Employee Skills Assessment: How Much Weight to Give Self-Assessments

An employee skills assessment measures how well people perform the skills required of their roles. Organizations use several methods to do it, including manager ratings, 360 feedback, knowledge tests, work samples and assessment centers. Most often, employees rate their own skill level and usage.

Self-assessments are the least expensive and quickest of these measurement methods. Self-assessments involve people in their own development from the start. This makes self-assessments a fixture in skills gap work, development planning, 360 processes and program evaluation. Yet research suggests that people are not very accurate in their self-assessments, which raises a fair question for talent and L&D teams: why rely on self-assessments, and how much weight should these assessments carry?

This article looks at how accurate self-assessments are, where they add value, when to pair them with other measurement techniques and how to make self-assessments more accurate. Throughout the article, we will draw on research on self-reported data and on our work at PPS International Limited, where we design, administer, report, debrief and coach around assessments. Our employee assessment tools page describes the assessments we offer.

Employee skills assessment methods

There are a variety of methods that can be used in order to assess employee skill levels. Each method has benefits and limits. Here is how the most common methodologies compare to one another:

The same methods can be used in an employee competency assessment, which measures people against the behaviors in the organization's competency model. Most organizations combine two or more methods, and self-assessment is often one of them. Our guide to leadership assessment tools covers how 360s, style profiles and simulations fit different priorities, and our sales assessment center case study shows several methods working together specifically to support business-to-business sales.

How accurate are self-assessments?

On average, self-assessments of job performance agree only modestly with how others rate the same person. A meta-analysis of 115 samples and more than 37,000 people found that self-ratings and supervisor ratings of job performance correlated at .22 [1]. A correlation that low means the two ratings tend to move in the same direction, but loosely, so a high self-rating says relatively little about how a supervisor sees the same work. In Western samples, self-ratings also tended to be higher than supervisors' ratings, and how much higher depended on the context and on the format and content of the rating scale [1].

A classic set of studies by Justin Kruger and David Dunning helps explain why [2]. People who scored in the bottom quarter on tests of grammar, logic and humor estimated their performance at the 62nd percentile when their scores put them at the 12th. The researchers linked this to metacognition: the knowledge needed to perform a skill well is often the same knowledge needed to recognize good performance. When participants were trained in the skill, they became better at recognizing the limits of their own performance [2]. This is why sometimes a pre- and post-learning self-assessment may show scores going down, rather than up. With greater understanding and practice of "what effective looks like," people may realize their skill level isn't as high as originally thought.

Accuracy varies from skill to skill and from person to person. Accuracy tends to be higher for concrete, observable skills and when people have had regular feedback against a clear standard, and lower for skills that are difficult to observe or that people are still learning. The rest of this article looks at best practices in self-assessments.

Why self-assessments still matter

Self-assessments have an important role in most assessment processes, for reasons beyond cost. Here are a few:

• The gap between self-perception and others' experience is valuable information. In a 360, the difference between how people rate themselves and what their direct reports, peers and managers experience often points to where development can start. A leader who rates their delegation highly while their team describes being closely supervised on every task has information that neither rating would show on its own.
• Self-assessment starts people self-diagnosing and learning. Rating oneself against a standard prompts reflection on strengths and areas to grow, and people tend to own a development plan more when they need it. Because self-insight tends to grow with skill [2], repeated self-assessments can also help people recognize their own progress.
• A behavioral inventory with examples helps to set the standard. An inventory written in specific behaviors, with examples of effective performance, teaches people the standard at the same time as it measures them against it. The Lockwood Leadership Assessment® 360 we administer, for example, can be tailored from a competency library of more than 2,000 behavioral items used in organizations around the world.
• Only the person using certain skills can report on some skills. How engaged someone felt in a program, how confident they are using a new skill and how much they intend to apply it are internal experiences. Self-reporting is the best way to measure these internal processes, which is why the reaction level of Donald Kirkpatrick's classic evaluation model asks participants directly [3].

As an example within our own team, we pair self-ratings with an observer's ratings for our facilitators. They rate themselves on the observer's form within a day of a session, before seeing the observer's scores. Our article on the facilitator evaluation form describes the process.

When to rely on a self-report and when to add other measures

Whether self-reported data can be trusted depends largely on three things: what you are trying to measure, what you ask and when people answer. Knowing what you are trying to measure determines whether a self-report is the right methodology to use; what to ask and when to have people self-report are covered in "Self-reporting bias: what distorts self-ratings" below.

A few principles from research methods and training evaluation form the basis of the information in this table. To show that a program caused a change in performance, collect several data points before and after it, so that conclusions do not rest on one unusual week or month [4]. When you measure learning, test before and after the program and, where practical, compare the results with a group that did not take part [3]. For behavior on the job, allow time for change to take place and ask the people who observe it [3].

Our article on training effectiveness measurement covers the full evaluation framework, and our article on how to measure leadership development applies it to leadership programs.

Self-reporting bias: what distorts self-ratings

Self-reporting bias, also called self-report bias, is the difference between what people report about themselves and what is true, introduced by the way questions are asked, the setting and the limits of memory and self-knowledge. Even when people try to answer truthfully, their responses blend accurate assessment with biases they cannot separate on their own [5]. Common sources include:

• Social desirability. People tend to answer in ways that make them look capable and likable, especially when they believe their manager or the organization will see the results [5].
• Similarity to the facilitator and the group. Ratings of a program can be colored by how similar participants feel to the facilitator and to the other participants, and by the associations they bring to the subject.
• Answering beyond what people can know. Asked how they will do something differently six months from now, people tend to answer clearly and with confidence even though they cannot know yet, often reporting what comes to mind first [4].
• Wording and question order. Leading wording invites agreement, and earlier questions can shape how people read later ones [4, 5].
• Recency and salience. When a multi-day program has a single evaluation at the end, ratings tend to reflect the most recent sessions [6], and the most emotionally memorable moments can outweigh the most important ones. The middle of a program may be the part people recall least.

Most of these biases operate without people noticing them. The remedies lie in how the assessment is designed, worded and timed, which we cover below.

How to make self-reporting assessments more accurate

Talent and L&D teams often ask how to get more accurate self-assessments. Here are a few practices we recommend.

Anchor the rating scale in behaviors. A scale labeled only from "poor" to "excellent" leaves each person to decide what the levels mean. Describing what someone does at each level, a format known as a behaviorally anchored rating scale, gives everyone the same reference point.

Here is an example one client from a global data processing organization used to self-report on behaviors and skills associated with their leadership competencies:

In another example from a global industrial technology and engineering enterprise, known for its ability to develop people, they also used a three-point scale. Here is an example around communication:

The format of the scale matters too, as the meta-analysis by Heidemeier and Moser found [1]. The LVI®, a 360 we administer and debrief, uses a Too Little / Too Much rating scale so leaders can distinguish their strengths from strengths they overuse.

Other practices that help include:

• Show examples of effective and ineffective behavior. Before people rate themselves, show short written or video examples of the skill done well and done poorly. Examples give people a reference point beyond their own experience, which tends to help most for those still learning the skill [2].
• Make self-assessment routine. When people rate themselves only at annual review time, the rating can feel high-stakes. Short self-assessments at regular development check-ins build the habit and the skill of judging one's own performance.
• Discuss differences in perception as a matter of course. Make the conversation about gaps between self-ratings and others' ratings a standard part of every debrief and development discussion, so people expect differences and explore them with curiosity. Our article on helping employees have a realistic view of their performance covers how managers can handle that conversation.
• Explain who will see the results and how they will be used. People tend to rate themselves more candidly when they know who will have access to the results and that the results serve their development [5]. Participants also tend to trust external assessors and coaches more than internal ones, because an outside party has no history with them and no authority over their career.
• Collect ratings during a program in the form of practice feedback, as well as at 90 or 120 days post-program. Keep in mind that for multi-day programs, a short rating after each day or module captures the early and middle content that an end-of-program survey tends to underweight [6].
• Write neutral questions and vary their order. "How much did this program improve your coaching?" assumes that it did, while "How, if at all, has your coaching changed since the program?" leaves room for any answer. Varying the order of items also reduces the influence of earlier questions on later ones [5]. If possible, we seek to have participants share not only that their behavior changed (or not), but for them to attribute results to it. For instance, in a manufacturing organization as part of a front-line leader academy, leaders described which skills they implemented and how it saved them time, improved quality, reduced expenses, improved engagement or other aspects of business results.
• Preview the self-assessment as part of prework, during program launch when discussing learning and business outcomes, and throughout the program. This ensures that individuals are expecting to apply what they learn and encourages them to look for opportunities to use the skills to their advantage.

Using self-reported information in a skills gap and development process

Self-assessments tend to add the most value inside a process that brings in other views and ends in a plan. Here is a common sequence. In practice the steps overlap, and a calibration conversation can send people back to gather another view.

1. Define the skills and the standard. Agree on the skills that matter for role success and describe each one in behaviors, using the organization's competency model where one exists. Tie the use of behaviors to business outcomes and improvements. Share examples of how they connect.
2. Collect the self-assessment. Ask people to rate themselves against the behavioral standard, with examples, and explain how the results will be used. Ask people to share how the use of the skills impacted business outcomes and results.
3. Add other views. Bring in manager ratings, 360 feedback, knowledge tests or work samples, depending on the skill and the desired outcome.
4. Consider holding a calibration conversation. The employee and their manager or coach discuss where the views agree and where they differ, and what might explain the differences.
5. Agree on development priorities. Choose no more than three development options and document them in an individual development plan.
6. Reassess. Repeat the self-assessment and at least one other measure at an agreed interval to see what has changed.

Decide before the assessment who will see individual results and how they will be used, and tell participants. Most organizations separate the assessment and debrief from decisions about pay or promotion, and we recommend sharing results with the people who completed the assessment.

When the goal is to identify needs across a team or organization, individual self-ratings can feed a broader needs analysis. The needs assessment survey we offer asks managers to rate both how useful 29 common workplace success points are and their current perceived skill in each, for their own role or for their direct reports. Our article on training needs assessment covers that process.

Common mistakes with self-assessments

• Using self-ratings for pay or promotion decisions. Self-ratings tend to be lenient [1], and people may rate themselves differently when they know what is at stake. In most cases, self-ratings are not appropriate for employment-related decisions such as pay and promotion. Keep self-ratings for measuring learning outcomes and structuring development conversations, as with 360s, which measure perceptions.
• Skipping the debrief. A self-assessment or 360 report needs a conversation and a plan to lead to change. Schedule the debrief when you schedule the assessment.
• Treating gaps as failures. A gap between self-ratings and others' ratings is information about how behavior is experienced. People who are criticized for their gaps may rate themselves more defensively the next time.
• Asking people to rate what they cannot observe. Questions about how others see them or how they will perform in the future invite confident guesses. Ask about behaviors people can observe in themselves and experiences only they can report.

Frequently asked questions

Is self-report data reliable?

Self-report data is reliable for some purposes more than others. It is the best source for what only the person can know, such as how engaged they felt in a program or how much they intend to apply what they learned. For skills and performance, self-ratings agree only modestly with supervisors' ratings [1], so pair them with other measures and design the questions carefully.

How accurate are employee self-reported assessments?

Employee self-assessments are only modestly accurate on average. Across 115 samples, self-ratings and supervisor ratings of job performance correlated at .22, and self-ratings tended to be higher [1]. Accuracy tends to improve with behavioral anchors, examples of effective and ineffective behavior, regular practice and conversations about differences in perception.

How often should employees complete a self-assessment?

How often employees complete a self-assessment depends on its purpose. For development, a short self-assessment at each development check-in keeps it routine. A fuller assessment such as a 360 is often repeated after 12 to 18 months, which gives behavior time to change. For training programs, collect self-ratings before, during and after the program.

Who should see self-assessment results?

Who sees self-assessment results should be decided before the assessment and explained to participants. If you are using self-assessment data to measure learning outcomes, it may only be necessary to share trends from the cohort or group. If the results are being used for development purposes, we recommend sharing results with the person who completed the assessment and involving their manager or coach in the development conversation.

Should self-assessments count in performance reviews?

Self-assessments can be a useful part of a performance review as a starting point for the conversation, especially when the employee and manager rate the same behaviors and discuss the differences. We recommend keeping self-ratings out of pay and promotion decisions on their own, because they tend to be lenient and people may rate themselves differently when the stakes are high.

Where to start

Self-reported measures are a great entry place for measuring results of learning. They are also excellent for development. If using for development, start with a self-assessment your organization already uses. Review its rating scale, add behavioral anchors and examples for the skills that matter most and plan the conversation about differences in perception before the next cycle begins. Our employee assessment tools page describes the assessments PPS administers, our guide to choosing the right leadership assessments compares 360s and other instruments and our article on measuring training success covers evaluation beyond self-report. If you are using self-reported data to measure learning outcomes, start by adding a post-learning data collection point at 90 days post-program that asks individuals to share a single example of what they've learned and how applying it helps on the job. To learn more about measuring learning outcomes, see our article How to Measure Leadership Development: Metrics, Methods and a Measurement Plan.

References

1. Heidemeier, H. and Moser, K. (2009). Self-other agreement in job performance ratings: A meta-analytic test of a process model. Journal of Applied Psychology, 94(2), 353-370. https://doi.org/10.1037/0021-9010.94.2.353
2. Kruger, J. and Dunning, D. (1999). Unskilled and unaware of it: How difficulties in recognizing one's own incompetence lead to inflated self-assessments. Journal of Personality and Social Psychology, 77(6), 1121-1134. https://doi.org/10.1037/0022-3514.77.6.1121
3. Kirkpatrick, D. L. (1994). Evaluating Training Programs: The Four Levels. Berrett-Koehler.
4. Morling, B. (2025). Research Methods in Psychology: Evaluating a World of Information (5th ed.). W. W. Norton & Company.
5. Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y. and Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879-903. https://doi.org/10.1037/0021-9010.88.5.879
6. Dickey, D. and Pearson, C. (2005). Recency effect in college student course evaluations. Practical Assessment, Research, and Evaluation, 10, Article 6. https://doi.org/10.7275/8fdy-vr38