Teacher retention and principal tenure

Eleven West Philadelphia K-8 schools and four 6-12 schools, computed from the School District of Philadelphia's published employee extracts.

Every October snapshot, 2015–2025. Source: SDP District Employees, OpenDataPhilly.

The claim being checked

Someone argued that principal stability drives school performance, that it takes about five years for a principal to move the needle, and that Lea's principal — who arrived in 2020 — was “now six years in, which is precisely when that literature says effects should be visible.”

In the payroll extracts, Aaron Gerwer is listed as Lea's principal in the October 2020, 2021, 2022 and 2023 snapshots — 4 school years. He does not appear anywhere in the district's employee extracts from October 2024 onward, including the most recent (April 2026), and does not appear in October 2019. He is listed as principal of Science Leadership Academy in the 2015, 2016 and 2017 extracts.

years Gerwer is listed as Lea’s principal

Oct 2020 to Oct 2023; absent from the district extracts from Oct 2024

+0.004

correlation, principal churn vs teacher retention

across 11 K-8 schools; 0 would be no relationship

schools with one principal for all 11 years

Powel and Greenfield

68%

lowest pooled teacher retention (Comegys)

against 89% at Penn Alexander

Principal turnover and teacher retention

Across the eleven K-8 schools, the correlation between number of principal changes and pooled teacher retention is +0.004. Individual cases: Powel and Greenfield each had one principal across all eleven years and pooled 84.4% and 86.0%. Comegys had one principal for ten years and pooled 67.9%, the lowest of the eleven. Meredith had five principals and pooled 84.9%, against a district K-8 median of 80–85% depending on year. Eleven schools is a small sample.

Principal changes against pooled teacher retention. One dot per K-8 school, 2015–2025. Horizontal is the number of principal changes over the window; vertical is the share of teachers retained year to year, pooled. Correlation across the eleven: r = +0.004.

Teacher retention, school by school and year by year

Teacher retention by school and year. Each cell is the share of that October's classroom teachers still at the same school a year later — darker is higher. The last row is the median across ~148 district K-8 buildings.

school	’15→’16	’16→’17	’17→’18	’18→’19	’19→’20	’20→’21	’21→’22	’22→’23	’23→’24	’24→’25	pooled
Penn Alexander	100	91	92	95	83	83	78	92	87	95	89.2%
Greenfieldno longer accepts out-of-catchment	91	91	94	86	80	77	85	81	85	93	86.0%
Meredith	82	84	81	77	86	83	84	94	91	87	84.9%
Powel	87	94	67	82	94	94	89	74	75	95	84.4%
Nebinger	67	79	84	91	89	83	82	86	78	94	83.1%
Bache-Martin	79	72	66	86	88	86	78	79	90	90	81.9%
Alain Locke	72	56	91	71	92	94	82	76	93	89	81.1%
Marian Andersonformerly Chester A. Arthur	67	95	86	71	91	79	92	75	83	75	81.0%
Lea	67	74	90	88	85	83	89	70	74	74	79.4%
Harringtoncatchment K-8 for 4914 Pentridge St	64	58	61	81	79	76	84	77	83	71	73.1%
Comegys	56	61	67	75	69	76	65	62	68	89	67.9%
district K-8 median	81	81	81	82	85	85	80	82	83	84	—

Principal tenure

Principal of record in each October extract. One segment per principal, sized by tenure. Hover a segment for the name and years.

Harrington

MARY DIGIACOMOJOE DIXONALONZO FULTONMARLITA THORN

Penn Alexander

SHEILA SYDNORMICHAEL FARRELLMEGAN WAPNERLAUREN OVERTON

Lea

JENNIFER DUFFYJENNIFER DUFFY + SHAVON SAVAGESHAVON SAVAGEAARON GERWERJULIA SNYDER

Comegys

LISA WILMERRAUCHAUN DUPREE

Powel

KIMBERLY ELLERBEE

Alain Locke

KATHERINE MCKELLAR CARTERBARETTA MASSEY + KATHERINE MCKELLAR CARTERBARETTA MASSEY

Meredith

CINDY FARLINOLAUREN OVERTONTIFFANY OSEIBRIAN WALLACENINA BREVARD

Nebinger

ANH BROWNNATALIE CATIN-ST LOUISAYANA TOWNSEND

Greenfield

DANIEL LAZAR

Marian Anderson

KIMBERLY NEWMANVICTOR HUNTERMARY LIBBYNICOLE PATTERSON

Bache-Martin

MARK VITVITSKYMATTHEW TOSSMAN

SLA Middle School

not listedTIMOTHY BOYLE

SLA Center City

AARON GERWERCHRISTOPHER LEHMANN

SLA at Beeber

CHRISTOPHER JOHNSONCHRISTOPHER JOHNSON + GABRIEL KURILOFFCHRISTOPHER JOHNSON

Saul HS

TAMERA CONAWAYALEXA DUNNSAMUEL HOWELL

’15’17’19’21’23’25

Lea retention by principal tenure

Yes — but before Gerwer, not under him. Lea climbed from the 60s under Duffy to 73.8% under Shavon Savage (2017–2019), then fell to 79.1% across Gerwer’s four years and 74.4% since. The turnaround is real; it is just attributed to the wrong principal.

Teacher retention within each principal's tenure. Each segment is one principal, sized by how long they held the building and shaded by the retention over their years — darker is higher. A transition belongs to whoever held the starting October, since that is the staff who either stayed or left.

Harrington

DIGIACOMO61%DIXON70%FULTON80%THORN71%

Penn Alexander

FARRELL93%OVERTON87%

Lea

SAVAGE88%GERWER79%SNYDER74%

Comegys

DUPREE70%

Powel

ELLERBEE84%

Alain Locke

MCKELLAR CARTER79%MASSEY86%

Meredith

OVERTON82%OSEI84%WALLACE92%BREVARD87%

Nebinger

BROWN73%CATIN-ST LOUIS86%TOWNSEND85%

Greenfield

LAZAR86%

Marian Anderson

NEWMAN82%LIBBY84%PATTERSON79%

Bache-Martin

VITVITSKY81%TOSSMAN90%

SLA Middle School

BOYLE79%

SLA Center City

GERWER86%LEHMANN86%

SLA at Beeber

JOHNSON86%JOHNSON80%

Saul HS

CONAWAY86%DUNN85%HOWELL81%

The 2020 cohort

Share of the October 2020 teaching cohort still present. Percent of each school's October 2020 classroom teachers appearing at the same school in each later October.

school	2020	2021	2022	2023	2024	2025
Greenfield n=43	100	77	74	67	65	63
Powel n=17	100	94	88	65	59	59
Penn Alexander n=42	100	83	64	62	57	57
Meredith n=30	100	83	73	70	63	57
Alain Locke n=34	100	94	79	62	59	56
Nebinger n=40	100	83	70	57	45	45
Bache-Martin n=50	100	86	66	48	46	42
Lea n=41	100	83	78	56	46	41
Marian Anderson n=24	100	79	71	54	46	33
Harrington n=37	100	76	65	54	46	32
Comegys n=29	100	76	52	38	34	28

Student and contracted teachers

The headline figures count only PFT-represented classroom teachers, excluding student teachers and bought-in contracted positions. Including those lowers every school's figure. Penn Alexander falls 4.5 points, the largest drop of the eleven, and remains highest. Its share of such staff is 15% of the 2025 teaching bench, against 27% at Marian Anderson, 19% at Bache-Martin and 18% at Meredith.

Retention recomputed with student and contracted teachers included. The headline figures count only PFT-represented classroom teachers. The right-hand columns show how much of each school's 2025 teaching bench is student or CONTRACTED TEACHER staff.

school	PFT only	incl. student + contracted	change	alt staff, Oct 2025	share of bench
Penn Alexander	89.2%	84.7%	-4.5	7 of 47	15%
Greenfield	86.0%	83.3%	-2.7	1 of 45	2%
Meredith	84.9%	82.3%	-2.7	7 of 38	18%
Powel	84.4%	81.2%	-3.2	3 of 26	12%
Nebinger	83.1%	77.8%	-5.4	1 of 34	3%
Bache-Martin	81.9%	78.5%	-3.4	14 of 73	19%
Alain Locke	81.1%	79.8%	-1.3	4 of 30	13%
Marian Anderson	81.0%	79.5%	-1.5	10 of 37	27%
Lea	79.4%	77.3%	-2.2	5 of 48	10%
Harrington	73.1%	71.9%	-1.2	4 of 26	15%
Comegys	67.9%	66.2%	-1.7	1 of 22	5%

Where departing teachers go

Mostly, they don’t. Across all these buildings 56.5% of departures are exits from the district entirely rather than moves to another SDP school, so these schools are not trading staff with one another. The district-wide K-8 median also barely moves year to year (80%–85% across the whole window), so a bad year at one school is a fact about that school, not about the district.

Destination of departing teachers. Every departure across 10 transitions, split by where the person appears in the next October extract.

left the district moved to another SDP school stayed in SDP, left teaching

Harrington

4840

52% out

Penn Alexander

2710

64% out

Lea

6019

71% out

Comegys

4838

55% out

Powel

917

31% out

Alain Locke

3625

55% out

Meredith

2914

63% out

Nebinger

25258

43% out

Greenfield

2827

49% out

Marian Anderson

1730

34% out

Bache-Martin

5529

62% out

SLA Middle School

1812

58% out

SLA Center City

2912

71% out

SLA at Beeber

3315

62% out

Saul HS

3415

64% out

6-12 schools

Ranked against schools at their own org level rather than against elementary buildings.

school	level	pooled ’15–’25	pooled ’20–’25
SLA Middle School SLA-MS	middle school	79.5%	80.6%
SLA Center City SLA CC HS	high school	86.1%	86.7%
SLA at Beeber the other SLA campus	high school	83.3%	80.4%
Saul HS agricultural sciences, citywide admission	high school	84.0%	81.5%

Catchment

For 4914 Pentridge St, Philadelphia PA 19143, the assigned K-8 is Harrington, Avery (org 1300), and the assigned high school is Sayre, William (org 1100). Resolved by point-in-polygon against the city's catchment layer rather than by distance; school names are as that layer writes them.

Scope

These figures are counts of who appears in which extract. They describe staffing, not its causes.
Retention differences may reflect intake, building conditions, grade configuration or resources outside the district budget. Nothing here separates those.
High retention is not necessarily good and low retention is not necessarily bad; a school replacing staff shows low retention by that fact alone.

Method

Names are the only join key. The extracts carry no employee ID, so a teacher who changes their surname reads as a departure plus an arrival, and two people sharing a name read as one. This biases every school's number down by a similar amount, so comparisons hold even though levels are approximate.
“Teacher” means PFT-represented classroom staff. Contracted and student teachers are excluded — they turn over by design. Counselors, nurses and therapists are excluded as non-classroom.
A move between schools counts as a departure for the school being left. That is intended for school-level retention.
One snapshot per year means someone who leaves in November and is replaced by the next October is not counted as turnover.
Peer group is every district building at the same org level with at least 15 teachers — about 148 K-8 schools per year.

Built from 11 October extracts of the School District of Philadelphia's employee information dataset, which lists every active employee with name, title and home organization. Catchment from the city's SchoolDist_Catchments layers. Analysis code and the raw archives are in advice/sdp-retention. See also: how these schools compare with Tuscaloosa City · chronic absenteeism.