A manager’s rating of you may just be a rating of the meeting.
Nearly half of all workplace meetings are two people in a room, one of whom manages the other. Most firms treat that meeting as a courtesy, a check-in, a place to unblock. It is also, quietly, the main instrument by which a manager forms a view of a person, and that view later becomes a rating. So it is worth asking what the instrument measures.
Where this ends up: almost nobody has studied the meeting itself, the numbers that circulate about it come from companies selling software for it, and the two real findings say a manager’s rating tracks the manager’s impression rather than the work. There is also one thing I would refuse to do with the meeting, and it is the most important paragraph on the page.
The field has barely studied its own most common meeting
Observation The paper that first named this as a research gap says so in its own words: one-on-one meetings “have not been studied empirically as a focal topic,” despite making up close to half of all workplace meetings. Since then a handful of small studies exist, mostly from one lab or one country, none independent of each other. That is the entire direct literature on a meeting nearly every working adult sits through weekly.
What fills the gap instead is marketing. A search for statistics about one-on-ones returns figures like a 300 percent engagement lift, attributed to Gallup. None of it traces to a primary study. The sources are blogs run by companies that sell one-on-one software. If a number about this meeting sounds too clean, the company selling the software for it is a good first guess for where it came from.
What the meeting is actually built on
Evidence The closest research is not about one-on-ones directly. It is about the working relationship between a manager and a report, called leader-member exchange in the literature, and two findings from it are worth carrying into any conversation about what a one-on-one can honestly claim to measure.
First, the manager and the report do not agree about their own relationship. Asked separately to describe the same relationship they are both in, agreement between them was only moderate, a correlation of about .37. That is not a finding about anything external. It is two people describing one thing and landing in different places. Whatever a manager writes down after a one-on-one is one of those two descriptions, not the relationship itself.
Second, and sharper. When a manager rates the relationship and also rates the person’s performance, the two ratings correlate at .58. When the relationship rating is compared against performance measured by someone else, independently, the correlation drops to .14. Small sample on that second figure, so hold it loosely on precision, but the direction is not in doubt and the researchers report it themselves. A manager’s sense of how the relationship is going tracks the manager’s own account of the work far more than it tracks the work.
That is the finding stated as a number. The rating measures the rater’s impression, formed largely inside the room where only the rater and the person being rated are present, and it then gets treated as if it measured the person.
The one thing I would refuse to do with this meeting
Hypothesis The honest research response to weak measurement is usually to measure more, and more carefully. Here I think that response is the wrong one, and it is worth saying why in the open.
The only studies that exist on the meeting itself got their data by instrumenting the conversation: timing how much of it the supervisor spent talking, or feeding an AI system the notes from it. Nothing in what I found justifies recording, transcribing or scoring a one-on-one. A one-on-one is often the one place a person will raise a health problem, a difficulty at home, or a concern about the manager themselves. Turning it into a measured surface changes what can be safely said inside it, permanently, whether or not the measurement ever produces anything useful.
So the answer to weak measurement here is not better measurement. It is knowing the limits of what the meeting can honestly tell anyone, and building consequential decisions on something else.
What this changes practically
If you manage people, the specific claim to drop is that your sense of how someone is doing, formed across your one-on-ones, is a reliable account of their work. It is a reliable account of your relationship, which is worth having and is not the same thing. Where a real decision is riding on performance, the research points toward getting a second, independent view of the work itself rather than trusting the meeting to supply one.
If you are the one being managed, the same finding cuts the other way. A single bad stretch of one-on-ones is evidence about the relationship in that stretch, not a verdict on the work, and the .14 figure is the number to remember when it starts to feel like one.
Where this stops. The direct literature on one-on-one meetings is thin, and I have said so rather than smoothing over it. The relationship-versus- performance figures come from research on manager-report relationships in general, not from a study of the one-on-one format specifically, and applying them here is a reasonable inference rather than something the original researchers tested.
What I am not saying. That one-on-ones are pointless. Nothing here argues against having them. The argument is narrower: do not ask this meeting to be the instrument that measures someone’s work, because the evidence says that is not what it is measuring.