Love them or hate them, rating scales are everywhere. We rate our food delivery, our eCommerce products and our contractors. This article will examine why they are essential in performance management and how to select one for your organisation.
Rating employee performance
Rating scales are important because organisations need to make data-informed decisions on development plans, increases and promotions. This process also needs transparency and clarity to ensure that performance management is consistent and fair.
Using rating scales in performance management also means we can measure performance at an individual, team and company level. This speeds up the process and allows for great insights into the organisation's trends, weaknesses and opportunities.
Challenges when using ratings
They’ve been getting quite a bad reputation as they are often seen as reducing someone to a number or inaccurate in measuring performance. While these are common and valid complaints, they are primarily attributed to poor data and bad design.
It is very common for rater biases to skew the data. For example, Leniency Bias occurs when managers give all their reports higher ratings, and Centrality Bias occurs when managers rate everyone as satisfactory without recognising poor or excellent performance.
These are both damaging as they lead to the wrong outcomes from the process. If people are performing poorly, they need to be told and assisted in getting back to proper performance. If they are performing well, they need to be rewarded and recognised.
A combination of management training and the correct rating scale can address these challenges.
Types of rating scales
Most rating scales collect data that is ordinal, meaning that they are a list ranked from worst to best. For example, a rating of two is better than a rating of one.
Some less common types of collected data are binary (yes or no) or nominal (select all that apply).
This is an example of a balanced, five-point Likert scale which tests the respondent’s level of agreement with a statement. It can have different options and could be an even number of options by removing the Neutral option. This would be called a “Forced Choice”, as the respondent has to pick a positive or negative option.
Semantic Scales have a negative and positive extreme with a set number of options. Roslin uses these for Engagement Surveys, typically with Poor and Excellent as the two extremes.
Customising your organisation’s rating scale
Likert and Semantic scales are well-validated but don’t typically match what you’re trying to measure. That’s why most HR teams use custom scales tailored to their needs.
Taking the time to customise your choice of rating scale will yield more valuable data and reduce rater biases. Here are four questions you can ask yourself when doing so.
Does this scale differentiate performance?
Simply put, this scale needs to point out the differences in performance levels. For instance, when tackling leniency bias, you could extend your scale to four or five points with two points for excellent performance, where you can assess above-average performance with more granularity.
You typically want one point for low performance, one for on-par performance and then as many points on the scale as needed to capture above-average performance. This is why most rating scales are between three and five points scales.
Does this scale accurately measure performance?
It’s crucial to ensure that the scale you choose measures what the question or statement related to it is intending to measure. If the question is asking a reviewer to rate an employee’s coding skill, then the rating scale must correlate well with performance in that skill.
Is this scale easy to understand?
Managers who don’t truly understand the scale tend to stick to the safe, central or neutral option, which causes centrality bias. Making the scale simple for managers and their reports goes a long way to preventing this.
Is this scale transparent?
Fairness is a huge component of performance management. Designing a good rating scale, explaining it well and using it consistently will improve the perceived fairness of your performance management process.
Does this scale use your performance language?
Every business has its brand, language and objectives. Ensure that you capture that in the wording of your rating scale options.
Rating scale examples
To kick start building your custom rating scales, here are some best-practice examples you can bookmark and return to later.
Rating scale 1 to 3 examples
This is many organisations start. The idea is to measure the three expected outcomes from a performance review.
Some other examples are:
- Below Level | At Level | Above Level
- Poor | Good | Excellent
- Not Often Enough | From Time to Time | Most of the Time
Rating scale 1 to 4 examples
A common challenge when using a three-point scale is that rewarding everyone in the exceeding category is tricky. A significant improvement is to add a fourth option and thus have two options for above-average performance. This makes it much easier to reward and recognise your top performers.
Our recommendation would be the following four-point scale:
This scale starts tackling centrality bias as there is no middle option. Managers are forced to choose a more meaningful option. If managers are too lenient, you still have two options for above-average performance so that your top performers always stand out.
Some other examples are:
- Never | Sometimes | Often | Always
- Low Performer | Developing Performer | Highly Valued Performer | Top Performer
Rating scale 1 to 5 examples
You can add another above-average category if you still struggle to recognise your top performers.
This is still going to tell you who needs help, who is on track and who is performing exceptionally, with three categories for high performers.
Some other examples are:
- Poor | Needs Improvement | Meets Expectations | Exceeds Expectations | Exceptional Work
- Unacceptable | Needs Improvement | Meets Expectations | Exceeds Expectations | Outstanding
A rating scale for soft skills
When measuring performance related to soft skills like interpersonal skills, communication skills or living the company values, you should consider using an observational scale. This simply measures how often the employee demonstrates the desired behaviours.
For this to work, it’s important to phrase the question as a behaviour, not an attribute, skill or intention. For instance, instead of saying, “This person has integrity”, say, “This person acts with integrity.” A commonly used scale is the Behavioural Observation Scale (BOS).
The language is a bit boring, but it’s a great place to start because it can be applied to all your soft skill questions for consistency and simplicity. A four-point alternative would be: