Influence of Affect on Attention for
Non-Emotional Stimuli
Tristan Hoobroeckx
10534040
Hans R. Phaf and Lotte Mulder Word Count: 4614
Abstract
There seems to be a relation between affect and attention for non-emotional stimuli. This study aims to find out what this relation is by comparing slopes of reaction times on a
conjunction visual search task in positive and negative affect conditions. The affect-manipulation has not been entirely successful, however. There was a pop-out effect, resulting in low increases of reaction time per added distractor. As such the effect sizes were very small. Trials where the target was present showed that the positive affect condition produces the steepest slope,
indicating that the temporal hypothesis holds true. However the trials were the target was absent showed that the negative affect condition produces the steepest slope, indicating that the spatial hypothesis holds true. No to little conclusions could be drawn from this data.
Influence of Affect on Attention for Non-Emotional Stimuli
Emotions can get in the way sometimes. Eye-witnesses to a crime often have trouble identifying a suspect when the perpetrator was carrying a weapon. This is called the weapon focus effect. When a weapon is drawn, a highly emotional situation occurs. People are overcome with fear and panic and it seems that they only pay attention to the weapon, because this is the immediate threat that should be avoided. This leads to trouble identifying the person who carried that weapon. The appearance of a person does not have any emotional aspects. Yet when people are in a certain affect, the attention for such features seems to change. The encounter with a weapon like a gun puts most people in a state of fear, which is a negative emotion. This negative emotion seems to have some influence on the attention for non-emotional stimuli, and positive affect could have a different influence.
emotions can broaden an individual’s momentary thought-action repertoire. This means that individuals with positive emotions have a broader mindset. In contrast, Frederickson states that people with negative emotions may have a narrower mindset. These mindsets do not project to the outside world, but rather as the name suggest on the mind. What is more interesting is that according to Frederickson switching between action tendencies is also faster, indicating that attention for non-emotional stimuli could be more flexible. What this means is that positive affect should make it easier to direct your attention elsewhere, be that from one task to another or between different stimuli. Isen (2000) and colleagues found that people who experience positive emotions have patterns of thought that are flexible, creative, and open to information.
Frederickson’s work condenses into two major hypotheses about affect and attention: The spatial attention hypothesis, and the temporal attention hypothesis.
Gasper and Clore (2002) hypothesized that happier moods promote a greater focus on global aspects of stimuli and sadder moods greater focus on local aspects of stimuli. They have found that when participants are subjected to an image-based task (images of forests) and positive or negative affect, the participants who had positive affect focused more on the forest than on the individual trees. This was found by having the participants draw the image from memory. Gasper and Clore’s hypothesis was based on the idea that affective cues in task
situations may be experienced as task-relevant information. This would then influence the spatial aspect of attention: global vs local attention. Their second experiment, which used global and local attributes of geometric figures, confirms their first findings. Row, Hirsh, and Anderson (2006) found something similar with their study. Their hypothesis was that positive affect serves to broaden the scope of attentional filters and therefor reduce their selectivity. Row, Hirsh, and Anderson took measures in two different domains: semantic search and visual attention. In short
they used a remote associates test (3 words are presented, participant must find a fourth related word) and an Eriksen Flanker task. In this task participants must tell whether a target ‘flanked’ by distractors was one of two similar types such as ‘N’ and ‘H’. The distractors or ‘flankers’ could be either congruent (HHHHH and NNNNN) or incongruent (HHNHH and NNHNN). They found that participants who experienced positive affect had better access to remote associates, suggesting a broadening of the scope of semantic access. Their though-action
repertoire has increased. This is a good example of how affect influences the internal world. The current study however tries to investigate temporal attention and spatial attention. Spatial
attention, as the name suggests, is aimed at the external, spatial world. That’s why it is more interesting that participants who experienced positive affect were slower in determining the target in the Eriksen Flanker task (which is very heavily aimed at the external world), suggesting that they were more distracted by the flankers. The scope of attention could be broadened
according to Row, Hirsh, and Anderson, allowing more distractors to come into attention. These distractors cause interference which slows down reaction times. However, Bruyneel et al(2011) were not able to reproduce these findings as across the three experiments they performed, positive affect consistently failed to have any impact on selective attention.
One would notice that these studies are more aimed at the spatial component of attention. With spatial tasks such as image-based tasks and especially the Eriksen Flanker task results are expected to be found in terms of spatial attention and not in temporal attention. In the Eriksen Flanker task the target is always at the same location, eliminating the opportunity to switch between targets. Without some form of switching between targets a temporal component of attention (i.e. flexibility) cannot be measured. MacLean, Arnell, and Busseri (2010) have used the attentional blink paradigm to show that positive affect leads to more diffuse temporal
attention instead of a broader spatial attention. Participants’ affect was measured with the Positive Affect Negative Affect Schedule (PANAS) after which the trials started. Participants were shown a streamed sequence of letters and were asked to identify the red letter in that stream. Participants became more accurate in discerning targets that occur within a short time interval when they experienced positive affect. In addition, Heerebout, Todorovic, Smedinga, and Phaf (2013) have found that when flashing the participants with happy faces they became better at switching from an automated response routine. They also switched more easily from a dominant global spatial bias to a more local spatial bias. When angy faces were presented the participants became less able to switch from an automated response routine and switched less easily from a dominant global spatial bias to a local bias. This is in contrast to Gasper and Clore’s aforementioned claim that positive affect elicits a more global attention bias and not local attention. Phaf (2015) sought to compare temporal switching and spatial broadening and has subsequently found a complete reversal of the effect of positive affect on attentions when a masked flanker task is used. Participants would be given a target stimulus which would be masked beforehand by flankers (distractors). The interval between target and distractors was varied. When the interval was very short, participants had more trouble identifying the target when they experienced negative affect versus positive affect. In sharp contrast, the opposite happened when intervals were longer. These results indicate that attention for non-emotional stimuli contains a temporal component. This is at odds with the spatial hypothesis, but do support the temporal hypothesis, where attention becomes more flexible with positive affect.
In previous studies tasks were used that favored one of two hypotheses. These studies made use of tasks that were not capable of indicating both spatial attention and temporal attention. So now comes the question of whether positive affect has a broadening effect on
attention or makes attention more flexible. The current study aims to make use of a task that is generally unbiased. Phaf’s study from 2015 used an altered version of the Eriksen Flanker task in order to measure attention flexibility. The spatial relation between distractors and target and target position were not varied, however. In this study we will be presenting the participants with a conjunctive visual search task (Kristjánsson, 2015). The benefit of such a task is that we can alter the spatial relation between elements in addition to it being an assortment of different stimuli between which the participant must switch attention. This is in essence a serial search, which can measure the flexibility of attention, because participants must switch between the individual elements of the series of stimuli. And yet, with the varying spatial relations between elements it is possible to test the spatial attention hypothesis too. Also the benefit of a
conjunction visual search task over a feature visual search task is that the latter gives lower reaction times and slopes that are around zero. The search for the target is in that case not affected by the number of distractors (Kristjansson , 2015). This may lead to a ceiling effect because of which there is no differentiating between conditions. Conjunction visual search makes the search for the target inefficient resulting in a response time increase of 10 to 50 milliseconds with each added distractor (Wolfe, 1998). The stimuli differed on orientation and form. This study opted for orientation instead of color to try to reduce pop-out effects.
Participants’ affect was manipulated by listening recorded music, and writing an autobiographic story. Reaction times on the conjunction visual search task were measured.
In this study slopes representing the relationship between the amount of distractors and reaction times are of interest (Figure 1). When participants are experiencing positive emotion and the spatial hypothesis is true, we should expect to see a flattening of the slope representing the relation between amount of distractors and reaction times (versus when participants are
experiencing negative emotions). In contrast, when participants are experiencing positive emotions, we should see a steepening of the slope (versus when participants are experiencing negative emotions) when the temporal hypothesis is true.
Figure 1: In this figure two examples of graphs we expect can be seen. The graph on the left represents what to expect when the spatial hypotheses is true and the graph on the right represents what to expect when the temporal hypotheses is true.
Methods
Participants
Forty-four (psychology) students (age 21.5, 34 female) from the University of
Amsterdam with normal or corrected vision participated voluntarily, for free, in exchange for course credit, or in exchange for a monetary reward, after providing informed consent. Participants were excluded if they were injured or otherwise impaired.
Design
The visual search task is a (2x4) within subjects design. Manipulation of affect (positive or negative) was chosen as the first independent variable. The order of this manipulation was counterbalanced. Every odd participant started with the negative affect manipulation. The second independent variable was array-size (otherwise called ‘set-size’). Array-sizes varied between 4, 10, 18, and 30 elements. While the amount of trials per condition remained equal, the order of the trials was randomized. The target was one of four different conjunctions between orientation (vertical or horizontal) and form (rectangle or oval) and was randomly assigned to the participant at the start of the experiment. The distractors differed from the target on one of the
aforementioned aspects but not both. The target remained the same throughout the experiment. As the dependent variable, reaction times to correct responses where the target was present were chosen.
Materials and apparatus
The stimuli were presented in black on an lcd computer monitor (Asus VG248QE 24" 1920x1080p 144Hz 1ms response time) against a gray background. Distance to the screen was by approximation 60 cm. Music was played on an over-ear headphone (MONACOR stageline md5000dr) from a Tom-tec mp4 player (MP1035 4GB) at a confortable sound volume. Lights were kept constant between participants by turning the lights all the way up. Responses were measured from two one-button response boxes appropriately labeled ‘left’ and ‘right’.
Depending on handedness, the dominant hand button represented that the target is ‘present’ and the other that the target is ‘absent’. Participants were told to hold both their hands on their respective response boxes at all times and use their index fingers for pressing.
To measure mood, SAM characters were used. Participants could choose in two rows each which character best represents their current emotional state.
Figure 2
Affect was measured happy to sad (1-5) and Arousal was measured excited to calm (1-5). Arousal measurements were not used or analyzed in this study. They were collected as an extra measure.
Distractors and target were one of the possible four combinations of form (rectangle, oval) and orientation (horizontal, vertical). Distractors differed on one of two aspects from the target but not both. All stimuli were printed in black. All stimuli had a length of 1°.Stimuli were randomly positioned in a framework consisting of three concentric circles which contained from smallest circle to largest, 6, 12, and 18 possible stimulus-positions. Four array sizes were
randomly used, consisting of 4, 10, 18, and 30 stimuli. Targets had a 50% chance of being present. Affect manipulation was executed by instructing the participant to write about a positive of negative personal event, after having read a corresponding positive or negative story. All while listening to positive or negative music they had selected beforehand. The music was looped using youtubeonrepeat.com. At the start of each manipulation attempt, participants were
asked to indicate their mood on a mood-scale. After the manipulation they were asked again to indicate their mood.
Figure 3
Procedure
Participants were first made at ease and were informed about the intentioned affect manipulation. After signing the informed consent, handedness was determined. Participants were instructed to place their hands on the respective response-boxes and that they should try to react as fast as possible by pressing the appropriate button. They were instructed to press with their dominant hand if the target was present and with their non-dominant hand if the target was absent. The participant was told that there would first be twenty practice trials to get used to the task at hand. The participant was informed that the experimenter would be present in the room, but that they could not see what was displayed on their screen. Participants were explicitly told that any texts they produced were strictly private and could be taken home afterwards. Such texts
would not be read by any experimenters. The participant was asked to try to get in the intended emotional state as best as they could. After the practice trials the participants indicated their current emotional state using the SAM characters for the first time. At the start of each
manipulation attempt (baseline), after said manipulation attempt (post induction), and after each trial block (post block), participants were asked to indicate their mood using SAM. This totals to six measurements. After the first mood measurement the first affect manipulation took place. Participants were asked to choose a musical fragment and read an example text (Appendix) and write about a memory of their own that represented the same emotion that they read in the example text. All example texts were written in Dutch. The music was played on a loop during the rest of this condition. Participants began with the first set of trials. After, they indicated their current emotional state for the third time using the same SAM characters. The second affect manipulation took place. Participants again chose a musical fragment and wrote and read a story. They then began their second set of trials. After the second set of trials has been finished, they indicated their emotional state for a sixed and last time, again by using the SAM characters. During the debriefing the participant was asked how they feel and if they feel too negative, the experimenter will try to calm them down. The participants are also asked if they thought the affect manipulation had worked, if positive and negative mood had influence on their performance in finding the target, and if they had a specific strategy for finding the target.
Analysis Plan
In this study effect sizes were considered and because slopes were of interested, averages were also analyzed. The sample size of forty-four participants fell a little short of the target of sixty. Three selections of participants were made: all participants, participants whose
deviation. The last selection was made in an effort to eliminate participants who experienced pop-out effects. This yielded respectively 44, 31, and 10 participants. Manipulation was
successful if the participant had a lower SAM-score at the end of the positive-affect-block than at the end of the negative-affect-block. Outliers were removed from the data by applying the
boxplot method three times. The remaining data was averaged per participant per array-size. Using the sum of the least squares method, slopes and intercepts were calculated. Per condition a slope-average was calculated, otherwise called “the mean reaction time per distractor in
milliseconds”. Afterwards an independent samples t-test was performed to calculate the effect-sizes. The effect size is the difference between the two means from the two conditions divided by a standardizer. This is also known as Cohen’s d (Cumming, 2014). The standardizer is an
estimate of the standard deviation. In this study is opted to look at effect sizes because Null Hypothesis Significance Testing (NHST) isn’t very accurate (Hartgerink et al, 2016). NHST involves two distinct hypotheses, one of which is often the absence of an effect. When that hypothesis is rejected the alternative is accepted. The current study utilizes two distinct and mutually exclusive hypotheses, none of which involves the absence of an effect. Even if the “null hypothesis” is rejected, whichever hypothesis that may be, the alternative cannot be blindly accepted. Null Hypothesis Significance Testing is also highly subjective to false negatives as evidenced by Hartgerink et al (2016).
Results
Manipulation Check and Background Information
The initial manipulation check consisted adding the postinduction to the postblock and then subtracting two times the baseline. This resulted in a number that could both be positive and negative. A positive number meant a move from negative affect to positive affect (meaning the
positive affect manipulation had worked) and a negative number meant a move from positive affect to negative affect (meaning the negative affect manipulation had worked). However, manipulation had not succeeded. Only 17 participants had a success on both
affect-manipulations. Criteria were lowered, and now only post-trial measurements were compared. This resulted in 13 drop-outs, leaving 31 participants.
Results
There was no difference between the slope-averages. The mean difference for the remaining 31 participants (Figure 2) was 0.483 ms (d = 0.111, 95% CI [-1.719, 2.684]). The difference in increase of reaction time per array-element of 0.483 ms/distractor indicates that affect had no influence on the time the participants needed to find the target. This is confirmed by the small effect-size. However, participants were on average slower in the negative condition than in the positive condition (Figure 4).
Figure 4: In this figure a line graph is shown that illustrates the relation between reaction times and set-size for the remaining participants whose manipulation had succeeded.
Table 1
Mod success
Mean RT(ms) RT per distractor(slope)(ms) Intercept(ms)
Positive 688.92 4.91 600.18
Negative 722.56 4.43 628.48
The small mean difference of 0.483 ms and effect size of 0.111 could be due to pop-out effects. In the exit-interview participants told the experimenter that finding the target became easier as they went on with the trials and that it became a routine, almost automatic. Most participants had very flat slopes of only a few milliseconds per added item. These slopes point to a pop-out effect as adding items didn’t add to search-times. To test for this a new selection was made from participants who had slopes that were higher than one standard deviation (in each condition)
600 700 800 900 1000 1100 1200 1300 1400 1500 1600 4 10 18 30 Positive Negative Linear (Positive) Linear (Negative)
taken from an independent samples T-test that included all forty-four participants (Figure 3). The standard deviation for the positive affect condition was 4.349 and the standard deviation for the negative affect condition was 4.027. Any participants that had any slopes lower than these standard deviations in their respective conditions were not included in this selection. Consequently, only ten participants were included in the selection. This yielded results that showed a greater disparity between the two conditions (MD = 1.128, d = 0.314, 95% CI [-2.244, 4.500]), but due to the small sample size no conclusions could be drawn. However, the positive condition produces a steeper slope, indicating that participants were distracted by additional items than they were in the negative condition. Here too had the participants on average higher reaction times during the negative condition (Figure 5).
Figure 5: In this figure a line graph is shown that illustrates the relation between reaction times and set-size for participants whose slopes where higher than one standard deviation.
600 700 800 900 1000 1100 1200 1300 1400 1500 1600 4 10 18 30 Positive Negative Linear (Positive) Linear (Negative)
Table 2
No pop-out
Mean RT(ms) RT per distractor(slope)(ms) Intercept(ms)
Positive 767.70 9.10 626.63
Explorative Research: Absent Target and Order effects
In the main results section only data from trials where the target was present was analyzed. In these trials pop-out effects seemed to be present throughout. Pop-out effects
however do need the target to be present to ‘pop out’. It would be logical to assume that pop-out effects therefor do not exist in trials were the target was absent. An Independent Samples T-Test was performed on a selection of data where the target was absent and the affect manipulation was considered a success. The results were quite different from the results obtained from the data where the target was present. The mean difference between slopes was -0.701 (SE=3.488, d=-0.051, 95% CI [-7.678, 6.277]), meaning that the direction of the effect was reversed, though still very small. Also slopes were much steeper (table 3), which could be expected without pop-out effects. The negative condition produced the steepest slope in these trials.
Figure 6:In this figure a line graph is shown that illustrates the relation between reaction times and set-size on trials where the target was not present for the participants whose manipulation had succeeded.
600 700 800 900 1000 1100 1200 1300 1400 1500 1600 4 10 18 30 Positive Negative Linear (Positive) Linear (Negative)
Table 3
Absent Target
Mean RT(ms) RT per distractor(slope)(ms) Intercept(ms)
Positive 1000.13 26.83 584.33
Negative 1041.99 28.19 605.16
As manipulation had not succeeded, there was a need for exploration on how that could have happened. Order-effects could be a reason. As participants switch from positive to negative emotion or the other way around, it might be that the previous emotion inhibits the participant from coming into the next. This didn’t seem to be the problem as there were no order-effects at all (table 4).
Table 4
Order-effects
Positive after Negative
Negative after Positive
Mean Difference 1.090909
1.090909
Discussion
From the results it cannot be concluded that there is a difference in steepness of the slopes of the positive affect condition and the slopes of the negative affect condition. The manipulation has not succeeded as intended. Instead, the weaker, aforementioned criteria were
used. Therefor any conclusions based on this data are highly disputable. It cannot be concluded whether spatial attention is broadened or temporal attention becomes more flexible.
Manipulation not succeeded
Some participants indicated that they thought the manipulation had no effect. This could be because of the choice of music. Music is experienced by different people in different ways and it could be that the selection of music the participants were given did not contain music that would put them in a certain state. It could also be that participants chose sad music that they nonetheless appreciated. This appreciation could have had a different effect on the participants affect than just making them sad. Appreciation is positive affect, and the differences we saw in SAM-values could be due to demand characteristics or a different response bias. The
experimenters have used terminology such as ‘positive block’ and ‘negative block’ and it may be that participants have responded accordingly, not because that is their true current affect, but simply because they feel that they should respond in this way.
Several participants indicated they had trouble taking the story reading and writing serious. This might be of influence on how much effect this part of the manipulation had on the participants’ affect. Some indicated they felt restricted in writing, because the experimenter was present, even though they had been assured the experimenter would not read it, as the
participants were allowed to take it home.
Rhythm and pop
Some participants indicated that the music put them in a certain response rhythm. This could be an explanation for the low slopes. If participants responded rhythmically, the difference between set-sizes could be diminished, but the participants had an accuracy rating of above 95%.
This is in conjunction with the fact that reaction times don’t seem to be correlated with each other. There is a definitive pop-out effect, but it is not clear what is causing it. What could make the target pop out is the fact that we opted for form instead of color. This suggestion is based on the fact that a previous bachelors-project called ‘Is de Cirkel Rond?’ did not report such pop-out effects. They used color instead of orientation. It could be that people are even better at
distinguishing shapes than colors. Participants also had the same target for the entirety of the experiment. A learning effect could be the base of the pop-out effect. As participants have more experience with the target, recognizing it becomes increasingly easier.
Feature Visual Search
It might be that shapes and orientation are processed in a similar way by the brain. This could mean that two shapes in two different orientations are interpreted by the brain as four distinct shapes. If this is the case, the intended conjunction visual search task, in which targets differ on two aspects, might have been a feature visual search task in reality. This is a viable cause for the flat slopes as according to Kristjansson (2015) slopes are around zero on a single feature visual search task. The search is unaffected by the amount of distractors. Contrary to what Wolfe (1998) wrote on the subject, response times did not increase with 10 to 50 millisecond per added distractor, but sometimes even decreased with each added distractor. Shape and orientation may be considered the same aspect and as such render the conjunctive visual search task a feature visual search task. Kristjansson didn’t cover the performance of a conjunction visual search task that utilizes shape and orientation as target/distractor aspects, but it is now clear that this combination may have caused a lot of pop-out effects.
Closing Words and Future Studies
A somewhat failed manipulation and possibly rhythm and learning effects have led to low slopes and small effect-sizes. What is interesting though, is that participants had overall longer reaction times in the negative condition as opposed to the positive condition. Had this study’s manipulation succeeded it could have stated that negative affect had at least some influence on reaction times, but now it is just a curious observation. A future study could use irritating sounds during the negative condition, instead of sad music, as these might be less appreciated. Different elements should be used in stimuli for the conjunction visual search task, such as color, as these yielded better results in the past. It might also be a good idea to switch up the target sometimes to combat learning effects. When the target was present the positive
condition produced the steepest slope, indicating that the spatial hypothesis might be true and that the attention scope becomes broader with positive affect, allowing more distractors to be processed at once, and become narrower with negative affect, allowing fewer distractors to be processed at once. However, there was no serial search in these trials and mean difference and effect sizes were very small, so any results obtained from this data remains inconclusive. When the target was absent the negative condition produced the steepest slope. In these trials there was serial search as pop-out effects were impossible. This would indicate that the temporal
hypothesis might be true and switching between items becomes easier when positive affect is experienced and harder when negative affect is experienced. Though slopes were much steeper, the mean difference and effect size remained very small, so any results obtained are also
inconclusive. The data from the trials were the target was absent do represent the best what the current study was trying to achieve in terms of design, since serial search was a vital part of the
study. It may be that the trend that was found with this data (negative condition produces the steepest slope) will persist in future studies.
References
Bruyneel, L., van Steenbergen, H., Hommel, B., Band, G. P., De Raedt, R., & Koster, E. H. (2013). Happy but still focused: failures to find evidence for a mood-induced widening of visual attention. Psychological research, 77(3), 320-332.
Cumming, G. (2013). The new statistics why and how. Psychological science, 0956797613504966
Fredrickson, B. L. (2004). The broaden-and-build theory of positive emotions. Philosophica transactions-royal society of london
series b biological sciences, 1367-1378.
Gasper, K., & Clore, G. L. (2002). Attending to the big picture: Mood and global versus local processing of visual information. Psychological science, 13(1), 34-40.
Hartgerink, C. H. J., van Assen, M. A. L. M., & Wicherts, J. M. (2015). Too good to be false: Non-Significant results revisited. Retrieved from osf. io/qpfnw.
Heerebout, B. T., Todorović, A., Smedinga, H. E., & Phaf, R. H. (2013). Affective modulation of attentional switching. The American
journal of psychology, 126(2), 197-211.
Isen, A. M. 2000 Positive affect and decision making. In Handbook of emotions, 2nd edn (ed. M. Lewis & J. M. Haviland-Jones), pp. 417–435. New York: Guilford.
Kristjánsson, Á. (2015). Reconsidering visual search. i-Perception, 6(6), 2041669515614670.
MacLean, M. H., Arnell, K. M., & Busseri, M. A. (2010). Dispositional affect predicts temporal attention costs in the attentional blink paradigm. Cognition and Emotion, 24(8), 1431-1438.
Phaf, R. H. (2015). Attention and positive affect: Temporal switching or spatial broadening?. Attention, Perception, &
Psychophysics, 77(3), 713-719.
Rowe, G., Hirsh, J. B., & Anderson, A. K. (2007). Positive affect increases the breadth of attentional selection. Proceedings of the
National Academy of Sciences, 104(1), 383-388.
Appendix
Positive songs
- Pharrel Williams – Happy
- Bob Marley – Don’t worry be happy - Jackson 5 – ABC
- Katrina & The Waves – Walking on sunshine - The Jacksons – Blame it on the sunshine Negative songs
- Adele – Someone like you - REM – Everybody hurts - Passenger – Let her go
- Eric Clapton – Tears in heaven - Sam Smith – Stay with me
Example text negative induction (all texts where in dutch)
Een aantal jaar geleden zou ik met drie vriendinnen op vakantie gaan in de zomer. Het plan was om een maand te gaan interrailen in Oost-Europa. We waren net klaar met de middelbare school en dit was dus onze examenreis. De tickets waren al geboekt en alles was geregeld, toen een van mij vriendinnetje een week voor ons vertrek gekke bultjes in haar nek ontdekte. Ze besloot om voor de zekerheid toch maar naar de huisarts te gaan. De huisarts stuurde haar meteen door naar het ziekenhuis omdat hij het niet vertrouwde. Het bleek lymfeklier kanker te zijn. Ze moest per direct behandeld worden om de schade zo beperkt mogelijk te houden. Een week nadat we zouden vertrekken begon zij aan haar chemokuur. Onze vakantie waar we zo naar uit hadden gekeken ging uiteraard niet door en er volgde een zeer moeilijke periode.
Example text positive induction
Ik zeil van jongs af aan met mijn ouders en heb veel ervaring met verschillende boten. Vaak gingen wij in het weekend naar het IJsselmeer, vaak ook met vrienden, om het hele weekend lekker te zeilen. Via de zeilschool heb ik ook zeildiploma’s uit verschillende disciplines behaald. Twee jaar geleden slaagde in voor mijn propedeuse van mijn studie en behaalde ik ook het diploma voor het zeilen op een catamaran. Mijn ouders wilde mij als beloning verrassen met een mooi cadeau. Ze namen mij geblinddoekt mee naar Schiphol. Slechts enkele minuten voor vertrek kreeg ik te horen dat we naar de Griekse Ionische kust zouden vliegen.. Hier zouden wij gedurende twee weken gaan zeilen op onze eigen catamaran! Een
mooier cadeau kon ik mij niet voorstellen. Ik had veel ervaring in Nederland, maar had nog nooit gezeild op zo’n mooie plek als Griekenland. Azuur blauw water, witte stranden en onbewoonde eilanden. Bovendien was het ontzettend gaaf om dankzij mijn net behaalde zeildiploma zelf de catamaran te kunnen besturen. Deze vakantie koester ik als een van de fijnste herinneringen die ik heb.
