Self-assessment, manager, or peers — what each source contributes

Individuals may know the reasoning behind their decisions. Their manager may understand the expected scope and observe results over several months. Their peers may describe how they coordinate work when day-to-day dependencies arise. None of these positions guarantees a sound assessment. Each offers a different window.

The useful question is not who occupies the most important position. It is who could observe the manifestation we need to assess and what evidence they can provide.

Choose the source after defining what you need

A skill never appears in full to an observer. What can be observed are behaviors, decisions, outcomes, procedures, explanations, and applications situated in a particular context.

Before selecting assessors, three elements should be defined:

1. The relevant manifestation. Do we need to understand how the person decides, coordinates, executes, explains, or responds under a specific condition?
2. The observation context. Where, and how often, could that capability have appeared?
3. The subsequent decision. Will the result support development, mobility, an assignment of responsibility, or a performance conversation?

These definitions change which source is appropriate. When assessing coordination across functions, a peer involved in those dependencies may contribute more than a distant manager. When examining scope and sustained results, that same peer may see only a fraction of the work.

Research on multisource assessment reaches a compatible conclusion. A study by Batista-Foguet and colleagues proposes treating assessor context as part of what is being measured. In their sample, external sources and self-assessments seemed to apply different criteria or capture different nuances of the same competency. The study also cautions that comparing averages across sources requires establishing that the scale functions equivalently for each. Batista-Foguet et al. (opens in a new tab)

The finding comes from developmental assessments in an academic context. By itself, it does not validate an organizational skills architecture. It does challenge the assumption that all sources produce interchangeable versions of the same data.

What self-assessment can contribute

Self-assessment has access to information that others rarely see in full: the reasoning, alternatives considered, difficulties encountered, and perceived degree of autonomy. It can also help reconstruct less visible episodes and indicate where to look for additional evidence.

That access does not turn intention into performance. Memory can be selective. Self-concept may influence the response. A person may also be unaware of effects that are clear to those who work with them.

Self-assessment therefore becomes more valuable when it provides concrete situations, explainable decisions, and verifiable outcomes. It becomes weaker when a general claim—“I coordinate well” or “I have mastered this”—replaces a reconstruction of the work.

Its most useful function is often to provide context and formulate testable hypotheses. It also allows comparison between the person’s understanding of their own performance and observations from other relationships. Dismissing self-assessment as inherently biased discards information. Treating it as sufficient confirmation eliminates comparison.

What the manager can contribute

A manager may connect performance to role expectations, scope of responsibility, continuity over time, and relevant outcomes. They may also observe change, increasingly complex assignments, and responses to different priorities.

That position adds value only when the manager has sufficient contact with the actual work. A manager may receive results without knowing how they were produced. Their assessment may rely on reputation, highly visible episodes, or comments from third parties.

What peers can contribute

Peers often have access to day-to-day coordination. They may observe how someone shares information, honors commitments, manages dependencies, transfers knowledge, or affects other people’s work. This perspective is particularly valuable for skills expressed across functions or teams.

It is also partial. A frequent collaborator and an occasional one will not have seen the same things. People in the same group may share expectations, preferences, or biases. Proximity does not guarantee clear criteria.

Peer selection should correspond to the manifestation being assessed. Consider the duration and frequency of the interaction, the diversity of situations observed, and the connection between those situations and the intended conclusion. Gathering several nearly identical opinions may improve stability. It does not necessarily broaden the evidence.

A conditional match

The following table does not assign permanent competencies to each source. It organizes questions for determining whether a source’s contribution is relevant.

This match changes with the skill. A manager who works closely with someone may observe collaboration better than some peers do. A peer with subject-matter expertise may provide information about technical quality beyond the manager’s expertise. Function matters more than label.

When to combine sources

Combining perspectives is useful when a skill appears in several relevant contexts and no single source can see them all. It may also show whether a behavior remains consistent across different stakeholders or demands.

Multisource assessment does not improve simply because the number of assessors increases. A review of its use finds that its value depends on the purpose, the instrument, how respondents are selected and prepared, and the quality of the feedback process. Lockyer et al. (opens in a new tab)

A defensible combination requires at least three conditions:

• complementary access to relevant manifestations;
• known and sufficiently comparable criteria;
• the ability to preserve differences before consolidating them.

If sources observed different situations, their disagreement may contain information about the context. If they observed a comparable situation but applied different criteria, the divergence may reveal ambiguity in the scale. If one source had almost no access, the problem may be insufficient evidence. These situations should not receive the same mathematical response.

Interpret differences before turning them into an average

When ratings differ, the first step is to reconstruct what supports each one. This means separating the score from the evidence and making five pieces of information visible:

1. which manifestation each source assessed;
2. which situations each source could observe;
3. for how long and at what proximity;
4. which specific evidence supports the judgment;
5. which part is interpretation.

The perspectives can then be classified as comparable, complementary, or insufficient. Only comparable perspectives justify direct consolidation. Complementary perspectives should remain distinct because they describe different parts of the performance. Insufficient perspectives should lower confidence in the conclusion, not automatically become a gap.

This does not prevent a consolidated result when the decision requires one. It prevents consolidation from becoming the starting point.

A short selection sequence

The choice can be made in five moves:

1. Define the manifestation and the assessment’s intended use.
2. Identify who had a genuine opportunity to observe it.
3. Review the quality, breadth, and recency of that access.
4. Select one sufficient source or several complementary sources.
5. Interpret divergences before calculating a shared result.

A single source may be sufficient when it has relevant access and strong evidence for the decision. Multiple sources add value when they broaden the observation. If they merely multiply opinions about the same fragment, they add work without resolving the design’s limitations.

The right source is not defined by hierarchy or proximity. It is defined by the relationship between what we need to conclude and what each person was actually able to observe.