Slave of our Behaviour: The Influence of Strategy and Intrinsic
Habit Tendency on Changing Unwanted Habits
Master thesis
by
Sarah Knot
Name: Sarah Knot
St. number: 6072496/10003058
Supervisor: Dr. S. de Wit
Journal: Psychology & Health Words abstract: 245
Words total: 8546
2
Slave of our Behaviour: The Influence of Strategy and Intrinsic Habit
Tendency on Changing Unwanted Habits
Despite the intention and effort to break unwanted habits, many people fail in changing these behaviours. In this study we have investigated the effect of implementation and goal intentions and intrinsic habit tendency, the dominant use of the habitual control system, on changing unwanted habits. Using a within-subject design, participants (N=19) kept a sport and snacking diary for five weeks. Based on their diary entries during the first week, participants were coached to form two types of intentions: an implementation intention and a goal intention. One was formulated with regard to increasing healthy snacking and the other with regard to increasing sport behaviour. Intention types were counterbalanced across participants. Participants carried out the intended behaviour for four weeks while continuing the snack and sport diary. Intrinsic habit tendency was measured with the Slips of Action Task (de Wit et al., 2012). In contrast to our expectations, the preliminary analyses of the first 19 participants indicated that goal intentions and implementation intentions were equally effective in increasing healthy snacking and exercise behaviour. Furthermore, habit tendency did not significantly predict the increase in healthy snacking and exercise behaviour. Nevertheless, the directions of the correlations were as expected: task performance correlated positively with the increase in healthy snacking and exercise behaviour caused by goal intentions and negatively with the increase caused by implementation intentions. These findings suggest that the effectiveness of implementation intentions might be moderated by habit tendency and other potentially important factors, like daily life structure.
Keywords: health, eating, habit, implementation intention, goal intention, intrinsic habit tendency
Unhealthy snacking behaviour and lack of exercise can cause serious health problems, such as obesity, cardiovascular diseases and in some cases even premature death (van Gaal, Mertens & de Block, 2006). Even though people are aware of these negative consequences, and
3 despite the explicit intention and effort to get rid of them, many people fail in breaking these bad habits.
The inability to break unwanted habits is often explained by dual-system theories, which state that behaviour can be controlled by two different systems; the goal-directed system or the habitual system. Goal-directed control results in behaviour that is driven by the belief of the agent that this behaviour will achieve a desired goal or is able to prevent an undesired outcome (de Wit & Dickinson, 2009). If the same behaviour is performed
frequently because it successfully attained one’s goal, a strong association develops between the stimulus (e.g., entering the canteen during lunchtime) and the behavioural response (e.g., eating a sandwich; Thorndike, 1911). As result, a slow transition occurs in which the habitual system takes over control (de Wit & Dickinson, 2009; Rhodes, de Bruijn & Matheson, 2010). Consequently, the presence of the stimulus (e.g., entering the canteen during lunchtime) will automatically activate the behavioural response (e.g., eating a sandwich; Thorndike, 1911). Whereas the response was originally a deliberate choice to attain a specific goal, it has become an automatic response to a stimulus, which could be executed fast and without extensive cognitive effort (Aarts, Custers & Marien, 2008; Danner, Aarts & De Vries, 2008).
Although the habitual control system is beneficial when behaviour is performed regularly, it can hinder behaviour when someone’s intentions change. For example, during lunch time, someone might find herself eating a big sandwich even though she planned to eat a low-fat meal because of the desire to lose weight. Thus, in order to achieve her goal of losing weight, the habitual response of eating a sandwich needs to be adapted. Unfortunately, even when someone’s intentions have changed, the habitual stimulus-response association might remain intact, regardless of the fact that the response will no longer result in a desirable outcome (Aarts & Dijksterhuis, 2000b).
4 Despite the persistency of habits, it remains possible to change these automatic
responses. In order to break unwanted habits, motivation to change these behaviours is necessary. However, even though motivation is essential, it is insufficient to change habitual behaviour on its own (Hall, Fong, Epp & Elias, 2007; Verplanken & Wood, 2006; Web & Sheeran, 2006). The successful achievement of breaking a habitual response depends also on how an intentions to change the behaviour is specified (Gollwitzer, 1999). Two different kind of intentions are often used. One of these intentions are goal intentions, which specify the goal that the person aims to achieve. For example, ‘I am going to eat two pieces of fruit every day’. However, since habitual behaviour is more or less insensitive to the direct outcome of a response, it is likely that goal intentions are limited in their effectiveness. Namely, changing the outcome will often not immediately affect the stimulus-response association. Furthermore, goal-intentions might fail to change undesired habits because they depend on goal-directed control in which cognitive effort such as memorizing, attending, and self-control is needed to guide behaviour. Because cognitive capacity is limited, dependence on goal-directed control could hinder the execution of the intended behaviour (Gollwitzer & Sheeran, 2006). For example, if attention is focused on something else, someone could easily forget to perform the intended behaviour that is needed in order to attain the goal.
A more promising strategy to change habits is the use of implementation intentions (Gollwitzer & Sheeran, 2006), which concretize how the intended goal will be achieved. Implementation intentions specify a cue that will be associated to a response by using an ‘if-then plan’ (Gollwitzer, 1999). For instance, ‘If I am going to eat lunch, ‘if-then I will eat an apple’. The effectiveness of implementation intentions is considered to result from the reduced need for goal-directed control. Namely, when and what behaviour needs to be performed is determined in advance. Furthermore, it has been proposed that implementation
5 intentions affect behaviour via the same mechanism that underlies habits (i.e.,
stimulus-response association; Gollwitzer, 1999). Hence, in essence implementation intentions might generate a healthy ‘instant habit’ and share the same benefits as habits, namely, robust and automatic behaviour.
The effectiveness of implementation intentions in achieving behavioural changes has already been confirmed by various studies. Implementation intentions have been found to successfully reduce undesired health habits, such as unhealthy eating (Sheeran, Milne, Webb & Gollwitzer, 2005) and lack of exercise (Rise, Thompson & Verplanken, 2003).
Nevertheless, some studies have not found implementation intentions to be effective (de Vet, Oenema, Sheeran & Brug, 2009; Jackson et al., 2005; Verhoeven, Adriaanse, de Vet, Fennis & de Ridder, 2014). Still, it is unlikely that these negative findings are able to discard all promising literature about the effectiveness of implementation intentions and its strong
theoretical foundation. It is more likely that we still lack complete understanding of habits and how implementation intentions address their underlying mechanism.
One factor that may contribute to the aforementioned heterogeneity in findings
regarding implementation intentions, is the individual variability in intrinsic habit tendency of people. Normally, a balance is considered to exist in healthy people between goal-directed and habitual control systems (de Wit, Watson, Harsay, Cohen, van de Vijver & Ridderinkhof, 2012). If this balance is disrupted, behaviour could become dysfunctional, such as the
development of obesity (Volkow & Wise, 2005) and obsessive-compulsive disorder (OCD; Gillan et al., 2011). In these dysfunctional behaviours, people tend to rely more on the habitual control system than the goal-directed system (Robbins, Gillan, Smith, de Wit & Ersche, 2012).
6 In line with the proposed disbalance of control systems in obesity and OCD, it is plausible that even in healthy people one of these systems is slightly more active than the other. Consequently, people who have a stronger habit tendency might be more vulnerable to develop unwanted habits (de Wit et al., 2012). On the other hand, because implementation intentions are considered to affect behaviour via the same mechanism that underlies habits, people with stronger habit tendencies might particularly benefit from the way in which implementation intentions are formulated. Namely, implementation intentions provide the fundaments (i.e., cue and response) to easily develop new behaviour that is controlled by the habitual system. These expectations are in line with the suggestions of Gollwitzer (2006) who proposed that especially people who have problems with regulating their behaviour might benefit from implementation intentions.
In the present study, we examined how the ability to change one’s snacking and exercise habits is influenced by (i) strategy to change unwanted habits and (ii) intrinsic habit tendency. Based on one-week snacking and sport diaries, participants were coached to change their unhealthy habits by formulating an implementation intention and a goal intention. One intention was formulated with regard to increasing healthy snacking and the other with regard to increasing sport behaviour. During the subsequent four weeks, participants continued to register their snacking and sport behaviour in a diary. We expected implementation intentions to be more effective in changing unwanted snacking and sport habits than goal intentions. Furthermore, we expected that the improvement of unwanted snacking and sport habits, would be reflected by an increased habit strength of the healthy behaviour caused by the intentions. Finally, we expected implementation intentions to be particularly effective in people with a strong intrinsic habit tendency as assessed with the Slips-of-Action task.
7
Method
Participants
Thirty-nine participants with a mean age of 23.50 (SD=2.99) in the range of 18 to 34 years were recruited via posters at study and sport locations (e.g., universities, sport facilities) and via advertisements in student journals and newsletters of student sport facilities. Participation was on voluntary basis, no monetary compensation was given. Participants who met the following criteria were included: motivated to increase sport, healthy snacking and reduce unhealthy snacking behaviour, female, between 18 and 35 years, a normal to high BMI (20-35), no eating disorders or a history of eating disorders, maximal four days absence from home during the study. Although the World Health Organization (WHO) classifies BMI’s from 18.50 as normal (World Health Organization, 2015), we decided to use a lower bound of 20 to prevent potential underweight of participants after taking part in this study. Participants who consumed unhealthy snacks less than four days during baseline or exercised three or more times during the baseline, were excluded from further analysis due to a lack of improvement opportunities.
Materials
Entry questionnaire
The Entry Questionnaire consisted of 15 questions about demographic variables, age, height, and education. Furthermore, it verified whether participants had a history of eating disorders, obsessive behaviour related to food or weight, food restrictions and sport injuries, weight fluctuations of more than five kilos in the past six months and current sport behaviour. Having an eating disorder or a history of eating disorders functioned as an exclusion criterion. Other questions were used for explorative analyses. Furthermore, the Entry Questionnaire examined
8 participants’ motivation. On a 7-point Likert scale from ‘Not motivated’ to ‘Highly
motivated’ participants indicated whether they were motivated to reduce unhealthy snacking, increase healthy snacking and increase sport behaviour. At the end of the questionnaire, participants indicated what personal goals they would like to achieve by taking part in this study.
Slips-of-Action task (SOAT)
The SOAT is a computer task which aims to measure the balance between goal-directed and habitual control (de Wit et al., 2012). The task consists of two phases, as illustrated in Figure 1.
Figure 1. Overview of the SOAT. (a), illustrates the learning stage. The closed box with
grapes signal to press the right key in order to receive an orange, which is shown inside the box, and credits. No reward is provided by pressing the left key. (b), illustrates the test phase. In the instruction screen, the pineapple and peer are depicted with a red cross superimposed, indicating that making the previously learned responses to the stimuli of these outcomes will lead to subtraction of credits. Making correct responses to still-valuable stimuli (e.g., grapes) will still lead to credits. After the instruction screen, all stimuli are again presented in
9 Learning phase: participants learned, by trial and error, the correct associations between six
stimuli and their correct response response association) and outcome (stimulus-outcome association). The stimuli consisted of closed boxes with fruit, the correct response was the left or right key (i.e., Z and M), and the outcomes were open boxes with fruit. If the correct response was given, the box opened and another fruit was shown inside the box while a counter indicated that participants were rewarded with credits. If an incorrect response was given, the box remained closed and no credits were given. Outcomes were accompanied with sounds indicating a correct or incorrect response. For example, when a box with a banana was shown, participants had to press the left key. As a result, the box opened and a pineapple was shown while the counter indicated that the participant was rewarded with credits. Participants were instructed to make fast and accurate responses. The learning phase consisted of eight blocks with each 12 trials. Each stimulus was presented twice during a block. Stimuli and outcomes pairs were counterbalanced across participants by using six different versions. Stimuli were presented until a response was made. Fast correct responses were rewarded with more credits than slower correct responses, with a maximum of five credits per time. The outcomes were presented for 1000 milliseconds with an inter trial interval of 1500
milliseconds.
Test phase: all six outcomes from the learning phase were presented simultaneously for 10
seconds. Two of these outcomes were devalued, indicated by a red cross superimposed on stimuli. Making a response to stimuli of devalued outcomes resulted in subtraction of credits. Participants were instructed to memorize which outcomes were devalued, so that they could subsequently continue to make the previously learned responses to the stimuli that signalled still-valuable outcomes while refraining from making a response when a stimulus was shown that signalled one of the devalued outcomes. After the six outcomes had disappeared, each of
10 the six stimuli was presented in succession for 1000 milliseconds or until a response was made. The inter trial interval was 1500 milliseconds. During a block each stimulus was
presented twice. In contrast to the learning phase, after responses no outcome or feedback was provided. After each block of 12 trials, all six outcomes were shown again and two different outcomes were devalued. In total, each outcome was devalued for three times in a random order. The devaluation phase consisted of nine blocks in total.
After completing the SOAT, participants filled in a questionnaire to verify whether they had successfully learned the associations between stimuli, responses and outcomes.
In the performance on the SOAT, goal-directed control tendency was reflected by successfully refraining from making responses to stimuli of devalued outcomes. Habitual control tendency was reflected by a difficulty to refrain from making a response to stimuli of devalued outcomes due to their more or less insensitivity to changes in outcome value. SOAT scores were calculated by subtracting the responses on the devalued trials from the responses on the valued trials. This resulted in a total score between 0 – 100. Total scores below zero were excluded since they indicated that participants had not successfully learned the S-O-association or had misunderstood the instructions. A low total score indicated a stronger habitual tendency whereas a higher total score indicated a stronger goal-directed tendency.
Self-Report Habit Index (SRHI)
The SRHI aims to measure the strength of a habit by addressing three subscales: the frequency of the behaviour, the automaticity of the behaviour and to what extent the behaviour is part of one’s identity (Verplanken & Orbell, 2003). Participants were asked to rate 12 items on 7-point scale from 1 (totally disagree) to 7 (totally agree). An example that addresses the subscale ‘frequency’ is: ‘Behaviour X is something I do frequently’, an example
11 that addresses the subscale ‘automaticity’ is: ‘Behaviour X is something I start doing before I realise I’m doing it’, and an example that addresses the subscale ‘identity’ is: ‘Behaviour X is something I feel weird if I don’t do it’. Participants replaced ‘Behaviour X’ with the
behaviour of their implementation intention or goal intention. A high score on the SRHI indicates a strong habit whereas a low score indicates a weaker habit. The SRHI is considered to be valid in measuring the strength of habits (Cronbach’s α = .95) and has a high test-retest reliability (r= 0.90; Verplanken & Orbell, 2003).
Snack and exercise diary
The first seven days of the diary were used as a baseline measure. Via a web page, which was send by text message, participants had to fill in each snack occasion, what they consumed, how many, when and where it was consumed. For sport behaviour, participants had to fill in every sport activity that took longer than 20 minutes, the sort activity, the duration, where and when it took place.
From day eight, participants had to fill in the same information about snacking and sporting, however, from that moment participants did not have to indicate where and when the
behaviour took place but instead whether it was in line with their formulated intentions.
Implementation and goal intervention
The baseline diary functioned as a guideline to formulate a suitable implementation intention and goal intention. Participants were allocated in alternating order to the II-snack condition, in which an implementation intention was formulated for snacking behaviour and a
goal-intention for sport behaviour, or to the II-exercise condition, in which an implementation intention was formulated for sport behaviour and a goal-intention for snacking behaviour.
12 First, participants were coached by the experimenter to form a goal intention, for example: ‘My goal is to exercise three times a week’/ ‘My goal is to eat two pieces of fruit every day’. To make the goal easily accessible, participants were asked to write the intention on paper three times and say the intention aloud five times (e.g., ‘My goal is to exercise three times a week’). After the goal intention, the experimenter directed the attention to the other part of the diary (i.e., sport or snacking behaviour). Here the experimenter and participant identified the cue that caused daily regularities in unhealthy snacking or weekly sport occasions that did not occur even though the participants had planned to go. To make the intentions comparable among participants, only situational cues were used (e.g., ‘watching television’) and no internal cues (e.g., ‘feeling bored’) or time cues (e.g., ‘four o’clock’). Subsequently, healthy behaviour that could replace the unwanted habitual response was discussed. For example, if a participant consumed unhealthy snacks every evening when watching television, the
implementation intention could be ‘If I am watching television in the evening, then I will eat one piece of fruit’. After formulating the implementation intention the participant wrote the intention on paper three times, and said the intention aloud five times.
No negative intentions were formulated in which unhealthy behaviour was decreased (e.g., ‘My goal is to snack half a bag chips (instead of the whole bag)’). After formulating each intention, participants were motivated to carry out the intended behaviour by briefly discussing the benefits that would follow from these behaviours.
Exit questionnaire
In the Exit Questionnaire participants were asked how they had experienced taking part in the study, whether they succeeded in correctly filling in the diary, their motivation and whether the text messages where helpful.
13
Procedure
Via a telephone call, participants were asked whether they satisfied the inclusion criteria (see sample characteristics). The study had a length of five weeks in which three individual contact moments took place at one of the locations of the University Sports Centre (USC) of the University of Amsterdam. An overview of the five weeks is displayed in Figure2.
Day 1 (appointment 1):
Participants signed an informed consent and completed the entry questionnaire. Then, the waist and the weight of participants were measured. Furthermore, participants were informed about the procedure of filling in a daily diary about their snacking and sport behaviour.
Day 2-8 (baseline):
During these days participants received a text message twice a day (morning and evening) and an e-mail once a day with an address to a web page to fill in their snacking and sport behaviour.
Day 9 (appointment 2):
Participants performed the SOAT. Then participants were coached to form a goal intention and implementation intention with regard to sport and snacking behaviour. After forming the intentions, participants filled in the SRHI two times, one for the sport intention and one time for the snacking intention.
Day 10-37 (test phase):
During these days, participants received a text message twice a day with an address to a web page to fill in their snacking and sport behaviour. On day 16, 23, 30, 37 participants received two extra text messages with an address to a web page to rehearse their intentions.
Day 38 (appointment 3):
Participants were asked to rehearse their intentions and filled in the SRHI twice, one with regard to their sport intention and one with regard to their snacking intention. Subsequently,
14 participants filled in the exit questionnaire. Additionally, the waist and the weight of
participants were measured. Finally, participants were debriefed1.
Figure 2. Overview of the five week procedure
Results
From the 39 participants who had a first appointment, 33 participants actually started in the study. From the 33 participants, 19 participants completed the study whereas 14 participants (42 %) stopped before the end of the study. An overview of dropouts and reported reasons for dropout is provided in Appendix 1. Descriptive statistics identified one extreme score in the percentage difference in sport behaviour between baseline and test phase of the
II-intervention. Since this outlier had a distance of 2.80 standard deviations from the mean, the data of this participant are not used for further analysis. This resulted in a total sample of 18 participants, from which seven participants were assigned to the II-exercise condition and 11 to the II-snack condition. All participants were either undergraduates or had an academic degree. Participants had a mean BMI of 24.96 (SD=3.28). On average, participants indicated
1 In this study, participants performed as well a Stop Signal Task and completed the Barrett’s Impulsiveness
Scale, Dutch Eating Behaviour Questionnaire and General Self-Efficacy Scale. The results of these measures go beyond the scope of this thesis and are therefore not displayed nor analysed.
15 to be highly motivated to change their behaviour (M= 5.76, SD= 0.44). A repeated measures ANOVA with one within-subject variable Category (sport, healthy snacks, unhealthy snacks) revealed no significant difference in the mean motivation between increasing sport (M= 5.67,
SD= 0.59), increasing healthy snacking (M= 5.83, SD= 0.71), and reducing unhealthy
snacking (M= 5.78, SD= 0.55). Furthermore, three independent t-testswith between variable Condition (II-exercise versus II-snack) on the dependent variables age, BMI and mean motivation showed a successful randomized allocation of participants. As displayed in Table 1, no difference in mean age, BMI, and motivation was found between the II-exercise
condition and II-snack condition.To verify whether ‘occasions’ were a good reflection of the total amount of sport activity, we calculated the Pearson correlation between sport occasions and sport minutes. It showed that occasions and sport minutes were correlated significantly,
r= 0.82, p < 0.001.
Table 1.
Means and Standard Deviations for Age, BMI and Motivation for the II-snack Condition and II-exercise Condition
II-snack condition II-sport condition
M SD M SD t-test
Age 23.55 3.05 23.71 2.81 t(16) = -0.12, p = 0.91
BMI 25.23 2.65 24.89 3.93 t(16 )= 0.22, p = 0.83
Motivation 5.67 0.45 5.90 0.42 t(16) = -1.13, p = 0.28
II/GI manipulation check
To verify whether the manipulation was successful, the rehearsed intentions during the test phase were compared to the initially formed intentions on appointment 2. From the 18
16 (SD=0.48) out of eight times their intention. From these rehearsed intentions, intentions of eight participants were equal to the initially formed intentions. However, six participants had slightly modified their intentions. One participant had changed the cue of the implementation intention towards an internal cue. A few modifications were made in the formulation of the goal intentions, with three participants making their goal intention more specific. One
participant had in fact made her goal intention less specific. Another participant had made her implementation intention less specific. Taken together, these observations warrant caution in interpreting the effectiveness of implementation versus goal intentions. This issue will therefore be re-visited in the Discussion section.
Main analyses
Effect of intervention on snack and exercise behaviours
In Figure 3, the mean healthy and unhealthy daily snack occasions are displayed during baseline and after the intervention. A large reduction in unhealthy snack occasions can be observed, while the increase in healthy snack occasions is minimal.
Table 2.
Mean Healthy, Unhealthy Snacking, Sport Occasions and Minutes and Standard Deviations during Baseline and after the Intervention
Baseline After intervention
M SD M SD
Healthy snacking 0.96 086 1.06 0.73 Unhealthy snacking 1.78 0.75 0.83 0.53 Sport occasions 2.17 1.10 2.40 1.22 Sport minutes 146.11 99.06 147.64 86.35
17 Figure 3. Mean healthy and unhealthy daily snack occasions and weekly sport occasions
before and after the intervention
These observations were confirmed with a mixed ANOVA with the within-subject variables Time (baseline versus afterwards) and Snacks (healthy versus unhealthy), and one between-subjects variable Condition (II-snack versus II-exercise) on the amount of snack occasions, see for mean scores and standard deviations Table 2. A significant interaction effect of Time and Snack, F(1,16)= 27.77, p< 0.001, ƞ2= 0.63 was further investigated with separate post hoc-analyses, adjusted with a Bonferonni correction. While there was a highly significant decrease in unhealthy snack occasions after the intervention compared to before, D(0.78), p= 0.007, ƞ2= 0.37, there was surprisingly only a marginally significant increase in healthy snack occasions after the intervention, D(0.29), p= 0.07, ƞ2= 0.20. To verify whether a larger
increase in healthy snacking was associated to a larger decrease in unhealthy snacking, we calculated the Pearson correlation between the difference score between baseline and after the intervention for healthy snacking and unhealthy snacking. In contrast to our expectation that people would reduce their unhealthy snacking because of an increase in healthy snacking, a
0 0,5 1 1,5 2 2,5 3
Baseline After intervention
O
cca
sio
ns
18 larger increase in healthy snacking did not significantly correlate with a larger decrease in unhealthy snacking, r= 0.13, p = 0.60.
Importantly, in contrast to our expectation that implementation intentions would be more effective than goal intentions in achieving behavioural change, the main analysis did not reveal a significant interaction between Time and Condition, F(1,16)= 1.77, p= 0.68, ƞ2= 0.01 nor a significant three-way interaction between Time, Snack and Condition, F(1,16)= 0.20, p= 0.67, ƞ2= 0.01. This indicated that implementation intentions and goal intentions were
equally effective in changing healthy and unhealthy snack occasions.
In Figure 3, the mean weekly sport occasions are displayed as well during baseline and after the interventions. It shows almost no increase in sport occasions. To test whether there was a significant increase in sport occasions, a mixed ANOVA was performed with one within-subject variable Time (baseline versus afterwards) and one between-subjects variable Condition (II-snack versus II-exercise) on the amount of sport occasions. In contrast to our expectations that exercise behaviour would increase, especially due to the implementation intentions, the analysis did not yield a significant main effect of Time, indicating that there was no significant increase in sport occasions between the baseline and after the intervention
F(1,16)= 0.96, p= 0.34, ƞ2
= 0.06. It also yielded no interaction between Time and Condition,
F(1,16)= 1.35, p= 0.26, ƞ2= 0.08, indicating no difference between GI-interventions and II-interventions in increasing sport behaviour.
Comparing implementation and goal intervention
To further investigate the relative effectiveness of implementation and goal intentions in forming healthy habits, we performed a combined analysis of the percentage daily increase in healthy snacking and exercise behaviour. Percentage increase was used in order to control for
19 differences in general occurrence of sport and snacking behaviour (e.g., weekly versus daily). A mixed ANOVA was performed with one within-subject variable Strategy (II-intervention versus GI-intervention) and one between-subjects variable Condition (snack versus II-exercise) on percentage increase.
Figure 4. Percentage increase in healthy snack (dark grey bars) and sport occasions (light grey
bars) for implementation intentions (left) and goal intentions (right) compared to baseline
Figure 4 displays these percentages separately for healthy snack occasions and sport occasions for both interventions. The effect of GI-interventions on healthy snacking seems much more apparent than the effect of II-interventions on healthy snacking while the effect of GI-interventions for sport seems also slightly larger than the effect caused by the
II-interventions. Nevertheless, in line with the previous analyses, it yielded that GI-interventions and II-interventions were equally effective in increasing healthy snacking and sport
behaviour. This was confirmed by the absence of a significant main effect of Strategy,
F(1,16)=0.66, p= 0.43, ƞ2= 0.04. Furthermore, the effect of both intentions on snacking seems larger than the effect on sport behaviour. However, this observation as well failed to reach
0% 20% 40% 60% 80% 100% 120%
Implementation intentions Goal intentions
Per cen ta ge i nc rea se Snack Sport
20 significance. In line with one of our previous findings, it did not yield a significant interaction between Strategy and Condition, indicating that both intentions had the same effect on healthy snacking and sport behaviour, F(1,16)=0.95, p= 0.34, ƞ2= 0.06.
Self-Reported Habit strength
To test whether the mean scores on the SRHI, as depicted in Table 3, significantly increased over time and differed between GI-intervention and II-intervention, a mixed ANOVA was performed with two within-subject variables Time (Baseline versus Afterwards) and Strategy intervention versus GI-intervention) and one between-subjects variable Condition (II-snack versus II-exercise) on the mean SRHI scores. To illustrate the overall results, the mean percentage of increase in SRHI scores (after the intervention compared to before) are depicted in Figure 5. In line with our expectations there was an increase in mean SRHI scores between the baseline and after the intervention, this was confirmed by significant main effect of Time,
F(1,16)= 5.96, p=0.027, ƞ2= 0.27. Although it appears that there was a larger increase in SRHI scores for sport behaviour than for snacking behaviour, this effect failed to reach
significance, indicated by a non-significant interaction between Condition and Time, F(1,16)= 0.48, p=0.50, ƞ2= 0.03. More importantly, as can be seen in Figure 5, there was a tendency towards higher increase in automaticity with implementation intentions relative to goal intentions. However, this observation was again not supported by statistical analyses since no significant interaction effect was found between Time and Strategy, F(1,16)= 0.84, p=0.37,
ƞ2
21 Table 3.
Mean SRHI Scores and Standard Deviations for II-intervention and GI-intervention during Baseline and after the Intervention
Figure 5. Percentage increase in mean SRHI scores for II-intervention (left) and
GI-intervention (right). Dark grey bars indicate an increase for snacking and light grey bars for sport occasions.
To further examine the relation between the percentage increase of the SRHI scores and the percentage increase in healthy snack and sport behaviour, we calculated the Pearson correlations between these variables. In line with our previous finding, we did not find a significant correlation between the percentage increase of the SRHI-II and the percentage increase of the II-intervention, r= -0.10, p= 0.70. Neither did we find a significant correlation
0% 10% 20% 30% 40% 50%
Implementation intentions Goal intentions
Pe rce nt age ch an ge in SRH I Snack Sport
Baseline After intervention
M SD M SD
II-intervention 3.41 0.25 4.28 0.25 GI-intervention 3.44 0.27 3.99 0.33
22 between the percentage increase of the SRHI-GI and the GI-intervention increase, r= 0.09, p= 0.73. The percentage increase is however not necessarily the behaviour that is directly
initiated by the intentions, it also includes the general improvement that is indirectly caused by the intentions. Therefore, we calculated also the Spearman’s Rho correlation between the percentage increase of the SRHI scores with the number of times that participants indicated that their snack and sport behaviour was in line with their implementation intention (M= 0.23,
SD=0.05) or goal-intention (M= 0.44, SD= 0.12). The results indicated no significant
Spearman’s Rho correlation between the percentage SRHI-II increase and the number of times participants indicated that their behaviour was in line with their implementation intention, rs= -0.36, p= 0.15. Neither did we find a significant correlation between the percentage SRHI-GI increase and the amount of times that participants indicated that their behaviour was in line with their goal intention, rs= 0.17, p= 0.49. This suggested that habit strength of the intended behaviour was not clearly related to the increase in healthy snaking and exercise caused by the interventions.
Individual differences in habit tendency
To examine whether the II-interventions were in particular effective among participants who had an intrinsic habit tendency the data of the SOAT were analysed. One participant was excluded because of a negative SOAT score, indicating that she did not understand the task. As a result, 17 participants were included in these analyses. As illustrated by Figure 6(a), on average participants responded correctly on 97 % (range: 83%-100%) of the trials at the end of the learning phase (i.e., block 8), indicating that all participants had successfully learned the stimulus-response associations.As illustrated by Figure 6(b), during the test phase, on average participants responded correctly on 76 % of the valued trials, whereas, participants
23 were able to successfully withhold their response on 67% of the devalued trials. However, on average participants failed to withhold their response on 33 % of the devalued trials. Thus, in 33 % of the devalued trials participants made previously learned responses despite the
instruction to refrain from making a response (i.e., ‘slips of actions’). Furthermore, the SOAT questionnaire indicated that on average participants learned 5.70 (SD= 0.59) of the six
stimulus-response associations, 5.30 (SD= 0.71) of the response-outcome associations and 5.18 (SD= 1.07) of stimulus-outcome associations. Therefore, we concluded that participants successfully learned the associations between stimuli, responses and outcomes.
Figure 6. Percentage correct on the SOAT during the blocks of the learning phase (6a);
Responses on valued and devalued trials during the devaluation phase (6b).
In the top and bottom panels of Figures 7, the correlation of the SOAT scores with the percentage increase in healthy snacking and exercise behaviour of the II-intervention and the GI-intervention are displayed, respectively. This figure suggests that, although minimally, the percentage increase of the II-intervention might be negatively correlated with the SOAT
50% 60% 70% 80% 90% 100% 1 2 3 4 5 6 7 8 Per cen ta ge co rr ect Blocks (a) 0% 20% 40% 60% 80% 100% Valued Devalued Re po ns es o n t ria ls (b)
24 scores whereas the percentage increase of the GI-intervention might be positively correlated with the SOAT scores.
Figure 7. Pearson correlations between SOAT scores and the percentage increase in healthy
snacking and exercise behaviour. Lower scores on the SOAT indicate a habit tendency while higher scores indicate a goal directed tendency. (7a) Pearson correlation (r= -0.22) between SOAT scores and percentage increase of II-intervention. (7b) Pearson correlation (r= 0.20) between SOAT scores and percentage increase of GI-intervention.
-100% 0% 100% 200% 300% 400% 0 10 20 30 40 50 60 70 80 90 II c ha ng e c om pa re d to ba se line SOAT scores
(a)
-50% 0% 50% 100% 150% 200% 0 10 20 30 40 50 60 70 80 90 GI c ha ng e co m pa re d t o ba se line SOAT scores(b)
25 To statistically test whether implementation intentions were in particular effective in participants with a high intrinsic habit tendency, we performed a multivariate regression analysis with as independent variable SOAT score and dependent variables percentage II-intervention and GI-II-intervention increase of healthy snacking and exercise. In contrast to our expectations, the results showed that the SOAT scores was no significant predictor of the percentage increase, F(2,14)= 0.92, p=0.41. SOAT scores did not significantly predict the percentage II-intervention increase, β= - 0.01, p= 0.39 nor did it significantly predict the percentage GI-intervention increase β= 0.04, p= 0.45. This finding suggests that
implementation intentions might not be more effective in people with a habit tendency than in people with a goal directed tendency.
Since percentage increase is not necessarily the behaviour that is directly initiated by the intentions, we calculated the correlation between the SOAT scores and the number of times that participants had indicated that their behaviour was in line with their intentions. In contrast to our expectations, the results showed again no significant Spearman’s Rho
correlation between the SOAT scores and the number of times participants indicated that their behaviour was in line with their implementation intention, rs= -0.06, p= 0.83. Neither did we find a significant correlation between the SOAT scores and the amount of times that
participants indicated that their behaviour was in line with their goal intention, rs= 0.12, p= 0.65.
Discussion
In this study we have examined the influence of strategy and intrinsic habit tendency on breaking unwanted snacking and sport habits. A large reduction in unhealthy snacking was
26 observed whereas healthy snacking did only marginally increase. Surprisingly, no increase in sport behaviour was observed. In contrast to our expectations, implementation intentions were equally effective as goal intentions in reducing unhealthy snacking and in increasing sport behaviour. Furthermore, in contrast to our expectations, implementation intentions were not more effective in people with a higher intrinsic habit tendency. Yet, there still seems to be an association, although minimally, between habit tendency and an increased improvement of implementation intentions, whereas a goal-directed tendency seems more closely related to the improvement caused by goal intentions. In line with our expectations, habit strength of the intended behaviour increased after the intentions. However, implementation intentions did not result in a higher habit strength than goal intentions nor in a strong relation between the increase of healthy snacking and sport behaviour and the habit strength of the intended behaviour.
As noted before, the research presented here is part of a larger project (Healthyways) that will continue until a minimum of 80 participants is tested. Since, this thesis only presents interim analyses, the results should be treated with caution because of the small sample size. Nevertheless, this study revealed some interesting patterns that deserve mention.
In contrast to the general line within the literature (Gollwitzer & Sheeran, 2006), we did not find implementation intentions to be more effective than goal intentions. Three reasons can be given to explain this remarkable finding. First, implementation intentions depend much more on an accurate formulation than goal intentions. In implementation intentions, the cue needs to be identified that elicits the automatic behaviour (Adriaanse, De Ridder & De Wit, 2009; Verhoeven, Adriaanse, Evers & De Ridder, 2012). Because people have poor insight in their own behaviour (Nisbett & Wilson, 1977), finding the right trigger of their behavioural response is in general complicated, especially because habitual responses
27 occur automatically (Gollwitzer & Moskowitz, 1996). Unfortunately, for implementation intentions to be effective, it is crucial to identify the cue that truly elicits the behaviour. Studies that used cues other than the cue that truly elicited the behaviour, did not find implementation intentions to be effective (Adriaanse et al., 2009). A restriction of our study might have elevated the misidentification of the right trigger. In this study, we only used situational cues and no time or motivational cues in formulating the ‘if’ of the ‘if-then-plan’. Especially in our sample, people might have relied more often on internal cues because they were mostly students or people who recently graduated and looking for a job. Therefore, their daily activities might have differed from day to day. Consequently, the lack of a recurrent daily structure might have added to the reduced effectiveness of implementation intentions. Namely, if the recurrent cue (e.g., coming home from work) is not encountered, participants are unable to perform their intended behaviour. Even though previous studies also used samples consisting of students, they often did not solely include situational cues which are based on recurrent structure (e.g., Achtziger, Gollwitzer & Sheeran, 2008). Moreover, a study that did use solely situational cues failed to find implementation intentions to be effective (Adriaanse et al., 2009). In sum, it is possible that we did not always identify the cue that truly elicited the response. Nevertheless, in situations in which participants came up with
motivational or time cues we coached participants to identify an external cue that was closely related or consistent with the occurrence of the motivational or time cue. In further research, it is recommended to focus on identifying the trigger that elicits the habitual response than to use only comparable situational cues. Nevertheless, it is important to consider to what extent people are able to identify internal cues. Since people will often be unaware of these triggers, the proposed internal cues could be biased by conscious appraisals of participants which not necessarily equals the real trigger. Furthermore, if situational cues are used, it is
28 recommended to measure the level of recurrent daily structure and examine whether the amount of daily structure functions as a moderator in the effectiveness of implementation intentions. To our best knowledge, there is yet no useful measure to examine the level of daily structure in someone’s life. Therefore, a new measure, for example a questionnaire, needs to be developed that examines the recurrence of daily structure (e.g., day-night rhythm, daily activities, strictness of consuming three main meals).
A second reason that could explain why implementation intentions were not more effective than goal intentions has to do with the within-subject design of this study. To our best knowledge, no previous study has examined the effect of implementation intentions and goal intentions by formulating an implementation intention and goal intention at the same time in one participant. A potential risk of using such a within-subject design is that
participants have insight in the two different formulations of the intentions. Consequently, the chance exists that participants have made both intentions comparable in their formulation. Especially, goal intentions could become more specific like the formulation of
implementation intentions. For example, the goal intention ‘My goal is to eat one piece of fruit daily’ may become ‘If I am home from work, then I will eat a piece of fruit’.Since participants did not consequently rehearse their intentions every week, we were unable to verify whether participants had significantly changed their intentions. Nevertheless, from participants who did rehearse their intentions, it becomes clear that people indeed change their intentions over time. Further research should examine the exact modification of the intentions to conclude whether a within-subject design is an adequate design for these research
questions.
A third reason for finding implementation intentions equally effective as goal
29 size of implementation intentions proposed by previous studies might be inflated because of inadequate control conditions in which goal intentions were unclearly specified (Adriaanse, et al., 2011). In some control conditions of these studies, participants did not specify an
intentions at all, but instead performed activities associated with the healthy behaviour. For instance, solely filling in questionnaires, receiving information about the healthy behaviour or listing their favourite healthy snacks. In this study we used a strict control condition in which a clearly specified goal intention was formulated. Therefore, it might be that specifying a concrete intention already causes an improvement irrespective of the kind intentions. This speculation is supported by the study of Verhoeven et al., (2014) which did not find implementation intentions to be more effective when using a strict control condition.
Some other characteristics of this study, which can also be observed in other studies that did not find implementation intentions to be more effective, might provide more insight in why no difference is found between the effectiveness of implementation intentions and goal intentions. Some previous studies that did not find an effect were long-duration studies (de Vries, Kremers, Smeets, Brug & Eijmael, 2008). It might be that the effect of implementation intentions is more pronounced shortly after the intention is specified. Nevertheless, other long-duration studies did find implementation intentions to be more effective than goal intentions (e.g., Luszczynska, & Haynes, 2009). Finally, another study that did not find an effect of implementation intentions focused solely on exercising behaviour (de Vet, et al., 2009). In line with our results, it might be that the effect of implementation intentions is smaller for exercising behaviour than for snacking behaviour. In sum, there seems a variety of factors that might have caused implementation intentions to be less effective.
Beside the absence of a difference in effectiveness between the interventions, the results yielded three other interesting findings that deserve mention. First, the results yielded,
30 although minimally, a potential trend in which intrinsic habit tendency seemed to be
negatively correlated to healthy snacking and sport increase of implementation intentions whereas a goal-directed tendency seemed positively correlated to the snacking and sport increase of goal intentions. Although we did not find intrinsic habit tendency to be a significant predictor of implementation intentions’ and goal intentions’ improvement, the direction of the correlation suggests that people who have an intrinsic habit tendency might benefit more from implementation intentions whereas people who have a goal-directed
tendency might benefit more from goal intentions. Moreover, it is important to emphasize that this study used a sample containing healthy, young adults. Therefore, the variance in control tendencies within this sample is minimal. Consequently, it might be that a sample, in which a larger variance of control tendencies between participants exists, would show a stronger association between tendencies of control systems and the improvement of goal- and
implementation intentions. Our findings and speculations are in line with suggestions from the literature, which state that people who experience difficulties in regulating their behaviour, might benefit even more from implementation intentions (Gollwitzer, 2006). Therefore, implementation intentions might provide promising results for clinical populations with a dominant use of the habitual control system, such as obsessive-compulsive disorders (Gillan et al., 2011). To our best knowledge, we are the first to cautiously relate intrinsic habit and goal-directed tendency to the effectiveness of implementation intentions and goal intentions.
A second interesting finding is that we did not find a difference in increase in habit strength of the intended behaviour between goal intentions and implementation intentions. If implementation intentions are effective because they generate a healthy habit, a strong correlation between the increase in healthy snacking and decrease in unhealthy snacking is expected. Namely, the cue that originally triggered the unhealthy snacking will now trigger
31 healthy snacking. This could be illustrated by the ‘horse race’ metaphor (Logan, 1988).
According to the ‘horse race’ metaphor, people should be able to react when a response is retrieved. Since implementation intentions provide a second response option, two responses will be retrieved resulting in a ‘horse race’. The winning response will determine the action. If implementation intentions generate an instant habit, it is expected that the association between the cue and the intended is behaviour is stronger than the association between the cue and the old behavioural response and therefore the intended behaviour will be executed. So, if
implementation intentions generate an instant habit, they will also cause a larger increase in habit strength than goal intentions. We did however not find a difference in habit strength between goal intentions and implementation intentions. Moreover, even though a previous study did show a clear association between larger increase in healthy snacking and a larger reduction in the consumption of unhealthy calories (Adriaanse et al., 2009), we did not find such correlation. In sum, our results do not provide support for the hypothesis that
implementation intentions generate an instant habit. According to Web, Sheeran and
Luszczynska (2009) the intended behaviour will not always be selected. They suggested that the strength of the association between the situation and the responses determines the winner. If a strong counter intentional habit exists, implementation intentions are less effective to break the old habit. As a result, the old habitual response will be selected. Since we did not measure the strength of the old habit, it remains unknown whether implementation intentions failed to break unwanted habits due to the strength of the unwanted habit. Another
explanation could be that the formulation of an intention caused a conflict that eliminates the cognitive advance of the old habit (Adriaanse, Gollwitzer, de Ridder, de Wit & Kroese, 2010). In line with this idea, it seems plausible that the activation of both responses, gave the
32 conflict of two activated responses might have caused the goal-directed control system to take over control and thereby provided the opportunity to choose the intended behaviour. This idea is in line with the suggestion that solely specifying a concrete goal intention might be effective as well, as long as it will result in a conflict between two responses. Further research could investigate this hypothesis by measuring the habit strength of the old habit and examine whether the switch in control systems is observable in the activity of the corticostriatal pathways which are found to be related to goal-directed and habitual performance. Namely, goal-directed control is reflected by a corticostriatal network of the ventromedial
prefrontalcortex and caudate while habitual control is reflected by a corticostriatal network of the premotor cortex and posterior putamen (de Wit et al., 2012). If implementation intentions create a window for goal-directed control, it is expected that implementation intentions disrupt the activity that relates to habitual control and establish activity that relates to goal directed control.
Another, theory about the effectiveness of implementation intentions is that they improve someone’s general planning behaviour (Luszczynska, Sobczyk & Abraham, 2007). It has been found that stimulating people to make concrete plans of how the intended behaviour could be achieved causes a larger weight reduction than when no such a facilitation occurs. In this study we used a within-subject design in which we coached all participants to make an implementation intention. Thus, if facilitating planning behaviour causes the effectiveness of implementation intentions, we would expect to find goal- and implementation intentions equally effective. Namely, an increase in planning generally would facilitate both intended behaviours. In line with this idea, we did find no difference between the effectiveness of goal intentions and implementation intentions.
33 A third interesting result is that intentions seem to have different effects on sport compared to snacking behaviour. In this study we found both intentions to be more effective on snacking behaviour than sport behaviour. Our finding is in line with previous findings that suggest that the effect of intentions is smaller with regard to exercise behaviour than to snacking behaviour (Luszczynska, & Haynes, 2009). One explanation for this dissimilar effect is the difference in frequency with which the habit is encountered. Snacking habits are often habits, which someone encounters more frequently than sport habits. Consequently, the intention to change the unwanted behaviour might remain more accessible in high frequently encountered habits. According to Gollwitzer (2006), the inability of remembering to act is an important problem in changing habits. If the intention to act is not activated, it becomes hardly impossible to perform the intended behaviour. Consequently, these findings are important in the generalization of the effectiveness of implementation intentions. Namely, it suggests that the mean effect size of these intentions could not be generalized to different habits but that these intentions could have smaller or larger effects depending on the kind of habit. It is therefore interesting to analyse both habits separately and examine whether
frequency of the intended behaviour moderates the decrease of unwanted habits. For example, for implementation intentions, this could be done by asking participants to indicate whether they have encountered their cue. From this, one can examine whether there is a difference between participants who have more frequently the possibility to perform their intended behaviour compared to participants who have less frequently the possibility to perform their intended behaviour.
Finally, it is important to mention the high percentage dropouts that concern long-duration studies like this one. This study had a long-duration of five weeks, for which no monetary compensation was given. Therefore, completion depended entirely on the amount of intrinsic
34 motivation of participants to change their snacking and exercise behaviour. Yet, since all participants needed to be highly motivated at the beginning of the study, it is interesting why some participants remained motivated whether others did not. It might be for instance that only those participants who experienced an improvement of the interventions remained highly motivated and completed the study. If that would be correct, then studies like these might slightly bias our conclusion about the effectiveness of these intentions by taking into account only the positive findings. It would therefore be recommended to further analyse factors that have led to reduced motivation.
To conclude, although the results are based on an interim analysis from an ongoing study, it revealed many interesting results. First it suggests that the superiority of the effectiveness of implementation intentions compared to goal intentions is not self-evident. Namely, the effectiveness of intentions might be influenced by whether someone has a tendency to use the habitual or goal-directed control system and external factors, such as the level of daily structure. Furthermore, the results did not provide direct support for the idea that implementation intentions generate instant healthy habits. Therefore, it is important to analyse the data of more participants and include experimental studies in laboratories to control for the diversity of factors that currently provide a broad range of potential explanations of the mechanism that underlies implementation intentions and habits. In sum, it is clear that we still lack complete understanding of the effectiveness of implementation intentions and which factors need to be present in order to make these intentions effective. Nevertheless, the interim analysis of this study has provided many interesting findings to further extend our knowledge of breaking unwanted habits.
35
References:
Aarts, H., Custers, R., & Marien, H. (2008). Preparing and motivating behavior outside of awareness. Science, 319, 1639. doi: 10.1126/science.1150432.
Aarts, H., & Dijksterhuis, A. (2000a). Habits as knowledge structures: Automaticity in goal-directed behavior. Journal of Personality and Social Psychology, 78, 53–63. doi: 10.1037/0022-3514.78.1.53.
Aarts, H., & Dijksterhuis, A. (2000b). On the automatic activation of goaldirected behavior: The case of travel habit. Journal of Environmental Psychology, 20, 75–82.
Achtziger, A., Gollwitzer, P. M., & Sheeran, P. (2008). Implementation intentions and shielding goal striving from unwanted thoughts and feelings. Personality and
Social Psychology Bulletin, 34, 381–393.
Adriaanse, M. A., de Ridder, D. T. D., & de Wit, J. B. F. (2009). Finding the critical cue: Implementation intentions to change one’s diet work best when tailored to personally relevant. Psychology & Health, 35, 60-71. DOI:
10.1177/0146167208325612
Adriaanse, M., Gollwitzer, P. M., de Ridder, D. T. D., de Wit, J. B. F., & Kroese, F. M. (2010). Breaking Habits With Implementation Intentions: A Test of Underlying Processes. Personality and Social Psychology Bulletin, 37, 502– 513.
Adriaanse, M. A., Vinkers, C. D. W., de Ridder, D. T. D., Hox, J. J., & de Wit, J. B. F. (2011). Do implementation intentions help to eat a healthy diet? A systematic review and meta-analysis of the empirical evidence. Appetite 56, 183–193.
Ajzen, I. (2001). Nature and operation of attitudes. Annual Review of Psychology, 52, 27–58. Allan, J. L., Johnston, M., & Campbell, N. (2008). Why do people fail to turn good intentions
36 eating intentions into action in young Scottish adults. BMC Public Health, 8,
doi:10.1186/1471-2458-8-123.
Bélanger-Gravel, A., Godin, G., & Amireault, S. (2013). A meta-analytic review of the effect of implementation intentions on physical activity. Health Psychology Review, 7, 23–54. doi:10.1080/17437199.2011.560095.
Congdon, E., Mumford, J. A., Cohen, J. R., Galvan, A., Canli, T., & Poldrack, R. A. (2012). Measurement and Reliability of Response Inhibition. Front Psychology, 3,
doi: 10.3389/fpsyg.2012.00037.
Danner, U. N., Aarts, H., & de Vries, N. K. (2008). Habit versus intention in the prediction of future behavior: The role of frequency, context stability and mental accessibility of past behavior. British Journal of Social Psychology, 47, 245–265. doi:
10.1348/014466607X230876.
de Vet, E., Oenema, A., Sheeran, P., & Brug, J. (2009). Should implementation intentions interventions be implemented in obesity prevention: The impact of if-then plans on daily physical activity in Dutch adults. International Journal of Behavioural
Nutrition and Physical Activity, 6, doi:10.1186/1479-5868-6-11.
de Vries, H., Kremers, S. P. J., Smeets, T., Brug, J., & Eijmael, K. (2008). The effectiveness of tailored feedback and action plans in an intervention addressing multiple health behaviors. American Journal of Health Promotion, 22, 417–425.
de Wit, S., Corlet, P. R., Aitken, M. R., Dickinson, A., & Fletcher, P. C. (2009). Differential engagement of ventromedial prefrontal cortex by goal-directed and habitual behavior toward food pictures in human. The Journal of Neuroscience, 29, 11330-11338.
37 de Wit, S., & Dickinson, A. (2009). Associative theories of goal-directed behaviour: a case
for animal–human translational models. Psychological Research, 73, 463–476. doi: 10.1007/s00426-009-0230-6.
de Wit, S., Watson, P., Harsay, H. A., Cohen, M. X., van de Vijver, I., & Ridderinkhof, K. R. (2012). Corticostriatal Connectivity Underlies Individual Differences in the
Balance between Habitual and Goal-Directed Action Control. The Journal of
Neuroscience, 32, 12066 –12075.
Gillan C. M., Papmeyer, M., Morein-Zamir, S., Sahakian, B. J., Fineberg, N. A., Robbins, T. W., & de Wit, S. (2011). Disruption in the balance between goaldirected behavior and habit learning in obsessive-compulsive disorder. American Journal of
Psychiatry, 168, 718 –726.
Glascher, J., Hampton, A.N., & O’Doherty, J.P. (2009). Determining a role for the ventromedial prefrontal cortex in encoding action-based value signals during reward related decision making. Cerebral Cortex, 19, 483-495.
Gollwitzer, P. M. (1999). Implementation intentions: Strong effects of simple plans. American
Psychologist, 54, 493-503.
Gollwitzer, P. M., & Moskowitz, G. B. (1996). Goal effects on action and cognition. In E. T. Higgins & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 361–399). New York: Guilford Press.
Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. Advances in Experimental Social
38 Hall, P. A., Fong, G. T., Epp, L. J., & Elias, L. J. (2007). Executive function moderates the
intention-behavior link for physical activity and dietary behavior. Psychology and
Health, iFirst:1-18.
Jackson, C., Lawton, R., Knapp, P., Raynor, D. K., Conner, M., Lowe, C., & Closs, S. J. (2005). Beyond intention: Do specific plans increase health behaviours in patients in primary care? A study of fruit and vegetable consumption. Social Science and
Medicine, 60, 2383–2391. doi:10.1016/j.socscimed.2004.10.014
Logan, G. D. (1988). Toward an instance theory of automatization. Psychological Review, 95, 492–527.
Luszczynska, A., & Haynes, C. (2009). Changing nutrition, physical activity and body weight among student nurses and midwives: effects of a planning intervention and
self-efficacy beliefs. Journal of Health Psychology, 14, 1075–1084. Luszczynska, A., Sobczyk, A., & Abraham, C. (2007). Planning to Lose Weight: Randomized
Controlled Trial of an Implementation Intention Prompt to Enhance Weight
Reduction Among Overweight and Obese Women. Health Psychology, 26, 507-512. Nisbett, R., & Wilson, T. (1977). Telling more than we can know: Verbal reports on mental
processes. Psychological Review, 84, 231–259. doi:10.1037/0033-295X.84.3.231. Patton, J. H., Stanford, M. S., Barratt, E. S. (1995). Factor structure of the Barratt
Impulsiveness scale. Journal of Clinical Psychology, 51, 768–764. Rana, D. K., & Rao, S. L. (2013). Development and Standardization of Tests for the
Measurement of Inhibition. Delhi Psychiatry Journal, 16, 90-97.
Rise, J., Thompson, M., & Verplanken, B. (2003). Measuring implementation intentions in the context of the theory of planned behavior. Scandinavian Journal of Psychology,
39 Rhodes, R., De Bruijn, G. J., & Matheson, D. H. (2010). Habit in the physical activity
domain: integration with intention temporal stability and action control. Journal of
Sport and Exercise Psychology, 32, 84e89.
Robbins, T. W., Gillan, C. M., Smith, D. G., de Wit, S., & Ersche, K. D.
(2012). Neurocognitive endophenotypes of impulsivity and compulsivity: towards dimensional psychiatry. Trends in Cognitive Sciences, 16 (1), 81-91. doi:
10.1016/j.tics.2011.11.009
Sheeran, P., Milne, S. E., Webb, T. L., & Gollwitzer, P. M. (2005). Implementation
intentions. In M. Conner & P. Norman (Eds.), Predicting health behavior (2nd ed., pp. 276–323). Buckingham, UK: Open University Press.
Thorndike, E. L. (1911). Animal intelligence: Experimental studies. New York: Macmillan. Tricomi, E., Balleine, B. W., & O’Doherty, J. P. (2009). A specific role for posterior
dorsolateral striatum in human habit learning. European Journal of Neuroscience,
27, 4019-4026.
van Gaal, L. F., Mertens, I. L., & de Block, C. E. (2006). Mechanisms linking obesity with cardiovascular disease. Nature, 444, 875-880, doi:10.1038/nature05487.
Verbruggen, F., & Logan, G.D. (2008). Response inhibition in the stop-signal paradigm.
Trends in Cognitive Sciences, 12, 418-424.
Verhoeven, A. A. C., Adriaanse, M. A., Evers, C., & de Ridder, D. T. D. (2012). The power of habits: Unhealthy snacking behaviour is primarily predicted by habit strength. British Journal of Health Psychology, 17, 758–770.
doi:0.1111/j.2044-8287.2012.02070.
Verhoeven, A. A. C., Adriaanse, M. A., de Vet, E., Fennis, B. M., & de Ridder, D. T. D (2014). Identifying the ‘if’ for ‘if-then’ plans: Combining implementation
40 intentions with cue-monitoring targeting unhealthy snacking behaviour. Psychology
and Health, 29, 1476–1492, http://dx.doi.org/10.1080/08870446.2014.950658.
Verplanken, B., & Orbell, S. (2003). Reflections on past behavior: A self-report index of habit strength. Journal of Applied Social Psychology, 33, 1313–1330. doi:10.1111/j.1559 1816.2003.tb01951.
Verplanken, B., & Wood, W. (2006). Interventions to break and create consumer habits.
Journal of Public Policy & Marketing, 25, 90–103. doi:10.1509/jppm.25.1.90
Volkow N. D, & Wise R. A. (2005). How can drug addiction help us understand obesity?
Nature Neuroscience, 8, 555–560.
Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions engender behavioral change? A meta-analysis of the experimental evidence. Psychological Bulletin, 132, 249–268.
Web, T. L., Sheeran, P., & Luszczynska, A. (2009). Planning to break unwanted habits: Habit strength moderates implementation intention effects on behaviour change. British
Journal of Social Psychology, 48, 507–523.
World Health Organization (2015). BMI classification. Retrieved 24-05-215 from http://apps.who.int/bmi/index.jsp?introPage=intro_3.html
Yin, H. H., Knowlton, B. J., & Balleine, B. W. (2004). Lesions of dorsolateral striatum preserve outcome expectancy but disrupt habit formation in instrumental learning.
41
