Skip to main content
. 2020 Mar 25;15(3):e0230316. doi: 10.1371/journal.pone.0230316

Table 5. Effect of temperature over previous 30 days on the probability of self-reported mental health difficulties: By season.

Winter Spring Summer Fall
(1) (2) (3) (4)
A. By 3-month in a Season
Past 30-day average temperature (°F) 0.0030 0.0093** 0.0129*** -0.0010
(0.0039) (0.0039) (0.0048) (0.0024)
Temperature (°F)
Average (s.d.) 39.2 (15.0) 54.8 (14.0) 74.7 (8.5) 59.9 (13.4)
Minimum -28.3 -13.7 27.5 -8.2
Maximum 80.9 94.8 102.8 99.5
Interview Months Dec, Jan, Feb Mar, Apr, May Jun, Jul, Aug Sep, Oct, Nov
No. of Counties 2,392 2,392 2,396 2,392
N 739,138 779,834 761,498 774,985
B. By 2-month in a Season (5) (6) (7) (8)
Past 30-day average temperature (°F) 0.0003 0.0091* 0.0106** 0.0025
(0.0047) (0.0049) (0.0048) (0.0034)
Temperature (°F)
Average (s.d.) 38.4 (15.4) 59.3 (11.7) 76.2 (7.7) 55.2 (12.5)
Minimum -28.3 6.2 35.6 -8.2
Maximum 80.9 94.8 102.8 88.9
Interview Months Jan, Feb Apr, May Jul, Aug Oct, Nov
No. of Counties 2,386 2,382 2,391 2,387
N 485,205 517,384 508,684 526,762

Regressions follow the baseline regression in Table 2 column (5) but replace temperature-day bins with the average past 30-day temperature. Columns (1)-(4) in panel A include all interview months within the corresponding season while columns (5)-(8) in panel B exclude the first month of that season to ensure the previous 30-day temperature are within the same season as the interview date.

*** p<0.01,

** p<0.05,

* p<0.1.