Guide | Audience Finder Fieldworker Guidelines
A full guide to collecting questionnaire data that will help your fieldworkers to gain a clear, accurate and representative picture of your audiences.
Audience Finder fieldworker guidelines
Audience Finder is a free national audience data and
development tool, enabling cultural organisations to understand, compare and
apply audience insight. Quantitative survey data is collected by the
participating organisations using either face-to-face surveys or e-surveys.
These guidelines are intended to be used by fieldworkers i.e. those
staff/volunteers who are responsible for undertaking face-to-face surveys with
audience members, or by those responsible for training fieldworkers.
There are three key parts to the fieldworker’s role:
- Ensure an equal representation of respondents.
- Maximize the number of questionnaires collected (keep refusals to a minimum by engaging the respondent in the research process).
- Record responses accurately.
Each organisation should aim to collect 380 surveys over the 12-month
data collection period, or on average about seven to eight a week. This number
of surveys provides a robust sample and therefore ensures a low margin of
error. Furthermore, once a large sample has been achieved, it’s possible to
conduct insightful in-depth analysis of the data set.
- Do first time visitors behave differently?
- Which types of marketing communication channels do families respond best to?
For more guidance about how to interrogate your data in this
way, please see information on the Audience Finder enhanced survey dashboard.
The target group is anyone aged 16 or older and who has
engaged, on some level, with the event/organisation’s offer. If you are not
sure if someone is at least 16, or if they have engaged with the offer – always
ask. Please note that even if you collect data from under 16 respondents in
your sample, these survey responses will be automatically stripped out from
your Audience Finder dashboard.
The sample should be random, i.e. all audience members should have the same chance of
being approached to take part. Collecting a random sample is an effective
way of collecting responses from a wide range of people. The more people are
spoken to, the more we can be confident that their views are likely to be
representative of all visitors.
To ensure a good representation of visitors,
interviews should be spread throughout the week. For example, if you only
conduct interviews on Monday afternoons or Saturday mornings, it’s likely that
you’re only going to be sampling a particular type of visitor.
It’s a good idea to follow an interviewer schedule or sample framework which should reflect your visitor flow - i.e. if you know from visitor footfall counts that 50% of your visitors come on a Saturday, then 50% of your weekly target should be collected on this day. The Audience Agency can assist with creating bespoke sample frameworks to help you plan your fieldwork more effectively, please get in touch with one of our team for further information.
Conducting an audience survey
It’s important to be outgoing and keen to
interact with strangers - fieldworker body language and the survey introduction
are key! Start by offering a polite introduction and explain briefly what you
are doing. Avoid using questions or phrases that give the respondent a means of
quickly refusing (e.g.: may I… Can you spare a few minutes?).
As part of the introduction, communicate
the importance of the research, and convey the value of the respondent's
participation. Try to make the person being interviewed understand that their
opinions are very important and relevant to the organisation. It is important
to stress that there’s no need to be knowledgeable to participate in the study
– they don’t need to be an expert on the arts or a regular attender to take
Only one person per group should be
interviewed. If a group of visitors is approached, in order to ‘randomise’ which
member of the group is interviewed, interview the person whose birthday is
It’s important to avoid a biased sample, in which some groups are over-represented and others are under-represented. Common sources of bias include:
- Selection bias - fieldworkers only approaching visitors who look friendly or appear similar to them (e.g. only approaching people of the same age or gender).
- Refusal bias - certain types of people being less likely to want to take part (e.g. those visiting in large groups/families.)
- Interviews more likely to be conducted among people who have a closer relationship with the organisation/event.
To overcome interviewer bias, take the
decision making out of your own hands and approach every nth person
or group that walks past. Whether this is every 3rd person, or every 5th
person, or every 10th person etc., will depend on how busy the venue/event is.
Surveying large groups and families can be
difficult, they tend to have less time and/or have their hands full. Due to the
higher propensity to refuse, it can be tempting to just focus efforts on single
visitors or those without kids.
However, the quality of data collected is only
as good as the sampling and in order to be representative surveys must be done
with all types of visitors. Think about the number of families that come
through your doors, if it’s around half your visitors, then one out of two
interviews should be conducted with this group.
The interview should be led by the fieldworker. Respondents should not be handed a questionnaire to fill out themselves, unless they ask to do so. In this situation, the interviewer should be on hand to answer any questions and quickly check the questionnaire for missed questions on receiving it back from the respondent.
Each respondent should have the same experience regardless of who interviews them. In order to maintain this consistency, questions should be asked verbatim (exactly as they are written) or, where a question contains a long list of answer codes, they should be shown to the respondent, (e.g. “Which of the following describe your motivations for visiting today?”). Even when aware of all the options, respondents are often more likely to pick the option(s) towards the top of the list, so please ensure the respondent has considered all the options.
To get the best out of respondents give them time to think about the answer before prompting/assisting with responses but make sure assisting does not turn into leading. Often when a respondent doesn’t seem to understand the question, it is because they haven’t heard it properly. At this point repeat or re-framing the question and avoid reading out only some of the options.
Probe where necessary, e.g. when the respondent gives an insufficient answer. This most commonly occurs when asked open response questions – for example, ‘is there anything else you would tell us about your experience of the [venue/content]’.
It’s important to remain impartial throughout, otherwise respondents may be tempted to change their answers to be what you want to hear.
Make sure all questions are completed and
responses are legible. If a visitor prefers not to answer a question, just move
on to the next one. The survey will not be void if this happens and the data
from answered questions will be included in the sample.
Some questions within the survey are a little
more sensitive and personal (e.g. age, ethnic origin or disability). This data
is very important to understanding the demographic profile of audiences and
evidences which groups an organisation is serving, and which groups they
Most of these questions are designed to align
with the way the government collects Census data, which enables direct
comparison with local, regional and national population data.
The Audience Agency research shows very few
people decline to answer these questions. In the Audience Finder 2016/17
benchmark (sample of over 150,000 responses) only 2% of respondents ticked
‘Prefer not say’ when asked the ethnicity and disability questions and only 1%
gave the same answer to the age and sex questions.
Postcodes are also a key piece of audience
data and must be collected in full. Reassure respondents that postcode
information will not be used for direct marketing and will only be used for
research purposes (e.g. mapping and drive time analysis). On average, postcodes
identify 15 households, and are therefore not suitable for future contact with
It’s important that respondents know why they are being asked the question. In general, they’re happy to answer if they feel that:
- Something useful will be done with the results
- The information they give will not be used for direct marketing
- The process is direct and honest
If the respondent still doesn’t want to answer a specific question, that’s fine, just move on to the next one.
Common mistakes to avoid
In addition to collecting a representative sample it’s important to ensure that the data recorded is valid and of high quality. Ensure that the responses are unambiguous and easy for the data enterer to decipher. Avoid common mistakes such as:
- Ticking more than one box for single choice questions
- Only ticking one box for multiple choice questions (unless only one answer given)
- Leaving open-ended follow on questions blank (e.g. ‘please specify other’)
- Incomplete postcodes (e.g. SE1)
- Entering a range of values when a numeric open text question is asked (e.g. 5-6 times)
- Paraphrasing/re-wording the response
If you have access to the Audience Finder dashboard, review your results regularly to sense check the sample. Are you collecting surveys throughout the year and does your month-by-month count reflect your visitor footfall? Does your visitor profile support your observations of audiences coming through the door? Are fieldworkers collecting enough surveys from groups and families?
Data Protection & the Market Research Society Code of Conduct
Data protection legislation covers all aspects of how organisations might collect and store data on individuals. The key point for this project is around informed consent, which consists of two main elements:
- Transparency. Ensuring that individuals have a very clear and unambiguous understanding of the purpose/s of collecting the data and how it will be used.
- Consent. At the time that the data is collected, individuals agree to their data being collected, and have the opportunity to withhold their agreement to any subsequent use of data. In effect, researchers must always inform respondents about how their data is to be used and give them the opportunity to refuse permission for the data to be used in this way.
The MRS Code of Conduct is a voluntary code of practice that all members are obliged to follow and endorse within any research programmes they are involved with. The aim of the Code of Conduct is to ensure that professional standards are maintained at all stages of the research process. The fundamental principles concerning interviewers and underlying the MRS Code of Conduct are that research is founded upon the willing co-operation of the public, that it should be conducted honestly, objectively and without unwelcome intrusion/harm to respondents, and that the rights of respondents are paramount.
Key points of the Code of Conduct for this research project are:
- Respondents must not be misled when being asked for cooperation to participate in a research project.
- A Respondent’s right to withdraw from a research project at any stage must be respected.
- Respondents must be able to check without difficulty the identity and bona fides (credentials) of any individual and/or their employer conducting a research project.
- Interviewers must ensure that all of the following are clearly communicated to the Respondent: the name of the interviewer; the general subject of the interview; the purpose of the interview; if asked, the likely length of the interview.
- Respondents must not be unduly pressured to participate.
- Interviewers must delete any responses given by the Respondent, if requested, and if reasonable and practicable.
- Interviewers must not reveal to any other Respondents the detailed answers provided by any Respondent or the identity of any other Respondent interviewed.
- Where incentives are offered, Interviewers must clearly inform Respondent on who will administer the incentive and its conditions.
- The anonymity of Respondents must be preserved unless they have given their informed consent for their details to be revealed.
- Strict regulations apply for interviewing children.
you have any further questions whatsoever, please feel free to contact email@example.com