Behavioural Science Blog

The Science of Human Behaviour

Posts Tagged ‘Dijksterhuis

Unconscious perceptual processes

with 5 comments


This article was inspired by a lecture given by Pamela Smith at the Radboud University Nijmegen.

The purpose of perception is to support behaviour.

Why do we perceive things in the first place? What is the point of it? Is it so we can understand the world? That sounds logical, but that cannot be the whole story. For example a frog can taste, smell and see, but hardly understands. If you look at “lower level” animals you see the direct connection of perception and behaviour.
Frogs have two perceptual systems, one is for finding small objects (=get it and eat it) and the other one is for avoiding large objects (hop away and hide). Thus (at least in frogs) behaviour always follows perception.
But we humans are pretty elaborate creatures (at least we like to see ourselves that way). Do we also have that direct perception-behaviour link? When we see a cup, we do not necessarily have to grasp for it.

What makes us different?

There are two theories of why humans do not react directly to perception (at all times). One of them is called the facilitator-hypothesis and it states that there needs to be another source of energy (facilitator) in order (for the individual) to become active.However another hypothesis has received much more experimental attention and results: The inhibition-hypothesis: Perceptions always elicits action, but action can be inhibited (Gilbert, 1989).
The inhibition option makes sense, because important parts of the brain are the same as in reptiles. However we have some new structures as well. The neocortical area is sort of built on top of older parts of the brain, so it would make sense that it “adds” a function to the basic brain functions. Another important source of information are patients with prefrontal cortex damage. They often have problems controlling their actions and (re)act impulsively. It would seem that their “inhibition” system has somehow become damaged.

Experimental proof for the inhibition hypothesis

Seeing-Grasping
Tucker and Ellis (1998) looked at how we perceive objects. Participants had to say whether objects show where right up or upside down. All objects had handles either pointing to the right or the left. They found an interaction effect between handle side and object orientation. For example when the participants saw a teapot pointing to the right they were faster pushing the right key, even when object orientation was irrelevant to the task. The results of this experiment is usually interpreted as proof for the existence of automatic tendencies to interaction with the object. (If you find that interesting read some studies about embodied, embedded cognitions.)
But this is not the only experiment supporting the inhibition hypothesis. There is a brain circuit that is connected to a grasping motion (Chao and Martion 2000). Chao and Martion showed participants different tools. Objects that could be grasped elicited activated that specific brain circuit, even though grasping was not a relevant feature of the task.

Seeing – Taste
Simmons et al. (2005) showed that looking at food pictures activates taste perception. In a study by Zwaan and Taylor (2006), participants turned a button either clockwise or counter-clockwise to indicate whether a sentence makes sense or not. Some of the sentences proposed a clockwise or counter-clockwise motion. Those sentences that implied a certain motion facilitated turning the button in this direction. Again it is important to notice that turning the button was irrelevant to the task.

Chameleon Effect
What about if you see somebody performing a certain behaviour? Are you more likely to perform the same behaviour? Yes: It is called imitation and it makes other people like you. It is what Cartrand & Bargh (1999) call he Chameleon effect. A participant described different pictures together with a confederate. In one condition the confederate did rub their face and shook their feet. Participants did rub the faces and shook their feet more if the confederate did so. The best explanation of why they did so seems to be that through imitation they can build rapport with the other person.

Couples’ facial expressions
Another example is the imitation of emotional expressions. From a functional point of view this might help understand emotion (Zajonc et al. 1987). Zajonc looked at romantic partners. If you spend a lot of time with a certain person you naturally mimic that person a lot. If that happens for decade you develop the same sort of wrinkles, because you are making the same sort of facial expressions every day, again and again. Observers rated husband and wife as looking much more similar than random paired strangers.

Priming “elderly” reduces walking speed
Reading about old people makes you walk to the elevator much slower (Bargh, Chen, & Burruws, 1996). In their experiment the task was to make sentences with words that included words that primed participants with the concept of elderly. Thought the spreading of activation with activate the concept of elderly, which in turn activates the concept of elderly. This actually influences walking speed on the way back to the elevator (dependent measurement).

Smell of cleaning fluid makes you behave in a clean way
Rob Holland could show that the same effect can be shown using smell as the activator of a concept (Holland et al. 2005). People filled out an questionnaire in one room. Some participants sat in a room where there was a bucket of water with a little bit of cleaning fluid. Then all participants went into a room free of smells. They then had to eat a biscuit that crumbles a lot (very messy). Participants were recorded on video and observers counted how many times participants picked up the crumbs/cleaned the table. People who first sat in the room with the smell of cleaning fluid did pick up the crumbs significantly more often. Thus we see the concept of perception and action coming back once again. Presumably the smell of the cleaning fluid activated the concept of “clean”, which then became connected to action while eating the biscuit.

Conscious or unconscious?

Owen at al. (2006) put patients that were in a vegetative state in the FMRI and they were given instructions to visit different situations, such as playing tennis or visiting the rooms in their home. You can clearly see that the brain activity in the patient and the control show pretty much the same activation pattern. So the patient was able to understand the instructions and act upon it. However was the patient really conscious or could the effect also be explained by unconscious processes?

Unconscious Processes

Pierce and Jastrow 1884 were the first people to research unconscious processes. They had objects that differed in weight. They made themselves blind to what the weights were and took two weights and wrote down which one was heavier (you can try that yourself!). Even when they had a confidence rating of zero, they were doing better than chance. Even though they were reporting the differences, they were not consciously aware of the difference in weight.

Otto Pötzl (1917) showed participants pictures for 10 ms. Some of those pictures appeared in their dreams. More recently Bob Zajonc (1980) said that affect is primary, that it occurs before cognition and does not require cognition. Staple et al. (2002) elaborated on this idea and presented faces to participants on a computer screen. The faces were showing emotional expressions and had clear gender cues. What they found was that when things were presented very quickly people only picked up the emotional information (30ms). When presented a little bit longer (100ms) people were also able to pick up on the gender cue. This makes also sense form a evolutionary perspective: It is very handy to pick up fast on the emotional state of someone.

Murphy and Zajonc (1993) presented participants with happy and angry faces for either 4ms or 1000ms. After the picture participants saw a Chinese ideograph. They had to rate the ideograph on the emotional valence (“What does that mean?”). They showed that the presentation of the face did indeed influence the rating of the ideograph when the picture was presented subliminally. In the other condition (1000ms presentation) the effect was not present. So when people become consciously aware of the emotional stimulus they might correct for emotional influences. However as subliminally presented stimuli never react the conscious mind, one can not be aware of the effect it has on behaviour.
Whalen et al. (1998) presented participants with a happy or a fearful face for 33ms and a neutral face for 167ms. People cannot report seeing two faces, but they only see the neutral face. However they showed different reaction to the faces in the amygdale, a part of the brain that is involved in reactions to dangers.

From unconscious perception to unconscious behaviour

Bargh, Chen & Burrows (1996) subliminally presented participants either with Black faces or White faces. Participants were presented with the face for 26 ms and the immediately saw a bunch of circles. They had to judge whether it is an even or an odd numbers of circles on the screen. After 130 trials they got an error message. The experimenter came in and told them that they had to start all over again. The crucial measurement in the experiment was how the participant would react when told that he or she had to do the task all over again? The results show that people who saw the black faces acted more hostile than people who saw the white faces.

Fitzsimons et al. (2008) did research with subliminal brand perception, especially Apple and IBM computers. Participants saw numbers on the screen and had to keep track of the sum of numbers While participants did that they were primed with either the Apple or the IBM logo (or something unrelated). Then they did a creativity task (Something like “How many unusual uses of the item X can you think of?”). Apple primed people showed more unusual uses than the IBM primed group. Someone should repeat that experiment and test the brand perceotion of Linux…
Subliminal perception is closely related to subliminal advertising. Can we influence People to buy those products that we subliminally brand people with? James Vicary primed people with “drink Coke” or “eat popcorn” and told that people bought a lot more. However this most probably is a hoax, as no one has ever seen real data.
But people had already an attitude about Coke. Can we also change attitudes that people already have? You might remember the subliminal advertising that the Bush campaign used in the 2000 election? The gore administration as the bureauc-RATS. But that really was not subliminal.

So can you affect peoples behaviour? Johan Karremans looked at that in a consumer behaviour situation. He started out by making some participants thirsty and others not. Participants had to taste salty sweets (drop) – in an alleged tong letter detection test to make sure they were really sucking on the drop. Then were then primed with Lipton ice or Npeic Tol. There was a priming effect, but only to people who were thirsty. So subliminally priming works but only when people are in the right state.
Does subliminal self-help work?

Ap Dijksterhuis (2004) tried to improve participants self-esteem with a subliminal condition paradigm with the word I and positive adjectives such as warm, sweet, nice, sincere…
————–
Frame 1: Xxx
Frame 2: I
Frame 3: Nice
————–
When people see themselves paired with positive words they showed more positive self-esteem (tested with the name-letter-effect).
So what can we do with subliminal influencing? Well it is not sure, the field is just starting, Only in the last decade we have computers that can present stimuli in a very short time. There is good experimental software now that we can use and a lot of new test paradigms have been developed.

Advertisements

Learn to like – Implementation Intention & Affect

with 5 comments


1. Short Summary of the research proposal

This study investigates the role of affect in consciously created automatic behaviour tendencies (implementation intentions). The general hypothesis is that the frequency of a behaviour is closely connected to the activation of a specific behaviour tendency, but this model often fails when implementation intentions are used to change or override old habits. This may be due to the affective dimension, which is contained within the cognitive structure of the habit memory, but not in the implementation intention. By adding the affective domain to implementation intentions it might be possible to mimic “natural habits” much more efficiently.

2. Description of the proposed research

2a. Research Topic

Implementation intention refers to cognitive (if-then-)constructs that mimic the way of automatic behaviour (habits). A habit is formed by repeatedly experiencing the co-activation of a stimulus in the environment and a specific behaviour tendency. Intention implementations are formed by repeated cognitive effort instead of experience. In order to change behaviour, for example in order to override a bad habit, specific behaviour tendencies (the then-part) must be made salient at the right moment, which is achieved by associating the behaviour tendency to a relevant stimulus in the environment (the if-part of the statement). As the (if-)stimulus is perceived, the behaviour tendency is activated and eventually acted upon automatically. In fact implementation intentions are an attempt to mimic the effect of automatic behaviour, which could also be called habits. The process of habit formation is therefore of central value to this research. For goal enactment to become automated, a specific behaviour must be executed in a specific environment, leading to a specific goal. If these things occur in close temporal proximity the goal-means connection in memory grows stronger. In consequence priming the goal automatically activates the behaviour tendency. For example habitual bike riders were faster to respond to “bicycle” after being primed with the goal of travelling somewhere (Aarts & Dijksterhuis, 2000). Recently Sheeran et al. (2005) have shown that this fact is also true for health related behaviour. For example the concept of drinking is more active in heavy drinkers when primed with the goal of socialising.

Thus for automatic behaviour to occur, the spreading of the activation must have been prepared. If goals can be non-consciously activated than these goals must pre-exist in the mind of the actor. The pre-existing knowledge structure connected to these goals is thought to incorporate the goal itself, as well as the context and opportunities and actions that lead to goal fulfilment (Aarts & Dijksterhuis, 2000). But if goals can be activated automatically, how do people know which goals are desirable to attain and which states are worth working for? Ruud Custers and Henk Aarts (2005) have pointed out that mental accessibility is not enough for a goal to be acted upon. The idea of embodied embedded cognition proposes that information about a specific behaviour provided by the nervous system includes affective information. The information about a specific state not only includes information about the means necessary to attain it and important environmental cues, but also the information that this state is desired. This idea is supported by Aarts, Gollwitzer and Hassin (2004) who showed that the effect on effort were stronger if the goal state was more desired. There also needs to be a reason to come into action, which can be seen as a discrepancy between the actual state and the goal state. Furthermore the goal state must be associated with positive affect. Automatic goal-directed behaviour can arise only if all those three prerequisites are met (Custers & Aarts, 2005).

Priming has been investigated in many studies and has proven to be a very reliable effect. Much less research has been conducted into the other two factors, discrepancy with the actual state and association with positive affect. These two factors are much harder to study than priming effects, as people set their goals not based on objective value, but on their subjective interpretation of the variables involved (Kahneman & Tversky, 1979). But in order to change behaviour effectively we will have to fully understand the way automatic behaviour tendencies are naturally created.

I believe that processes that do only rely on association (as intention implementations do) are no match for the “real habits” that are enriched with affective information that tell the organism to follow through with that behaviour plan. In order to change behaviour effectively we will need to create habit-like structures that have a fair chance of competing with the automatic behaviours already in place. This might be especially important for addiction in which the situational stimulus, as well as the goal state are closely connected to physiological states that are regulated by the dopamine pathways of the mesolimbic system. Thus it is questionable if the strategy used by Holland, Aarts and Langedam (2006) to change recycling behaviour would be effective if it was used to change a behaviour that was strongly associated with positive affect. In this study I hope to show that affect is an important building material for the formation of habits and that it will interact with and enhance the effect of cognitive association training, such as implementation intentions.

Aarts, Custers and Holland (2007) showed that our mental apparatus is capable of ceasing goal directed activity when the goal is associated with negative affect. However this effect probably does not last for a long time and it is questionable if the priming paradigms used today have the power to compete with automatic behaviour tendencies strongly associated with positive affect, as they are found in addictions. The main question of this study is if affective information can be included into the implementation intentions. I hypothesize that affective priming just before engaging in forming intention implementations will include the affective information in the habit-like-memory of the intention implementations. When the if-stimulus of the intention implementations is perceived, the affective information will be activated together with the behaviour plan and in turn enable intention implementations to compete with the old habit.

2b. Methodological Approach (research methods, procedures, instruments, data collection)

Experiment 1. Affective priming and implementation intentions

Experiment 1 is designed to test the hypothesis concerning the influence of affective priming on implementation intentions. It is hypothesized that positive affective priming will affect implementation intentions. More specifically the affective priming is expected to enhance the effect of implementation intentions for participants with a weak prior habit, but not for participants with a strong habit already in place.

Method: Following the paradigm of Aarts en Dijksterhuis (2000), only university students who live in Nijmegen and own a bike are recruited. Each participant is randomly assigned to one of the conditions.

In a pilot study, locations for which a bike (bus, train, walking) would be a useful means of transportation are obtained. The five locations that have been indicated by a high percentage of participants to be suited for going to by bike are selected. Subsequently the most frequently mentioned reason to travel to this location is chosen. These reasons to travel constitute the description of the travel goals that are supplied to the participants during the experiment. In addition another five locations are chosen that have been judged to not be suitable for bike travel. This is repeated for all travel modes, so 40 travel mode-location pairs are created. Only the five pairs that consist of a biking-location and in which biking is the travel mode are selected as targets. The other 35 pairs are used as filler trials.

The experiment is run by a computer, which also provides all instructions. The students are told that the study consists of several experiments, designed by different experimenters and that one of the tasks will be to bike to one of the five (biking-congruent) travel goals, as part of an experiment that studies the influence of planning on behaviour. It is import to note that only the goal and the travel mode (biking) is primed, not the actual location. Participants then learn that the travel goal they will need to attain will be revealed in the last part of the experiment.

Then they are asked to complete an alleged reaction time experiment in which they have to zoom in or out pictures of transportation modes (bike, bus, train, walking), depending on the orientation of the picture, by the use of a joystick (pull or push). The approach-avoidance test (AAT) is usually used to measure implicit affective attitude, but it has been suggested, that it can also be used to change affective valence. Each participant completes two trials of 80 pictures (20 for each transportation mode). In the pull-bike condition participants have to pull 90% of the bike pictures and 50% of the images depicting other transportation modes. In the control condition images of each transportation mode must be pulled 50% of the time.

Each participant then engages in the planning task, in which they plan the five travel goals. They are asked to indicate the time of the day, the location of the district of the destination and the travel route suitable for the travel goal. Next, participants complete an association task in which they are presented with an images of 40 different locations (200 ms), that are followed by a row of asterisks (100 ms) and an image depicting a mode of transportation. Participants have to indicate as fast as possible if the depicted transportation mode is a suitable option for going to the depicted location. The dependent measurement is the mean response latency time across the five target location-bicycle pairs.

In the last task participants are presented with 10 different travel locations, which consist of the 5 target locations and 5 nearby locations. The are asked to indicate how often they used the bike to reach any of these location during the last 2 weeks. The median of the frequency is used to separate those participants that already have a habit of biking to the target locations from those that do not (or to a smaller extend). In later experiments it could be advisable to measure the habit-strength in advance and only admit those participants to the experiment that do not already have a habit formed.

It is expected that participants of the pull-bike condition react faster to the target location-bicycle pairs than participants of the control condition, but only if they did not have strong habits already.

Experiment 2. The duration of affective enhanced implementation intentions

Due to the close temporal proximity of the bike-AAT and the association task, the effect of the affective prime might be quite small. It is usually presumed that information that contains affective information is better retained by the memory. This makes sense because emotion was (and probably still is) the number one orientation for behaviour in humans. I would expect that the habit-like memory structure of the implementation intention is subject to the same effect. Therefore it would be interesting to have a repeated measure of the dependent variable.

Method: Experiment 2 uses the same procedures and materials as experiment 1 with the exception that the association task is not followed by the habit-strength measurement, but a filler task and a second association task are added into the experimental design. There could also be multiple association tasks. This could easily be realised if the filler tasks would be unrelated different “real” experiments.

I hypothesize that the difference in reaction time between the “normal” implementation intentions and the “affective enhanced” implementation intentions (for low-habit participants) will grow over time. The working mechanism of affect in habit might (simply) be a memory effect.

Experiment 3. Breaking the habit: Negative affect and implementation intentions

Förster, Liberman and Friedman (2007) proposed that a goal pursuit system is capable of structuring and restructuring the associative network in line with the actor’s goals and the situational affordances. By introducing a third condition (bike-push-condition) during the AAT, the effect of negative affect on implementation intentions can be studied.

Method: Experiment 3 uses the same paradigm as experiment 1 (experiment 2 can also be incorporated without problems), with the exception of the extra condition. Participants in the push-bike condition have to push 90% pf the bike pictures away from themselves. The images of the other modes of transportation receive an equal 50% push/pulling rate. I hypothesize that participants in the push-bike condition will show slower reaction times on the association task if they did not have the habit in advance. It would be remarkable if negative affect would also have an effect on reaction times on participants who already have a strong habit link. Yet the effect might be enhanced by multiple training session with the AAT.

2c. Societal Relevance

The relevance of this study can be explained best by the words of the famous William James:

“Ninety-nine hundredths or, possibly, nine hundred and ninety-nine thousandths of our activity is purely automatic and habitual, from our rising in the morning to our lying down each night.”.

Only in the last few years has behavioural sciences acknowledged that human behaviour is shaped by automatic tendencies in many and complex ways. However these automatic behaviour tendencies do not always elicit the behaviour we (consciously) want, but can lead to many problems and illnesses. Most problems western society faces today, such as drug abuse or eating disorders, can be explained in terms of addiction, which is just a name for a dysfunctional habit that displays certain properties. Being able to change these dysfunctional habits would benefit not only severe cases, but virtually everyone who wants to get rid of an annoying habit.

2d. Literature references

Aarts, H., Custers, R., Holland, R.W. (2007). The nonconscious cessation of goal pursuit: when goals and negative affect are coactivated. Journal of Personality and Social Psychology, 92, 165-178.

Aarts, H., Dijksterhuis, A., (2000). Habit as Knowledge Structures: Automaticity in Goal-Directed Behaviour. Journal of Personality and Social Psychology, 78, 53-63.

Aarts, H., Gollwitzer, P. M., & Hassin, R. R. (2004). Goal contagion: Percieving is for pursuing. Journal of Personality and Social Psychology, 92, 631-642.

Cuusters, R., Aarts, H., (2005). Positive Affect as Implicit Motivator: On the Nonconscious Operation of Behavioral Goals. Journal of Personality and Social Psychology, 89, 129-142.

Förster, J., Liberman, N., Friedman, R. S. (2007). Seven Principles of Goal Activation: A Systematic Approach to Distinguishing Goal Priming From Priming of Non-Goal Constructs. Personality and Social Psychology Review, 11, 211-233.

Holland, R. W., Aarts, H., & Langendam, D. (2006). On the power of implementation intentions: breaking and creating habits on the working floor. Journal of Experimental Social Psychology, 42, 776-783.

Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision und risk. Econometrica, 47, 263-291.

Sheeran, P., Aarts, H., Custers, R., Rivis, A., Webb, T., Cooke, R. (2005). The goal-dependent automaticity of drinking habits. British Journal of Social Psychology, 44, 47-63.