Science1 publisher3 min readPublished
Share of US workers saying their job harms their mental health rose to 37% in APA's 2026 survey
The Harris Poll asked 2,009 employed US adults about stress and also scored them on a five-item hope, efficacy, resilience and optimism scale, and the high scorers reported better conditions on every measure.
The Scientist · Science desk
What happened
- Nearly two in five US workers, 37%, told APA's 2026 Work in America survey that their work environment has a negative impact on their mental health, up from 32% a year earlier.
- The share of workers who would describe their workplace as healthy fell to 80% from 85% in 2025, even though workers say they want employers who value their well-being.
- Almost a third of workers, 32%, said they were concerned about losing their job within the next few weeks or months, up from 26% in the 2025 wave.
- The Harris Poll conducted the survey online for APA among 2,009 employed US adults aged 18 and older between May 21 and June 1, 2026.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- constraint Because each worker's psychological capital score and their answers about the workplace came from one sitting, an employer acting on the gaps is acting on a correlation whose direction the survey cannot settle.
- capability A five-item measure is cheap enough to add to an employee survey a company already fields, so an employer can see its own distribution instead of inferring from a national sample.
- decision Anyone writing a budget line for this gets a measure and a set of subgroup gaps from the survey, but no estimate of how many points of psychological capital a given program buys.
- exposure The low-scoring group carries both the highest stated intent to leave and the highest worry about artificial intelligence making their duties obsolete, so the retention problem and the reskilling problem sit in one population.
Hope, efficacy, resilience and optimism are the four resources in the construct, and APA scored them with a validated five-item scale [3]. Five items is short. It fits on a pulse survey a company already runs.
Everything else in the release compares workers who scored high on that scale with workers who scored low, and both the score and the outcomes came out of the same questionnaire [19]. Among high scorers, 95% described their workplace as healthy, against 63% of low scorers [8]. On mental health, 24% of high scorers said their work environment had a negative effect, against 47% of low scorers [6]. The same person supplied both halves of each comparison, so a supportive employer building psychological resources and a generally positive outlook lifting every answer fit these numbers about equally well.
Stress runs the same way. 48% of all workers said they typically feel tense or stressed out during the workday, including 55% of low scorers and 38% of high scorers [4].
The year-to-year movements come from separate waves. At a proportion near 37% on 2,009 respondents, the standard error of a single share is about 1.1 percentage points [16], which puts the five-point move well outside ordinary sampling noise if the sample behaved like a simple random one. Harris ran this online and weighted the data on characteristics including household size and work status, so that textbook figure is an approximation [2].
On retention, 51% of workers with low psychological capital said they intend to look for a new job at a different organization in the next year, against 19% of high scorers [11], a difference of 32 points and about 2.7 times the rate [17]. Both are stated intentions, given in the same sitting as the score.
"Employers who want to succeed in this environment need to look beyond productivity metrics and help cultivate the psychological resources workers need to thrive," said APA CEO Arthur C. Evans Jr. [12]. "Unlike credentials or technical skills, psychological capital often goes unrecognized and unaccounted for by organizational leaders," Evans said [13]. He added that research shows it "can be strengthened through intentional efforts, including steps employers can take" [14]. The work he points to there is not this survey, which measured each worker once.
APA's suggested steps are broadening the definition of performance to consider how workers build hope, efficacy, resilience and optimism, building the psychological capacities to adapt to new technologies alongside the investment in the technologies, and designing workplaces so workers feel valued and supported [15]. On that middle point, 53% of low scorers worried some or all of their duties could be made obsolete by artificial intelligence, against 37% of high scorers [10]. The release does not report how many workers landed in each group, so the survey's overall percentages cannot be split between them [18].
What to watch
- Whether APA's next Work in America wave follows the same workers over time, which would let a psychological capital score be tested against outcomes measured later.
- Publication of the full 2026 report with the size of the high and low groups and the cut points on the five-item scale.
- Any employer study measuring psychological capital before and after a specific change in management practice, which is the evidence Evans's claim about strengthening it would rest on.