Commuting stress is not captured by one number. In the United States, the average one-way commute reached 27.2 minutes in 2024, while 9.3% of workers traveled at least 60 minutes each way. Congestion measures add another dimension: in 2024, Los Angeles averaged 7:20 of congested freeway time per day, and its planning time index was 2.35. These figures describe exposure and reliability, not clinical stress diagnoses, but they show why the daily journey can feel demanding.
Contents
- How long commutes relate to stress
- Congestion and reliability in major metros
- April–June 2024 metro snapshots
- Travel modes and remote work
- Stress and safety behavior
- Evidence from a train-commute study
How long commutes relate to stress
The United States Commuting At A Glance: American Community Survey 1-Year Estimates reports a mean one-way commute of 27.2 minutes for U.S. workers in 2024. The comparable 2023 figure was 26.8 minutes. That is a national average, so it does not describe every worker’s route, mode, or daily variability.
Long commutes are a smaller but important part of the distribution. In 2024, 9.3% of U.S. workers had a one-way commute of at least 60 minutes. In 2023, the share was 8.9%. A one-hour threshold is a measure of commute exposure; it should not be read as a clinical definition of commuting stress.
The year-to-year figures provide context for interpreting survey headlines. The average rose from 26.8 to 27.2 minutes, while the share at or above 60 minutes rose from 8.9% to 9.3%. These are reported Census measures for U.S. workers in the stated years, not estimates of how stressed those workers felt.
Congestion and reliability in major metros
The Federal Highway Administration’s 2024 Urban Congestion Trends measures average daily congested freeway hours and two indices. The travel time index compares peak-period travel time with free-flow travel time. The planning time index compares 95th-percentile travel time with free-flow travel time. A higher planning time index indicates a wider reliability buffer in the reported measure; it does not directly measure emotional stress.
The 2024 report covered 52 metropolitan statistical areas. Among the listed metros, Los Angeles had 7:20 of average daily congested freeway hours, a travel time index of 1.35, and a planning time index of 2.35. Portland recorded 6:19 of congested freeway hours, with a travel time index of 1.31 and a planning time index of 2.24.
Washington, DC averaged 6:08 of congested freeway hours. Its travel time index was 1.33 and its planning time index was 2.30. Seattle averaged 6:07 of congested freeway hours, with a travel time index of 1.32 and a planning time index of 2.22. New York averaged 6:06 of congested freeway hours; its travel time index was 1.30 and its planning time index was 2.16.
| Metropolitan area | Average daily congested freeway hours | Travel time index | Planning time index |
|---|---|---|---|
| Los Angeles, CA | 7:20 | 1.35 | 2.35 |
| Portland, OR | 6:19 | 1.31 | 2.24 |
| Washington, DC | 6:08 | 1.33 | 2.30 |
| Seattle, WA | 6:07 | 1.32 | 2.22 |
| New York, NY | 6:06 | 1.30 | 2.16 |
These metro statistics describe freeway congestion and travel-time reliability during 2024. They are useful indicators of conditions that can make a trip unpredictable, but they do not establish that congestion caused stress for any particular commuter.
April–June 2024 metro snapshots
The Urban Congestion Report: Year-to-Year Trends in the U.S. for April through June 2024 provides a separate seasonal snapshot. During April–June 2024, Salt Lake City averaged 1:52 of daily congested freeway hours, with a travel time index of 1.12 and a planning time index of 1.64. San Antonio averaged 3:28, with indices of 1.22 and 1.99.
San Diego averaged 2:58 of daily congested freeway hours during the same period. Its travel time index was 1.24 and its planning time index was 1.81. San Francisco averaged 5:14, with a travel time index of 1.34 and a planning time index of 2.06. San Jose averaged 4:36, with indices of 1.32 and 1.97.
San Juan averaged 3:31 of daily congested freeway hours. Its travel time index was 1.38 and its planning time index was 2.37. Seattle averaged 6:06 during April–June 2024, with a travel time index of 1.38 and a planning time index of 2.17.
The seasonal figures should not be substituted for the full-year metro figures above. They cover April through June 2024 and come from the report’s U.S. urban congestion comparison. The different periods and report contexts matter when comparing values.
Travel modes and remote work
The Census American Community Survey commuting estimates also describe how workers traveled. In 2024, 69.2% of U.S. workers drove alone to work, the same percentage reported for 2023. Public transportation accounted for 3.7% in 2024, compared with 3.5% in 2023.
Working from home was reported for 13.3% of U.S. workers in 2024, compared with 13.8% in 2023. These figures place commuting stress in a broader travel context: most workers in the measure drove alone, a smaller share used public transportation, and a separate share worked from home. The figures do not identify which mode was most stressful for an individual or explain why a worker chose it.
| Census measure | 2023 | 2024 |
|---|---|---|
| Drove alone to work | 69.2% | 69.2% |
| Used public transportation | 3.5% | 3.7% |
| Worked from home | 13.8% | 13.3% |
| Mean one-way commute | 26.8 minutes | 27.2 minutes |
Mode shares and commute averages are population-level measures. They can help explain the scale of daily travel exposure, but they do not measure perceived stress, physiological stress, or a diagnosis.
Stress and safety behavior
A 2020 daily-diary study summarized in WORK AND COMMUTING SAFETY 267 examined 106 full-time private-vehicle commuters over 10 working days. Its coefficients are standardized regression estimates describing associations, not causal effects.
In the study, commuting stress had a standardized coefficient of β=.11 for risky commuting safety behaviors, with p<.01. Work-related affective rumination had a larger coefficient, β=.34, for risky commuting safety behaviors, also with p<.01. Job strain had a coefficient of β=.42 for work-related affective rumination, with p<.001. Together, these reported associations connect work strain, lingering work-related thought, and safety behavior within the study’s model; they do not prove that one factor caused another.
Several other reported coefficients were small or lacked a reported significance marker. Job strain had β=.06 for risky commuting safety behaviors. Average commute-home time had β=.02, and commute distance had β=.01, for that outcome. Problem-solving pondering had β=.01 with p=.70. The commuting-stress-by-affective-rumination interaction had β=.01 without a reported significance marker.
In the level-2 model of risky commuting safety behaviors, schedule control had β=.08 without a reported significance marker, while job control had β=.01 without a reported significance marker. These values belong to the specific 2020 study design and sample, so they should not be generalized to all commuters or interpreted as population percentages.
Evidence from a train-commute study
An older, specific study gives a different kind of evidence. The American Psychological Association report “Longer train commutes are more stressful, study finds” described a New Jersey–Manhattan train-commute study that included 208 commuters aged 25 to 60. The report is from 2006, so its period should remain explicit when it is compared with 2023 and 2024 population statistics.
Participants’ one-way train trips ranged from 45 to 180 minutes. During their usual morning train ride, they rated perceived commuting stress using 5-point Likert-scale items. This is a direct measure of perceived stress within that sample, unlike the Census commute-duration and mode measures or FHWA congestion indices.
The study’s narrow geography, rail mode, age range, and older measurement period limit what can be inferred about today’s commuters. It remains useful for showing that commute research can measure perceived stress directly, while national and metro transportation datasets mainly measure time, mode, congestion, and reliability.