Part 3: Can we measure culture – or are we measuring something else?
At work, we behave as if “culture” is something we can capture in a dashboard. We run surveys, plot index scores, and compare ourselves to benchmarks. This series asks what those numbers are actually telling us. In this part, the aim is to help leaders, HR, and OD practitioners distinguish between measuring “culture” in the abstract and measuring concrete constructs such as fairness, trust, or role clarity that you can genuinely act on.
​
Introduction
If culture matters, we should be able to measure it. That seems reasonable.

Across organisations, we run:

• Culture diagnostics
• Engagement surveys
• Values alignment assessments
• Competing values frameworks
• Proprietary culture tools

We generate dashboards, heat maps, and index scores.
But a more uncomfortable question sits beneath all of this:
Are we measuring culture or are we measuring something adjacent to it?

The Science
The CIPD evidence review examined how organisational culture is assessed across the research literature.
Several issues stand out:

• Over 70 diagnostic instruments exist.
• There is no consensus on the best method.
• Some tools show predictive validity.
• But construct validity remains unclear.

Construct and predictive validity matters. Predictive validity tells us that a measure correlates with an outcome. Construct validity tells us whether we are actually measuring the thing we claim to be measuring.
In some cases, prominent tools such as OCAI and DOCS were originally designed to measure organisational effectiveness. They were later positioned as culture instruments. This creates a circular problem. If a tool measures effectiveness-related variables and then correlates with performance, are we observing:

Culture → Performance
Or simply:
Effectiveness indicators → Performance

The distinction is not academic nit-picking. It goes to the heart of what we believe we are managing. The CIPD review also notes that many studies rely on cross-sectional surveys and self-report data, often completed by senior managers.

That introduces further concerns:

• Single-source bias
• Perception distortion
• Lack of temporal ordering

If we cannot clearly define culture, and we struggle to measure it with construct clarity, then management confidence should be cautious.

Key Findings
From the evidence:

• There is no agreed method of assessing culture.
• Scholars remain divided between qualitative deep-dive approaches and quantitative survey-based approaches.
• Many instruments demonstrate reliability and some predictive validity.
• However, it remains unclear what exactly they are capturing.

That final point is critical.

It is possible that culture surveys are measuring:

• Leadership behaviour
• Fairness perceptions
• Trust levels
• Psychological safety
• Clarity of strategy
• Incentive alignment
• Workload stress
• Management quality

All of which are meaningful. But those are not identical to culture as defined by Schein’s deep assumptions model. We may be aggregating perceptions of management systems and calling the result “culture”.

What Does This Mean in Practice?
This is where it becomes practical. When you run a culture survey, ask:

What constructs are actually being assessed?
If employees respond to items about:

• Feeling listened to
• Fair treatment
• Clear expectations
• Accountability
• Recognition

Are you measuring culture? Or are you measuring:

• Managerial competence
• Justice perceptions
• Psychological climate
• Performance management quality

These distinctions matter because the interventions will differ.

If low scores reflect poor role clarity, the intervention is structural.
If low scores reflect pay inequity, the intervention is economic.
If low scores reflect leadership inconsistency, the intervention is behavioural.

If everything is labelled culture, solutions become vague. Measurement clarity forces intervention precision.
This does not mean stop surveying. It means: be explicit about what your instrument captures and resist the temptation to let the word culture absorb multiple system variables without differentiation.

If You’re a Leader or HR Practitioner, Try This Week
To turn this into action, you can take a more forensic look at your “culture” data:

1. Deconstruct your latest survey.
   Take your most recent culture or engagement survey and list the main themes it measures (for example, trust, workload, leadership, clarity, recognition). Relabel each theme with the most concrete construct you can: “role clarity”, “perceived fairness”, “manager availability”, “psychological safety”.
2. Match constructs to interventions.
   For one low-scoring area, ask:
   • If this is really about workload, what structural changes would we consider?
   • If it is really about fairness, what policy or pay decisions are needed?
   • If it is really about leadership behaviour, what support or consequences are required?
     Write one potential intervention for each interpretation. Notice how different they are.
3. Change the way you present results.
   In your next slide deck or report, replace a generic phrase like “our culture scores are down” with a more precise line such as: “Scores on role clarity and perceived fairness have dropped; these are our current hypotheses about why.”
4. Ask your vendor harder questions.
   If you use an external tool, ask the provider:
   • Which constructs does this instrument validly measure?
   • Which constructs does it not measure?
   • What evidence do you have that this is about culture rather than, for example, leadership or climate? Use the answers to refine how you talk about the results internally.
5. Pilot a more targeted measure.
   Before commissioning another broad “culture” survey, consider a small, focused pulse on one construct you care deeply about (for example psychological safety or role clarity). Make the construct explicit from the start, and design your interventions to match it.

These steps will not eliminate all ambiguity but they will reduce the gap between what your numbers claim to measure and what you actually act on.

A Quote to Reflect On
“Although these instruments have often provided good predictive validity… there is no evidence of construct validity, meaning that it is unclear what exactly is being measured.”
Organisational Culture and Performance: An Evidence Review

A Question to Reflect On
When your organisation receives its next culture report:
Do you know which underlying constructs are being measured?
And do your interventions match those constructs?

If the answer is “not really”, your most impactful next step may be to tighten the link between what you measure, what you call it, and what you actually do.

Further Reading

• Barends, E. & Rousseau, D. (2022). Organisational culture and performance: An evidence review. CIPD
• Chatman, J. & O’Reilly, C. (2016). Paradigm lost: reinvigorating the study of organisational culture.
• Schneider, B., Ehrhart, M., & Macey, W. (2013). Organisational climate and culture.

 

Part 2: If Culture Improves Performance, How Exactly?

At work, we often say “culture drives performance” as if the pathway is obvious. This series explores what the evidence actually shows about that claim. It is for leaders, HR, and OD practitioners who are under pressure to “use culture” as a performance lever and want more clarity before investing time and money. The aim of this part is to help you move from culture as a slogan to culture as a clearer cause-and-effect story you can scrutinise and, where appropriate, act on.

Introduction

We often hear that culture drives performance.

It is said with confidence:

• Strong cultures outperform weak ones.
• Winning companies have distinctive cultures.
• Culture is a competitive advantage.

But once you ask a simple question, the certainty becomes harder to sustain:
How, exactly, is culture supposed to improve performance?

What is the mechanism?
Not the slogan, the mechanism

The Science

The CIPD evidence review explicitly examined the assumed logic model linking organisational culture to performance. Their conclusion is striking. It is unclear how organisational culture enhances performance.

The claim that culture affects performance rests on three assumptions:

1. Organisations have an identifiable culture.
2. Culture is related to performance.
3. Culture can be changed to improve performance.

But when researchers look for a coherent, unified theory explaining how culture produces performance gains, they do not find one. Instead, what exists is a collection of loosely related hypotheses.

For example:

• An “adhocracy” culture is said to enhance innovation by encouraging autonomy and risk-taking.
• A “market” culture is said to improve financial performance by emphasising results and competitiveness.
• A “clan” culture is associated with collaboration and internal cohesion.

These are plausible stories. But plausibility is not the same as causal clarity. There is no single, integrated model explaining:

• Through which pathways culture affects behaviour.
• Under what conditions it matters more or less.
• Whether it precedes performance or follows it.

In fact, some longitudinal studies suggest the direction of causality may run the other way: performance can shape culture, not just the reverse.

Key Findings

The evidence on culture and performance shows:

• Correlations are moderate to low (mean around 0.16 across studies).
• Associations are substantially weaker when objective performance measures are used.
• Culture is a weaker predictor of performance compared with factors such as general mental ability, personality traits, or structural variables.

Importantly, correlations do not explain mechanism. If culture correlates with innovation in certain contexts, that still leaves open critical questions:

• Is it culture that drives innovation?
• Or do innovative firms retrospectively describe themselves in certain ways?
• Or do both emerge from underlying leadership, incentives, and strategy?

Without a clear cause-and-effect story, culture risks becoming an explanatory shortcut.
When something works, we attribute it to culture. When something fails, we diagnose a culture problem. But that may be post hoc reasoning.

What Does This Mean in Practice?

In HR and leadership conversations, culture is often treated as a performance lever.
The implicit model looks like this:
Define desired values
→ Communicate them clearly
→ Align behaviours
→ Improve performance

But if the mechanism is unclear, the intervention pathway becomes fragile. So before launching a culture initiative, it may be worth asking:

• What behaviour are we trying to change?
• What system currently stabilises that behaviour?
• What incentives reinforce it?
• What leadership signals normalise it?
• What consequences make it costly to act differently?

If culture influences performance, it likely does so indirectly:

• By shaping coordination.
• By reducing friction.
• By influencing trust and psychological safety.
• By stabilising expectations.

In other words, culture may not be the engine, it may be the operating conditions. Operating conditions matter, but they are not the same thing as horsepower. The risk for practitioners is not believing culture matters, the risk is assuming you understand the pathway without interrogating it.

If You’re a Leader or HR Practitioner, Try This Week
To move from slogans to mechanisms, you can run a few focused experiments:

1. Name the specific outcome, not just “performance”.
   Choose one outcome that matters this year (for example, faster decision-making, fewer safety incidents, better cross-team collaboration). Write it down in a single sentence.
2. Sketch your current “culture story”.
   For that outcome, write how you currently believe culture helps or hinders it. For example: “Our culture of openness helps us spot risks early” or “Our blame culture slows reporting.” Then ask: what is the actual behaviour in meetings, emails, and decisions that makes this true?
3. Map system levers behind the story.
   For the same outcome, list:
   • How are decisions really made?
   • Who has authority to say “yes” or “no”?
   • What gets rewarded or penalised (formally and informally)?
   • What happens when someone raises a problem or makes a mistake? This starts to separate “culture” from the concrete levers you can adjust.
4. Test one assumption about cause and effect.
   Take one belief such as “we need a culture of accountability to improve results” and reframe it as: “We think clearer goals, regular follow-up, and consequences for missed commitments will improve results.” Run a small experiment (for example, weekly progress reviews in one team) and see what changes before you launch a broad “accountability culture” programme.
5. Be explicit in your language.
   In your next communication or slide deck, replace one phrase like “our culture drives performance” with a more precise statement: “These three practices (X, Y, Z) are how we expect to improve performance; over time, if we repeat them, they will shape what people experience as our culture.”

These steps will not produce a perfect logic model. But they will force you to be clearer about how you believe culture connects to performance - and where you might be relying on story rather than evidence.

A Quote to Reflect On
“The logic model for the culture–performance link is not based on a single coherent theory, but rather a number of separate, loosely related hypotheses.”
— Organisational Culture and Performance: An Evidence Review

A Question to Reflect On
When you say culture drives performance in your organisation:
Can you clearly articulate the mechanism? or are you inferring cause from correlation?

If your answer is closer to the second, your next step may be to pause the next broad “culture” initiative and first map the specific pathways you are actually trying to change.
​
Further Reading

• Barends, E. & Rousseau, D. (2022). Organisational culture and performance: An evidence review. CIPD
• Chatman, J. & O’Reilly, C. (2016). Paradigm lost: reinvigorating the study of organisational culture.
• Schneider, B., Ehrhart, M., & Macey, W. (2013). Organisational climate and culture.