Skip to main content
. 2021 Sep 1;4(9):e2123412. doi: 10.1001/jamanetworkopen.2021.23412

Table 5. Association Between Unionization Status and Program Outcomes, Stratified by Location in the New York–Newark CSAa.

Outcome Programs in New York–Newark CSA Programs outside New York–Newark CSA
Program typeb Linear regression, mean difference or OR (95% CI)c,d Program typeb Linear regression, mean difference or OR (95% CI)c,d
Unionized Nonunionized Unionized Nonunionized
Salary, mean (SD), US$e 63 369 (3059) 65 525 (6437) Mean difference, −2723 (−6145 to 700) 60 164 (5532) 57 045 (3671) Mean difference, 213 (−1895 to 2322)
Vacation length, wk
No. 17 15 12 155
<4 0/17 4/15 (26.7) NE 2/12 (16.7) 114/155 (73.5) OR, 11.05 (1.83 to 66.70)
4 17/17 (100) 11/15 (73/3) 10/12 (83.3) 41/155 (26.5)
Subsidized childcare 0/17 1/15 (6.7) NE 0/12 15/155 (9.7) NE
Housing stipend 4/17 (23.5) 8/15 (53.3) OR, 0.24 (0.05 to 1.20) 6/12 (50.0) 2/155 (1.3) OR, 1.71 (0.11 to 25.98)
Relocation stipend 0/17 0/15 NE 2/12 (16.7) 20/155 (12.9) OR, 2.75 (0.27 to 28.33)
Technology stipend 9/17 (52.9) 10/15 (66.7) OR, 0.57 (0.13 to 2.45) 4/12 (33.3) 66/155 (42.6) OR, 0.54 (0.12 to 2.46)

Abbreviations: CSA, combined statistical area; NE, not estimable; OR, odds ratio.

a

Because 56.7% of unionized programs are within the New York–Newark CSA, this sensitivity analysis assesses unionization stratified by program location within vs outside that area. Separate models were used to estimate each outcome among programs inside and outside the New York–Newark CSA. Nine programs without available salary data were excluded, resulting in 277 programs available for analysis. For benefits, only programs for which data were available through the American Medical Association FREIDA database or a publicly available union contract were included, resulting in 199 programs available for analysis.

b

Unless otherwise indicated, data are expressed as number/total number (percentage) of programs.

c

Linear regression model of salary for unionized vs nonunionized programs including program size, program type, and county median household income as covariates. For programs outside the New York–Newark CSA, geographic region and urban-rural classification were also included as covariates.

d

Logistic regression models of unionized vs nonunionized programs including program size and program type as covariates. For programs outside the New York–Newark CSA, geographic region and urban-rural classification were also included as covariates. For the housing stipend model, county-level median rent was also included.

e

Includes 16 unionized and 22 nonunionized programs within the New York–Newark CSA and 13 unionized and 226 nonunionized programs outside the New York–Newark CSA.