A policy is not a practice
When a company says it has adopted a hybrid model, that usually describes a policy: a rule, expectation or guideline about where people work. What employees actually do on a given day is a separate matter. Policies can be loosely enforced, unevenly applied, or interpreted differently by different managers. Reports about hybrid working often blur the two, and that blur is where reporting bias enters.
The point is not that organisations are being dishonest. A policy is a reasonable thing to report. The difficulty is that it can be mistaken for a description of behaviour.
Why statements and practice diverge
Several ordinary reasons explain a gap.
Rules are written at one level, applied at another. A company may set a general expectation for attendance, while team leaders decide how strictly to apply it. Individual arrangements may be agreed informally and never appear in the written policy.
Exceptions can be common. Travel, client visits, illness, caring responsibilities, business needs and special projects all change who is where. A policy of a fixed number of office days can coexist with a wide spread of actual patterns.
Policies change. A rule announced with fanfare may be softened or tightened later. Reports based on the announcement can quickly go out of date, while the original headline lingers.
Aspiration and description get mixed. Statements such as “our people value flexibility” or “we are a flexible employer” may express intent or branding rather than describe measured arrangements.
Where the information comes from
To understand how reporting bias arises, it helps to look at the main sources of information about hybrid working, each of which sees only part of the picture.
Employer statements and press releases. These describe official positions. They have a natural interest in presenting the organisation favourably, whether as flexible and modern or as committed to in-person collaboration. They seldom include the exceptions.
Employer surveys. An organisation asked about its working arrangements may answer in terms of policy. Those who respond may be the ones with a clear policy to report, while organisations with no formal approach may not reply. Questionnaires also depend on question wording: “do you allow remote work?” and “how many staff work remotely on a typical day?” are very different questions.
Employee surveys. These capture how people say they work. They are subject to the usual limits of self-reported behaviour: people recall and round, and may describe the pattern they are supposed to follow. Social desirability can play a part, particularly if employees think their answers might reach management or if there is an expectation about attendance.
Attendance and access data. Badge swipes, desk bookings and building entry records can seem more objective, and they often are for what they cover. But they record presence at a door or desk, not the quality of work, and they miss anyone working at another site. Different systems count different things, such as booked versus occupied desks.
Household and labour force surveys. Official surveys of the population ask people where they work. These reach a broader sample than employers do, though they depend on how the question defines home working, for example whether occasional days count.
Commuting and mobility data. Transport and location data show movement patterns, a useful clue, though one that is affected by many other factors. Such evidence connects naturally to commuting patterns, yet it is indirect for questions about workplace rules.
Types of bias to keep in mind
- Selection into reporting. Organisations that publicise their approach may differ from those that do not. Firms proud of a progressive policy are more likely to talk about it, so public examples are not a random sample.
- Survivorship in case studies. Success stories get told, while workplaces that tried a model and quietly abandoned it appear less often.
- Averaging over differences. A single figure for “days in the office” can hide very different patterns by role, seniority, team and sector.
- Timing. A survey taken at one moment captures a rapidly changing situation, and early data may not describe later stabilised arrangements.
- Definition drift. “Hybrid” has no single fixed meaning. It may refer to a few days at home each week, a flexible choice, or a mix across the year.
Claims about outcomes
Reports about hybrid work often go beyond describing arrangements and make claims about productivity, culture or retention. These need extra caution, because those who work hybrid can differ from those who do not in role, seniority and preferences, so any difference could stem from who is able to work flexibly rather than from the arrangement itself. That is a problem of confounding variables, and it affects comparisons that sound simple. The same applies to productivity measures, which are hard to define for many jobs and often rely on manager ratings or self-assessment.
Employees’ feelings about flexibility are also bound up with broader ideas such as work-life balance, which is itself measured in varied ways.
Reading the next report
Useful questions include:
- Is the statement about a policy, intended practice or observed behaviour?
- Who supplied the information: the employer, employees or an automated record?
- How was “hybrid” defined, and was the same definition applied to everyone?
- How were respondents chosen, and who might be missing?
- When was the information gathered, and has the policy changed since?
A sound conclusion is usually modest: policies tell us what organisations intend or announce, records tell us about presence, and surveys tell us what people say. Combining several sources gives a fuller picture than relying on any one, and where a single source is all there is, treating it as a partial view is the safer approach.