Bosses see less than they think: a gap in OHS and collaboration

In industry, 70 percent of employees saw a near-miss situation in the past year, while among management only 41 percent did. The authors of a study on leaders’ collaboration styles describe a similar gap. We looked into where it comes from and how to reduce it.

The number of people injured in workplace accidents in Poland is rising. According to preliminary data from Statistics Poland (GUS), in the first half of 2026, 33,081 people were injured, 7.9 percent more than a year earlier. In serious accidents, 214 people were harmed, and 71 were killed. The accident rate, meaning the number of injured people per 1,000 employed persons, rose from 2.23 to 2.41. This is not a one-off spike: already in the first quarter, the number of injured people was 19.3 percent higher than a year earlier.

Accident statistics have one drawback, however: they describe what has already happened. A new report by the Coalition Safe at Work Association shows that warning signals do exist in companies. The problem is that they often do not reach the people who make decisions.

Employees see much more than their supervisors

The study asked managers and employees about potentially accident-related events, known in English as near-misses. These are situations in which no one was injured, but an accident was very likely. Over the past 12 months, 70 percent of employees noticed such events in their company, compared with only 41 percent of managers.

It is similar with accidents that ended in injury: 62 percent of employees and 43 percent of managers confirm them. The biggest difference concerns serious accidents. They are indicated by 60 percent of employees, while among management the figure is 29 percent, or roughly half as much.

These figures need to be treated with two reservations. The study was conducted by the SW Research agency on June 18-26, 2026, among 300 managers (computer-assisted telephone interviews, CATI) and 300 employees (surveys). It covered industry only: food and beverage production, chemicals and pharmaceuticals, heavy industry, mining, and energy. So it does not describe the entire economy. In addition, both groups assessed their own companies, and the study description does not indicate that these were the same plants, so the result shows a difference in perspectives rather than a measurement of specific accidents.

Why information from the shop floor does not reach the office

The report shows that industrial companies analyze individual safety areas most often once or twice a year. 15 percent of organizations deal with accidents and near-miss situations only after something happens. Only 45 percent of managers and 46 percent of employees believe that there were fewer safety-related incidents in the past two to three years.

Both groups agree on the causes. Most often, they point to downplaying risk, the belief that “nothing will happen,” and being used to working “their own way.” One in four respondents mentions time pressure, productivity pressure, or pressure to meet targets. Statistics Poland data fits this picture: “incorrect employee behavior” accounted for 39.8 percent of accidents. If behavior is the cause, the first warning signals are seen by those working alongside it, not by those reading quarterly reports.

A partial explanation for the difference in assessments is provided by the previous report from the same coalition, from 2025. Managers then also rated preventive measures more positively than employees. Jacek Dolibóg, president of the association, explained it this way: “Managers participate more often in planning and implementing solutions - therefore they have greater awareness of the initiatives being undertaken.” Those who plan procedures easily assume that they work.

The same mechanism at the conference table

A similar gap was described by Ivan Matviak and Heidi K. Gardner in an article published on September 15 in “Harvard Business Review.” This time, it is not about safety but about collaboration. The authors studied more than 10,000 professionals and 400 senior leaders. Leaders usually see themselves as good collaborators who listen and give people autonomy. The people who work with them describe it differently.

The authors’ conclusion: the higher a leader is in the hierarchy, the more often they underestimate two harmful behaviors. The first is dominating conversations, the second is micromanagement, meaning controlling the details of subordinates’ work. Gardner notes that this does not stem from bad intentions. The source lies in the traits thanks to which leaders were promoted: demonstrating expert knowledge and striving for high-quality work.

This is a different study, in a different field, and connecting the two topics is our synthesis. However, the pattern also repeats in independent data. In a 2024 Epicor study conducted in industry among 1,400 managers and 1,200 employees, 52 percent of managers considered their companies modern, while 39 percent of employees agreed. Managers also rated job satisfaction more highly (57 versus 45 percent).

What a manager can do before something happens

The coalition’s board members are managers from companies in the workwear, fabrics, and protective clothing industry, so we treat their recommendations as a voice from the sector. Still, they describe the direction of change well:

Ewa Gawrysiak from the association’s board points out that indicators related to mental well-being and safety culture rank relatively low among monitored areas. In her opinion, this is often where the first signs of future problems can be seen. Elżbieta Rogowska, in turn, reminds us that accidents also mean absences, downtime, repairs, compensation, and legal consequences. Safety therefore has its own line in the budget, not only in the rulebook.

From the comparison of both studies, we draw several practical conclusions for managers, on the shop floor and in the office:

• Measure signals, not just damage: the number of near-miss reports and the response time to them.
• Make reporting easier: a short form or a phone app will work only if submitting a report does not risk causing trouble.
• Show what changed after a report was made, because without a response people quickly stop reporting.
• Check whether your goals do not encourage shortcuts, since one in four respondents points to productivity pressure.
• Ask your team about your work style, preferably anonymously, and compare the answers with your own assessment.

The common denominator is simple. Neither infrequent reviews nor management’s sense of well-being can replace information from the lower levels of the organization. Usually, this information is available, but no one collects it regularly. A company that starts doing so has a chance to notice a problem before it appears in Statistics Poland data or in a conversation about leaving the job.