures such as length of teaching experience and stability of a school (Boyd et al., 2010; Fetler, 1999; Murnane & Phillips, 1981; Watlington, Shockley, Guglielmino, & Felsher, 2010).

Quantitative studies examining the possible benefits of teacher retention find that increased teaching experience generally has a positive impact on student learning, but researchers also find that this effect levels off after the first 3 to 5 years of teaching (Ferguson & Ladd, 1996; Gordon, Kane, & Staiger, 2006). If teachers in high poverty/high minority urban schools leave before their 3rd year of teaching, not only are they unlikely to have fulfilled their full potential but the schools also have to consistently replace teachers—often with teachers who have not yet reached the critical 3-year mark. Teacher retention, although complex and far from an explicit measure of teacher quality, is still a necessary element in understanding how well teachers are prepared for teaching and remaining in high poverty/high minority urban schools.

Retaining teachers is especially challenging for schools in high poverty/high minority urban schools (Boyd, Grossman, Lankford, Loeb, & Wykoff, 2007; Boyd et al., 2010; Sass, Seal, & Martin, 2011). Teacher turnover has been quantitatively linked to crucial challenges that are especially relevant to and difficult for high poverty/high minority urban schools (Hanushek & Rivkin, 2010). Ingersoll (2004) found that schools in an “urban poverty” category had the highest rates of teacher turnover in a study using 1999-2000 Schools and Staffing Survey (SASS) and Teacher Follow-Up Survey (TFS). His findings, which include indicators from urban teachers such as high levels of student discipline problems, poor student motivation, inadequate time, and classroom intrusions, create a vivid description of what it means to teach in a high poverty/high minority school. These factors are either more severe and problematic in urban schools or (and) teachers are not prepared to cope with these factors. In a recent study of teachers in a high-needs urban district in Texas, Hanushek and Rivkin (2010) found that, although veteran teachers who chose to leave the district tended to be “less effective” than their counterparts who chose to stay, 1st-year teachers who changed schools or moved to another district were often more effective than those 1st-year teachers who chose to stay.

Retention of 1st-year and novice teachers takes on a slightly different and more complicated meaning than the discussion of teacher retention writ large, and the topic is worthy of unpacking further. First-year teachers who leave high poverty schools may be the more effective teachers and, for example, have the opportunity to go elsewhere. Furthermore, according to previously cited research, teacher effectiveness improves over the first few years of teaching; thus, it is necessary to understand the preparation experiences and perceptions of those who stay and those who leave.

In addition, a great deal of qualitative research offers more contextual and nuanced information related to teacher retention. Many studies, for example, document the importance of administrative support in the 1st year of teaching (Gonzalez, Brown, & Slate, 2008; Waddell, 2010), whereas others reference qualitative program factors related to developing certain dispositions (Freedman & Appleman, 2009). In this study, I address the issue of novice teacher retention from the perspective of teacher credentialing to assess the relationship between HQT status, amount of Education coursework, perceptions of preparedness (known as teacher efficacy), and the decision to remain in an urban school—a topic that has yet to be studied in this manner.

Method

To gain a greater understanding of the chain of evidence that links teacher preparation, teacher efficacy, and teacher retention, I conducted a mixed-methods sequential-explanatory study, which involved the collection and analysis of quantitative data followed by the collection and analysis of qualitative data aimed at specifically answering questions generated by the first phase of empirical analysis (Ivankova, Creswell, & Stick, 2006). In regard to the chain of evidence, the quantitative phase of research established the linkages, whereas the qualitative phase brought nuance, context, and understanding to each link in the chain. For the first, quantitative phase, I used nationally representative survey data from the SASS (2007-2008), TFS (2009), and the Common Core of Data (CCD), all obtained from the National Center for Education Statistics (NCES). For the second, qualitative phase, I conducted interviews with novice teachers who were currently or previously teaching in high poverty/high minority urban schools. Not only were the qualitative questions generated during the quantitative phase but the qualitative sample was also situated as a subset of the quantitative sample through administering surveys similar to the SASS to the qualitative sample. Figure 3 provides an overview of the design of the study as a whole.

Figure 3.

Design of the study

Phase 1: Quantitative Analysis

Research questions

To address the overarching question, “To what extent do accepted measures of incoming teacher qualification predict teacher efficacy and teacher retention in high poverty/high minority urban schools?”, I focus on the following specific questions:

Research Question 1: What is the relationship between the amount of teacher preparation (as defined by number of education courses completed and length of student teaching) and teacher efficacy among novice teachers in high poverty/high minority schools?

Research Question 2: What is the relationship between HQT status and teacher efficacy among novice teachers in high poverty/high minority schools?

Research Question 3: What is the relationship between the amount of teacher preparation (as defined by number of education courses completed and length of student teaching) and 1-year teacher retention among novice teachers in high poverty/high minority schools?

Research Question 4: What is the relationship between HQT status and 1-year teacher retention among novice teachers in high poverty/high minority schools?

Based on literature previously cited regarding the nature of teacher efficacy and teacher retention (Chester & Beaudin, 1996; Gibson & Dembo, 1984; Hanushek & Rivkin, 2010; Hoy, 2000; Ingersoll, 2004), I hypothesized that this research would find a positive, significant relationship between amount of teacher preparation and the level of teacher efficacy among novice teachers in high poverty/high minority schools. I further hypothesized that I would uncover a significant relationship between both teacher preparation and teacher efficacy and teacher retention among 1st-year teachers.

Data and sample

To empirically evaluate these research questions, I analyzed data from the 2007-2008 SASS, which is conducted by the NCES, to empirically connect teacher preparation, teacher efficacy, and teacher retention. In 1987, NCES began collecting survey data from teachers and principals on school conditions, teacher shortages, and teacher and principal characteristics, among other things. Since that time, NCES has been through six cycles of the SASS and the TFS, which they administer 1 year after SASS to a sample of teachers who participated in SASS. SASS is administered to representative samples of public, private, tribal, and charter schools drawn from the CCD and the Private School Universe Survey (PSS), which include all elementary and secondary schools in the United States. Because of this sampling frame, where teachers are drawn from a representative sample of schools, these data technically represent nested data (teachers are nested within schools), which became relevant during analysis (Tourkin et al., 2010). I was able to link SASS data to the NCES CCD, in which the U.S. Department of Education annually gathers demographic and fiscal data from all public school districts and state education agencies, to accurately measure school-level variables (Chen, 2009). Accurate school-level variables are an important consideration, namely, because the stated research questions focus explicitly on the urban context. This study focused on public (including charter) schools, and data were mostly drawn from the teacher questionnaire, which examines preparation, attitudes, teaching methods, and experiences.

Because of my focus on novice teachers and the relationship between preparation, practice, and retention, I chose to restrict the sample to those teachers who began teaching after the 2003-2004 school year (those with less than 4 full years of full-time teaching: full sample, n = 37,709; novice sample, n = 9,134). Descriptive statistics (Table 1) reveal that there are some noticeable differences between the novice and full samples. Notably, only 80% of novice teachers are “highly qualified,” as compared with 88% in the full sample. Furthermore, 15% of novice teachers teach in at-risk schools, as compared with 11% of the full sample. The novice sample, furthermore, is 70% female, 83% White, and has a mean age of 31.5 years. For the longitudinal retention analysis, I elected to only use 1st-year teachers. I made this decision for two reasons; first, because TFS data were collected for all 1st-year teachers (and only a subsample of other teachers), therefore, it represents a more complete picture. Second, I made this decision because those teachers in their 2nd and 3rd years of teaching have already passed the 1st-year threshold. If I were to elect to look at all 3 years, it would cause inconsistencies, as I cannot measure how many teachers leave each year. Teachers, therefore, who began teaching in 2005 and 2006 but left after the 1st year of teaching would not be captured in retention data, so as an accurate measure of 1-year retention, I chose to only use 1st-year teachers (n = 1,759).

Table 1.

Descriptive Statistics for Variables Asked in Analysis of the 2007-2008 Schools and Staffing Survey and 2008-2009 Teacher Follow-Up Survey

	Entire sample (n = 37,710)		Novice sample (n = 9,130)
Variable	M	SD	M	SD
Female	0.69	0.46	0.70	0.46
White	0.87	0.34	0.83	0.37
Highest degree	2.58	0.65	2.33	0.57
Age	42.74	11.68	31.49	8.88
Years teaching experience	12.98	10.48	1.67	0.14
Northeast	0.14	0.35	0.13	0.34
Midwest	0.26	0.44	0.24	0.43
West	0.25	0.43	0.26	0.44
Southeast	0.34	0.47	0.37	0.48
Elementary	0.36	0.48	0.35	0.48
Secondary	0.52	0.50	0.52	0.50
Combined	0.12	0.33	0.12	0.33
HQT	0.88	0.32	0.80	0.40
Preparation^a	0.00	0.39	−0.09	0.43
High poverty/high minority	0.11	0.32	0.15	0.36
PTE^a	—	—	0.00	0.91
GTE^a	0.00	0.80	−0.02	0.82
	N = 3,830		n = 1,760
Stays	0.57	0.50	0.74	0.44

Note: HQT = highly qualified teacher; PTE = personal teacher efficacy; GTE = general teacher efficacy.

Factor scores.

Major Variables

In the following section, I describe the major variables used in the analysis and provide a description of how each is measured. The goal of the study was to uncover empirical connections between several of the variables outlined below.

High poverty/high minority urban school

I created this variable by combining urbanicity, percentage of minority students in the school, and the percentage of students receiving free- and reduced-price lunches (a very basic measure of poverty). In this case, I label a school as “high poverty/high minority urban school” if it is located in or on the outskirts of a large or midsized city, if it has more than 50% minority enrollment, and if more than 50% of students receive free- or reduced-price lunch.

Preparation level

This is a factor score, which combines self-report information regarding the number of education courses taken and the length of student teaching into one continuous variable. Higher scores indicate a higher level of preparation. This variable fails to capture many of the qualitative aspects of teacher preparation, and it does not include very specific elements of coursework; however, the listed information is often the most easily available information for administrators evaluating possible hires.

HQT status

This is a dichotomous variable that indicates whether the teacher meets the federal requirements for a HQT. This question was asked directly on the survey and was defined for the teachers as follows: “Generally, to be highly qualified, teachers must meet requirements related to (a) a bachelor’s degree, (b) full state certification, and (c) demonstrated competency in the subject area(s) taught. The HQT requirement is a provision under NCLB” (SASS).

Teacher efficacy

This is measured using two variables to capture the two types of teacher efficacy discussed above. PTE is a factor score specifically measuring a teacher’s perceptions of his or her abilities to orchestrate the necessary actions to encourage learning. Low scores on PTE indicate that a teacher felt less prepared to organize and run successful lessons. GTE is also a factor score but measures a teacher’s perceived ability to affect outcomes regardless of circumstances (Tschannen-Moran et al., 1998). A low score on the measure of GTE indicates that a teacher has a low expectation regarding his or her capability to produce student outcomes in the current context.

Teacher retention

This is a dichotomous variable that measures whether a 1st-year teacher has remained in the same school into his or her 2nd year of teaching.

Control variables

Following previous research on teacher retention (Ingersoll, 2001) and teacher efficacy (Chester & Beaudin, 1996; Greenwood, Olejnik, & Parkay, 1990), I controlled for teacher sex (male/female), teacher race (White/non-White), highest degree (ranging from bachelor’s to PhD), teacher age, years of teaching experience, school level (elementary/secondary), and region of the country.

Data Analysis

To evaluate the relationships between these variables, I utilized multinomial ordinary least squares (OLS) and logistic regression. However, because of the nested structure of the data (teachers drawn from a nationally representative sample of schools), my final analyses are based on multilevel models that I created using the “xtmixed” and “gllamm” commands (Rabe-Hesketh, Skrondal, & Pickles, 2005) in STATA to observe and account for any school-level variation.

After quantitative analysis found support for the linkages in the chain of evidence, I was able to craft questions for participant interviews to bring deeper meaning to the linkages (see below).

Phase 2: Qualitative Phase

Research question

How do novice teachers in urban schools explain each of the relationships found in quantitative research?

Participant selection

The, second, qualitative phase of the study was designed specifically to answer why the above research questions were either supported or not supported. To answer these “why” questions, I conducted interviews with novice urban teachers over the phone and in person. Although I was unable to directly contact teachers who had participated in the SASS survey, I was able to situate my sample as a subsample of the nationally representative SASS sample. To compare these teachers within the context of the SASS sample, 25 novice teachers in high poverty/high minority urban schools were asked to participate in an online survey containing all relevant items from the SASS teacher questionnaire. I created composite scores for measures of preparation level, PTE, and GTE for the data I collected as well as for the SASS data to compare the two samples. To locate participants, I used convenience and snowball sampling methods (Maxwell, 2005). I relied on personal contacts of known teachers in urban schools and I contacted alumni offices of teacher education programs in urban, suburban, and rural areas surrounding one northeastern city, to achieve a diverse sample. Alumni offices were able to connect me with graduates who had accepted jobs in urban schools. Unfortunately, as I could only interview teachers who responded to requests to complete the electronic survey, there is a response bias in this sample. In addition, as I utilized convenience and snowball sampling, the qualitative sample is not random and may not be entirely representative of 1st-year teachers in urban schools.

Of the 25 teachers who completed the online survey and agreed to participate in follow-up interviews, I selected 14 teachers for initial phone interviews. In selecting the initial pool of 14 teachers, I stratified participants based on scores for PTE and GTE as well as age, gender, preparation type, and school level. Because the majority of respondents were White, all non-White teachers were contacted and participated in phone interviews. Teachers in the initial pool had varying levels of preparation, PTE and GTE, and different student teaching experiences (urban and nonurban placements, for example). Table 2 outlines the characteristics of the 14 teachers participating in interviews. The highlighted scores indicate that the teacher is below the average score on a given measure for the SASS sample. From the initial pool of 14 teachers, I chose five teachers for in-person follow-up interviews (key informants¹). I selected these teachers for diversity of experience and perspective as well as responsiveness during the initial interview. The five teachers who I selected as key informants, therefore, were stratified not only on the previous factors but also qualitatively on diversity of experiences and openness to discussion. After interviewing the fifth key informant, it was clear that I had achieved data saturation. It was clear that I had achieved data saturation when new concepts did not emerge in the last two initial interviews. In other words, the last two participants were answering questions in similar ways to previous participants. In Table 2, the five teachers selected for intensive interviews are marked with asterisks. I chose not to restrict the sample to a specific grade level or subject area, again to better connect qualitative data with SASS data.

Table 2.

Participant Scores on PTE/GTE

Participant	Grade level/subject	Preparation	PTE	GTE
Nell^a	Fifth grade	0.80	2.29	1.67
Shelly	Ninth-grade English	0.80	2.14	2.33
Brad^a	High school Spanish/social studies	0.92	2.29	2.33
Hadley	Fifth grade	1.00	2.57	2.33
Rachel	High school, special education math	1.00	2.57	2.67
Laura^a	Third grade	1.00	2.86	2.33
Tom	High school physics	1.00	2.43	3.00
Hana	Second grade	0.92	3.14	2.67
Natalie	Fourth grade	1.00	3.00	3.00
Noelle^a	High school English	1.00	3.14	3.00
Rhea^a	Fourth- and fifth-grade science and social studies	1.00	3.00	3.33
Annie	High school math	1.00	3.43	3.00
Kent	Ninth-grade English	1.00	3.29	3.33

Note: PTE = personal teacher efficacy; GTE = general teacher efficacy. Shaded scores indicate that the participant scored above the average in the novice population of SASS.

Key informants.

Procedures

Participants were asked to partake in two separate interviews. During the initial phone interviews (which lasted around 20 min), 14 teachers were asked to comment on their experiences during their first years of teaching. During the second, (in person) interview, 5 of those 14 teachers were asked to spend 1 hr discussing the specific results found in the quantitative phase of the study. Teachers were given definitions of each of the teacher efficacies, asked as the following questions: Why do novice teachers in urban schools have lower feelings of teacher efficacy?; What explains the relationship between general teacher efficacy and teacher retention?; And what elements of teacher preparation affect feelings of teacher efficacy and retention? After interviews were transcribed and analyzed by the researcher, the five key informants were sent a two-page summary of findings and asked to comment on the interpretation given. The comments and reactions that these teachers provided were then integrated into an updated assessment of findings. Interviews were analyzed through coding and re-coding using a priori categories drawn from the hypotheses that were refined during analysis. Themes were developed to directly to respond to the quantitative findings; therefore, analysis was divergent from traditional qualitative analyses. Initial line-by-line coding of interviews identified themes without a priori categories, but themes were assessed for relationship to quantitative findings (supporting, explaining, refuting) and initial codes were sorted according to those relationships. Unrelated data are excluded from these analyses as they are outside the scope of this article. Because teachers were asked to respond directly to findings, the results presented in the following section are organized such that the findings from quantitative analysis are stated and then explained using the qualitative interview data.

Results

Table 3 reveals that, controlling for teacher and school characteristics, there is a positive, significant relationship between level of preparation and PTE. In other words, although teachers in high poverty/high minority schools generally have lower levels of PTE, those who have completed more coursework and had lengthier student teaching experiences are likely to have higher feelings of self-confidence related to day-to-day teaching tasks (PTE). Table 4, however, shows that there is a weaker, but still positive and significant, relationship between preparation level and GTE. Despite the negative relationship between teaching in a high poverty/high minority school and GTE, those teachers who have taken more education courses and completed lengthier student teaching practices are also slightly more likely to report feeling able to encourage success in students in spite of outside obstacles.

Table 3.

Results of a Multi-Level Linear Regression of Teacher Background Characteristics, Preparation and At-Risk Teaching on PTE (n = 9,130)

	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Fixed effects
Intercept	.00 (.01)	−.01 (.06)	.06 (.07)	.09 (.07)	.12 (.07)	.15* (.07)
Teacher characteristics
Female		−.02 (.02)	−.02 (.02)	−.04 (.02)	−.02 (.02)	−.04 (.02)
White		.13*** (.03)	.10*** (.03)	.06 (.03)	.06* (.03)	.02 (.03)
Highest degree		.07*** (.02)	.07*** (.02)	.04 (.02)	.08*** (.02)	.04* (.02)
Age		−.00*** (.00)	−.00*** (.00)	−.00 (.00)	−.00*** (.00)	−.00 (.00)
Total experience		−.04*** (.01)	−.04*** (.01)	−.06*** (.01)	−.04*** (.01)	−.05*** (.01)
HQT			—	.08** (.02)	—	.07** (.02)
Preparation			—	.52*** (.03)	—	.52*** (.02)
School-level factors
Elementary (comparison)			—	—	—	—
Secondary			.01 (.02)	.06** (.02)	−.01 (.02)	.04 (.02)
Combined			−.03 (.03)	.01 (.03)	−.05 (.03)	−.01 (.03)
Northeast (comparison)			—	—	—	—
Midwest			.02 (.03)	−.02 (.03)	.02 (.03)	−.02 (.03)
Southeast			−.11*** (.03)	−.04 (.03)	−.11*** (.03)	.04 (.03)
West			−.10*** (.03)	−.16*** (.03)	−.10* (.03)	−.15*** (.03)
At risk					−.17*** (.03)	−.16*** (.03)
Random effects
Intercept	.03 (.01)	.02 (.01)	.02 (.01)	.01 (.01)	.02 (.01)	.01 (.01)
Residual	.80 (.02)	.80 (.02)	.80 (.02)	.76 (.02)	.80 (.02)	.76 (.01)
Model fit statistics
AIC	24,220.87	24,177.36	24,179.64	23,642.98	24,148.7	23,614.36
BIC	24,220.87	24,234.32	24,272.19	23,749.77	24,248.38	23,728.27
ICC	.04
Variance—Teacher	—	.01	.00	.05	.01	.06
Variance—School	—	.10	.26	.62	.32	.66

Note: HQT = highly qualified teacher.

p < .05. **p < .01. ***p < .001.

Table 4.

Results of a Multi-Level Linear Regression of Teacher Background Characteristics, Preparation and At-Risk Teaching on GTE (n = 9,130)

	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Fixed effects
Intercept	.03** (.01)	.10 (.06)	.63*** (.07)	.62*** (.07)	.80*** (.07)	.79*** (.07)
Teacher characteristics
Female		.13*** (.02)	.04 (.02)	.04 (.02)	.04 (.02)	.04 (.02)
White		.08** (.03)	.05 (.03)	.05 (.03)	−.04 (.03)	−.04 (.03)
Highest degree		−.05* (.02)	−.03* (.02)	−.04 (.02)	−.03 (.02)	−.03 (.02)
Age		−.00 (.00)	.00 (.00)	.00 (.00)	.00 (.00)	.00 (.00)
Total experience		−.03*** (.01)	−.03*** (.01)	−.04*** (.00)	−.04*** (.01)	−.04*** (.01)
HQT				.02 (.03)		.02 (.03)
Preparation				.06* (.02)		.06* (.03)
School-level factors
Elementary (comparison)			—	—	—
Secondary			−.61*** (.03)	−.60*** (.03)	−.67*** (.03)	−.66*** (.03)
Combined			−.37*** (.04)	−.37*** (.04)	−.42*** (.04)	−.42*** (.04)
Northeast (comparison)			—	—	—	—
Midwest			−.02 (.04)	−.02 (.04)	−.02 (.04)	−.02 (.03)
Southeast			−.28*** (.04)	−.27*** (.04)	−.26*** (.04)	−.25*** (.04)
West			−.23*** (.04)	−.24*** (.04)	−.21*** (.04)	−.22*** (.04)
At risk					−.58*** (.03)	−.58*** (.03)
Random effects
Intercept	.37 (.02)	.24 (.01)	.28 (.02)	.28 (.02)	.24 (.02)	.24 (.02)
Residual	.68 (.01)	.43 (.01)	.68 (.01)	.68 (.01)	.67 (.01)	.67 (.01)
Model fit statistics
AIC	25,770.65	25,743.39	25,161.94	25,169.29	24,874.21	20,849.99
BIC	25,792.01	25,800.35	25,254.5	25,276.09	24,973.89	20,963.91
ICC	.36
Variance explained: Teacher		.00	.00	.00	.01	.01
Variance explained: School		.03	.24	.24	.35	.35

Note: HQT = highly qualified teacher.

p < .05. **p < .01. ***p < .001.

Research Question 2: What is the relationship between HQT status and teacher efficacy among novice teachers in high poverty/high minority schools?

Table 3 displays a significant, positive relationship between HQT status and PTE. In other words, teachers classified as “highly qualified” (those who are certified in their subject area) are more likely to report higher levels of confidence related to day-to-day teaching tasks. Table 4, on the other hand, demonstrates that there is no significant relationship between HQT status and GTE. Those teachers who are deemed “highly qualified” are no more or less likely to report confidence in their abilities to encourage student success despite outside influences. Teachers in high poverty/high minority urban schools, therefore, who are certified, are no more or less likely to display high feelings of GTE than those who are uncertified.

Qualitative Results for Research Questions 1 and 2

When teachers were asked to address the relationship between readily available measures of incoming teacher qualification (preparation level and HQT status) and the cognitions underlying PTE or GTE, teachers outlined different factors that can build teacher efficacy. Analysis of their comments indicates that entirely different sets of experiences are responsible for developing PTE and GTE, in which those experiences that shape GTE are much more dependent on context. This would explain why PTE is more likely than GTE to be influenced by preparation level and certification status. Teachers indicated that some of these experiences are present in preparation programs, whereas others are not. In regard to developing GTE (“What I do matters for student outcomes”), participants cited three major influences: qualitative program differences, individual disposition and, school-level factors (none of which are captured in the HQT or preparation variables). They further explained that PTE (“I can do it!”), alternatively, is more likely to be influenced by the length of student teaching and increased subject-area knowledge—two elements that are directly tied to HQT status and preparation level. The elements teachers associated with developing positive feelings of GTE, on deeper analysis, were either not captured in the variables measuring preparation level and highly qualified teaching status or have little to do with the preparation at all and are specific to the teaching context.

One of the qualitative teacher education program differences responsible for building a sense of GTE, for example, was an explicit philosophy of education. In discussing this particular program element, teachers stressed that it was not just the presence of an educational philosophy but also that the philosophy had meaning for the urban teaching context. The two following comments exemplify the essential components of the program philosophies discussed by Kent and Noelle (who both expressed above average GTE scores)—In both cases, the teacher explains that it was crucial to know him or her on a critical level and to understand how his classroom fits into broader society.

That’s where I have gotten a real sense of purpose from my preparation—and it goes back to feeling like I know why I am teaching on a really critical level, and I have a good sense of what my racial identity is and what my cultural identity is—and what the cultural identity of the students are, I feel like I can go in there and have some success in terms of opening the students minds. (Kent, high school English)

My Philosophy of Ed—borrowed from school—is transforming hearts and minds . . . their philosophy of Ed is really focused on having us know who we are when we walk into a classroom, before deciding how you are going to teach. If you don’t know who you are and what you value, it is hard to uphold that in a classroom when you have kids who question you. So my approach was that kids had to be constantly making connections, so connections across like curriculum and connections across cultures so that they can start to see how things are not totally separated. (Noelle, high school English)

One of the major commonalities between these statements is the shared goal is of opening students’ minds, a goal that helps a teacher find a sense of purpose and leads a teacher to believe that all students can achieve success. Having a philosophy of education, like these two teachers explain, allows a teacher to feel more confident in his or her ability to encourage student success, because his or her teaching has meaning.

In discussing PTE, on the other hand, participants expressed that training programs provided them with tangible resources (especially lesson plans), which acted to decrease uncertainty and increase personal teacher efficacy. These more tangible sources of teacher efficacy, which they felt had a greater impact on PTE than on GTE, could be learned in a student teaching placement or during education coursework. Teachers specifically discussed the importance of collecting tested lesson plans and management tools as sources of higher levels of teaching confidence. These two elements were essential for feelings of PTE because they helped teachers to feel mentally prepared and in control each day as they walked into a classroom. Brad, for example, stated,

The thing that was most successful for me was that my preparation was very practical, very targeted towards teaching in an urban school. I felt more prepared then I expected for a lot of things, like planning lessons especially . . . I feel like the specific tangible stuff I found very useful. We had to do a lot of lesson planning—and they got graded and handed back go us, like really rigorous. (Brad, high school Spanish and Social Studies)

Lesson planning was necessary to participants because, from their perspective, if they walked into the classroom knowing what they were going to do, they were more likely to maintain control of the classroom. Hana further explained, “having resources from student teaching made it that much easier for me . . . A lot of people didn’t realize that it was my 1st year teaching” (Hana, second grade). The sources of PTE, according to comments, were easily built into teacher education programs—and were, at least partially, captured by the HQT credential and the preparation level variable.

Teachers in high poverty/high minority urban schools, according to the participants in this study, encounter a number of stressors specific to the urban environment that they feel unprepared to cope with. These stressors and a lack of preparation geared toward urban teaching can lead to lower levels of teacher efficacy. Because PTE is more easily generated during teacher preparation and is less dependent on context, HQT status and preparation level are more likely to have a positive influence on those cognitions—the “I can do it!” kind of teacher efficacy. GTE, on the other hand, is not only more vulnerable to school-level factors but positive GTE is also more likely to be developed through nonquantifiable elements of teacher training programs. The “What I do can make a difference” kind of teacher efficacy, therefore, is less likely to be related to commonly measured qualifications.

Research Question 3: What is the relationship between the amount of teacher preparation (as defined by number of education courses completed and length of student teaching) and teacher retention among 1st-year teachers in high poverty/high minority schools?

Table 5 demonstrates that there is no relationship between the level of teacher preparation and teacher retention. Those 1st-year teachers in high poverty/high minority schools who have completed more education coursework and engaged in lengthier student teaching practices are no more or less likely to remain in their original school after 1 year. Although this is troubling, a possible explanation for the lack of relationship is found when teacher efficacy is added into the equation (see Research Question 4).

Table 5.

Results of Multilevel Logit Analysis of the Influence of School and Individual Predictors, Including Teacher Efficacy, on 1st-Year Teacher Retention

	Stay in same school (n = 2,190)
	Null	Conditional
Background characteristics
Female		−.09 (.13)
White		.16 (.16)
Age		.01 (.01)
Highest degree		−.21 (.11)
School characteristics
Northeast (comparison)		—
Midwest		.06 (.21)
Southeast		.02 (.20)
West		−.06 (.21)
Elementary (comparison)		—
Secondary		.56*** (.15)
Combined		.07 (.18)
At risk		.26 (.18)
Preparation factors
Preparation		.27 (.14)
HQT		.01 (.13)
Teacher efficacy
PTE		.02 (.07)
GTE		.27*** (.08)
Constant	1.12*** (.12)	.91* (.42)
Random effects
Variance	.41 (.47)	.37 (.48)
Model fit statistics
AIC	2,024.549	2,018.631
BIC	2,035.494	2,106.192

Note: HQT = highly qualified teacher; PTE = personal teacher efficacy; GTE = general teacher efficacy.

p < .05. **p < .01. ***p < .001.

Research Question 4: What is the relationship between HQT status and teacher retention among 1st-year teachers in high poverty/high minority school?

Table 5 also reveals that there is no relationship between HQT status and teacher retention among 1st-year teachers in high poverty/high minority schools. Those 1st-year teachers in high poverty/high minority schools who are labeled as highly qualified, therefore, are no more or less likely to remain in their original school after 1 year. Although neither preparation level nor HQT status is related to 1st-year teacher retention in high poverty/high minority urban schools, GTE is positively related to teacher retention. In other words, those 1st-year urban teachers who have confidence in their abilities to overcome external influences and encourage positive outcomes from students are more likely to remain in their original school after 1 year: GTE matters for teachers in urban schools.

Qualitative Results for Research Questions 4 and 5

When asked to discuss possible reasons for changing schools or leaving teaching altogether, participants overwhelmingly cited cognitions associated with low GTE (“What I do matters”) as having the potential to influence teacher retention. When teachers experienced low feelings of GTE they reported feeling completely powerless. Feeling powerless or meaningless to students and administrators was often cited as a reason for teachers to change schools or leave teaching. Noelle provides a cohesive summary of the impact that low teacher efficacy can have on teacher retention:

Those who resign in the middle of the year or plan on leaving often look at the situation as: “I am doing everything that I can, and the kids are not performing and I can’t do it anymore, what am I supposed to do here? My boss is telling me that it is all on me . . . I am the teacher; I need to make things happen. I am doing everything that I can humanly do, but the kids are not putting in what they need to.” (Noelle, high school English)

In this statement, although there are hints of low PTE (“What else am I supposed to do?”), a low GTE related to students and administrators is clearly the major factor. The hypothetical teacher described by Noelle feels that students do not try and that administrators are making the task far more difficult than it already is; for this reason, she feels powerless (low GTE) and leaves the school or teaching altogether.

Kent, in the following statement, highlights how feeling like “You are making a difference” can help to prevent attrition:

As often, you know, when I am with friends and I talk about the experience, I tend to talk about the negative aspects, just because it is so strong, also, there are students who are there that want to learn. Even in that crazy 6th period class—who want to learn. You know I got an apology letter from a student, she said, you know “I want to move my seat” and that makes you want to stay there. So . . . that’s why, I want to do it for those types of students.

When student apathy or administrative pressures cause teachers to feel as though they cannot make a difference in the lives of their students (low GTE), they sometimes will choose to leave the school where they are teaching. Kent explains, however, that when a teacher feels empowered and that he or she is making a difference—he or she may be more inclined to keep working in that school. The perspective offered by the qualitative sample provides support and contextualization for the quantitative results regarding the impact of GTE on 1-year teacher retention. Teachers, they explain, need to feel that they can make a difference to their students to keep trying and to remain in teaching. From their perspective, perceived student apathy and unsupportive administrators, sources of low GTE, could lead to teacher attrition. Because GTE is not related to HQT status and only weakly associated with preparation level, it makes sense that the commonly used qualification measures will not influence retention because it does not measure the school level (like level of administrative support) and preparation factors that matter for GTE and retention.

Discussion and Conclusion

The results presented in this article make a strong argument for the lack of clarity regarding the traditional means of measuring teacher qualification. although taking into account the number of education courses and the length of student teaching and looking for “highly qualified teachers,” school districts and administrators are able to predict a teacher’s level of PTE, those measures do a poor job of predicting the GTE of incoming teachers, and an even worse job of predicting retention. Readily available measures of quality of teacher preparation can predict teachers’ feelings of confidence in the classroom but do not predict feelings of confidence in producing student outcomes in the face of outside influences. The latter statement is especially troubling in high poverty/high minority urban schools—where the context of poverty creates significant challenges for teachers. Furthermore, neither preparation level nor highly qualified teaching status predicts 1-year teacher retention for 1st-year teachers in urban schools. Therefore, these easily accessible measures of preparedness do not provide necessary information for principals and administrators attempting to locate teachers able to remain in their schools. Interestingly, GTE (the feeling that you can produce positive outcomes in students despite outside influences), which is not measured by the variables evaluated under NCLB, does predict teacher retention. Regarding this finding, novice teachers in high poverty/high minority schools indicate that retention and GTE are most likely influenced by preparation factors such as program quality and a focus/philosophy on educating students living in poverty. Therefore, in addition to instruments based on federal qualifications, nonquantifiable screening interviews are an essential tool in the evaluation of incoming teacher candidates. More specifically, it is important that incoming teachers believe that they can help students in any context succeed at high levels.

In large urban school systems, these findings are especially troubling. Not only do those systems cope with the challenges of students living in poverty but also are responsible for staffing the largest number of schools—schools that, notably, have very high turnover rates. The findings presented herein pose a challenge to these districts because they indicate that those credentials and qualifications that are most commonly reviewed do not actually present the whole picture of that teacher’s likelihood of success. These districts have the least amount of time and resources with which to evaluate their applicants and are rarely able to “dig in” and investigate program quality or qualities, but they should.

In addition, although not a focus of the quantitative study, qualitative participants consistently recognized the importance of strong leaders in holding students to high standards and for attracting teachers to remain in the school. This is an issue that may not be addressed through altering teacher preparation and teacher hiring but lends support to research on the effects of administrators. Although preparation programs cannot address these leadership issues, they can work to build strong philosophies of education and continue to give teachers tangible resources for teaching (lesson-planning and classroom management experience). Preparation programs can also encourage preservice teachers to engage in student teaching in urban schools to give them experiences with urban students and urban school systems prior to entering the field. In this sense, teacher education programs can learn a good deal from urban teacher residencies, and when applied to traditional teacher preparation programs, this process could have a greater impact.

Based on these findings, I conclude that there is clearly a problem regarding the measurement of quality of preparation for teachers entering high poverty/high minority urban schools that desperately need a quality teaching force. Future research needs to identify measurable elements of teacher quality that will enable districts to accurately predict teacher retention and GTE, confidence regarding abilities to overcome obstacles to ensure positive student outcomes. If a teacher does not believe that students in a given school can succeed, it is much less likely that he or she will remain in that school. To retain teachers, it is important that high poverty/high minority schools hire teachers who can articulate a belief in the success of the students in that school.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by The Pennsylvania State University Research Initiation Grant.

Notes

About the Author

Sarah Anne Eckert, PhD, is an assistant professor of education and Co-Coordinator of Professional Development Schools at Notre Dame of Maryland University. Her research focuses on education policy in general, and more specifically on teacher preparation policy for high poverty/high minority urban schools currently and historically.

References

Achinstein

Ogawa

R. T.

Speiglman

(2004). Are we creating separate and unequal tracks of teachers? The effects of state policy, local conditions, and teacher characteristics on new teacher socialization. American Educational Research Journal, 41(3), 557-603.

Allgood

Rice

J. K.

(2002). The adequacy of urban education: Focusing on teacher quality. In Roelke

C. F.

Rice

J. K.

(Eds.), Fiscal policy issues in urban education (pp. 155-180). Greenwich, CT: Information Age Publishing.

Anderson

Greene

Loewen

(1988). Relationships among teachers’ and students’ thinking skills, sense of efficacy, and student achievement. Alberta Journal of Educational Research, 34(2), 148-165.

Armor

Conroy-Oseguera

Cox

King

McDonnell

Pascal

Zellman

(1976). Analysis of the school preferred reading programs in selected Los Angeles minority schools (No. R-2007-LAUSD). Santa Monica, CA: RAND.

Ashton

P. T.

Webb

R. B.

(1986). Making a difference: Teachers’ sense of efficacy and student achievement. New York, NY: Longman.

Boyd

Grossman

Ing

Lankford

Loeb

Wykoff

(2010). The influence of school administrators on teacher retention decisions. American Educational Research Journal, 20(10), 1-31.

Boyd

Grossman

Lankford

Loeb

Wykoff

(2006). How changes in entry requirements alter the teacher workforce and affect student achievement. Education Finance and Policy, 1(2), 176-219.

Boyd

Grossman

Lankford

Loeb

Wykoff

(2007). Who leaves? Teacher attrition and student achievement. Albany, NY: Teacher Policy Research.

Bruce

C. D.

Esmonde

Ross

Dookie

Beatty

(2010). The effects of sustained classroom-embedded teacher professional learning on teacher efficacy and related student achievement. Teaching and Teacher Education, 26(8), 1598-1608.

10.

Buddin

Zamarro

(2009). Teacher qualifications and student achievement in urban elementary schools. Journal of Urban Economics, 66, 103-115.

11.

Burgess

R. G.

(1982). Field research: A sourcebook and field manual. London, England: Allen & Unwin.

12.

Burgess

R. G.

(1983). Experiencing comprehensive education: A study of Bishop McGregor School. New York, NY: Methuen.

13.

Chen

C.-S.

(2009). Documentation to the NCES Common Core of Data Local Education Universe Survey Preliminary File: School Year 2008-09 (No. NCES 2009-354). Washington, DC: Institute of Education Sciences, U.S. Department of Education.

14.

Chester

M. D.

Beaudin

B. Q.

(1996). Efficacy beliefs of newly hired teachers in urban schools. American Educational Research Journal, 33(1), 233-257.

15.

Clotfelter

C. T.

Ladd

H. F.

Vigdor

J. L.

(2010). Teacher credentials and student achievement in high school: A cross subject analysis with student fixed effects. Journal of Human Resources, 45(3), 655-681.

16.

Cochran-Smith

(2005). Studying teacher education: What we know and need to know. Journal of Teacher Education, 55(1), 3-7.

17.

Constantine

Player

Silva

Hallgren

Grider

Deke

(2009). An evaluation of teachers trained through different routes to certification, final report (NCEE 2009-4043). Washington, DC: National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences, U.S. Department of Education.

18.

Darling-Hammond

Green

(1990). Teacher quality and equality. In Goodlad

Keatings

(Eds.), Access to knowledge: An agenda for our nation’s schools (pp. 237-259). New York, NY: College Entrance Examination Board.

19.

Desimone

Long

D. A.

(2010). Teacher effects and the achievement gap: Do teacher and teaching quality influence the achievement gap between black and white and high- and low-SES students in the early grades? Teachers College Record, 112(12), 3024-3073.

20.

Duncan

(2009). Teacher preparation: Reforming the uncertain profession. New York, NY: Teachers College, Columbia University.

21.

Earley

(2005). Standards alignment and government policy: Unresolved policy dilemmas. In After standards alignment (pp. 29-68). Amherst, MA: National Evaluation Systems. Retrieved from http://www.pearsonassessments.com/pai/ea/ES_Pub/ES_Publications.htm?ResourceId=es_pub_2005

22.

Ferguson

R. F.

Ladd

H. F.

(1996). How and why money matters: An analysis of Alabama schools. In Ladd

H. F.

(Ed.), Holding schools accountable: Performance-based reform in education (pp. 265-298). Washington, DC: The Brookings Institution.

23.

Fetler

(1999). High school staff characteristics and mathematics test results. Education Policy Analysis Archives, 7(9). Retrieved from http://epaa.asu.edu/ojs/article/view/544

24.

Freedman

S. W.

Appleman

(2009). “In it for the long haul”: How teacher education can contribute to teacher retention in high-poverty, urban schools. Journal of Teacher Education, 60(3), 323-337.

25.

Gibson

Dembo

(1984). Teacher efficacy: A construct validation. Journal of Educational Psychology, 76(4), 569-582.

26.

Goddard

R. D.

Hoy

A. W.

Woolfolk Hoy

(2004). Collective efficacy beliefs: Theoretical developments, empirical evidence, and future directions. Educational Researcher, 37(3), 3-13.

27.

Goe

(2007). The link between teacher quality and student outcomes: A research synthesis. Washington, DC: National Comprehensive Center for Teacher Quality.

28.

Gonzalez

Brown

M. S.

Slate

J. R.

(2008). Teachers who left the teaching profession: A qualitative understanding. Qualitative Report, 13(1), 1-11.

29.

Gordon

Kane

T. J.

Staiger

D. O.

(2006). Identifying effective teachers using performance on the job. Washington, DC: Hamilton Project (Brookings Institution).

30.

Greenwood

G. O.

Olejnik

S. P.

Parkay

F. W.

(1990). Relationships between four teacher efficacy belief patterns and selected teacher characteristics. Journal of Research & Development in Education, 23(1), 102-106.

31.

Gregory

Skiba

R. J.

Noguera

(2010). The achievement gap and the discipline gap: Two sides of the same coin. Educational Researcher, 39(1), 59-68.

32.

Hanushek

E. A.

Rivkin

S. G.

(2010). Constrained job matching: Does teacher job search harm disadvantaged urban schools? (Working Paper No. 15816). Cambridge, MA: National Bureau of Economic Research.

33.

Hollins

Guzman

M. T.

(2005). Research on preparing teachers for diverse populations. In Cochran-Smith

Zeichner

(Eds.), Studying teacher education: The report of the AERA panel on research and teacher education (pp. 477-549). Mahwah, NJ: Lawrence Erlbaum.

34.

Horng

E. L.

(2009). Teacher tradeoffs: Disentangling teachers’ preferences for working conditions and student demographics. American Educational Research Journal, 46(3), 690-717.

35.

Hoy

A. W.

(2000, April). Changes in teacher efficacy during the early years of teaching. Paper presented at the Annual meeting of the American Educational Research Association, New Orleans, LA.

36.

Ingersoll

R. M.

(2001). Teacher turnover and teacher shortages: An organizational analysis. American Educational Research Journal, 38(3), 499-534.

37.

Ingersoll

R. M.

(2004). Why do high-poverty schools have difficulty staffing their classrooms with qualified teachers? Washington, DC: Center for American Progress.

38.

Ivankova

N. V.

Creswell

J. W.

Stick

S. L.

(2006). Using mixed-methods sequential explanatory design: From theory to practice. Field Methods, 18(1), 3-20.

39.

Jacob

B. A.

(2007). The challenge of staffing urban schools with effective teachers. Future of Children, 17(1), 129-153.

40.

Knight

Wiseman

(2005). Professional development for teachers of diverse students: A summary of the research. Journal of Education for Students Placed at Risk, 10(1), 23-37.

41.

Lankford

Loeb

Wykoff

(2002). Teacher sorting and the plight of urban schools: A descriptive analysis. Educational Evaluation and Policy Analysis, 24(1), 37-62.

42.

Levin

(2009). Enduring issues in urban education. Journal of Comparative Policy Analysis, 11(2), 181-195.

43.

Levine

(2006). Educating school teachers. Washington, DC: The Education Schools Project.

44.

Marszalek

J. M.

Odom

A. L.

LaNasa

S. M.

Adler

S. A.

(2010). Distortion or clarification: Defining highly qualified teachers and the relationship between certification and achievement. Educational Policy Analysis Archives, 18(27), 1-29.

45.

Maxwell

J. A.

(2005). Qualitative research design: An interactive approach (Vol. 42, 2nd ed.). Thousand Oaks, CA: SAGE.

46.

Milanowski

A. T.

Longwell-Grice

Saffold

Jones

Schomisch

Odden

(2009). Recruiting new teachers to urban school districts: What incentives will work? International Journal of Education Policy and Leadership, 4(8), 1-13.

47.

Murnane

R. J.

Phillips

(1981). Learning by doing, vintage, and selection: Three pieces of the puzzle relating teaching experience and teacher performance. Economics of Education Review, 2(4), 453-465.

48.

No Child Left Behind Act of 2001, 20 U.S.C. § 6319 (2009).

49.

Noguera

Wells

(2011). The politics of school reform: A broader and bolder approach for Newark. Berkley Review of Education, 2(1), 1-21.

50.

Peske

H. G.

Haycock

(2006). Teaching inequality: How poor and minority students are shortchanged on teacher quality. Washington, DC: The Education Trust.

51.

Rabe-Hesketh

Skrondal

Pickles

(2005). Maximum likelihood estimation of limited and discrete dependent variable models with nested random effects. Journal of Econometrics, 128(2), 301-323.

52.

Ross

J. A.

(1992). Teacher efficacy and the effect of coaching on student achievement. Canadian Journal of Education, 17(1), 51-65.

53.

Sass

Seal

A. K.

Martin

N. K.

(2011). Predicting teacher retention using stress and support variables. Journal of Educational Administration, 49(2), 200-215.

54.

Shidler

(2009). The impact of time spent coaching for teacher efficacy on student achievement. Early Childhood Education Journal, 36(4), 453-460.

55.

Skaalvik

E. M.

Skaalvik

(2010). Self-efficacy efficacy and teacher burnout: A study of relations. Teaching and Teacher Education, 26(4), 1059-1069.

56.

Sleeter

C. E.

(2001). Preparing teachers for culturally diverse schools: Research on the overwhelming presence of whiteness. Journal of Teacher Education, 52(2), 94-106.

57.

Talbert

J. E.

(1990). Teacher tracking: Exacerbating inequalities in the high school. Stanford, CA: Center for Research on the Context of Secondary Teaching, Stanford University.

58.

Talbert-Johnson

(2006). Preparing highly qualified teacher candidates for urban schools: The importance of dispositions. Education and Urban Society, 39(1), 147-160.

59.

Thernstrom

(2003). No excuses: Closing the racial gap in learning. New York, NY: Simon & Schuster.

60.

Tourkin

Thomas

Swaim

Cox

Parmer

Jackson

Cole

(2010). Documentation for the 2007-08 Schools and Staffing Survey (No. NCES 2010332). Washington, DC: Institute for Education Sciences, U.S. Department of Education.

61.

Tremblay

M. A.

(1957). The key informant technique: A nonethnographic application. American Anthropologist, 59(4), 688-701.

62.

Tschannen-Moran

Hoy

A. W.

(2001). Teacher efficacy: Capturing an elusive construct. Teaching and Teacher Education, 17, 783-805.

63.

Tschannen-Moran

Hoy

A. W.

Hoy

W. K.

(1998). Teacher efficacy: Its meaning and measure. Review of Educational Research, 68(2), 202-248.

64.

U.S. Department of Education, Office of Postsecondary Education. (2004, March). No Child Left Behind Act of 2001: Qualifications for teachers and professionals. Washington, DC: Author.

65.

U.S. Department of Education, Office of Postsecondary Education. (2006). The secretary’s fifth annual report on teacher quality: A highly qualified teacher in every classroom. Washington, DC: Author.

66.

Waddell

J. H.

(2010). Fostering relationships to increase teacher retention in urban schools. Journal of Curriculum and Instruction, 4(1), 70-85.

67.

Watlington

E. M.

Shockley

Guglielmino

Felsher

(2010). The high cost of leaving: An analysis of the cost of teacher turnover. Journal of Education Finance, 36(1), 22-37.

68.

Wilson

S. M.

Floden

R. E.

Ferrini-Mundy

(2001). Teacher preparation research. Seattle, WA: Center for the Study of Teaching and Policy Center for the Study of Teaching and Policy, University of Washington.