I Surveyed Clubbers at Ministry of Sound in 2009. The Methodology Did Not Account for the Bass.

In 2009, as part of my MA in Marketing Management, my fellow student Sercan and I went to Ministry of Sound to conduct market research.

This meant approaching people who had paid money to drink, dance and listen to extremely loud music, then asking whether they would mind helping us with an academic questionnaire.

In retrospect, there may have been easier research environments.

The project was multi-purpose. We wanted to understand clubbing behaviour, the audience, branded club nights and perceptions of Housexy, the Ministry of Sound-associated brand we were studying.

So we did what students did at the time.

We printed questionnaires.

We took pens.

We went to where the clubbers were.

And then we attempted to conduct structured research in a building specifically designed to make conversation difficult.

The Venue Was a Problem. It Was Also the Point.

One of the easiest criticisms to make now is that a nightclub is a terrible place to interview people.

It is noisy. People are distracted. Some have been drinking. Nobody has bought a ticket because they were hoping to discuss questionnaire design.

All true.

But there is a rather important counterargument:

when else are you going to find lots of clubbers?

That is the compromise at the heart of field research. The environment that gives you access to the right people is not always the environment that gives you ideal measurement conditions.

A pristine survey completed quietly at home may give you cleaner answers from the wrong sample. A messy conversation in a nightclub may give you access to exactly the people you need, but under conditions that affect how they respond.

The interesting question is not whether one method is perfect.

It is which imperfection matters most for the decision you are trying to make.

Some of the Questionnaire Has Aged Better Than Other Bits

Looking back at the original questionnaire now is slightly like finding an old photograph of yourself wearing something you once considered entirely reasonable.

There is nothing catastrophic in it.

There are also several things I would do differently.

The questionnaire starts with age and gender, then asks about frequency of clubbing, how important clubbing is, how people hear about events, what they expect to pay and how they perceive Housexy.

That logic is understandable. We were trying to build a profile of the audience and understand the brand.

But survey design has a habit of exposing assumptions that seemed invisible when you wrote them.

The Three-Times-a-Month Problem

The easiest flaw to spot is almost comically small.

We asked:

• Once a month
• 2–3 times a month
• 3–4 times a month
• 5+ times a month

Three times a month appears twice.

It is exactly the sort of thing that looks harmless until you try to analyse the responses.

Good closed-ended response options should be mutually exclusive and should cover the reasonable possibilities. Both Pew Research Center and the American Association for Public Opinion Research make this point explicitly in their survey-design guidance.

The irony is that the more structured the questionnaire, the more authoritative the resulting spreadsheet can look.

You can calculate percentages to two decimal places from a question whose categories overlap.

The decimal places do not rescue it.

Open Questions Were Harder When You Had to Write the Answer Down

This is one area where I would be careful about criticising the 2009 questionnaire using 2026 expectations.

Open-ended questions are useful because they allow respondents to tell you what matters to them rather than forcing them into the choices you thought of beforehand.

They are also inconvenient.

In 2009, if somebody gave us a long answer in a noisy club, somebody had to hear it, write it down and later interpret the handwriting.

That naturally encouraged structured questions.

Today, with appropriate consent, you could record an interview on a phone, transcribe it and analyse the responses later. You could mix structured survey questions with short qualitative interviews far more easily.

Technology changes what counts as practical methodology.

That does not make open questions magically better. They still take more work to code and compare. AAPOR notes the same trade-off: closed questions are easier for respondents and easier to interpret quantitatively, while open questions allow people to answer in their own words but require coding afterwards.

The difference is that the operational cost of capturing those words has fallen enormously.

Paper Could Not Skip the Irrelevant Bits for You

The original questionnaire contains several questions that only apply if somebody has heard of Housexy or is attending for Housexy.

On paper, the interviewer has to route the respondent manually.

Today a digital questionnaire can do that automatically.

If the answer is “No”, the next three irrelevant questions simply never appear.

This sounds trivial, but good routing reduces burden on respondents and reduces the opportunities for inconsistent data.

Digital tools can also randomise answer options, enforce mutually exclusive categories, flag missing responses and record timestamps without us standing in a nightclub holding a clipboard like the least exciting members of the evening’s entertainment.

The Interviewer Changes the Interview

Another problem is more fundamental.

We were there.

Face-to-face research gives you things self-completion surveys do not. You can clarify a question. You can notice confusion. You can ask a follow-up.

But an interviewer also changes the social situation.

Pew’s survey-methodology guidance notes that social desirability bias can be greater when an interviewer is present than when respondents complete a survey themselves.

That feels particularly relevant when asking somebody in a nightclub how important clubbing is to them, whether music is an important part of their life, or what they think of a club brand while they are physically standing at the event.

The context is doing some of the questioning for you.

Again, that does not make the data worthless.

It means the method should be part of the interpretation.

Housexy Was the Point, Not Just the Setting

The questionnaire becomes much more interesting when read alongside the wider consultancy project.

We were not wandering around Ministry of Sound conducting a national census of nightlife for the intellectual thrill of it.

We were trying to understand a real audience and a real brand problem.

That is why questions appear about Housexy awareness, where people first encountered the brand, whether they had attended previous events, what drew them to the night and how they rated competing club brands such as Hed Kandi, Kinky Malinki, Kissdafunk and Pukka Up.

There are things I would change about the execution.

The commercial logic is still recognisable.

Who are these people?

How do they behave?

How do they hear about events?

What do they value?

How do they see us relative to alternatives?

Those are still perfectly sensible marketing questions.

The Biggest Risk Is Data That Looks More Certain Than It Is

This is probably the lesson I care about most now.

Bad research does not always look bad.

Sometimes it produces an extremely tidy chart.

If the wording is ambiguous, the options overlap, the sample is distorted by the setting or the question has quietly primed the answer, the spreadsheet still calculates perfectly.

The software does not put a red warning above the bar chart saying:

These percentages may be nonsense, but they are beautifully formatted nonsense.

Pew describes questionnaire design as a multi-stage process because wording, answer options, question order and survey mode can all influence responses. Pre-testing matters precisely because the problems are often obvious to respondents before they are obvious to the person who wrote the survey.

Would I Do It Again?

Yes.

I would design the questionnaire differently.

I would probably start with more engaging behavioural questions rather than demographics. I would fix the response bands. I would use conditional logic. I would use a better combination of structured questions and short open responses.

And, where appropriate and consented to, I would record qualitative interviews rather than trying to annotate somebody’s answer while a sound system attempts to reorganise my internal organs.

But I would still go to Ministry of Sound.

Because however inconvenient it is to interview clubbers while they are actually clubbing, the methodology retains one considerable advantage:

What I Would Keep From 2009

It is easy to review old work and treat everything as evidence of how much cleverer we have become.

That is not especially useful.

The questionnaire did something important: it left the classroom.

Sercan and I went into the market we were studying, spoke to the people actually using the category and tried to connect academic research to a live marketing problem.

I would take a better questionnaire today.

I would take better technology.

I would probably take earplugs.

But I would still take the research to the customer.