Profiling research and profiling interview: meaning and application

Data shows you that people drop out, hesitate or choose with conviction, but it does not show you why. Profiling research looks for that explanation in in-depth interviews with supporters and opponents alike. This reveals how associations feed into trade-offs and behaviour, and how people justify their choice to themselves and others.

Profiling research is a Neurofactor research format in which you use in-depth interviews to examine how people turn their associations into trade-offs and behaviour, and how they justify that choice. The profiling interview is the instrument for this: a semi-structured interview that explains an attitude rather than measuring it.

This is a Neurofactor working definition, not a standardised and validated method. A short association test beforehand shows a tendency and focuses the interviews. "Profiling" here does not mean assessing individuals, which is what the General Data Protection Regulation (GDPR) means by profiling.

At Neurofactor, profiling research is the in-depth, qualitative step in association research. An association measurement shows which meanings people link to an offer, an organisation or a change. It does not show how those associations are weighed in a specific choice. Two people may both link an e-bike with freedom and with hassle, while one buys and the other drops out.

The research format has three parts:

• Preparatory association test: a short measurement that shows which associations dominate and where groups differ.
• Profiling interviews: individual, semi-structured interviews with people who made the choice, did not, or are still undecided.
• Analysis and reporting: themes about trade-offs and justifications, compared with the association data.

"Profile" here refers to the pattern of trade-offs within a choice context. It is neither a personality profile nor a clinical or psychological diagnosis of individuals.

In data protection law, "profiling" has a fixed meaning. Article 4(4) of the GDPR defines profiling as any form of automated processing of personal data used to evaluate personal aspects of a natural person, for example to analyse or predict their behaviour, preferences or reliability (Regulation (EU) 2016/679).

Profiling research is different: it does not assess or decide about individual participants and reports on themes and patterns. You do process personal data, because recordings and transcripts fall under the GDPR. Linking individual interview results to a customer database to select or approach individuals may move towards profiling in the GDPR sense. Have such use legally assessed beforehand; this article is not legal advice.

The broader services are described on the method page Profiling and data strategy.

The profiling interview combines documented techniques from qualitative research. These sources provide context; they do not validate the Neurofactor approach as a whole.

• Semi-structured in-depth interview: an interview with a fixed guide and room for follow-up questions. Brinkmann and Kvale describe interview research in seven stages, from thematising and designing to reporting (Brinkmann & Kvale, 2015).
• Laddering and means-end chain: Gutman described a model that links perceived product attributes to personal values (Gutman, 1982). Reynolds and Gutman developed laddering as a one-to-one interviewing technique for understanding how people translate attributes into meanings for themselves (Reynolds & Gutman, 1988). You ask step by step why something matters.
• Critical incident technique: Flanagan described an approach that records concrete behaviour in specific situations rather than general opinions or estimates (Flanagan, 1954).
• Thematic analysis: Braun and Clarke describe six phases for finding patterns of meaning in qualitative data (Braun & Clarke, 2006).

The preparatory association test is a short measurement among a broader group, for example with free associations or ratings of preselected attributes. For careful practice, see How do you measure associations?. The test has three functions:

1. Showing a tendency: which associations occur often, and do they differ between buyers, drop-outs and the undecided?
2. Supporting the selection: you choose interviewees from groups that respond differently.
3. Focusing the guide: striking or contradictory associations become interview themes.

Treat the outcome as a tendency. An association that is mentioned often is not automatically decisive in the choice. That is what you explore in the interviews.

1. Formulate the choice question. Which decision do you want to understand, in which context and for which next step?
2. Define the contrast groups. Interview supporters, opponents and the undecided.
3. Plan the size per group. Qualitative research works with purposive selection and saturation: the point at which new interviews no longer yield new themes. In their own dataset of sixty interviews, Guest, Bunce and Johnson found saturation within the first twelve interviews, with basic elements of metathemes present after six (Guest et al., 2006). This applies to one relatively homogeneous group and is not a fixed rule.
4. Write an interview guide with themes, opening questions and possible probes, leaving room for what the participant brings up.
5. Arrange consent and data handling before the first interview takes place (see below).

A profiling interview follows a fixed structure, with room to deviate:

• Opening: confirm the purpose, confidentiality and consent; there are no right or wrong answers.
• Concrete moment: ask about a specific situation, such as the test ride or the moment of dropping out, following the logic of the critical incident technique.
• Associations: let the participant say what comes to mind about the offer, and if relevant present associations from the test.
• Probing: move from attribute to consequence and to what makes that consequence important, as in laddering. Avoid leading questions.
• Justification: ask how the participant would explain the choice to someone else. This shows which reasons feel socially acceptable.
• Closing: summarise, check that you understood correctly and ask what has not yet been discussed.

The interview explains an attitude. It does not measure how strong it is or how often it occurs.

Work from verbatim transcripts and follow the phases of thematic analysis: familiarising yourself with the data, coding, searching for themes, reviewing, naming and reporting. You can work inductively, from the data, or theoretically, from preselected themes such as the associations from the test (Braun & Clarke, 2006).

Look for two kinds of themes:

• Trade-offs: which associations clash or reinforce each other, and what tips the balance?
• Justifications: which reasons do people use to explain their choice, and do they differ from what they said earlier?

Braun and Clarke stress that the importance of a theme does not simply follow from its frequency. Also describe why a theme matters for the choice question, and ideally have a second researcher check part of the coding.

For each theme, the report describes its meaning, the groups in which it occurs and illustrative quotations, along with the selection, the number of interviews per group and the limitations.

Then compare the themes with the association data. Does a frequent association return in the interviews as a trade-off, or does it carry little weight? This comparison of sources is called triangulation. When sources support each other, confidence in an interpretation grows; when they clash, you have a follow-up question.

This matters because people do not always have access to the processes behind their judgements. Nisbett and Wilson showed that people often explain choices on the basis of assumptions about what a plausible cause is (Nisbett & Wilson, 1977). What someone gives as a reason mainly shows how they understand and justify the choice. That is valuable, but not direct evidence of the cause.

This example is fictional and contains no research results.

Situation: an e-bike manufacturer sees that many people do not buy after a test ride at a dealer. The figures show that undecided customers drop out, not why.

Choice question: which trade-offs lead someone not to buy after a positive test ride, and how do buyers differ?

Available information: test ride records from dealers and a short association test among recent test riders.

Approach: suppose the test shows that buyers more often link the e-bike with convenience and drop-outs more often with weight and maintenance. That tendency sets the interview themes. The manufacturer interviews buyers, drop-outs and the undecided, each time centred on the test ride. When probed, a drop-out might move from "heavy frame" to "not being able to lift it into the shed myself" and finally to "not wanting to depend on others".

Possible interpretation: drop-outs may mainly mention the price, while independence carries more weight in their story. In that case, price may be a socially easy justification. That is a hypothesis, not a finding.

Next step: check with a survey among more test riders whether this theme is widespread, and only then discuss changes to the test ride or communication.

Profiling interviews deal with personal considerations; care belongs in the design.

• Informed consent: explain in advance the purpose, what you record, who has access, how long you keep data and that participation is voluntary.
• Legal basis: determine your GDPR legal basis. Consent is one option; the GDPR describes it as a freely given, specific, informed and unambiguous indication of wishes (Article 4(11)).
• Data minimisation: collect only what you need (Article 5(1)(c)) and pseudonymise transcripts as early as possible.
• Special categories: avoid questions about, for example, health or religion unless they are necessary and you have checked the requirements of Article 9.

For broader ethical considerations, see also Neuromarketing ethics.

• Using interviews as a measurement. "Most drop-outs find the e-bike too expensive" does not fit a small, purposive selection. Describe themes and test their size separately.
• Reading statements as causes. Stated reasons are explanations after the fact; compare them with association and behavioural data.
• Treating the association test as a final conclusion. Whether an association weighs in the choice only follows from interviews and further research.
• Creating a personal profile. Profiling research describes patterns in a choice context, not personality, and is no basis for assessing individuals.

Profiling research explains why a group does or does not make a choice. The association test shows which associations are at play; profiling interviews with supporters and opponents show how they are weighed and justified. The results explain but do not measure size, and they become stronger alongside association and behavioural data.