Semi-Quantitative Risk Assessments – Are the Numbers and Matrices Useful?

Risk assessments are essential in workplace health and safety. They allow organisations to identify hazards, evaluate risks, and implement appropriate control measures. 

There are different methods to conduct a risk assessment, ranging from qualitative, semi-quantitative to quantitative risk assessments.

In this blog, we will focus on semi-quantitative risk assessments and the scoring system and risk matrices that they use to determine the risk rating.

While they attribute a clear, easy-to-understand, numerical value to a risk, we need to consider the methodology to arrive at these numbers to judge if they really are as useful as they seem.

Semi-Quantitative Risk Assessment Scoring Matrix

Semi-quantitative risk assessments involve assigning scores to two key factors:

• Likelihood: How likely is it that an incident will occur? This is often scored on a scale from 1 to 5, with 1 being very unlikely and 5 being highly likely.
• Consequence: If the incident does occur, how severe would the consequences be? This too is scored from 1 to 5, where 1 indicates minor harm and 5 indicates severe harm or fatalities.

We can then calculate the Risk as the product of these two scores:

Risk = Likelihood × Consequence

Once the score is calculated, the risk is usually classified into the following three categories:

• Low (1-8)
• Medium (9-15)
• High (16-25)

This system allows risk assessors a reliable framework to fall upon a number that will help them determine the next steps to be taken regarding the risk. However, by attributing a score and a number to a risk, you can open yourself up to not viewing the situation from a completely objective view.

The Problem of Subjectivity in Semi-Quantitative Risk Assessments

While the attributing of scores and numerical approach may appear to add a layer of objectivity to semi-quantitative risk assessments, they are, paradoxically, highly subjective. 

The scores for both likelihood and consequence are often based on the assessor’s opinion rather than hard data or research. For example, two different assessors may have very different views on how likely an incident is or how severe the consequences might be, leading to inconsistencies in the risk rating.

These subjective scores can vary based on the assessor’s experience, biases, or even external pressures. This subjectivity raises an important question: if the risk rating is based on opinion rather than fact, how reliable are the numbers?

To understand the issue of subjectivity in more depth, let’s take an example of a construction company assessing the risk of workers operating on scaffolding. The site manager, who has seen few incidents in recent months, rates the likelihood of a fall as 2. The consequence of a fall is rated as 5, as a fall from this height could result in death or serious injury.

Therefore, the score attributed to the risk in this instance is 2 x 5 = 10.

However, we can see the site manager may be biased towards a lower likelihood of risk as they haven’t seen an incident like this in months. Another assessor, familiar with a recent incident on a similar site, might rate the risk likelihood as 3, acknowledging different factors such as high winds that may have been a cause of the recent incident.

In this case, the risk score would be higher at 3 x 5 = 15.

As we can see, just increasing the risk likelihood by 1 in this example raises the risk rating by 5 points. This shows how subjective opinions and experiences have the potential to affect the scores by an important margin.

The Potential for Manipulation

Another significant issue with semi-quantitative risk assessments is that they are open to abuse. 

Some organisations implement policies that require work to stop if a risk is rated as “High”. In such cases, there may be pressure on managers and risk assessors to assign lower scores to ensure that the job isn’t halted, even if the risk is objectively high. 

This “fudging of the figures” can undermine the entire purpose of the risk assessment and put workers at unnecessary risk. We’re left to ask: if risk ratings can be manipulated to avoid inconvenient conclusions, how effective are they at improving health and safety?

To see an example of manipulation in action, let’s imagine a company using a semi-quantitative risk assessment to evaluate the risk of forklifts moving heavy pallets in a busy warehouse.

The assessor rates the likelihood of an accident involving a pedestrian and forklift as 4, while the consequence is rated as 4 as serious injury could occur.

In this instance, we have a high risk score of 4 x 4 = 16. Operations may have to stop.

However, during the assessment, the warehouse manager hopes to pressure the assessor to lower the likelihood to 3. They argue their forklift drivers are very experienced and state incidents are extremely rare in their warehouse. Due to the pressure and anecdotal evidence, the assessor agrees to lower the risk rating to 3 x 4 = 12.

Now falling within the medium risk category, the operations can continue, even though there has been no objective change to the risk. Again, we see the volatility of the matrix where a change of 1 point has changed the final score by 4 and the severity of the risk. 

This demonstrates the effectiveness of employing manipulation – a seemingly minor change in one matrix can have a massive impact on the final score and keep operations running.

Overemphasis on Scores Rather Than Hazards and Controls

Due to the semi-quantitative risk assessment’s nature of assigning a number which dictates the severity of a risk, more attention can often be paid to the score itself than to the actual hazards or necessary controls.

Ultimately, risk matrices are not an end in themselves—they are a tool to help prioritise actions. Yet in practice, they sometimes become the focal point of the assessment, reducing the focus on what truly matters.

This fixation on getting the “right number” can overshadow the real goal: assessing the situation thoroughly and implementing practical measures to control the risks. The matrix can become more of a box-ticking exercise to fall on the right number than a meaningful, objective, research-led evaluation of workplace hazards.

What’s the Purpose of Semi-Quantitative Risk Assessments?

Throughout the blog, we have shown the potential risks of using semi-quantitative risk assessments. Why, you might ask, are they used? Should they be used?

The main purpose of using a scoring system like this is to sort priorities for action. By classifying risks as low, medium, or high, organisations can allocate resources more efficiently, focusing attention on higher-risk activities that require more immediate or stringent controls. In this way, semi-quantitative risk assessments can be useful for prioritising actions.

However, it’s important not to lose sight of the bigger picture. While the numbers allow for quick decision making and identification of priorities, it’s important to assess the methodology behind the risk rating and if it was open to subjective biases or manipulation

Where Does The Scoring Matrix Come From?

Despite the prevalence of semi-quantitative risk assessments, it’s important to note the Health and Safety Executive (HSE) in the UK does not use scoring systems or matrices in its risk assessments.

You can view a sample risk assessment provided by the HSE under resources here: http://www.hse.gov.uk/risk/

Instead of scoring systems, you will see the focus is placed directly on identifying hazards, evaluating risks based on professional judgement, and putting controls in place without relying on numerical calculations.

The HSE’s approach emphasises practical risk management over the theoretical manipulation of numbers. By focusing on real-world conditions and the implementation of controls, this methodology avoids many of the pitfalls of semi-quantitative risk assessments, such as subjectivity, manipulation, and over-reliance on scores.

Conclusion

While semi-quantitative risk assessments and matrices can help prioritise actions, they are not without their problems. The subjective nature of assigning scores, the potential for manipulation, and the overemphasis on scores rather than real hazards suggest that we need to evaluate the numbers for what they are.

The real value in any risk assessment comes not from the numbers but from the actions taken to manage and reduce risks. Whether or not a matrix is used, the focus should always remain on identifying hazards, understanding the risks, and implementing effective controls to ensure the safety of everyone involved.