The importance of semantic associations

The following publications  stress the importance of subconscious semantic associations in purchase decision making. 

Thinking Fast and Slow

Daniel Kahneman, psychologist and 2002 Nobel prize winner Behavioral Economics, was  named 7th most influential economist in 2015 by “The Economist”.

Kahneman’s Thinking Fast and Slow is considered a hallmark publication when it comes to the impact of our instinctive “system 1” brain processes.

Together with his colleague, Amos Tversky, Kahneman conducted extensive research on the impact of our subconscious thinking.   He mentions two systems of thinking. System 1 is fast, intuitive, impressionistic, associative and fully automatic. System 2 is slower, reflective, logical and rational. Both systems influence each other and influence the decisions that we make. Nonetheless, 95% of these decisions are made by the subconscious System 1 processes!


In ‘Decoded’, published in 2013, Phil Barden shows us a practical framework, backed by neuroscience,  explaining “why people buy what they buy”. Based on insights from psychology, neuroscience and behavioural economics Mr. Barden found that purchase decision making is not so much influenced by signals (colours, shapes and logos) but by the mental concepts (semantic associations) these signals trigger in our brain.  Signals such as colours, shapes and logos, are translated in our brain into mental concepts based on learned associations from the past. These mental concepts give value to a brand or product and influence the purchase decision making. 


The image underneath shows how marketing communication stimuli, such as logos or advertisements are recoded into certain mental concepts. The semantic network, that is based on associations between words, offers the context wherein this recoding takes place. Mental concepts may or may not lead to certain emotions and finally to purchase decisions.

Kahneman and Barden leave us with two main take-away messages:

  • A semantic network in our brain creates a context giving meaning to marketing communication stimuli continuously being processed by consumers at the subconscious level.
  • Subconscious semantic associations are the precursor of emotion, the latter being more difficult to measure objectively

Semantic database

Mindspeller uses a unique semantic database.

A semantic database is focussed on defining the semantic meaning by connecting the data with other meaningful data (putting to a context).   A commonly used metric is Forward Association Strength (FAS): respondents are asked what word (or words) pops up when seeing a word or image (fee association task). The FAS is the percentage of participants in the survey that answers a certain association towards a given word. Vice versa, the share of respondents answering the first word in response to the second, is referred to as Backward Association Strength (BAS). Since we know that semantic associations are active in the brain during the early stages of semantic processing, even before the subject is consciously aware of this, it’s essential to use metrics based on spontaneous associations like FAS and BAS as proxy for the semantics neural relations in the brain, provided they can be validated. Mindspeller’s founding team was the first to correlate with statistical significance the  FAS with the magnitude of brain activity registered with EEG  (van Vliet et al. 2016) when processing target and prime words (semantic priming experiment).

Research shows that a semantic network is an excellent tool to predict cognitive behavior (De Deyne, Navarro & Storms, 2013De Deyne et al, 2013b). Since subconscious associations influence pruchase decisions, a semantic network can be used to predict purchase behavior.

As the meaning of a concept is represented as a node linked to other nodes in the network, it becomes possible to position a new concept in the network (and, as such,  organically grow the network). A limited number of respondents suffices to provide you with sponteaneious associations. Mindspeller’s survey and positioning protocol can reliably postition concepts as from 50 respondents (depending on the concreteness/abstractness level of the concept) With this small dataset, a new node can be positioned in the network allowing to visualize the complete subconscious meaning of the concept in the brain.

Source of standard wanted associations

Accurately measuring distances between brands and wanted associations in a network is only part of the story. When the chosen wanted associations are not correct, the results can be deceiving. Therefore Mindspeller suggests to use following set of wanted associations, which is based on scientific research:

Aspired Associations



care, trust, proximity, security, warmth


Relaxation, openness, pleasure


Freedom, courage, rebellious, discovery, risk


Pride, success, power, superiority, recognition


Regular, hones, healthy, happy


Reliable, intelligent, successful


Hard, outside

Each of these wanted associations consists of different sub-associations to facilitate a more accurate measurement of distances. The wanted associations Safety, enjoyment, adventure and Autonomy were suggested in the bood Decoded (Barden, 2013). The wanted associations Honesty, Competence and Durability were suggested by Jennifer Aaker in “Brand Personality Framework (Aaker, 1997). Mindspeller has the tools to better focus these associations to your sector and wishes.

Priming effect

Kahneman says: “You know far less about yourself than you feel you do.” Our environment and the circumstances influence our actions in a subtle way. This is called “priming effect”.

Exemplary for this is that if you see words about aging, you automatically start walking slower.

Focussing on semantics research has shown that showing certain words trigger associations or can bring recognition. Kahneman mentions in his book a research where the letters ‘SO_P’ are completed differently depending the context the subjects are in. Subjects who had been shown the word ‘EAT’, completed the word to “SOUP”. Subjects who had been shown the word “WASH”, completed the word to “SOAP”.

Priming can be used positively. When mergers or take-overs are difficult and co-workers are not enthusiastic, texts are adjusted with a lot of words associated with ‘working together’, to create a positive vibe. Also advertising campaigns use priming, often without us realising it.

In market research, uncontrolled priming effect is a disturbing factor. In traditional market research respondents are questioned about f.e. 5 different banks at the same time. They get a list of names of these banks and say if they associate these banks to certain features or values. The order in which the banks are shown, has a great influence on the answers that are given. The result of this research is thereby influenced according to the order in which the brands are shown. This priming effect is evitable. In some cases uncontrolled priming is useful, f.e. when you want to investigate what consumers think of your brand after showing them a commercial message from another brand.

It is of importance that in market research we formulate the question in a way priming effect is under control. With Mindspeller we ask spontaneous associations about all kinds of concepts to avoid respondents knowing the goal of the study.

Showing certain features, followed by “yes or no” answers whether they are valid for what is shown (as done a lot in market research), is also a way of priming: certain associations are shown, which you possibly would not make beforehand. It could be that you do not spontaneously associate a certain brand with ‘reliable’, but when asked “is this brand reliable or not?”, you would only answer ‘no’ because of a past experience. With “Yes or no” questioning the brand will be described as ‘reliable’.

Mindspeller only searches for spontaneous associations so this form of priming will not influence the results of market research with Mindspeller.