Ties That Bind? The Teaching and Post-Teaching Trajectories of Black and Latino/a Community Insiders and Elite College Graduates

Structural Changes to the Profession and the Historical Exclusion of Minoritized Teachers

For much of the 20th century, most U.S. teachers, inclusive of mathematics teachers, remained in teaching for an entire career (Ingersoll & Merrill, 2011). In part, this was because opportunities for career advancement in K–12 education were largely restricted to middle-class, White, Protestant males (Lortie, 1975). As with others in their age cohorts (AbouAssi et al., 2021), recent generations of teachers expect teaching to be one of several jobs they will have in their lifetimes (Bayer et al., 2009). Social and economic changes in the latter half of the 20th century, including greater opportunities for college-educated women and minorities outside of teaching, were accompanied by changes to teachers’ career trajectories. Currently, many teachers stay in teaching for only a few years, others leave teaching only to later return, others transition in and out of school leadership positions, and still others have long gaps in employment. Furthermore, rather than leave teaching entirely, the new generation of teachers often migrate from school to school in search of a supportive workplace (Simon & Johnson, 2015).

Social and economic changes in the 20th century also altered the composition of the U.S. teacher workforce. For example, the Black teacher population decreased dramatically following the 1954 Brown vs. Board of Education decision as White administrators resisted hiring them to teach in integrating schools (Hudson & Holmes, 1994). This was particularly harmful to Black students as, prior to this decision, Black teachers often were better prepared than White teachers and many either taught in the communities in which they were raised or “when not, they came from communities that were similar in culture and beliefs” (Walker, 2001, p. 769).

Early-entry alternative teacher certification programs, which emerged in the 1980s, have leveraged and contributed to changes to the teaching profession and workforce, having had an outsized effect on the composition of the workforce in low-income neighborhood urban schools. Through providing a fast track to paid teaching, and often also incentives like a subsidized master’s degree, these programs have facilitated the entry of “non-traditional” entrants to teach core subjects like mathematics (Carver-Thomas & Darling-Hammond, 2017; Donaldson & Johnson, 2010), including both community insiders, who tend to resemble the students they teach, and ECGs, who tend not to.

Community Insiders’ and ECGs’ Reasons for Entry

Maier (2012) asserts that ECGs choose to enter teaching through selective programs like TFA because they further credential them as nation’s best and brightest recent college graduates. Elite academic credentials, coupled with a 2-year experience in a neighborhood urban school, seem to enhance their post-teaching prospects of finding high-status, remunerative employment. To be certain, not all ECGs fit this profile (Donaldson & Johnson, 2010; Lovison, 2022). Some stay in neighborhood urban schools citing issues of social justice and altruism such as “seek[ing] to have direct impact on social inequality by helping young children become successful despite all the harsh circumstances that are part of their daily reality in the inner-city neighborhoods” (Tamir, 2009, p. 524).

Black and Latino/a teachers, including many who are community-based, more often enter teaching through alternative certification programs than other teachers (Carver-Thomas & Darling-Hammond, 2017). As with ECGs, many cite issues of social justice and altruism as motivating their entry, although their conceptions may differ. Black teachers tend to view teaching as community work and seek to be role models for their students (Callahan & Brantlinger, 2022; Kokka, 2016). Latino/a teachers similarly tend to be motivated by a commitment to working with and serving as role models for Latino/a students (Valenzuela, 2016). Reflective of this, Black and Latino/a teachers are at least twice as likely as White teachers to work in schools serving majority Black and Latino/a students (Ingersoll et al., 2019).

The Career Trajectories of Community Insiders and ECGs

In explaining teacher turnover, quantitative studies frequently distinguish between personal factors, school factors, and external factors (Borman & Dowling, 2008; Nguyen et al., 2019). Personal factors include teacher characteristics such as race and age, subject-matter qualifications, and the socioemotional benefits of working with students (Nieto, 2003). School factors include student demographics and organizational climate issues such as administrator support and teacher collegiality (Kokka, 2016; Nguyen et al., 2019; Simon & Johnson, 2015). External factors include accountability and other policies (Nguyen et al., 2019; Nieto, 2003) and economic realities such as teacher salaries and regional labor market conditions (Goldhaber & Theobald, 2022). Personal, school, and external factors are interactive; for example, accountability policies can influence school climate (Valli & Buese, 2007).

ECGs seem to leave teaching at higher rates than others for a range of reasons including dissatisfaction with teaching as a career (Donaldson & Johnson, 2010; Kelly & Northrop, 2015). Their turnover also may be driven by the fact that so many are community outsiders who have weak ties to the high-minority neighborhood schools that they teach in or from their having high-status career options (Maier, 2012). Moreover, selective fast-track programs may facilitate their turnover by placing them into difficult, full-time teaching positions with limited training, asking them to only commit to only 2 years of teaching, and doing little to incentivize them to stay (Brantlinger et al., 2022; Donaldson & Johnson, 2010).

Some posit that, due to their local ties and cultural congruence, community teachers will exhibit higher rates of retention than other teachers, particularly in low-income neighborhood schools (Carver-Thomas & Darling-Hammond, 2017; Gist et al., 2019). However, to our knowledge, no quantitative study has examined the retention of teachers who are community insiders and outsiders. Evidence about the high levels of commitment and retention of community insiders comes principally from qualitative studies or small-scale evaluations of individual certification programs (e.g., Brantlinger et al., in press; Brantlinger & Grant, 2022; Gist et al., 2019; Kokka, 2016). There also is a robust literature on Black and Latino/a teachers (e.g., Dixon et al., 2019; Valenzuela, 2016) that shows them to be rooted in the communities they teach, to share similar life experiences with their students, and to be committed to working in low-income Black and Latino/a communities.

Many quantitative retention studies allow for comparisons to be made between the retention of White, Black, and other non-White teachers. These studies suggest that, although there are differences, their retention behaviors and career decisions may not be that different. Nationally representative studies show that White teachers have modestly better rates of retention than Black and other non-White teachers (Carver-Thomas & Darling-Hammond, 2017; Ingersoll et al., 2019), whereas regional studies show that, in certain contexts at least, Black teachers exhibit better retention than White and other teachers (Sun, 2018). White and non-White teachers also have been shown to cite similar phenomena as factoring in their career decision-making. For example, as with other teachers (Simon & Johnson, 2015), Black and Latino/a teachers cite working conditions, particularly administrator support, as exerting a strong influence on their decisions to remain in certain schools (Ingersoll et al., 2019).

However, qualitative and mixed-methods studies (Dixon et al., 2019; Farinde et al., 2016; Frank et al., 2021) show that Black and Latino/a teachers face unique organizational asks and aggressions which factor in their career decisions. Cormier and Scott (2021) present the cases of a female Latino/a teacher tasked with being the school’s Spanish-speaking interpreter and a Black male special education teacher asked to serve as a disciplinarian for Black and Latino/a students in their schools. Both teachers “struggle to balance instruction and forced obligations to serve as racial experts, as racial justice advocates, and in nonacademic positions” (p. 236). Moreover, whereas White teachers are generally presumed to be competent by administrators and policymakers (Cooley et al., 2021), Black and Latino/a teachers find that they need to prove that they are competent to White administrators and colleagues (Dixon et al., 2019) and have comparatively limited opportunities for career advancement (Farinde et al., 2016).

Finally, it frequently is assumed that the career trajectories and retention of mathematics (and science) teachers will differ from those of other teachers because they can find professional and remunerative jobs in the science, engineering, and technology sectors. However, there is little evidence in support of this. For example, drawing on nationally representative data from 1999 to 2000, Ingersoll and Perda (2010) find that, although “the educational system does not enjoy a surplus of new mathematics and science teachers relative to losses” (p. 589), the turnover of these teachers was comparable with that of most other teachers.

Sample and Context

Launched in 2000 by the NYC Department of Education (NYCDOE) in collaboration with TNTP, NYCTF was a flagship program, resulting in more than two dozen replica Teaching Fellows programs nationally. At that time, policymakers expressed concern about both teacher shortages and teacher quality and “selective” programs were seen as addressing both concerns (Brantlinger et al., 2022). Over the past two decades, more than 50,000 teachers, including over 5,000 secondary mathematics teachers, have entered teaching through NYCTF. Many Fellows were the recent graduates from very selective colleges and very few of them had attended public schools in NYC as youths.

The study participants became paid teachers of record after successfully completing NYCTF’s 7-week summer preservice program in either 2006 or 2007. This program included 120 hours of university coursework, 40 hours of NYCTF-delivered training, and 40 hours of practice teaching (Brantlinger & Smith, 2013). Following this, they began paid teaching in either a middle or high school in NYC public schools while they continued to take master’s certification courses. NYCTF teachers were restricted to teaching positions in “high-needs” neighborhood schools in different regions throughout the city—which regions depended on their place of residence. The students they taught were from low-income backgrounds with, on average, 78.4% receiving free or reduced-price lunch. Most also were Latino/a and Black; according to NYCDOE, on average, the students in their first schools were 48.5% Latino/a, 37.6% Black, 6.9% Asian, and 6.5% White. Students categorized as Black included African Americans as well as those with family roots in Jamaica, other Caribbean nations, and sub-Saharan Africa. Students categorized as Latino/a included those of Puerto Rican, Dominican, Colombian, Mexican, Ecuadorian, and Salvadoran descent. Survey data showed that approximately 18% of these teachers considered themselves to be Black or part-Black.

Local labor market conditions might have influenced the teachers’ decisions to enter and remain in teaching (see, for example, Goldhaber & Theobald, 2022). They were in their first or second year when the Great Recession of 2008 began. At that time, NYC added jobs until September 2008 and, following that, lost jobs in a range of sectors including those that, like teaching, required a college degree (e.g., investment banking, accounting, law; DeFreitas, 2009). For example, these economic trends seemed to have spurred some of the career changers in the study to enter and remain in teaching until more attractive employment opportunities reappeared (Brantlinger, 2021; Hurst & Brantlinger, 2022).

Data Sources

Measures

Analyses

Using Stata software, we began the analysis by producing descriptive statistics for the four teacher subgroups. To examine our first research question about whether members of the four teacher subgroups who took the 2016 survey reported entering for similar or different reasons, we analyzed their interval-ratio scores on the four reasons for entry subscales using ANOVA. We then used Kruskal–Wallis tests to examine whether their ordinal scores on the individual reasons for entry items differed by subgroup. In cases where these tests were significant, we conducted post hoc tests comparing the reasons of entry responses of one teacher subgroup with those of the other three subgroups combined.

To answer our second research question about the full sample of teachers’ retention in district schools, we began with a descriptive analysis of their rates of first- and second-year attrition. We tested for group differences using chi-square tests for each of the four attrition outcomes of interest. When that was significant at the .05 level, we conducted two-sided t tests comparing the proportion of teacher leavers from each subgroup with the proportion of leavers in the remaining teacher population. We also employed logistic regression to model the retention of the four different subgroups in their first schools, the district, and teaching at 5 years (again using a cut point of 4.9 years). These models were estimated as follows:

ln ( p ∧ 1 − p ∧ ) = β 0 + β 1 X 1 + β 2 X 2 + β 3 X 3 .

As shown in Table 4, the coefficients β _i represent the expected change in the retention outcome (in log-odds units) for each vector of variables: their membership in one of the four teacher subgroups (X₁), additional teacher demographic characteristics (X₂), and first school contexts (X₃). Because 43.9% of the mathematics Teaching Fellows were clustered in the same schools, we calculated the intraclass correlations for the three retention outcomes. These were .221, .150, and .034, respectively, for school, district, and K–12 retention. The first two of which were sufficiently large that we used clustered robust standard errors in all the retention models to adjust for the clustering of the variance within schools. Missing data were eliminated listwise; the models included 97% (or 598) of the 617 observations, with missing data appearing to be random.

To answer our third research question on career trajectories, we analyzed the teachers’ distributions to different types of occupations about a decade after they entered NYCTF by subgroup. We specifically used a chi-square test of independence. As this was significant, we used two-sided comparison tests (t tests) of proportions as post hoc tests to understand whether any of the subgroups’ distributions to different occupational categories differed significantly from the three other subgroups combined.

Finally, to understand the teachers’ career decision-making, we conducted qualitative analyses of their responses to a series of open-ended survey prompts included on the 2016 survey. In particular, the teachers responded to different, but parallel, prompts depending on their status as either district leavers or district stayers. One set of prompts asked the stayers and leavers about the school and district characteristics that they believed explained their retention or attrition, respectively. In this case, open coding was used to develop six codes for the leavers’ responses (e.g., administrator support, classroom management) and five codes for the stayers’ responses (e.g., administrator support, collegial relationships). A follow-up item inquired about the personal attributes that the district stayers believed had contributed to their retention. In this case, open coding was used to develop seven codes (e.g., knowledge of mathematics, relationships with students).

Descriptive Profiles of the Teacher Subgroups

The ECGs, BLIs, White-Asian Insiders, and Nonelite Outsiders differed significantly on several demographic characteristics. Some of this was by design as college selectivity, high school location, and race/ethnicity were used to construct the teacher subgroups. However, as Table 1 shows, they also differed significantly by social class, age, gender, and career status at entry. In terms of their mathematical backgrounds, the subgroups did not seem to be that different as indicated by their rates of science, technology, engineering, and mathematics (STEM) degree and mathematics coursework completion prior to NYCTF.

OPEN IN VIEWER

Table 1

Descriptive Statistics by Teacher Subgroups

Teacher characteristics	Elite college graduate	Black-Latino/a insider	White-Asian insider	Nonelite Outsider	Total
N	205	83	55	274	617
College selectivity**
Very selective	100%	0%	0%	0%	33%
Moderately selective	0%	29%	44%	42%	26%
Less selective	0%	71%	56%	58%	41%
High school location**
In NYC	16%	100%	100%	0%	28%
<150 miles to NYC	29%	0%	0%	40%	27%
>150 miles to NYC	55%	0%	0%	60%	44%
Race/ethnicity**
White	55%	0%	69%	66%	54%
Asian	21%	0%	31%	14%	16%
Black	18%	66%	0%	14%	21%
Latino/a	6%	34%	0%	6%	9%
Postsecondary degree
Mathematics	8%	16%	7%	15%	12%
Other STEM field	22%	17%	17%	20%	20%
Postsecondary
Mean math courses	5	5	4	5	5
Median math courses	3	4	3	4	4
High school**
Public	74%	79%	66%	84%	81%
Social class*
Working/poor	22%	68%	52%	29%	29%
Middle	49%	30%	44%	49%	55%
Upper/upper-middle	29%	1%	4%	22%	17%
Gender**
Female	51%	60%	57%	48%	51%
Age at entry*
Mean (years)	28	30	32	29	29
Median (years)	24	26	26	25	25
Career status**
Recent college graduate	70%	56%	51%	67%	65%
Career changer	30%	44%	49%	33%	35%

Note. The descriptive statistics are for the full sample (N = 617) apart from career status which pertains to the survey sample (N = 389). Analysis of variance was used to test for significant differences on the mathematics course and teacher age variables. NYC = New York City; STEM = science, technology, engineering, and mathematics.

*

Significant at the p < .05 level. **Significant at the p < .01 level based on chi-square tests of independence for each nominal variable.

ECGs comprised a third (33%) of the NYCTF mathematics teacher population (Table 1). Recruited from the nation’s most selective colleges, they hailed from all over the United States and, to some extent, the globe (e.g., Chile). Most were community outsiders; only 16% were NYC high school graduates. Compared with all others, a significantly higher proportion of ECGs were graduates of private high schools. Although 22% had completed a postsecondary degree in a STEM field, only 8% had completed a postsecondary degree in mathematics. ECGs were the youngest subgroup typically entering teaching when they were 24 years old and only 30% were considered career changers by NYCTF. Although somewhat racially and ethnically diverse, ECGs were majority White (55%). A slim majority were female (51%). They were significantly more likely to be raised in upper or upper-middle class households than other NYCTF mathematics teachers. As this suggests, ECGs generally came from different social and economic circumstances than the predominantly Black and Latino/a students they would teach in NYC public schools.

BLIs comprised about 13% of the NYCTF mathematics teachers. They were approximately two-thirds Black and one-third Latino/a. A comparatively high proportion attended an undergraduate institution that Barron’s ranked as less selective or nonselective. Although not significantly different from all others, a modestly higher proportion of BLIs entered NYCTF with a postsecondary degree in mathematics. The majority (79%) graduated from a public high school (in NYC). BLIs skewed female and were few years older on average than ECGs and Nonelite Outsiders—but not White-Asian Insiders. Reflective of this, a comparatively large proportion of BLIs (44%) entered with prior career experience. BLIs also were more likely than others to have been raised in a working-class or low-income household. Thus, in many regards, BLIs seemed to resemble the low-income Black and Latino/a students they taught in neighborhood NYC public schools.

White-Asian Insiders comprised just under 9% of the NYCTF mathematics teachers. They were majority White (69%) and minority Asian (31%). The majority of White-Asian Insiders (56%) were graduates of less selective or nonselective colleges. Less than a quarter entered teaching with a postsecondary degree in mathematics or another STEM field. White-Asian Insiders were significantly more likely to have attended private high schools than the other teachers. As with BLIs, White-Asian Insiders skewed female. White-Asian Insiders were significantly older on average than all others and almost half (49%) were career changers. They were significantly more likely than ECGs and Nonelite Outsiders to have grown up in a working-class household—and, in this sense, were like BLIs.

Nonelite Outsiders were the largest subgroup, comprising about 44% of the NYCTF teachers. By subgroup definition, they graduated neither from a very selective college nor from a NYC high school. Most (84%) were public high school graduates. More than a third (35%) entered teaching with a postsecondary STEM degree and they were significantly more likely than ECGs and White-Asian Insiders—but not BLIs—to have earned postsecondary degree in mathematics. They were majority White (66%), tended to come from upper or upper-middle class (22%) and middle class (49%) backgrounds, and were equally split by gender. The typical Nonelite Outsider entered NYCTF when they were 25 years old, and most were recent college graduates (67%).

Reasons for Entry

NYCTF mathematics teachers cited many of the same reasons for becoming teachers irrespective of their subgroup membership. As the descriptive statistics in Table 2 show, the most highly cited reason-for-entry was “want[ing] a job with purpose or meaning” under the meaningful job subscale. The mean score for all four subgroups was well above “4,” which meant that most of the teachers identified this reason for entry as “very important” or “absolutely essential.” Two reasons-for-entry items that all four groups cited at similarly high levels of importance were: “wanting to make a social difference,” under the altruism subscale, and “the subsidized master’s degree,” under the alternative certification subscale.

OPEN IN VIEWER

Table 2

NYCTF Mathematics Teachers’ Reasons for Entry, Reported by Subgroup

Teacher characteristics	Elite college graduate	Black-Latino/a insider	White-Asian insider	Nonelite outsider
N	125	55	37	172
Meaningful job subscale	0.03 (1.05)	−0.27* (0.94)	0.33* (1.00)	0.00 (0.98)
I wanted a job with purpose or meaning.	4.37 (0.84)	4.45 (0.74)	4.54 (0.56)	4.25 (0.82)
I was tired of my previous job.	2.10* (1.47)	2.91** (1.60)	2.62 (1.48)	2.26 (1.45)
Altruism subscale	0.00 (0.95)	0.71* (0.94)	−0.20* (0.99)	−0.18* (0.98)
I wanted to make a social difference.	4.14 (0.94)	4.33 (0.75)	3.92 (1.09)	3.96* (1.03)
I wanted to work with students of color.	3.25 (1.25)	3.95** (1.25)	2.76* (1.14)	2.99 (1.28)
I wanted to give back to my community.	2.86 (1.50)	4.25** (0.93)	3.35 (1.32)	2.74** (1.48)
Alternative certification subscale	−0.09 (0.99)	−0.06 (1.02)	0.12 (1.00)	0.06 (1.00)
I liked the subsidized master’s degree.	3.87 (0.97)	4.22* (0.99)	3.89 (0.91)	4.04 (0.91)
I wanted to start paid teaching before completing state certification requirements.	3.54 (1.33)	3.62 (1.39)	3.97 (1.19)	3.72 (1.30)
Job benefits subscale	−0.20* (1.02)	0.27* (1.07)	0.24 (0.97)	0.01 (0.93)
I needed a job with a decent salary.	3.49 (1.09)	3.60 (1.26)	3.57 (1.30)	3.46 (1.07)
I liked the job benefits and/or security of teaching.	3.18* (1.18)	3.65 (1.17)	3.70 (1.15)	3.59 (1.02)
Additional items—not on a subscale	N/A	N/A	N/A	N/A
I wanted to become a career teacher or teach until retirement.	2.47** (1.24)	3.40** (1.24)	3.51** (1.15)	2.90 (1.32)
I thought that being a Teaching Fellow would look good on my resumé.	2.36 (1.32)	2.09 (1.18)	2.00 (0.93)	2.36 (1.30)
I saw teaching as a stepping stone to educational administration.	1.60* (0.98)	2.13** (1.25)	1.68 (0.97)	1.77 (1.08)

Note. The table reports means and standard deviations (in parentheses) for the four continuous subscales (all standardized) and their constituent Likert–scale items. Teachers responded to the items on a 5-point subscale where “1 = not important at all” and “5 = absolutely essential.” NYCTF = New York City Teaching Fellows.

*

Significant at the .05 level. **Significant at the .01 level determined by post hoc two-way comparisons—subgroup responses to all other responses following a significant analysis of variance test for the subscale scores or a significant Kruskal–Wallis test for the individual survey items.

As a group, ECGs expressed the least clarity about why they had entered teaching. As Table 2 shows, their scores were at or below the group mean on all four subscales and their mean scores on individual survey items were uniformly low. In terms of subgroup differences, their scores were significantly lower than all others on the following items: “[being] tired of my previous job,” “want[ing] to become a career teacher or teach until retirement,” and “[seeing] teaching as a steppingstone to educational administration.” The only item that ECGs collectively cited as more important than others was, “I thought that being a Teaching Fellow would look good on my resumé.” However, even on this item, their mean score was closest to a “2”—“of little importance.” Given the literature (e.g., Maier, 2012), one might have expected the ECGs to have emphasized resume building more than they did. The ECGs’ apparent lack of clarity about their reasons for becoming teachers might have been a function of their comparative youth and limited career experience (Table 1).

BLIs seemed to be much clearer about why they entered teaching. They were particularly motivated by altruism, scoring significantly higher on the altruism subscale than all others (p < .01); 0.71 SDs higher than ECGs, 0.91 SDs higher than White-Asian Insiders, and 0.89 SDs higher than Nonelite Outsiders. Under the altruism subscale, BLIs indicated that the following reasons were very or extremely important: “wanting to work with economically disadvantaged students,” “wanting to work with students of color,” and “wanting to give back to my community.” The BLIs’ reasons for entry also differed from those of the other NYCTF teachers in other regards. Consistent with the literature on Black and Latino/a teachers (e.g., Carver-Thomas & Darling-Hammond, 2017), they voiced a particular concern with the financial considerations of becoming a teacher, scoring significantly higher on the job benefits subscale than all others. And, under the alternative certification subscale, they cited the importance of “subsidized master’s degree” at a significantly higher rate than others. BLIs also were significantly more likely to have become teachers because they saw “teaching as a steppingstone to educational administration,” suggesting that public education promised to be a site for their upward career mobility.

White-Asian Insiders’ motivations for entry also were unique, scoring significantly higher on the meaningful job subscale (0.33 SDs, p < .05) and significantly lower on the altruism subscale (−0.20 SDs, p < .05). On the altruism subscale, White-Asian Insiders cited “want[ing] to work with students of color” as less important than all other subgroups—a clear difference between BLIs and White-Asian Insiders. A post hoc analysis showed that White Insiders and Asian Insiders alike cited this item at similarly low rates. White-Asian Insiders also reported “want[ing] to become a career teacher or teach until retirement” at a significantly higher rate. A post hoc analysis indicated that this latter result was due, in large part, to the high proportion of career changers in their ranks.

Early-Career Retention

Collectively, as shown in Table 3, the NYCTF mathematics teachers exhibited considerable early-career turnover. A sizable proportion (15.1%) left their first school at some point during their first school year—at least 6 weeks before it had ended. Of those that did, about four in five left the district rather than move to a second district school. About 3 in 10 (30.9%) of the teachers left their first school prior to completing the 2-year commitment to the district, and only a minority moved to a second district school. Table 3 also shows that the subgroups exhibited markedly different early-career turnover rates with ECGs exhibiting significantly higher rates than all others. For their part, both Black-Latino/a and White-Asian Insiders exhibited comparatively low rates of first- and second-year turnover. To be clear, these early-career turnover results do not control for school context and teacher background.

OPEN IN VIEWER

Table 3

Descriptive Rates of First- and Second-Year Turnover by Teacher Subgroups

Teacher characteristics	Total	Elite college graduate	Black-Latino/a insider	White-Asian insider	Nonelite outsider
N	615	205	83	54	273
<1 year
Exited first school	15.1%	19.0%*	8.4%	7.4%	15.8%
-Exited district	12.0%	17.6%*	7.2%	5.6%	10.6%
-Stayed in district	3.1%	1.4%*	1.2%	1.8%	5.2%*
<2 years
Exited first school	30.9%	37.1%*	26.5%	22.2%	29.3%
-Exited district	19.3%	27.8%*	12.0%	9.3%*	17.2%
-Stayed in district	11.6%	9.3%	14.5%	12.9%	12.1%

Note. A chi-square test of independence was run for each retention outcome. As all four outcomes were significant, we used two-sided t tests to compare the subgroup proportion to that of three other subgroups combined.

*

Significant at the p < .05 level based on the post hoc two-sample t tests of proportions.

Table 4 provides models of the NYCTF mathematics teachers’ retention at 5 years with controls for school context and teacher background included. These models estimate the odds ratios for teachers remaining based on each variable. Beginning with the first row, the models estimate that, in comparison with Nonelite Outsiders (the comparison group), ECGs had lower odds of remaining in—and thus higher odds of leaving—their first schools, the district, and the profession within their first 5 years. Specifically, the first model estimates that the odds of ECGs remaining in their first schools for 5 years were about 60.7% the odds of Nonelite Outsiders remaining (p < .05). The second model estimates that the odds of ECGs remaining in the district at 5 years were 52.6% of the odds of Nonelite Outsiders remaining (p < .01). The third model estimates that, 5 years after entry, the odds of ECGs working in any K-12 setting as a teacher or administrator were 50.8% of the odds of Nonelite Outsiders doing the same (p < .01).

OPEN IN VIEWER

Table 4

Logit Models of Retention in First School, District, and K–12 Education at 5 Years

Teacher characteristics	First school retention(odds ratios)	District retention(odds ratios)	K–12 retention(odds ratios)
Elite college graduate (nonelite outsider)	0.630*	0.528**	0.487**
Black-Latino/a insider (nonelite outsider)	1.851*	1.977*	1.728
White-Asian insider (nonelite outsider)	1.262	1.629	0.887
Entry age, mean centered	0.994	0.997	1.025
Male (female)	1.337	1.035	0.657***
STEM degree (other degree)	1.053	0.809	1.014
School leadership	1.013	1.013	0.898
Teacher collegiality	1.100***	1.100	1.239
School safety	1.041	1.042	1.027
Student attendance (%)	1.044*	1.008	0.995
Subsidized lunch (%)	1.006	0.996	0.998
Black student (%)	0.986*	0.998	0.991
Latino/a student (%)	0.982**	0.994	0.990
Constant	1.356	1.356	1.878**
N	598	598	384

Note. Logistic regression models estimated with robust standard errors adjusting for the clustering of teachers at the school level. Reference group for categorical variables in parentheses. STEM = science, technology, engineering, and mathematics.

†

p < .10. *p < .05. **p < .01. ***p < .001.

In contrast, BLIs had higher estimated odds of remaining for 5 years in their first schools, the district, and the profession than Nonelite Outsiders. Specifically, the first model estimated that the odds of BLIs remaining in their first schools were 85.1% higher than those of Nonelite Outsiders (p < .05). The second model estimated that the odds of BLIs remaining in the district were 97.7% higher than those of Nonelite Outsiders (p < .05). As Table 4 models used Nonelite Outsiders as the comparison group, as a post hoc step, we used Stata’s linear combination (i.e., “lincom”) command to test whether the retention outcomes of ECGs, BLIs, and White-Asian Insiders differed significantly from each other under the modeled conditions. We found that the (a) difference between ECGs and BLIs was significant at the .01 level for all three retention outcomes, (b) difference between ECGs and White-Asian Insiders was significant at the .01 level for first school retention and at the .05 level for district retention but not significant for professional retention, and (c) difference between BLIs and White-Asian Insiders was not significant for any outcome.

We conducted two post hoc analyses to check the robustness of the 5-year retention results. First, to determine how our classification of the graduates of a very selective college and a NYC high school as ECGs might have affected the results, we reclassified these teachers as either BLIs or White-Asian Insiders and reran the regression analyses. This classification decision seemed to have little effect on the main results as the magnitude and significance levels of the odds ratios of the post hoc models were very similar to those for Table 4 models. In fact, the post hoc models would have modestly strengthened the results about the ECGs’ comparatively low levels of retention, decreasing the odds ratio estimates for their first school and district retention from 0.630 to 0.580 and 0.528 to 0.453, respectively.

Second, we conducted a separate post hoc analysis that ran the district retention model for the subsample of teachers who left their first school within their first 5 years to estimate the first school leavers’ odds of leaving the district by subgroup. Although not reported in Table 4, the odds ratios in the post hoc model were similar in magnitude and significance to those in the district model for the full mathematics teacher population. The restricted model specifically estimated that, among the first school leavers, the estimated odds of ECGs remaining in the district at 5 years were 0.538 times those of Nonelite Outsiders (p < .05) and the estimated odds of BLIs remaining were 1.756 times those of Nonelite Outsiders (not significant). Consistent with this, BLIs were much more likely than ECGs to move to another district school (p < .05) rather than leave the district.

Career Trajectories

The teaching and post-teaching trajectories of ECGs and BLIs were significantly different. As the descriptive statistics in Table 5 show, 10 years after entry, in comparison with all others, a significantly lower proportion of ECGs (31.0%) remained in the district. In contrast, and in comparison with all others, a significantly higher proportion of BLIs (63.6%) and White-Asian Insiders (62.1%) remained in the district in some role. Among those who stayed in the district, comparatively high proportion of BLIs became teacher leaders or school administrators, whereas a comparatively high proportion of White-Asian Insiders remained in a teaching-only role. These modest differences suggest that the Insiders’ career experiences and decisions may have been racialized to some extent or that BLIs and White-Asian Insiders had different career aspirations.

OPEN IN VIEWER

Table 5

Career Status a Decade After Entry by Teacher Subgroup

Teacher characteristics	Elite college graduate	Black-Latino/a Insider	White-Asian insider	Nonelite outsider
N	126	55	37	171
Paid role in district, Fall 2016	31.0%**	63.6%**	62.1%**	42.8%
• Teacher	14.3%**	32.7%	45.9%**	24.0%
• Teacher leader	11.1%	18.2%	10.8%	9.4%
• School administrator	5.6%	12.7%	5.4%	9.4%
In other K–12 setting	18.3%	9.1%	8.1%	15.8%
• Teacher	11.9%	7.3%	8.1%	11.1%
• Teacher leader	4.0%	1.8%	0%	2.9%
• School administrator	2.4%	0%	0%	1.8%
Education sector (non-K–12)	10.3%	7.3%	2.7%	12.9%
Working outside of education	31.7%**	14.5%	21.6%	18.7%
Not in workforce (personal, school)	8.7%	5.5%	5.5%	9.8%

Note. Based on surveyed subsample, N = 389. A chi-square test was used on a table that included observed (nonpercentile) data for the indented rows and bottom three rows combined. This showed that the teachers’ distributions to different occupational sectors were nonrandom, χ² (24, N = 389) = 37.39, p = .040.

**

Significant at the p < .01 level based on two-sided t tests adjusted using a Bonferonni correction (with an individual subgroup compared with all others for each row).

Many of the NYCTF mathematics teachers who left the district, but particularly the community outsiders, migrated to a public school or other K–12 organization (e.g., private school, international school). By the end of the decade, 18.3% of ECGs (who were mostly community outsiders) and 15.8% of Nonelite Outsiders were working as a K–12 teacher or school administrator outside of NYC public schools whereas this held for only 9.1% of BLIs and 8.1% of White-Asian Insiders. On the 2016 survey, some clarified that they left the district to be closer to their childhood homes or extended families.

A decade after entry, higher proportions of ECGs and Nonelite Outsiders were working in the non-K–12 education sector than BLIs and White-Asian Insiders, although this difference was not statistically significant. Although a few were in leadership positions, ECGs in the non-K–12 education sector generally were working in mid-level positions in educational nonprofit, charter, and philanthropic organizations 10 years after entering NYCTF. Smaller proportions were working as tenure-track professors, full-time instructors, and part-time adjuncts in institutions of higher education. In contrast, very few Insiders of any race had found employment in nonprofit, philanthropic, and other (non-K–12) education sector organizations, except for one Black female who had launched her own education technology company. A few Insiders were working as tutors and adjunct instructors in postsecondary institutions. Among those in the non-K–12 education sector, the Insiders generally seemed to be in low-status positions than the ECGs.

ECGs also were significantly more likely than the other subgroup members to be working outside of education a decade after entering NYCTF. Of those, approximately 20% were in finance (e.g., wealth management consultant, financial research analyst), approximately 15% to 20% were in engineering or architecture, 10% to 15% were in information technology or software development, and 5% to 10% were in research as data analysts or archivists. Smaller proportions were in the health sector, as nurses and doctors, or in the legal sector, as lawyers or legal clerks. In general, and consistent with the “return on investment” of elite college degrees, ECGs who had left teaching were working in higher status occupations and earning higher incomes. In contrast, smaller proportions of BLIs and White-Asian Insiders were working outside of education. Many of those were had found jobs in engineering, information technology, and other technical fields. A few also were in the health sector as nurses or nurses’ aides. In contrast to ECGs, none of the Insiders were in finance a decade after teaching. In general, irrespective of race, Insiders working outside of education were earning salaries that were comparable with those of 10-year teachers. The Insiders either had different career opportunities or had different career aspirations than ECGs.

Reasons for Leaving or Staying

Although they left at different rates, members of the different teacher subgroups reported similar reasons for leaving their last NYC public school and hence the district. (For 52% of the district leavers, the last school was their first district school, for 35% it was their second district school, and for 13% it was their third, fourth, or fifth district school.) The teachers only differed statistically by subgroup on two of 29 items on the reasons-for-leaving inventory. In general, the district leavers reported that they left their last NYC public school due to a mix of organizational issues including dissatisfaction with the principal (32.7%), other school administrators (26.9%), staff dynamics (22.2%), student discipline (32.7%), and school safety (11.4%). About 20% of district leavers cited a “disenchantment with teaching mathematics” as a “very important” or “extremely important” reason for leaving. Some reported leaving their last school due to frustrations with accountability policies, namely, how “student assessments and/or school accountability measures impacted my teaching” (17.5%) and “the influence that standardized tests had on the curriculum” (16.8%). A small proportion of leavers reported leaving due to dissatisfaction with job benefits and their teacher salary (9.8%).

Although generally citing similar reasons for leaving, there were two clear differences by subgroup. First, a significantly higher proportion of ECGs and Nonelite Outsiders cited dissatisfaction, “with the large number of students I taught” as an “extremely important” or “very important” reason for leaving; specifically, 11.1% of ECGs and 9.0% of Nonelite Outsiders did so, whereas 2.9% of White-Asian Insiders and 2.1% of BLIs did so. Second, a significantly higher proportion of ECGs and Nonelite Outsiders also cited dissatisfaction, “with the influence standardized tests had on the curriculum” as an “extremely important” or “very important” reason for leaving; specifically, 17.6% of ECGs and 19.5% of Nonelite Outsiders did so, whereas only 8.6% of White-Asian Insiders and 8.3% of BLIs did so.

Finally, the 2016 survey prompted those teachers who had remained in their first NYC public school for close to a decade to list their school’s organizational attributes that explained their retention. Members of all four subgroups named “administrator support” and “teacher collaboration” as their top two reasons for staying. At the same time, a higher proportion of BLIs cited “teacher collaboration” and “teacher collegiality,” whereas a higher proportion of ECGs cited “teacher talent” and “student diversity” as explaining their longevity in their first NYC public school.

Limitations

There were several study limitations. First, the survey data relied on teachers’ retrospective reporting about their career decisions. Teachers’ memories about past events might change, and in addition, some might provide socially desirable responses. Second, NYCTF teachers who left the district during the first year were undersampled on the 2016 survey. Thus, selection bias may have influenced the descriptive results. However, the retention results (Tables 3 and 4) were based on the full NYCTF mathematics teacher population and thus not subject to selection bias. Third, due in part to space limitations, the current study did not examine influence of training and induction on the career trajectories of the teachers although we examine this elsewhere (Brantlinger et al., 2022). Fourth, being specific to NYCTF mathematics teachers, the results may not generalize to teachers from other initial certification programs. However, the results about the early-career retention of ECGs are consistent with extant research on teachers from selective alternative route programs like TFA (e.g., Donaldson & Johnson, 2010) and the results about BLIs are consistent with results—and assumptions about superior rates of retention—in the literature on community-based teachers (e.g., Gist et al., 2019). Fifth, we ran multiple tests of comparisons which meant that Type 1 error was an issue. Thus, some significant differences reported in the descriptive statistics (e.g., Table 2) may have been identified as statistically significant when they should not have been.

Sixth, being quantitative, the study did not produce a robust critical counternarrative that a qualitative CRT study might provide, nor did it give qualitative depth to participants’ voices, nor did it capture the institutional racism that many teachers of color face. That said, the results about ECGs run counter to dominant narratives about how academically elite, and predominantly White, teachers help to address staffing issues and improve the organizational functioning of neighborhood schools that serve low-income, minoritized students (see, for example, Higgins et al., 2011). Further, consistent with QuantCrit, we understand that the subgroup typology used in this study glosses the considerable within-group heterogeneity of those subgroups and, more broadly, that the ethnoracial identity categorizes we used—like those used in many quantitative studies of teacher retention—are dramatic oversimplifications of people’s actual ethnoracial identities. However, given the teacher recruitment presented in the Introduction, this typology is of practical import. Our assumption is that, for the most part, the teachers included in the BLI subgroup would be the most likely to be committed to remaining as teachers in schools that serve predominantly low-income Black and/or Latino/a student populations. This assumption is rooted in prior project research on NYCTF mathematics teachers (e.g., Cooley et al., 2021) and the literature on community-based teachers (e.g., Gist et al., 2019). A related assumption, drawn from Boyd et al.’s (2005) “draw of home” study, is that community outsiders on average seem to be less interested in remaining in NYC public schools for the long term than community insiders.

Implications for Policy and Practice

The costs of recruiting, developing, and replacing alternative route teachers are considerable. In the 2000s, NYC public schools spent approximately $50,000 on the recruitment, training, and induction of individual mathematics Teaching Fellows (Brantlinger, 2020). The search for replacements when NYCTF teachers exited district schools was at least $20,000. Thus, even modest improvements to teacher retention would have saved the district and the public a considerable amount of money, some of which could have been reinvested to improve teaching and learning in the district’s neighborhood schools. As important, teachers who stay can benefit the organizational functioning of low-income schools and the experiences and learning outcomes of students in them (e.g., Sorensen & Ladd, 2020).

A clear implication of this study is that NYC public schools and other diverse urban school districts would be wise to invest in teachers who are rooted in the community or school district in which they teach, that is, to invest in community teachers like BLIs and White-Asian Insiders and to divest in community outsiders like most ECGs. After controlling for first school contexts, the estimated odds of BLIs remaining in the district are at least double those of ECGs remaining in their first schools, the district, and the profession at 5 years (Table 4). An investment in BLIs also seems warranted as financial considerations appeared to weigh heavily on BLIs (Table 2) which is consistent with the literature on the financial barriers that many teachers of color face in becoming teachers (e.g., Carver-Thomas, 2018).

Some argue that ECGs’ high levels of turnover are acceptable under the assumption that they are more effective at raising student achievement than the mathematics and other teachers they work alongside in low-income neighborhood urban schools (e.g., Higgins et al., 2011; Lovison, 2022). But the evidence in support of this is mixed. Taken as a whole, this literature suggests that ECGs who enter through selective programs are not more effective than their colleagues, including many community insiders of color (Boyd et al., 2006; Brantlinger & Griffin, 2019). High levels of early-career turnover of ECGs in the kinds of low-income neighborhood schools many teach in likely have adverse effects on student learning beyond those attributable to individual teachers (Sorensen & Ladd, 2020). Even if those ECGs who stay in teaching were to offset the negative effects on student growth of those ECGs who leave (see Lovison, 2022), there are other outcomes to consider.

As noted, supporters of selective alternative route programs argue that, at least in the short run, ECGs will help to stabilize the teaching staffs in “high-needs” neighborhood urban schools, in particular, by replacing uncertified and substitute teachers with the “best and the brightest” college graduates (see Brantlinger, 2020). However, the fact that, in this study, more than a third (37.1%) of ECGs left their first school before fulfilling their 2-year commitment to the district (Table 3) suggests that they either maintain or exacerbate teacher staffing problems in these schools, forcing administrators to spend valuable time and resources searching for replacements.

Three factors seem to contribute to the ECGs’ high rates of early-career attrition. First, the majority entered teaching in their early 20s, had little prior work experience, and expressed uncertainty about their reasons for becoming teachers (Table 2). Second, as pertinent scholarship suggests (Maier, 2012), this study shows that ECGs have high-paying and high-status career alternatives, which likely makes exiting teaching easy and attractive when they find teaching is more difficult than or is not what they expected. As pass-throughs to future career opportunities, ECGs can capitalize on a short stint in urban school teaching in making their next career moves. Third, recruited nationally from the very selective colleges, most ECGs appeared to have little in common with the students they taught in NYC public schools. Their lack of prior social ties to students and others in the local community likely made their exits from NYC public schools less painful or costly socially than they otherwise might have been.

Proponents of selective alternative route programs (e.g., Higgins et al., 2011; TFA, 2018) argue that ECGs should not be judged solely on their high rates of turnover, positing that their post-teaching trajectories also be considered. They specifically claim that, after teaching, many ECGs move into the broader education and public sectors where they devote themselves to working on behalf of students who attend low-income public schools. There may be some truth to this. A decade after entering NYCTF, some 10% to 15% of ECGs were working in educational nonprofits and philanthropic organizations. However, once in these positions, it is unclear the extent to which their efforts benefit Black and Brown students like those they taught. However, their overrepresentation in non–public school leadership organizations demonstrates that ECGs benefit professionally from their urban teaching experience in a way that other NYCTF mathematics teachers do not. We view this as particularly problematic as educational nonprofits—often launched and run by elite “community outsiders”—hold considerable power in certain educational ecosystems inclusive of urban districts like NYC public schools (Au & Ferrare, 2015; see also Brantlinger, 2020). This is an issue that future research should investigate more deeply as it is an argument that sustains district and philanthropic investment in selective alternative route programs and their elite teacher recruitment efforts (Brantlinger, 2020).

Implications for Research

The study adds to a small number of quantitative studies that draw on CRT and fall under the umbrella of QuantCrit (Frank et al., 2021; Khalil & Brown, 2020). As with that research, this study specifically attempts to move quantitative research on teacher recruitment and retention beyond social- and racial-neutrality. CRT (e.g., Ray, 2019) informs our understanding of NYCTF as a racialized teacher recruitment organization that enables ECGs to capitalize on their racial and social class privilege to monopolize teaching opportunities in low-income, high-minority schools, despite their high rates of turnover. Wielding racial and social class power, ECGs retain the right to “use” these schools as stepping stones to advance their careers and, in some cases, to advance the prospects of the educational nonprofit and other organizations (e.g., NYCTF) that they have launched after finishing teaching (Au & Ferrare, 2015; Brantlinger, 2020). As this study shows, CRT and QuantCrit frameworks provide relevant lenses for other researchers to question policies that are identified or analyzed as if they are race-neutral. Replicating the study design using data from other teacher recruitment and certification programs could also provide valuable insights and support large-scale changes to teacher hiring and retention policies.

CRT and the growing scholarship on community teachers also moved us to focus on teachers with ties to the communities in which they teach which, in this study, were predominantly Black, Latino/a, and lower income. As indicated, a recent exhaustive literature review on community teachers and grow-your-own programs indicate that this study of teachers’ retention and career trajectories is unique in its focus on the relationship between teachers’ ties to the communities in which they teach. As noted, the retention gaps between BLIs and ECGs seem to be rooted, at least in part, by the nature of the ties they have to local communities and schools. At the same time, teacher race and ethnicity also are not unimportant, as the BLIs had moderately lower estimated odds of turnover than White-Asian Insiders after controlling for the climate and student demographics of the teachers’ first schools (Table 4).

Our results also contrast with extant research that shows that, nationally in the United States, Black and other minoritized teachers have higher rates of attrition than White teachers (e.g., Carver-Thomas & Darling-Hammond, 2017; Ingersoll et al., 2019). There are several explanations for this including: First, that nationally representative results may fail to adequately account for local school context, whereas our 5-year retention results came from models that included controls for both student demographics and school climate (Table 4). Second, and related, most NYCTF teachers started in low-income schools with high proportions of Black and/or Latino/a students, whereas nationally, Black and Latino/a teachers are two to three times more likely than White teachers to start in such schools (Carver-Thomas & Darling-Hammond, 2017; Ingersoll et al., 2019). Although White teachers may exhibit better rates of retention nationally, that does not mean that they will exhibit better retention than non-White teachers in schools that are predominantly non-White. Third, rather than simply distinguishing between teachers of different races, our teacher subgroup categories accounted for teachers’ ties to local communities (i.e., as graduates of NYC high schools) and the selectivity of their college in addition to race.

Finally, methodologically, this study illustrates to the value in collecting longitudinal retention and survey data on teachers. By tracking teachers’ career trajectories for a decade, this study was able to shed light on how long different types of teachers stay in teaching and the kinds of occupations that they move into if they leave teaching. Existing retention studies tend to be either cross-sectional, studying the year-to-year retention of teachers or short-term-longitudinal in the sense that they track cohorts of teachers through their first or second years. While both study designs provide valuable insights about teacher retention, they leave gaps in our understanding.