Home

About Us

Membership

Professional Development

Knowledge Hub

Events

Awards

Careers

Coronavirus Guidance

Industry News
IoIC
 
 
 

Peter Hutton of BrandEnergy Research provides insights on sample size and composition when carrying out employee research.


Surveys can be a good way of identifying internal communications needs and evaluating the effectiveness of your communications. But how many employees should you include in your sample?

The short answer is ‘as many as possible’ since this increases the statistical reliability of the sample – its ability to reflect accurately what all staff think – and increases your options for breaking the results down by smaller subgroups such as department or job function.

It is rare to find the same issues manifesting themselves across the business. The job content and working conditions in the warehouse are likely to be very different to those in the accounts department or marketing. The overall results may show 75% satisfied with their job, but this may hide the fact that 85%+ are satisfied in accounts but only 45% in the warehouse. Breaking down the sample so you can analyse by subgroups is therefore important to enable you to identify where your issues lie so you can take effective action.

Technology means that it is much more cost-effective to approach a very large sample of employees than it used to be – it costs little more to send 1000 emails with a link to an online survey than to send 100. However, not everyone has easy access to a computer; manual workers and others who are rarely in a fixed office will need alternative arrangements such as a printed questionnaire, which will increase data collection costs significantly.

As important as the sample size, however, is the representativeness of the sample. If you get, say, only 20% to 30% of the people invited to participate actually completing the questionnaire, then it raises questions about whether these employees are the same kinds of people as those 70% to 80% who did not complete the questionnaire. The lower the response rate, the more likely that biases have crept into the sample that favour certain kinds of respondent. For example, if the culture is fairly negative it is likely that the more negative people are the ones who have opted out of completing the questionnaire; access to online surveys will favour managerial and clerical rather than non-clerical staff; those working in clerical jobs are also more likely to feel comfortable with filling out forms compounding this bias; if participation is in some way dependent on managers or department representatives distributing printed questionnaires, or if they are important in communicating messages about the survey, then areas of the business with managers who are less effective in these roles will be under-represented in the sample.

It is therefore important to take as many steps as possible to maximise the response rate. This includes communicating effectively to staff about the survey using multiple channels over a reasonable period of time, making it as easy as possible for them to take part and ensuring that the questionnaire is interesting and feels relevant to staff throughout the organisation.

Under Market Research Society rules, in order to preserve respondent anonymity, the results should not be broken down into sub-groups of fewer than ten. Aiming to include all employees in the survey, rather than just taking a sample, and maximising your response rate, is therefore important to ensuring you get sufficient respondents in important subgroups, particularly in departments or areas of the business that employ relatively few staff e.g. HR or marketing. An industrial company with 2000 employees may have only 15 staff in marketing. If only half the employees are asked to participate, even with a 100% response rate there will not be enough in the marketing department sample to analyse the results. If 2000 are targeted then a 2/3rd response will be needed to ensure the minimum of ten staff in the marketing subgroup.

Focus groups of staff are useful as a complement to a major quantitative survey. Their value is not in being able to say x% of staff think or feel this or that, but in identifying issues to explore in, and shedding light on what lies behind the findings from, a quantitative survey. The process will also help to make sure that the questionnaire covers issues that staff feel are important and ensure buy-in to the survey.

Normally groups of 6-9 are recruited – sufficiently small for everyone to have a reasonable say and large enough to spark interesting discussion and generate different perspectives on a range of issues. Normally it is good practice that each group consists of people from the same level in the organisation; otherwise more junior staff will tend to defer to the more senior staff. To get a good spread of views a number of focus groups should be conducted among people from different parts of, and levels of seniority within, the organisation.

Peter Hutton is managing director of BrandEnergy Research Ltd, an agency providing research and consultancy services. He was previously deputy head of leading research agency MORI, and has written many articles, papers and books on research.


www.brandenergyresearch.com
Resources & Guides

Latest Jobs

Our Sponsors

Room Booking

Thanks for staying with us! Please fill out the form below and our staff will be in contact with your shortly. The see all of our room options please visit the link below.
See All Rooms