Managing contextual complexity in an

experiential learning course: a dynamic

systems approach through the
identification of turning points in
students emotional trajectories
Gloria Nogueiras1*, E. Saskia Kunnen2, Alejandro Iborra1

1
Educational Sciences, University of Alcal, Spain, 2Developmental psychology, University of Groningen,

w
Netherlands

i e
Submitted to Journal:

v
Frontiers in Psychology

re
Specialty Section:
Educational Psychology

In
Article type:
Original Research Article

Manuscript ID:
251813

Received on:
31 Dec 2016

Revised on:
07 Apr 2017

Frontiers website link:

www.frontiersin.org
Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial
relationships that could be construed as a potential conflict of interest

Author contribution statement

GN was responsible for research design, data collection, data analysis and writing of the manuscript.
AI was responsible, together with GN, for research design and data collection. He also supported GN in the elaboration of the
discussion and in the revision of the manuscript by providing feedback and comments.
SK was responsible, together with GN, for data analysis. She also supported GN in the revision of the manuscript by providing
feedback and comments.

Keywords

i ew
Contextual complexity, cognitive conflict, complexity management, emotional trajectories, dynamic systems, turning points,
change detection, Monte Carlo permutation tests, process-oriented experiential learning, higher education

v
re
Abstract

In
Word count: 310
This study adopts a dynamic systems approach to investigate how individuals successfully manage contextual complexity. To that
end, we tracked individuals emotional trajectories during a challenging training course, seeking qualitative changes turning
points - and we tested their relationship with the perceived complexity of the training. The research context was a five-day
higher education course based on process-oriented experiential learning, and the sample consisted of 17 students. The students
used a five-point Likert scale to rate the intensity of 16 emotions and the complexity of the training on 8 measurement points.
Monte Carlo permutation tests identified 30 turning points in the 272 emotional trajectories analyzed (17 students * 16 emotions
each). 83% of the turning points indicated a change of pattern in the emotional trajectories that consisted of: a) increasingly
intense positive emotions or b) decreasingly intense negative emotions. These turning points also coincided with particularly
complex periods in the training as perceived by the participants (p = 0.003, and p = 0.001 respectively). The relationship between
positively-trended turning points in the students emotional trajectories and the complexity of the training may be interpreted as
evidence of a successful management of the cognitive conflict arising from the clash between the students prior ways of meaning-
making and the challenging demands of the training. One of the strengths of this study is that it provides a relatively simple
procedure for identifying turning points in developmental trajectories, which can be applied to various longitudinal experiences
that are very common in educational and developmental contexts. Additionally, the findings contribute to sustaining the
assumption that complex contextual demands lead unfailingly to incomplete learning by individuals learning is incomplete. Instead,
it is how individuals manage complexity which may or may not lead to learning. Finally, this study can also be considered a first
step in research on the developmental potential of process-oriented experiential learning training.

Funding statement

The first author of this paper was generously funded through a grant for the Training of University Teachers from the University
of Alcal (Spain).

Ethics statements

(Authors are required to state the ethical considerations of their study in the manuscript, including for cases
where the study was exempt from ethical approval procedures)
Does the study presented in the manuscript involve human or animal subjects: Yes

Please provide the complete ethics statement for your manuscript. Note that the statement will be directly added to the
manuscript file for peer-review, and should include the following information:

Full name of the ethics committee that approved the study

Consent procedure used for human participants or for animal owners
Any additional considerations of the study in cases where vulnerable populations were involved, for example minors, persons with
disabilities or endangered animal species
As per the Frontiers authors guidelines, you are required to use the following format for statements involving human subjects:
This study was carried out in accordance with the recommendations of 'name of guidelines, name of committee' with written
informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki.
The protocol was approved by the 'name of committee'.
For statements involving animal subjects, please use:
This study was carried out in accordance with the recommendations of 'name of guidelines, name of committee'. The protocol
was approved by the 'name of committee'.

If the study was exempt from one or more of the above requirements, please provide a statement with the reason for the
exemption(s).
Ensure that your statement is phrased in a complete way, with clear and concise sentences.
This study was carried out in accordance with the recommendations of University of Alcal Research Committee with written
informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The
protocol was approved by the University of Alcal Research Committee.

vi ew
In re
 1 Managing contextual complexity in an experiential learning
2 course: a dynamic systems approach through the identification of
3 turning points in students emotional trajectories
4
5 Gloria Nogueiras*,1 Saskia Kunnen,2 Alejandro Iborra1
1
6 Department of Educational Sciences, University of Alcal, Alcal de Henares, Spain
2
7 Department of Developmental Psychology, University of Groningen, Groningen, The Netherlands
8 * Gloria Nogueiras, Department of Educational Sciences, University of Alcal, C/ San Cirilo, s/n, 28801, Alcal
9 de Henares, Spain. E-mail: gloria.nogueiras@edu.uah.es
10
11 Abstract
12
13 This study adopts a dynamic systems approach to investigate how individuals successfully
14 manage contextual complexity. To that end, we tracked individuals emotional trajectories
15 during a challenging training course, seeking qualitative changes turning points - and we
16 tested their relationship with the perceived complexity of the training. The research context

w
17 was a five-day higher education course based on process-oriented experiential learning, and

e
18 the sample consisted of 17 students. The students used a five-point Likert scale to rate the
19
20
21

re i
intensity of 16 emotions and the complexity of the training on 8 measurement points. Monte

v
Carlo permutation tests identified 30 turning points in the 272 emotional trajectories analyzed
(17 students * 16 emotions each). 83% of the turning points indicated a change of pattern in

In
22 the emotional trajectories that consisted of: a) increasingly intense positive emotions or b)
23 decreasingly intense negative emotions. These turning points also coincided with particularly
24 complex periods in the training as perceived by the participants (p = 0.003, and p = 0.001
25 respectively). The relationship between positively-trended turning points in the students
26 emotional trajectories and the complexity of the training may be interpreted as evidence of a
27 successful management of the cognitive conflict arising from the clash between the students
28 prior ways of meaning-making and the challenging demands of the training. One of the
29 strengths of this study is that it provides a relatively simple procedure for identifying turning
30 points in developmental trajectories, which can be applied to various longitudinal experiences
31 that are very common in educational and developmental contexts. Additionally, the findings
32 contribute to sustaining the assumption that complex contextual demands lead unfailingly to
33 incomplete learning by individuals learning is incomplete. Instead, it is how individuals
34 manage complexity which may or may not lead to learning. Finally, this study can also be
35 considered a first step in research on the developmental potential of process-oriented
36 experiential learning training.
37
38 Keywords: contextual complexity, cognitive conflict, complexity management,
39 emotional trajectories, dynamic systems, turning points, change detection, Monte Carlo
40 permutation tests, process-oriented experiential learning, higher education.
41

1
42
43 1. Introduction
44
45 This study is based on two assumptions: first, that learning and development occur in
46 response to contextual demands that challenge individuals ways of meaning-making and
47 lead to the creation of more adapted ones (Piaget, 1975/1985); second, that individuals
48 encounters with conflicting contextual demands are usually associated with the experience of
49 negative emotions (Carver & Scheier, 1990; Frijda, 1986; Inzlicht, Bartholow, & Hirsch,
50 2015). Within this framework, we find that a promising way to grasp individuals successful
51 management of challenging environmental demands is to track their emotional experience
52 over time from a Dynamic Systems perspective (Kunnen, 2012; Thelen, 1989; Van Geert,
53 1994). This process-oriented approach supports the idea that it is the dynamic patterns of
54 positive and negative emotions over time which can be positive or negative for learning
55 (Sansone and Thoman, 2005). Our assertion is that individuals initial emotional responses to
56 conflicting and potentially unpleasant demands will be replaced by a qualitatively different
57 and more pleasant emotional response, in the event of successful management. This change
58 will be indicated by turning points (Hayes, Laurenceau, Feldman, Strauss, & Cardaciotto,
59 2007; Eubanks-Carter & Muran, 2012) in the individuals emotional trajectories. We also

w
60 assume that a particularly complex contextual input may be a trigger for the management of

e
61 complexity.
62
63
64

re vi
As higher education teachers, we find process-oriented experiential learning methodologies
to be particularly appropriate for challenging individuals ways of meaning-making, and thus

In
65 for contributing to their learning and development. In this study we therefore focused on the
66 17 participants in a five-day higher education course based on process-oriented experiential
67 learning. We aimed to investigate how these students successfully managed contextual
68 complexity by tracking their emotional trajectories during the training, in search of turning
69 points and determining their relationship with perceived training complexity.
70
71 1.1. Cognitive conflict as a trigger for meaning-making
72 From a constructivist standpoint, human beings make meaning of reality by creating our own
73 personal theories or models of the world, which give us a provisional framework for
74 understanding. As a result, meaning-making entails a dynamic process of continuous
75 updating of these models in order to create adapted responses to an ever-changing
76 environment (Piaget, 1975/1985). This process is triggered by conflicts that arise in the event
77 of discrepancies between our experience and our model of the world, or in other words,
78 between our way of creating meaning and the results we obtain (see Piaget, 1952, and
79 cognitive disequilibrium).
80
81 Conflicts are acknowledged as being the trigger for learning (Piaget, 1975/1985), although
82 whether this happens depends on the strategies we adopt to cope with those conflicts
83 (Kunnen, 2006). In Piagetian terms, when we are first confronted with a conflict, our most
84 economical reaction is to try to solve it through assimilation. In other words, we change our
85 interpretation of the situation, or if possible, the situation itself, so that it once again fits in
86 with our model of the world. The more challenging response of accommodation is only
87 applied if assimilation is unsuccessful. Accommodation entails making significant changes in
88 our model of the world, which reduces the discrepancy between our way of meaning-making
89 and the contextual demands. These accommodational changes lead to learning and
90 development (Kunnen & Bosma, 2000). By contrast, if we are unable to resolve the

2
 91 discrepancy, our confidence in our way of creating meaning is undermined and the potential
92 for learning is narrowed.
93
94 1.2. Cognitive conflict and negative emotions in learning contexts
95 Emotions enhance our meaning-making processes by boosting what we attend to and by
96 providing us guidance for adaptive action (Bradley, 2009; Frijda, 1988; Lazarus, 1991;
97 Solomon, 2007). In particular, conflictive situations tend to be linked with negative emotions
98 (Carver & Scheier, 1990; Frijda, 1986; Inzlicht, Bartholow, & Hirsch, 2015). In this context,
99 Kunnen and Wassink (2003) state that unpleasant emotions are drivers for learning, as they
100 motivate us to react in order to reduce the discrepancy between the meaning we create in a
101 given situation and the demands of that situation. However, like conflict, emotional arousal
102 does not automatically lead to learning (Weiss, 2000). Instead, emotions may either impede
103 or motivate learning depending on how individuals become aware of those emotions and how
104 they manage them (Taylor & Cranton, 2013). As an example, D'Mello and Graesser (2011)
105 argue that it is not confusion itself, but the effortful cognitive activities aimed at resolving
106 the confusion that presumably are beneficial to learning (p. 1307).
107
108 Research in higher education has proved that learning settings that lead students to question

w
109 their accepted ways of knowing tend to be unsettling (Apte, 2009; Antonacopoulou &

i e
110 Gabriel, 2001; Cranton, 2002; Kegan, 1994; McEwen, Strachan, & Lynch, 2010) and elicit

v
111 emotions such as fear, grief, and anger (Dirkx, Mezirow, & Cranton, 2006; Dirkx, 2011). As

re
112 examples, the following studies investigated the relationship between conflicting learning
113 demands and students unpleasant emotional experiences: D'Mello, Craig, and Graesser

In
114 (2009) found a predominance of confusion and frustration in a learning context that was
115 aimed at facilitating deep learning. Nogueiras, Herrero, and Iborra (2016), and Nogueiras and
116 Iborra (2016) found disorientation, insecurity and frustration as a common initial response to
117 a training course that promoted students self-direction.
118
119 1.3. Studying emotions from a dynamic systems perspective: identifying turning points
120 It has long been widely accepted that negative emotions are bad for learning and positive
121 emotions are good for learning, to the extent that the former should be controlled or
122 eliminated (Lepper & Henderlong, 2000; Noddings, 2003). However, some authors (see for
123 example D'Mello, Lehman, Pekrun and Graesser, 2014) argue that this assumption is
124 simplistic and inaccurate. Interestingly, Sansone and Thoman (2005) point out that it is the
125 dynamic patterns of positive and negative emotions at certain points in time in a given
126 context what can be considered good or bad for learning (p. 509). This emphasis on dynamic
127 patterns is at odds with the widespread static approach to the study of emotions, which is
128 based on cross-sectional research designs underlined by a unidirectional model of causation.
129 However, if emotions are considered to be processes that dynamically evolve over time due
130 to interactions between individuals and context (Barrett, 2009; Fogel et al., 1992; Frijda,
131 2009), a paradigm shift is required. This paradigm shift points towards process-oriented
132 approaches that enable the study of within-person emotional patterns of change (Kuppens,
133 Oravecz, &Tuerlinckx, 2010; Larsen, Augustine, & Prizmic, 2009; Scherer, 2009). Dynamic
134 systems theory is particularly suited to this approach (Camras & Witherington, 2005; Lewis,
135 2005; Lichtwarck-Aschoff, Kunnen, & Van Geert, 2009).
136
137 Dynamic systems theory is a metatheoretical framework for understanding developmental
138 processes, which are conceived as non-linear dynamic systems (Van Geert & Van Dijk, 2015;
139 Witherington, 2007). These systems are formed by interconnected and interacting
140 components that affect each other and develop over time, due to interactions between

3
141 individuals and their context (Kunnen, 2012; Thelen, 1989; Van Geert, 1994). As
142 developmental processes are characterized by sudden changes and irregularities, their study
143 requires methodologies that enable to grasp variability and change while they occur, rather
144 than comparing pre- and post-change behavioural patterns (Fogel, 2011; Van Geert & Van
145 Dijk, 2002; Van Dijk & Van Geert, 2007). A dynamic systems approach therefore uses
146 individual microdevelopmental trajectories1 as the unit of analysis, and examines them using
147 as many measurements over time as possible (Yan & Fischer, 2007; Molenaar, 2004; Siegler,
148 2006).
149
150 Qualitative changes in individual microdevelopmental trajectories can be identified based on
151 the concept of the turning point. Turning points are points that mark meaningful deviations in
152 trends in time series data that involve discontinuous changes (Hayes, Laurenceau, Feldman,
153 Strauss, & Cardaciotto, 2007; Eubanks-Carter & Muran, 2012). These changes entail a
154 transition from one variability pattern to another variability pattern (Kunnen, Van Dijk,
155 Lichtwarck-Aschoff, Visser, & Van Geert, 2012) in such a way that the trajectories separated
156 by a turning point differ in direction or nature (Abbott, 1997).
157
158 1.4. Meaning-making from a dynamic systems perspective

w
159 From a dynamic systems perspective, contexts contribute to the emergence of individuals

i e
160 behaviour by providing both constraints and opportunities (Thelen & Ulrich, 1991; van Geert,

v
161 1994; Vallacher, Van Geert, & Nowak, 2015). The greater the contextual variability, the

re
162 greater the likelihood of experiencing conflicts that lead us to make changes to adapt (Bosma
163 & Kunnen, 2001; Hayes, Yasinski, Barnes, & Bockting, 2015). In other words, contextual

In
164 variability offers systems room to explore and to adapt to new situations (Thelen & Smith,
165 1994; Van Geert, 1994). As a result, if learning and development is to occur, a factor must
166 challenge our patterns of meaning-making so that they are reorganized on a more complex
167 level (Kloep, Hendry, & Saunders, 2009; Thelen, 2005). Within this framework, and in line
168 with Piagets argument, the stability of our model of the world due to the maintenance of our
169 meaning-making patterns would entail assimilation, whereas the destabilization of our model
170 of the world and the emergence of new meaning-making patterns would entail
171 accommodation. In this sense, the occurrence of any variability in developmental trajectories
172 of any kind such as emotional trajectories - enables us to identify developmental transitions
173 (Granott, Fischer, & Parziale, 2002; Van der Maas & Molenaar, 1992).
174
175 1.5. Experiential learning as a source of cognitive conflict
176 Our understanding of experiential learning differs from the model proposed by Kolb (Kolb,
177 1984; Kolb & Kolb, 2009). This model is based on a learning cycle that starts with students
178 reflection on the content of concrete experiences in order to create abstract concepts, which
179 are then tested by active experimentation. This in turn generates further concrete experiences.
180 By contrast, and based on the work of McWhirter (2002), we argue that the key to
181 experiential learning is to use of students own sensory and natural experiences to
182 subsequently structure them using detailed modelling distinctions. As a result, in the
183 process-oriented experiential learning model2 (McWhirter 2002), the emphasis is not on the
184 creation of abstract ideas or explanations of personal experiences based on the content of the
185 experience, or what happens. On the contrary, it is the creation and exploration of formal
186 distinctions that are tested as they may be useful in making sense of the individuals

1
The general course of change over time in a variable is described as a developmental trajectory (Bosma & Kunnen,
2001).
2
In the remainder of the article, and for the sake of brevity, we will use the label experiential learning to refer to a process-
oriented experiential learning model as described by McWhirter (2002).

4
187 experiences in a more complex way - considering the whole experience from a process
188 perspective, or what, how and why it happens.
189 In specific terms, the experiential learning model that we present is based on the following
190 sequence: 1) Creating an experience: students undertake an open exploration of their natural
191 experience or intuitive understanding about the phenomena being studied; 2) Reviewing:
192 students share the range of experiences they obtained in the exploration, identifying
193 similarities and differences in comparison with other classmates experiences - or even with
194 other previous personal experiences; 3) Formalising: the trainer introduces formal models and
195 distinctions; 4) Testing: students compare and test the formal model and distinctions against
196 their own personal and group experience.
197 The experiential learning model differs from the traditional learning model, labelled didactic
198 learning, which is based on the following sequence: 1) Description of a formal model: the
199 trainer tells; 2) Demonstration: the trainer shows; 3) Experience: the students do; 4) Provision
200 of feedback: the trainer tells the students what went wrong and right. This sequence is useful
201 for rote learning and learning protocols, so that there is an increase in knowing before the
202 uncertainty and risk involved in doing. However, it overlooks the fact that ready-made
203 techniques do not correspond with reality, which has variations that are ignored in favour of
204
205
206
w
the illusion of certainty that is provided by protocols. The didactic learning model therefore

i e
leads to a reduction towards the correct way, and fosters a dependent, repetitive and

v
unquestioning style of learning that prevents students from engaging in exploration,

re
207 creativity, and self-direction.
208

In
209 By contrast, experiential learning as understood by McWhirter (2002) contributes to
210 developing the competence of learning to learn, increases the depth of learning, encourages
211 an attitude of curiosity and wonder, and prepares students to take what they learn into the
212 world. Indeed, this learning sequence is similar to the one we all naturally follow as children
213 when making meaning of our surroundings: starting from a baseline of not knowing, we build
214 an understanding. However, the application of an experiential learning sequence also entails
215 little security in knowing, because there is no right answer or example to begin with.
216 This might initially lead students to feel insecure and uncomfortable, as they have to develop
217 an open orientation to new experiences. Furthermore, if students continue to apply the
218 didactic learning sequence they are accustomed to, there may have an additional sense of not
219 knowing, while believing that they should know before they continue to learn. Apart from
220 the learning sequence itself, the contents explored in process-oriented experiential exercises
221 also tend to be complex and challenging for students.
222
223 A formal model that can be useful for illustrating students typical responses to experiential
224 learning when they are first exposed to it is the Three sets model: set-up - upset - set-down
225 (McWhirter, 2000). This model describes three stages in the process of meaning-making of a
226 conflictive experience. The students initial set-up would be the learning sequence that they
227 are used to, and which they expect to find at the beginning of the training - the didactic
228 sequence in most cases. The upset would be the destabilization experienced by the students
229 caused by the challenging demands of the experiential sequence and the exploration of
230 personal experiential content. The set-down would involve potential changes made by the
231 students in response to the upset. This set-down could potentially be related to cognitive
232 accommodation.
233
234 In view of the above, the hypothetical emotional trajectory of students in a process-oriented
235 experiential learning context may be as follows. First, the students upset at the conflicting

5
236 training demands could lead them to the experience of intense negative emotions and a low
237 level of positive emotions. The students development of new meaning-making strategies for
238 coping with the challenging demands would entail a decline in the intensity of negative
239 emotions and an increase in the intensity of positive emotions. This could be identified
240 through the occurrence of turning points in emotional trajectories.
241
242 It is important to note the positive aspect of the upset in a process-oriented learning setting,
243 and how students are also supported to learn to learn in this way. In fact, the experiential
244 methodology itself revolves around learning management, which in this methodology is
245 something that can be taught and learnt. In fact, this learning model is intended to facilitate
246 qualitative changes in students way of learning, and specifically in the way they organize
247 information while they learn. In this process, both the recognition and the management of the
248 negative emotions that are expected to be experienced as a response to contextual challenge
249 are assumed to enhance individuals ability to manage learning effectively.
250
251 1.6. The present study
252 The main goal of the present study is to investigate the process of successfully managing
253 contextual complexity by the 17 participants in a five-day higher education course, based on

w
254 a process-oriented experiential learning model. To meet this goal, we aimed to: 1) identify

e
turning points in the participants positive and negative emotional trajectories, 2) test whether

i
255

v
256 those turning points coincide with periods in the training perceived by participants as

re
257 particularly complex. We also aimed to test the widespread assumption that contextual
258 complexity leads to the destabilization of individuals ways of meaning-making, and

In
259 consequently to learning.
260 2. Method
261
262 2.1. The learning context studied
263
264 2.1.1. Description of the training course
265 This study focuses on an intensive training course based on a process-oriented experiential
266 learning model (McWhirter, 2002). The course took place in a Faculty of Education at a
267 Spanish university. It was part of a summer training program for both university students and
268 non-students. The training course lasted 31.5 hours over five consecutive sessions. The first
269 four sessions lasted for 7 hours, and the fifth lasted for 3.5 hours. The training was conducted
270 by the developer of Developmental Behavioural Modelling (DBM)3 John McWhirter
271 (McWhirter, 2011).
272
273 The course was entitled Self-created learning throughout life, and aimed to develop
274 participants life competences of self-managing and self-directing their own learning
275 processes. Participants were therefore expected to: 1) gain a deeper understanding of the
276 learning to learn process by modeling, developing, and assessing their autonomous learning
277 processes; 2) explore changes in beliefs, values, self-concept, identity or vital aspirations that
278 take place as a consequence of self-created learning. The reason why we selected this training
279 course for this study was that both the course content (how to manage learning) and the
280 experiential learning methodology (intended to support students in their improved

3
Developmental Behavioural Modeling is a comprehensive field to systematically model modelling (McWhirter, 2002,
2011). It studies how human beings create our own models of the world through natural modelling skills, how effective our
models are, and how we change them in order to adapt them to new circumstances in an optimal way (McWhirter, 1998).

6
281 management of learning) were expected to be highly upsetting and challenging for the
282 participants. This made the training ideal for examining adaptation to cognitive conflict.
283
284 The first session of the course started with the participants exploration of their current
285 understanding about learning. The next sessions started with a review by the participants
286 revision of the content covered and the exercises performed during the previous session, in
287 order to reconnect with the experience, share their understandings, make connections and
288 identify issues for clarification. The remainder of all the sessions consisted of: 1) the trainers
289 proposal of experiential exercises and the students performance of them, with an emphasis
290 on trying different ways of doing things and on extending ones own attention; 2) sharing
291 experiences in order to identify what was similar and what was different, fostering the
292 students awareness of variety, 3) the trainers feedback, 4) the trainers introduction of
293 formal models and process distinctions so that the students could explore their experience
294 with further direction.
295
296 During the experiential exercises, three processes were encouraged: a) to attend (playing with
297 attention and moving it); b) to notice (paying attention to the content of what one notices and
298 to the way in which one notices: active or passive, detailed or not, rigid or open); c) to

w
299 explore and to investigate (going beyond what is initially recognized, opening things up and

e
300 creating things that one does not know). By way of an example, an experiential sequence
301
302
303

re i
which was used on the course to explore different personal learning experiences consisted of

v
the following steps: 1) Identifying the experience of learning to be explored; 2) Identifying
the resources available before the experience, both from the environment and from oneself; 3)

In
304 Reviewing what happened throughout the learning sequence, paying attention to what was
305 changing and what one was doing; 4) Reviewing how one checked that the learning had
306 happened; 5) Thinking about further development, in terms of changes that could be
307 introduced to improve learning.
308
309 2.1.2. Participants
310 The participants on course were 31 higher education students and professionals, mostly from
311 educational, psychological, and health disciplines. At the beginning of the course, the
312 participants were informed about the research and the confidentiality thereof, and all
313 contributed information and agreed to take part by means of informed consent.
314
315 The sample in this study consisted of 17 participants (12 women, mean age 33.53, age range
316 19-55) who completed all the measurements during the course. All the participants lived in
317 Spain, except one who lived in United Kingdom and came to Spain for the training. There
318 were 13 participants from Spain, 1 from Ireland, 1 from Mexico, 1 from the United Kingdom
319 and 1 from Romania. Of the 17 participants, 7 were undergraduate students on either Teacher
320 Training or Psychology courses, 2 were doctoral students in Education, 2 were psychologists,
321 2 were Psychology university lecturers, 1 was a Secondary Education English teacher, 1 was
322 a librarian, 1 was a veterinarian and 1 was a doctor.
323
324 2.2. Data collection
325 The first author of this article attended the course as a participant, while the third author was
326 an active observer of the training. The two authors and the trainer were responsible for the
327 data collection. Within the framework of a more extensive data collection, the focus in this
328 article is on a follow-up questionnaire distributed over eight measurement points (m.p.)
329 during the training course, i.e. twice per session in the first four sessions. The follow-up
330 questionnaires were strategically distributed at the beginning and at the end of the session or

7
331 at particular points that were expected to be more challenging for the course participants, so
332 that the potential emotional arousal associated with them was more likely to be recorded. See
333 Appendix 1 for an overview of the timeframe of the training course, including the distribution
334 of questionnaires.
335
336 In each questionnaire, the participants self-reported the intensity of their emotions and the
337 degree of perceived training complexity that they experienced at that point on the course.
338
339 The intensity of the emotions was assessed by means of a list rated on a Likert scale, from
340 very low (1) to very strong (5). The list was designed by the first and the third authors of this
341 article with the trainer, and included a range of emotions that they had observed in previous
342 process-oriented experiential learning training activities they had led. The 16 emotions in the
343 list were: joy, sadness, anger, fear, enjoyment, interest, distress, boredom, hope,
344 overwhelmed, overload, confusion, enthusiasm, dissonance, ignorance and curiosity.
345
346 The degree of perceived training complexity was assessed using two items, rated on a Likert
347 scale from very low (1) to very strong (5). The items were: (a) conceptual complexity, which
348 referred to the complexity associated with the students understanding of the formal models

w
349 introduced by the trainer, (b) performative complexity, which referred to the complexity

e
350 involved in doing the experiential exercises, in terms of both their structure and the personal
351
352
353 2.3. Data analysis

re i
experiential content explored by students.

v
In
354 In order to systematize the analysis of the emotions assessed by the participants, each
355 emotion was coded as either positive or negative, which resulted in: (a) 6 positive emotions
356 (joy, enjoyment, interest, hope, enthusiasm, curiosity), and (b) 10 negative emotions (sadness,
357 anger, fear, overwhelmed, boredom, distress, overload, confusion, dissonance,
358 ignorance).Emotional trajectories were created for each participant and for each emotion,
359 which consisted of the series of intensity scores provided for the eight measurement points.
360 This means that the trajectories are nested, i.e. there are several emotional trajectories for
361 each individual.
362
363 For the analysis of the complexity scores, an overall complexity score for every participant
364 was computed for each measurement point by averaging the scores given in the two items
365 included in the questionnaires - conceptual complexity and performative complexity. This
366 provided a score for complexity for each student and for each measurement point, with a
367 range of between 1 and 5.
368
369 2.3.1. Identifying turning points
370 Looking for an unexpectedly large peak in the data has proved to be effective in identifying
371 discontinuous patterns showing the emergence of qualitative changes in individual
372 developmental trajectories, (see for example Van Dijk & Van Geert, 2007, 2011; Van Geert
373 & Van Dijk, 2002). This technique inspired us when we designed our procedure to identify
374 turning points in the participants emotional trajectories. This procedure consisted of two
375 steps. First, the trajectories for the 16 emotions scored by the 17 participants (i.e. 272
376 emotional trajectories: 102 positive and 170 negative) were examined for points that fell
377 outside a computed confidence interval. These were labelled as exceptional points. Second,
378 exceptional points were examined using Monte Carlo permutation tests to determine whether
379 they showed a qualitative change in the pattern of the emotional trajectory. In this case they
380 were labelled as turning points. The complete procedure is detailed below.

8
381
382 2.3.1.2. The first step: looking for exceptional points
383 The regression line underlying every emotional trajectory was first computed. Based on the
384 values of that regression line, a confidence interval of 1.65 standard deviations around the
385 spread of the data was computed, so that the upper control limit (UCL) of the confidence
386 interval was 1.65 standard deviations above the regression line, and the lower control limit
387 (LCL) of the confidence interval was 1.65 standard deviations below the regression line.4 The
388 points on the emotional trajectories which fell outside the computed confidence intervals
389 were labelled exceptional points.
390
391 Table 1 shows an example of the computing of the confidence interval of the scores for
392 intensity of distress of one of the participants throughout the eight measurement points (m.p.).
393 In the example, m.p. 5 is an exceptional point, which means that the score in the intensity of
394 distress in that point is greater than the value of the upper control limit of the computed
395 confidence interval. Figure 1 graphically presents the emotional trajectory for distress, the
396 underlying regression line and the upper and lower control limits of the confidence interval
397 computed from the regression line. This graph shows the exceptional point at m.p. 5, which
398 can be seen above the upper control limit.

w
399

e
400 2.3.1.2. The second step: determining which exceptional points are turning points
401
402
403

re i
Once the exceptional points were identified, a statistical analysis was performed in order to

v
determine whether they indicated a qualitative change in the emotional trajectory, i.e. a
turning point. The analysis consisted of a comparison of the slope5 of the emotional trajectory

In
404 before and after the exceptional point, in search of significant differences. To that end, Pop
405 tools (Hood, 2010) in Microsoft Excel 2010 were used to perform Monte Carlo permutation
406 tests (see Todman & Dugard, 2001). These tests are also known as random permutations,
407 random sampling techniques or resampling techniques, and are included in the family of
408 bootstrap techniques (Efron & Tibshirani, 1993; Good, 1999). Resampling techniques are
409 well-suited to longitudinal research, and have great explanatory value for small or skewed
410 samples, and result in reliable P values, since they do not assume any underlying distribution,
411 or a minimum sample size (for this argument see Van Geert, Steenbeek, & Kunnen, 2012).
412 Standard tests such as t-tests are not allowed in these cases (Kunnen, 2006).
413
414 Monte Carlo permutation tests estimate the chances that an observed result is caused by
415 chance alone. They compare an empirical distribution of data with a random distribution that
416 is created by reshuffling the empirical data in accordance with a null hypothesis. In this case,
417 the null hypothesis stated that there was no significant difference between the slopes of an
418 emotional trajectory before and after an exceptional point. The reshuffling computes all
419 possible re-orderings of the empirical data set, by computing a very large number of
420 accidental distributions and counts how often the observed or a bigger difference occurs in the
421 random distributions. In this case, the reshuffling counted how often the difference between
422 slopes was the same or bigger than the observed difference. This frequency is then divided by
423 the number of random samples in order to produce a P value for the tested difference, which
424 is the probability of the observed difference occurring in the random distributions of the data.
425 If the probability is low, this means that the observed difference is not due to chance and
426 therefore that it is a legitimate difference (for more detail, see Van Geert, Steenbeek, &

4
In a normal distribution, 1.65 standard deviations from the samples mean represents around 90% of the population. We
therefore assumed that a confidence level of 90% was reasonable for identifying exceptional points in the trajectories.
5
The slope is a linear trend parameter that describes both the direction (increase or decrease) and the steepness (the strength
of such decrease or increase) of the changes in the variable studied.

9
427 Kunnen 2012). In this analysis, 10.000 random distributions were computed and a P value
428 lower than 0.05 was considered significant.
429
430 2.3.2. Testing the relationship between turning points and perceived training complexity
431 The next step in the analysis consisted of testing whether the perceived complexity of the
432 training was significantly higher at the measurement points where at least one turning point in
433 some emotional trajectories had been identified. Monte Carlo permutation tests were
434 performed to that end. The average complexity score for the measurement points where
435 turning points had been detected and the overall complexity score of the measurement points
436 where no turning points had been identified was computed in order to test whether the former
437 was higher than the latter. The perceived complexity scores of each participant at every
438 measurement point were then reshuffled, and the overall complexity score at the measurement
439 points with turning points, the average score of complexity in measurement points without
440 turning points and the difference between both were then computed. Monte Carlo simulations
441 (10.000 random distributions computed) were used to test whether the difference in the
442 degrees of perceived complexity was significant (a P value lower than 0.05) or due to chance.
443
444 3. Results

w
445

e
446

i
3.1. Detection of turning points

v
447 3.1.1. The first step: identification of exceptional points

re
448 In the 272 emotional trajectories analyzed, 142 exceptional points were identified, i.e. 52% of
449 the emotional trajectories had at least one exceptional point. There were 53 exceptional points

In
450 in positive emotion trajectories (52 % of these trajectories), and 89 exceptional points in
451 negative emotion trajectories (52% of these trajectories).
452
453 Two types of exceptional points were identified: 1) High exceptional point: a point on an
454 emotional trajectory above the upper control limit of the trajectorys confidence interval; 2)
455 Low exceptional point: a point on an emotional trajectory below the lower control limit of the
456 trajectorys confidence interval.
457
458 Low exceptional points were the most common in positive emotion trajectories (62 %) and
459 high exceptional points were the most common in negative emotion trajectories (75 %).
460 These two types accounted for 70% of the total number of exceptional points (23% and 47 %
461 respectively). The amount and percentage of high exceptional points and low exceptional
462 points in positive and negative emotion trajectories are presented in Table 2.
463
464 3.1.2. The second step: detection of turning points
465 30 of the 142 exceptional points (21%) were identified as turning points. This meant that 11%
466 of the total number of emotional trajectories analyzed (272) had a turning point. Of the 30
467 turning points, 12 turning points were found in positive emotion trajectories (12% of positive
468 emotion trajectories) and 18 turning points were found in negative emotion trajectories (11%
469 of negative emotion trajectories).
470 Two types of turning points were identified: 1) Pre-decrease turning point: a point on an
471 emotional trajectory above the upper control limit of the confidence interval of the trajectory,
472 indicating a change in the trajectory from an increasing pattern in the intensity of the emotion
473 towards a decreasing pattern; 2) Pre-increase turning point: a point on an emotional trajectory
474 below the lower control limit of the confidence interval of the trajectory, indicating a change
475 in the trajectory from a decreasing pattern in the intensity of the emotion towards an
476 increasing pattern.

10
477
478 Pre-increase turning points were the most common in positive emotion trajectories (83%),
479 while pre-decrease turning points were the most common in negative emotion trajectories (83
480 %). These two types accounted for 83% of the total number of turning points (33% and 50%
481 respectively).The amount and percentage of pre-decrease turning points and pre-increase
482 turning points in positive and negative emotion trajectories are presented in Table 3.
483
484 The following figures (Fig. 2, Fig. 3, Fig. 4, and Fig. 5) are examples of the two types of
485 turning points, pre-decrease and pre-increase, in positive and negative emotion trajectories.
486
487 The 30 turning points were distributed over 13 of the 17 participants (76%) and 9 of them
488 presented pre-increase turning points in positive emotions, pre-decrease turning points in
489 negative emotions, or both, which could be considered positively trended turning points. A
490 high percentage of turning points (63.33%) were located at m.p. 4 (9 turning points: 3 pre-
491 decrease turning points in negative emotions, 4 pre-increase turning points in positive
492 emotions, and 2 pre-increase turning points in negative emotions), and at m.p. 6 (10 turning
493 points: 1 pre-decrease turning point in a positive emotion, 3 pre-decrease turning points in
494 negative emotions, and 6 pre-increase turning points in positive emotions). A precise and
495 detailed distribution of the turning points for the participants and for each measurement point
496

497
is given in Appendix 2.

vi ew
3.2. The relationship between turning points and perceived training complexity

re
498 The measurement points at which turning points were identified had significantly higher
499 complexity scores than the other measurement points (p = 0.002). More detailed results were

In
500 found when the pre-decrease turning points and pre-increase points were analyzed separately
501 in both positive and negative emotion trajectories: significantly higher levels of complexity
502 were found at the measurement points with pre-increase turning points in positive emotion
503 trajectories (p = 0.003) and at the measurement points with pre-decrease turning points in
504 negative emotion trajectories (p = 0.001). Table 4 shows the average and the range of the
505 complexity scores for each measurement point. Table 5 shows detailed results for the
506 differences in complexity scores between the measurement points with turning points and the
507 measurement points without turning points. An overview of the trajectories of the complexity
508 scores for each student over the 8 measurement points is provided in Appendix 3.
509
510 3.3. Summary of results
511 11% of the emotional trajectories analyzed had a turning point. The two most common types
512 were: 1) pre-increase turning points in positive emotion trajectories (83% of the turning
513 points in these trajectories); 2) pre-decrease turning points in negative emotion trajectories
514 (83 % of the turning points in these trajectories). These two types of turning points accounted
515 for 83% of the total. The relationship between the occurrence of turning points and perceived
516 training complexity was significant at : 1) pre-increase turning points in positive emotion
517 trajectories (p = 0.003); and 2) pre-decrease turning points in negative emotion trajectories (p
518 = 0.001), which was consistent with the first result.
519
520

11
521 4. Discussion
522
523 We have organized our discussion around four issues: 1) The positive orientation of the
524 students emotional trajectories; 2) The relationship between contextual complexity and
525 positively trended turning points in the students emotional trajectories; 3) The concentration
526 of turning points in the students emotional trajectories around the middle of the training
527 course; 4) The apparent scarcity of turning points in the students emotional trajectories.
528
529 4.1. The positive orientation of the students emotional trajectories
530 Emotions play a key role in learning processes, since they enable us to make meaning of our
531 experiences and adapt to our environment (Bradley, 2009; Frijda, 1988; Lazarus, 1991;
532 Solomon, 2007). We therefore expected our participants emotional experience to fluctuate
533 over time as a result of their changing ways of making meaning of the new demands arising
534 from the process-oriented experiential learning setting. The predominant emotional trajectory
535 among our participants had a positive orientation, as evidenced by the two most frequent
536 types of turning points: pre-increase turning points in positive emotion trajectories, and pre-
537 decrease turning points in negative emotion trajectories. These turning points indicated that
538 an initial response consisting of either increasingly intense negative emotions or decreasingly

w
539 intense positive emotions was replaced by a pattern that consisted of decreasingly intense

e
540 negative emotions or increasingly intense positive emotions.
541
542
543

re vi
On the one hand, the students prevalent initial experience of intense negative emotions and
non-intense positive emotions in response to challenging demands is consistent with previous

In
544 findings in higher education settings (see for example Apte, 2009; Dirkx, 2011). Dirkx (2008)
545 argues that adults emotional responses in learning settings are usually related to the content,
546 the structure or the processes that they entail, so that an open structure can lead students to
547 feel overwhelmed and to complain of a lack of direction. This is consistent with the likely
548 experience of our participants.
549
550 Instead of the typical emphasis on learning protocols that would be expected in a didactic
551 learning sequence, an experiential learning sequence places the emphasis on the students
552 exploration of their natural experience. This means that there is a predominance of open an
553 exploratory exercises, and that no closed answers or procedures to follow are provided by the
554 trainer. This is potentially challenging for many students, who are mostly used to content-
555 based teaching practices (for related findings, see Nogueiras & Iborra, 2016), and can be a
556 source of upsets. In addition to the new experiential learning sequence, something that can be
557 upsetting for students is the content explored in the exploratory exercises. Newcomers may
558 become upset at the beginning of the training by the mismatch between their expectations and
559 the proposed learning sequence. Meanwhile, students who are used to the experiential
560 learning sequence are expected to have a maximized set-up, since they already expect that
561 they will not know from the beginning and know that they have to remain curious and open.
562 For them, the possible would not be related to the experiential learning sequence, but instead
563 to the possibility of dealing with complex experiential content.
564
565 On the other hand, the positive orientation of our students emotional trajectories over time is
566 similar to that found by Arpiainen, Lackus, Tks, and Tynjl (2013) in their research on
567 students emotions in an entrepreneurship learning program. In their thematic analysis of the
568 students in-depth interviews after the training, they found waves of emotions consisting of
569 frequent negative emotions at the beginning of the program, and positive emotions towards
570 the end. According to these authors, the students negative emotional experience was a

12
571 response to the new learning environment and to the challenging tasks they were set.
572 Conversely, the positive shift in the students emotional experience over time was considered
573 to be related to their increased ability to cope with uncertainty during their learning process.
574 This is something that we find also plausible in our study, as discussed in the paragraphs
575 below.
576
577 In situations where our ways of meaning-making are destabilized, managing both the
578 destabilization and the associated unpleasant emotions is necessary if learning is to occur
579 (Taylor & Cranton, 2013). Our participants might have developed two different responses to
580 the likely destabilization they experienced. One possibility is that students were successful in
581 their attempt to reduce the discrepancy between the demands of the training and their ways of
582 meaning-making, so that they effectively managed their learning process. In this case,
583 students would have moved from a period of emotional discomfort to a period of emotional
584 comfort (see Kunnen & Wassink, 2003). Another possibility is that students were unable to
585 manage the challenging demands of the training, so that the destabilization might have been
586 counterproductive for learning, and undermined the students confidence in their way of
587 creating meaning. In this case, the initial unpleasant emotional experience would have
588 persisted or become more profound over time.

w
589

e
590 The predominantly positive orientation of our students emotional trajectories over time can
591
592
593
i
be interpreted as evidence of their successful management of the challenging demands of the

v
situation. This may in turn indicate a greater possibility for students cognitive

re
accommodation. At this point, it is necessary to explicitly state what kind of cognitive

In
594 accommodation we are referring to. To do this, it is possible to distinguish between two types
595 of accommodation. On the one hand, an individual can change their cognitive structures
596 when they understand something. In this case, the accommodation is focused on the content.
597 On the other hand, the accommodation can be more focused on the process. In our case, we
598 refer to accommodation that is related to the students adaptation to the learning
599 methodology, rather than to the content of learning itself. The former would involve a change
600 in the way in which the students adapted to the course, modifying their initial expectations
601 and going beyond them. If this kind of accommodation took place, it would involve the
602 students changing their preference to a didactic learning sequence, and being more open to
603 investigation and development in an experiential learning sequence. This would necessarily
604 involve a higher degree of flexibility and tolerance of uncertainty.
605 In short, the two most common types of turning points found in emotional trajectories can be
606 taken as an evidence of the students positive set-down in response to the initial upset arising
607 from the mismatch between their expectations and the training demands. After the students
608 adapted to the new learning model, their emotional experience shifted from being
609 predominantly unpleasant to being predominantly pleasant.
610 In view of the above, the study of emotional patterns over time in both Arpiainen et al.s
611 (2013) research and our own research supports the idea that a dynamic approach to emotions
612 enable to overcome the simplistic claim that positive emotions are good for learning and
613 negative emotions are bad. Instead, and according to Sansone and Thomans (2005)
614 arguments, this study confirms that it is the dynamic patterns of positive and negative
615 emotions over time, in connection with individuals changing ways of managing contextual
616 complexity, which can be considered positive or negative for learning. This last point is
617 discussed further below.
618

13
619 4.2. The relationship between contextual complexity and positively trended turning
620 points in the students emotional trajectories
621 The coincidence between positively trended turning points and particularly complex periods
622 in the training is consistent with the idea that high levels of contextual complexity might act
623 as a catalyst for individuals new and more adapted behavioural patterns (Piaget, 1975/1985).
624 If an experience is not challenging for individuals, they will not become involved in a
625 meaning-making process aimed at creating a response that is adapted to their environment
626 and the experience will therefore not be developmental. The positively trended turning points
627 in the participants emotional trajectories are an example of a developmental orientation,
628 which leads us to assume that the complex experiential training triggered the participants
629 adaptation to the new contextual demands.
630
631 However, these assertions do not mean that contextual complexity always leads to
632 experiences of upsets, and that those upsets lead to learning. Instead, if this happens it
633 depends on how individuals manage complexity and the emotional upset associated with it.
634 If the reasons for contextual complexity were emotional, all the students emotional
635 trajectories would be similar, and this is not the case. Our hypothesis is therefore that there
636 are other issues involved apart from complexity, such as the individuals self-management

w
637 baseline and the different paces and phases that individuals can follow over time when

e
638 learning how to manage complexity, and which play an important role in learning. For
639
640
641

re i
example, some students may respond to contextual complexity by experiencing excitement

v
and enjoyment, and not necessarily by experiencing unpleasant emotions. Thus, contextual
complexity will not therefore always match upsets. Indeed, as people learn to manage

In
642 complexity, emotional upsets may no longer be an issue. As mentioned above, students who
643 expected complexity in the experience may be less upset than students who did not expect
644 complexity. In conclusion, we cannot generalize that complexity is upsetting for every
645 individual, and neither can we state that it will inevitably lead to a successful adaptation to
646 the context.
647
648 4.3. The concentration of turning points in the students emotional trajectories around
649 the middle of the training course
650 In connection with the above, it is important to acknowledge that the contextual complexity
651 identified at various points in the training may have led to different responses by the
652 participants. If very high levels of complexity had been elicited at the beginning of the
653 course, the participants destabilization could have been too high to be managed successfully.
654 However, the middle period of a training program is a more suitable time for participants to
655 become destabilized. This is potentially because they are more prepared after a prior period
656 of experience, in which when their ways of meaning-making might have been reorganized.
657 From this vantage point, facing high complexity might have led to the emergence of a
658 qualitatively different and more adapted emotional response, as indicated by positively
659 trended turning points in the emotional trajectories. Interestingly, not only were more
660 complex exercises likely to be proposed, but the participants perception of complexity might
661 also have varied over time. For example, at the beginning of the training the students
662 complexity scores were not still very high probably because they were creating their own
663 standards for the course, and not necessarily because the exercises proposed were less
664 complex. This makes sense when we recall that a large percentage (63.33%) of the 30 turning
665 points in emotional trajectories were located at m.p. 4 and at m.p. 6, and that furthermore,
666 most of these turning points were positively oriented. The middle period of the training might
667 therefore be considered the safest moment for the trainer to propose challenging input, both

14
668 because students are likely to be more accustomed used to the learning sequence, and because
669 they still have time to settle down in the context of the training course.
670
671 From the above, it can be concluded that contextual complexity alone is not a trigger for
672 development and learning, but also that when this complexity is faced is also relevant in
673 terms of the individuals resources for managing it. When discussing identity development,
674 Kroger (1993) interestingly points out that a certain readiness is needed if a conflict is to
675 induce change in individuals. This is consistent with cognitive structural theory and research,
676 which have provided evidence to suggest that individuals have to be at a certain stage for an
677 optimal period of time before change is possible. Accordingly, in order to explain a
678 qualitative change in students emotional response to the training demands, we have to take
679 into account the prior process of reorganization and its subsequent impact. In dynamic
680 systems, this is termed feedback delay (Van Geert, 1994; Kunnen, 2012).
681
682 4.4. The apparent scarcity of turning points in the students emotional trajectories
683 An issue arising from our findings that is also worth discussing is the apparent low number of
684 turning points detected. We found 30 turning points in 272 emotional trajectories - 87% of
685 which were positively trended. However, we believe that this number is reasonable, as
turning points in participants emotional trajectories are taken as evidence of the

w
686

e
687 reorganization of students prior ways of meaning-making. First, it must be acknowledged
688
689
690

re i
that in general, more stability than change would be expected, especially among adult

v
students. Second, sometime is needed for this re-organization to take place. According to the
literature on developmental turning points (Rnk, Oravala y Pulkkinnen, 2002), as they are

In
691 associated with changes in trajectories and internal reorganizations, sometime is needed to
692 process these changes. Regardless of the timeframe of the nature of the trajectories studied
693 (as highlighted by Litchtwarck-Asschoff, Van Geert, Bosma, & Kunnen, 2008, for example),
694 which could amount to years (as in identity changes, e.g. Stevens, 2012), months (in the case
695 of beliefs, attitudes, values, or commitment orientations, as in Kunnen, Sappa, van Geert, &
696 Bonica, 2008) or even days (for specific patterns of behavior, creation of habits, and
697 strategies, as in Siegler, 2006), turning points require some time to take place. We could
698 therefore expect a low frequency of turning points.
699
700 Nevertheless, the number of exceptional points in the participants emotional trajectories was
701 quite high. There were 142 exceptional points in the 242 emotional trajectories analyzed, of
702 which 70% were positively trended. Exceptional points provide also relevant information
703 about the predominant shape of the emotional trajectories during the course. In this study, we
704 have focused on the changes in the students emotional trajectories which were statistically
705 significant, i.e. on turning points. However, , as evidenced in the amount and quality of the
706 exceptional points, the emotional orientation of most of the participants trajectories is similar
707 to the one signaled by the most common types of turning points: a positive orientation
708 associated with points in time during the training which the participants perceived as
709 particularly complex.
710
711 5. Educational implications
712
713 Educational interventions that challenge individuals ways of meaning-making are usually
714 associated with emotional arousal (Apte, 2009; Antonacopoulou & Gabriel, 2001; Cranton,
715 2002; Kegan, 1994; McEwen, Strachan & Lynch, 2010). A detailed tracking of students
716 emotional trajectories over time in search of transition points marked by discontinuities the
717 turning points in our study could therefore enable teachers to identify periods in training

15
718 that might mobilize or inhibit students adaptation and learning (see Hayes et al., 2007 for the
719 same argument in the context of therapy). Similarly, a skilled teacher should take into
720 account the periods when students are more open to change, in order to provide appropriate
721 new input that destabilizes their current ways of meaning-making and facilitates the
722 emergence of more adapted and complex ones (see Seligman, 2005 and Thelen, 2005 for this
723 argument in therapy contexts).
724
725 We therefore believe that learning contexts can be structured to support and guide students
726 towards accommodation processes (for a similar argument referring to the role of therapy, see
727 Kunnen & Wassink, 2003), so that they become deliberately developmental learning contexts
728 (Kegan & Lahey, 2016). An endeavor for teachers in developmental learning contexts is to
729 undertake a careful follow-up of how learners emotionally respond to the new demands
730 during the training. Teachers must acknowledge that students may feel anxious, insecure or
731 overwhelmed when immersed for the first time in a learning context which no longer
732 provides them with the guidance they are used to. Awareness of the intra-individual
733 variability of students, as evidenced in different emotional responses to a challenging
734 learning context, is equally important in supporting individuals learning. The same group
735 may contain students who are able and willing to explore and enjoy new learning approaches,

w
736 and students who are reluctant and afraid to do so. An issue to bear in mind is how open

e
737 individuals are when responding to new inputs, i.e. how dominant and strong their patterns of
738
739
740

re i
meaning-making are, and how they interact with the new situation (Thelen, 2005). Kunnen

v
and Bosma (2000) argue that individuals differ both in their preference for either
accommodation or assimilation, and in their skills to apply these in a satisfactory manner,

In
741 which leads to different learning and developmental trajectories. Accordingly, we predict that
742 on the one hand, students with a greater preference for accommodation would be expected to
743 present more turning points in their emotional trajectories, and particularly positively-trended
744 turning points. On the other hand, students with greater preference for assimilation would be
745 expected to present a more stable emotional experience, i.e. fewer turning points, or none.
746
747 6. Limitations
748
749 This study has three main limitations: (1) the number and distribution of measurements; (2)
750 the grouping of emotions into positive and negative; (3) the specificity of the sample.
751
752 As regards the number of measurements, the shortcoming involved in asking the students to
753 assess their emotions at a few points in time is that some of their emotional fluctuations
754 during the training was inevitably lost. This limitation is inherent in the study of any
755 developmental process. Although the questionnaires were strategically distributed in order to
756 increase the probability of capturing expected emotional upsets, the students could have
757 experienced other upsets which they may have overcome by the time they completed the
758 questionnaire. One possible way of recording these would entail asking the students at the
759 end of every training session about the emotions they experienced most strongly, when they
760 felt them and what they were related to.
761
762 It could also be argued that the distribution of measurements during the training course
763 studies was not consistent. We agree that this is the case, but this was intentional since the
764 goal was to obtain data from the students after the periods of the course that we expected
765 would be most challenging for them. We therefore do not consider the issue of whether the
766 follow-up questionnaires were filled in after a phase in which students had been sharing and
767 reflecting on an experiential exercise, or the students had been doing an exploratory exercise,

16
768 or the trainer had been giving instructions for the development of an exercise, to be a
769 problem.
770
771 We acknowledge that the grouping of emotions into positive and negative emotions is an
772 oversimplification of the students emotional experience, which might lead to nuances in the
773 patterns of discrete emotions being overlooked. However, considering the aim of this study,
774 we found that valence sufficed for grasping the main emotional patterns.
775
776 Finally, it can be argued that the specificity of the sample may mean that the generalization of
777 the results is questionable. It is important to note that our aim was not to generalize these
778 results to other samples, but rather to examine a training course in which a process-oriented
779 experiential learning model was implemented. In fact, the trainer of the course, John
780 McWhirter, was the developer of the experiential learning model applied. Nevertheless,
781 further research with different samples and in different learning settings is recommendable in
782 order to explore possible differences in how students manage complexity.
783
784 7. Directions for Future Research
785

w
786 This study can lead to the formulation of new research questions and hypotheses about how

e
787 individuals manage contextual complexity. On the one hand, it is a first step in reviewing
788
789
790
i
some of the traditional ideas about learning, such as the assumption that contextual

v
complexity always leads to individuals learning.

re
In
791 It also opens the door to further research on cognitive accommodation. This would require
792 the inclusion of measures of students learning performance. To do so, a learning setting in
793 which performance could be followed over a much longer time span several weeks at least-
794 after potential turning points in students emotional trajectories would be needed. Time is
795 required before increases in individuals performance become visible. In this study it was
796 therefore not useful to include performance measures. From a systems perspective,
797 transformations have to resettle, and often, there may be a short dip in performance shortly
798 after the transformation, due to the system having to reorganize and getting used to new
799 patterns.
800
801 The collection of qualitative data both throughout the learning experience and at the end of it
802 is essential in an in-depth investigation on how different students make meaning of the
803 challenging demands arising from a process oriented experiential learning model. The
804 collection of time series data on variables such as the duration of the emotions experienced
805 and the degree of challenge and support perceived during the training are also important for
806 gaining greater insight into students experience during a training course. This latter would
807 provide more information than the variable of complexity used in this study.
808
809 Additionally, investigating traditional learning settings could show us whether the patterns in
810 participants emotional trajectories differ from those found in a process-oriented experiential
811 learning context. We anticipate that there would be no turning points in students emotional
812 trajectories, which would present a mostly flat shape. This would demonstrate that training
813 based on didactic learning does not upset students and that it is therefore not a supportive
814 context in terms of enhancing students adaptation to complex contexts and providing greater
815 opportunities for cognitive accommodation.
816

17
817 8. Conclusion
818
819 Using a dynamic systems approach, this study examined the emotional trajectories of the
820 participants in an experiential learning course in order to investigate how these individuals
821 managed conflicting training demands. The most frequent types of turning points identified -
822 pre-increases in positive emotion trajectories and pre-decreases in negative emotion
823 trajectories- coincided with particularly complex periods in the training. The positive trend
824 found in emotional trajectories is taken as evidence of the participants adapted responses to
825 the demands of the challenging learning approach.
826
827 These findings are consistent with the ideas that development and learning occur in response
828 to conflicting demands that challenge individuals ways of meaning-making, and that
829 cognitive conflict is usually associated with the experience of negative emotions. However,
830 our main contribution is providing empirical proof of these former claims, by establishing a
831 dynamic systems procedure that identifies turning points in developmental trajectories. This
832 procedure offers an alternative to standard research methods focusing on the production of
833 development and which overlook the process, and confirms that dynamic systems are a very
834 robust approach for achieving this. This procedure could be applied to a wide range of

w
835 longitudinal experiences that are very common in educational and developmental settings.

e
836
837
838
839
i
Our findings also support the idea that contextual complexity alone does not lead to

v
individuals adaptation and learning. Instead, in order to explain this adaptation, we need to

re
consider the way in which individuals manage contextual complexity. The quality and result

In
840 of this management can be influenced by several factors, such as the particular individuals
841 self-management baseline, and the timing when the contextual complexity is faced in terms
842 of the individuals resources to manage it.
843
844 Finally, this study confirms that the process-oriented experiential learning model is
845 particularly useful for supporting students in the process of managing contextual complexity.
846 It provides students with a wide range of experiential variation and demands they engage in a
847 different way of learning, in which exploration, curiosity and uncertainty are key factors. This
848 is a potential opportunity for individuals to revise their previous ways of learning, which are
849 usually rooted in didactic learning settings, and opens up the path to becoming more
850 autonomous and creative learners.
851

18
852 References
853
854 Antonacopoulou, E. P., & Gabriel, Y. (2001). Emotion, learning and organizational change:
855 Towards an integration of psychoanalytic and other perspectives. Journal of
856 Organizational Change Management, 14(5), 435-451.
857 http://dx.doi.org/10.1108/EUM0000000005874
858 Abbott, A. (1997). On the concept of turning point. Comparative social research, 16, 85-106.
859 Apte, J. (2009). Facilitating transformative learning: A framework for practice. Australian
860 Journal of Adult Learning, 49(1), 169189.
861 Arpiainen, R.-L., Lackus, M., Tks, M., & Tynjl, P. (2013). The Sources and Dynamics of
862 Emotions in Entrepreneurship Education Learning Process. Trames. Journal of the
863 Humanities and Social Sciences, 17(4), 331-346. http://dx.doi.org/10.3176/tr.2013.4.02
864 Barrett, L. F. (2009). Variety is the spice of life: A psychological construction approach to
865 understanding variability in emotion. Cognition & Emotion, 23(7), 12841306.
866 http://doi.org/10.1080/02699930902985894
867 Bosma, H. A., & Kunnen, S. (2001). Determinants and Mechanisms in Ego Identity
868 Development: A Review and Synthesis. Developmental Review, 21(1), 3966.
869 http://doi.org/10.1006/drev.2000.0514

w
870 Bradley, M.M. (2009). Natural selective attention: orienting and emotion. Psychophysiology,

e
871 46(1), 111. http://dx.doi.org/10.1111/j.1469-8986.2008.00702.x
872
873
874

re i
Camras, L. A., & Witherington, D. C. (2005). Dynamical systems approaches to emotional

v
development. Developmental Review, 25(3-4), 328350.
http://dx.doi.org/10.1016/j.dr.2005.10.002

In
875 Carver, C. S. & Scheier, M. F. (1990). Origins and functions of positive and negative affect: a
876 controlprocess view. Psychological Review 97(1), 1935. http://dx.doi.org/10.1037/0033-
877 295X.97.1.19
878 Cranton, P. (2002). Teaching for Transformation. New Directions for Adult and Continuing
879 Education (93), 6372. http://doi.org/10.1002/ace.50
880 Dirkx, J. M. (2008). The Meaning and Role of Emotions in Adult Learning. Emotion,
881 2008(120), 718. http://dx.doi.org/10.1002/ace.311
882 Dirkx, J. M. (2011). Romancing Tales From the Dark Side: Crisis, Emotion, and the
883 Construction of Meaning in Transformative Learning. In 9th International Transformative
884 Learning Conference - Athens 2011 - Book of Abstracts (pp. 7681).
885 Dirkx, J. M., Mezirow, J., & Cranton, P. (2006). Musings and Reflections on the Meaning,
886 Context, and Process of Transformative Learning: A Dialogue Between John M. Dirkx
887 and Jack Mezirow. Journal of transformative education, 4(2), 123-139.
888 http://dx.doi.org/10.1177/1541344606287503
889 D'Mello, S. K., Craig, S. D., & Graesser, A. C. (2009). Multimethod assessment of affective
890 experience and expression during deep learning. International Journal of Learning
891 Technology, 4(3/4), 165-187. http://dx.doi.org/10.1504/IJLT.2009.028805
892 D'Mello, S., & Graesser, A. (2011). The half-life of cognitive-affective states during complex
893 learning. Cognition & Emotion, 25(7), 12991308.
894 http://doi.org/10.1080/02699931.2011.613668
895 D'Mello, S., Lehman, B., Pekrun, R., & Graesser, A. (2014). Confusion can be beneficial for
896 learning. Learning and Instruction, 29, 153170.
897 http://dx.doi.org/10.1016/j.learninstruc.2012.05.003
898 Efron, B., & Tibshirani, R. J. (1993). An Introduction to the Bootstrap. New York and New
899 York: Chapman and Hall.
900 Eubanks-Carter, C., Gorman, B. S., & Muran, J. C. (2012). Quantitative naturalistic methods
901 for detecting change points in psychotherapy research: An illustration with alliance

19
902 ruptures. Psychotherapy Research, 22(6), 621-637.
903 http://dx.doi.org/10.1080/10503307.2012.693772
904 Fogel, A. (2011). Theoretical and Applied Dynamic Systems Research in Developmental
905 Science. Child Development Perspectives, 5(4), 267272.
906 http://dx.doi.org/10.1111/j.1750-8606.2011.00174.x
907 Fogel, A., Nwokah, E., Young Dedo, J., Messinger, D., Dickson, L. K., Matusov, E., & Holt,
908 S. A. (1992). Social process theory of emotion: a dynamic systems approach. Social
909 Development, 1(2), 122-142. http://dx.doi.org/10.1111/j.1467-9507.1992.tb00116.x
910 Frijda, N. H. (1986). The emotions: Studies in emotion and social interaction. Cambridge:
911 Cambridge University Press.
912 Frijda, N. H. (1988). The laws of emotion. The American Psychologist, 43(5), 349358.
913 http://dx.doi.org/10.1037/0003-066X.43.5.349
914 Frijda, N. H. (2009). Emotions, individual differences and time course: Reflections.
915 Cognition & Emotion, 23(7), 14441461. http://doi.org/10.1080/02699930903093276
916 Good, P. (1999). Resampling methods: A practical guide to data analysis. Boston:
917 Birkhauser.
918 Granott, N., Fischer, K. W., & Parziale, J. (2002). Bridging to the Unknown: A Transition
919 Mechanism in Learning and Development. In N. Granott and J. Parziale (Eds.),

w
920 Microdevelopment: Transition processes in development and learning (pp. 333). New

e
921 York: Cambridge University Press. http://dx.doi.org/10.1017/CBO9780511489709.006
922
923
924

re i
Hayes, A. M., Laurenceau, J. P., Feldman, G., Strauss, J. L., & Cardaciotto, L. (2007).

v
Change is not always linear: The study of nonlinear and discontinuous patterns of change
in psychotherapy. Clinical Psychology Review, 27(6), 715723.

In
925 http://dx.doi.org/10.1016/j.cpr.2007.01.008
926 Hayes, A. M., Yasinski, C., Ben Barnes, J., & Bockting, C. L. H. (2015). Network
927 destabilization and transition in depression: New methods for studying the dynamics of
928 therapeutic change. Clinical Psychology Review, 41, 2739.
929 http://doi.org/10.1016/j.cpr.2015.06.007
930 Hood, G. M. (2010) PopTools version 3.2.5. Available on the internet. URL
931 http://www.poptools.org
932 Inzlicht, M., Bartholow, B. D., & Hirsh, J. B. (2015). Emotional foundations of cognitive
933 control. Trends in Cognitive Sciences, 19(3), 126-132.
934 http://dx.doi.org/10.1016/j.tics.2015.01.004
935 Kegan, R. (1994). In over our heads: The mental demands of modern life. Cambridge, MA:
936 Harvard University Press.
937 Kegan, R., & Lahey, L. L. (2016). An Everyone Culture: Becoming a Deliberately
938 Developmental Organization. Harvard Business Review Press.
939 Kloep, M., Hendry, L. B., & Saunders, D. (2009). A new perspective on human
940 development. Conference of the International Journal of Arts and Sciences, 1(6), 332
941 343.
942 Kolb, D. A. (1984). Experiential Learning: Experience as the source of learning and
943 development. Englewood Cliffs, N. J.: Prentice-Hall.
944 Kolb, A., & Kolb, D. (2009). Experiential learning theory: A dynamic, holistic approach to
945 management learning, education and development. In Armstrong, S. J. & Fukami, C.
946 (Eds.) The SAGE Handbook of Management, Learning, Education and Development (pp.
947 1-59). London: Sage Publications. http://dx.doi.org/10.4135/9780857021038.n3
948 Kroger, J. (1993). Discussions on ego identity. Hillsdale, NJ: Erlbaum.
949 Kunnen, S. (2006). Are Conflicts the Motor in Identity Change? Identity, 6(2), 169186.
950 http://dx.doi.org/10.1207/s1532706xid0602_3

20
951 Kunnen, S. (2012). A Dynamic Systems Approach to Adolescent Development. London:
952 Psychology Press.
953 Kunnen, S., & Bosma, H. A. (2000). Development of meaning making: A dynamic systems
954 approach. New Ideas in Psychology, 18(1), 5782. doi:10.1016/S0732-118X(99)00037-9
955 Kunnen, E. S., Sappa, V., van Geert, P. L., & Bonica, L. (2008). The shapes of commitment
956 development in emerging adulthood. Journal of Adult Development, 15(3-4), 113-131.
957 Kunnen, S., Van Dijk, M., Lichtwarck-Aschoff, A.,Visser, M., & Van Geert, P. (2012). The
958 search for variability and change. In S. Kunnen (2012), The search for variability and
959 change in A Dynamic Systems Approach of Adolescent Development (pp. 53-71). London:
960 Psychology Press.
961 Kunnen, S., & Van Geert, P. (2012). General characteristics of a dynamic systems approach.
962 In S. Kunnen (2012), A Dynamic Systems Approach of Adolescent Development (pp. 15-
963 34). London: Psychology Press.
964 Kunnen, S., & Wassink, M. E. K. (2003). An Analysis of Identity Change in Adulthood.
965 Identity, 3(4), 347-366. http://dx.doi.org/10.1207/S1532706XID0304_03
966 Kuppens, P., Oravecz, Z., & Tuerlinckx, F. (2010). Feelings change: accounting for
967 individual differences in the temporal dynamics of affect. Journal of Personality and
968 Social Psychology, 99(6), 10421060. http://dx.doi.org/10.1037/a0020962

w
969 Laevers, F. (2000). Forward to Basics! Deep-Level-Learning and the Experiential Approach.

e
970 Early Years: An International Research Journal, 20(2), 2029.
971
972
973

re i
http://dx.doi.org/10.1080/0957514000200203

v
Larsen, R. J., Augustine, A. A., & Prizmic, Z. (2009). A process approach to emotion and
personality: Using time as a facet of data. Cognition and Emotion, 23(7), 14071426.

In
974 http://dx.doi.org/10.1080/02699930902851302
975 Lazarus, R. S. (1991). Emotion and Adaptation. New York: Oxford University Press.
976 Lepper, M. R., & Henderlong, J. (2000). Turning play into work: 25 years of research on
977 intrinsic versus extrinsic motivation. In C. Sansone and J. M. Harackiewicz (Eds.),
978 Intrinsic and extrinsic motivation: The search for optimal motivation and performance
979 (pp. 257307). San Diego, CA: Academic Press. http://dx.doi.org/10.1016/B978-
980 012619070-0/50032-5
981 Lewis, M. D. (2005). Bridging emotion theory and neurobiology through dynamic systems
982 modeling. Behavioral and Brain Sciences, 28(2), 169194.
983 http://dx.doi.org/10.1017/S0140525X0500004X
984 Lichtwarck-Aschoff, A., Kunnen, S., & Van Geert, P. (2009). Here we go again: a dynamic
985 systems perspective on emotional rigidity across parent-adolescent conflicts.
986 Developmental Psychology, 45(5), 13641375. http://dx.doi.org/10.1037/a0016713
987 Lichtwarck-Aschoff, A., van Geert, P., Bosma, H., & Kunnen, S. (2008). Time and identity:
988 A framework for research and theory formation. Developmental Review, 28(3), 370-400.
989 McEwen, L., Strachan, G., & Lynch, K. (2010). Shock and Awe or Reflection and
990 Change: stakeholder perceptions of transformative learning in higher education. Learning
991 and Teaching in Higher Education, 5, 3455. Available at http://eprints.uwe.ac.uk/20326
992 McWhirter, J. (1998). Re-Modelling NLP Part One: Models and Modelling. Retrieved from
993 http://sensorysystems.co.uk/sst/dbmremodellednlp/re-modelling-nlp-articles-in-pdf-
994 format/
995 McWhirter, J. (2000). Re-Modelling NLP. Part Six: Understanding Change. Retrieved from
996 http://sensorysystems.co.uk/wp-content/uploads/2012/07/RemodellingNLPPart6.pdf
997 McWhirter, J. (2002). Re-modelling NLP. Part Fourteen: Re-Modelling Modelling. Retrieved
998 from http://sensorysystems.co.uk/wp-
999 content/uploads/2012/07/RemodellingNLPPart14.pdf

21
1000 McWhirter, J. (2011) Behavioral remodeling. In Hall, L. M., Charvet, S. R., & Rose-Charvet,
1001 S. (Eds.) Innovations in NLP: for challenging times (pp. 95-114). Crown House
1002 Publishing.
1003 Molenaar, P. C. M. (2004). A Manifesto on Psychology as Idiographic Science: Bringing the
1004 Person Back Into Scientific Psychology, This Time Forever. Measurement:
1005 Interdisciplinary Research Perspective, 2(4), 201218.
1006 http://dx.doi.org/10.1207/s15366359mea0204_1
1007 Noddings, N. (2003). Happiness and education. Cambridge: Cambridge University Press.
1008 Nogueiras, G., & Iborra, A. (2016). Understanding and promoting self-direction in freshman
1009 and masters students: a qualitative approach. Behavioral Development Bulletin. Advance
1010 Online Publication. http://dx.doi.org/10.1037/bdb0000024
1011 Nogueiras, G., Herrero, D., & Iborra, A. (2016). Teaching for Epistemological Change: Self-
1012 direction through Self-assessment. In Cano, E., Ion, G. (Eds.), Innovative Practices for
1013 Higher Education Assessment and Measurement (pp. 207-225). Hershey, Pennsylvania:
1014 IGI Global. http://dx.doi.org/10.4018/978-1-5225-0531-0.ch011
1015 Overton, W. F. (2006). Developmental psychology: Philosophy, concepts, methodology. In
1016 W. Damon & R. M. Lerner (Series Eds.) & R. M. Lerner (Vol. Ed.), Handbook of child
1017 psychology: Vol. 1. Theoretical models of human development (6th ed., pp. 1888).

w
1018 Hoboken, NJ: Wiley. http://dx.doi.org/10.1002/9780470147658.chpsy0102

e
1019 Piaget, J. (1952). The origins of intelligence in children. New York: International Universities
1020
1021
1022
Press.

re vi
Piaget, J. (1985). The equilibrium of cognitive structures. Chicago: University of Chicago
Press. (Original work published 1975).

In
1023 Rnk, A., Oravala, S., & Pulkkinen, L. (2002). I met this wife of mine and things got onto
1024 a better track Turning points in risk development. Journal of adolescence, 25(1), 47-
1025 63.
1026 Sansone, C., and Thoman, D. B. (2005). Does what we feel affect what we learn? Some
1027 answers and new questions. Learning and Instruction, 15(5), 507515. Retrieved from
1028 http://dx.doi.org/10.1016/j.learninstruc.2005.07.015
1029 Scherer, K.R. (2000). Psychological models of emotion. In Joan C. Borod (Ed.), The
1030 Neuropsychology of Emotion (pp. 137162). New York: Oxford University Press.
1031 Scherer, K. R. (2009). The dynamic architecture of emotion: Evidence for the component
1032 process model. Cognition & Emotion, 23(7), 13071351.
1033 http://dx.doi.org/10.1080/02699930902928969
1034 Schwarz, N. (1990). Feelings as information: Informational and motivational functions of
1035 affective states. In E. T. Higgins & R.M. Sorrentino (Eds.), Handbook of motivation and
1036 cognition: Foundations of social behavior, Vol. 2 (pp. 528561). New York: Guilford.
1037 Siegler, R. S. (2006). Microgenetic Analyses of Learning. In W. Damon and R. M. Lerner
1038 (Eds.), Handbook of child psychology (pp. 464-510). New York: John Wiley and Sons.
1039 http://dx.doi.org/10.1002/9780470147658.chpsy0211
1040 Solomon, R. C. (2007). True to our feelings: What our emotions are really telling us. Oxford:
1041 Oxford University Press.
1042 Stevens, A. (2012). I am the person now I was always meant to be: Identity reconstruction
1043 and narrative reframing in therapeutic community prisons. Criminology and Criminal
1044 Justice, 12(5), 527-547.
1045 Taylor, E. W., & Cranton, P. (2013). A theory in progress? Issues in transformative learning
1046 theory. European Journal for Research on the Education and Learning of Adults, 4(1),
1047 3347. http://dx.doi.org/10.3384/rela.2000-7426.rela5000

22
1048 Thelen, E. (1989). Self-organization in developmental processes: Can systems approaches
1049 work? In M. R. Gunnar and E. Thelen (Eds.), Systems and development (pp. 77107). New
1050 York: Lawrence Erlbaum Associates.
1051 Thelen, E. (2005). Dynamic Systems Theory and the Complexity of Change. Psychoanalytic
1052 Dialogues: The International Journal of Relational Perspectives, 15(2), 255283.
1053 http://dx.doi.org/10.1080/10481881509348831
1054 Thelen, E., & Smith, L.B. (1994). A dynamic systems approach to the development of
1055 cognition and action. Cambridge, MA: MIT Press.
1056 Thelen, E., & Ulrich, B. D. (1991). Hidden skills: a dynamic systems analysis of treadmill
1057 stepping during the first year. Monographs of the Society for Research in Child
1058 Development, 56(1), 1-98, 99-103. http://dx.doi.org/10.2307/1166099
1059 Todman, J.B. & Dugard, P. (2001). Single-Case and Small-N Experimental Designs: A
1060 Practical Guide To Randomization Tests. Lawrence Erlbaum Associates: Mahwah, NJ,
1061 USA.
1062 Vallacher, R. R., Van Geert, P. L. C., & Nowak, A. (2015). The intrinsic dynamics of
1063 psychological process. Current Directions in Psychological Science, 24(1), 58-64.
1064 http://dx.doi.org/10.1177/0963721414551571
1065 Van der Maas, H. L. J., & Molenaar, P. C. (1992). Stagewise cognitive development: an

w
1066 application of catastrophe theory. Psychological Review, 99(3), 395417.

e
1067 http://dx.doi.org/10.1037/0033-295X.99.3.395
1068
1069
1070

re vi
Van Dijk, M., & Van Geert, P. (2007). Wobbles, Humps and Sudden Jumps: A Case Study
of Continuity, Discontinuity and Variability in Early Language Development. Infant and
Child Development, 16(1), 733. http://dx.doi.org/10.1002/icd.506

In
1071 Van Dijk, M., & van Geert, P. (2011). Heuristic techniques for the analysis of variability as a
1072 dynamic aspect of change. Infancia Y Aprendizaje, 34(2), 151167.
1073 http://dx.doi.org/10.1174/021037011795377557
1074 Van Geert, P. (1994). Dynamic systems of development: Change between complexity and
1075 chaos. New York and London: Harvester Wheatsheaf.
1076 Van Geert, P., Steenbeek, H., & Kunnen, S. (2012). Monte Carlo techniques. Statistical
1077 simulation for developmental data. In S. Kunnen (Ed.), A Dynamic Systems Approach to
1078 Adolescent Development (pp. 4351). London: Psychology Press.
1079 Van Geert, P., & Van Dijk, M. (2002). Focus on variability: New tools to study intra-
1080 individual variability in developmental data. Infant Behavior and Development, 25(4),
1081 340374. http://dx.doi.org/10.1016/S0163-6383(02)00140-6
1082 Van Geert, P. & Van Dijk, M. (2015). The intrinsic dynamics of development. In R.A. Scott
1083 and S. M. Kosslyn (Eds.) Emerging Trends in the Social and Behavioral Sciences: An
1084 Interdisciplinary, Searchable, and Linkable Resource. New York: Sage Publications.
1085 http://dx.doi.org/10.1002/9781118900772.etrds0343
1086 Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief
1087 measures of positive and negative affect: the PANAS scales. Journal of Personality and
1088 Social Psychology, 54(6), 10631070. http://dx.doi.org/10.1037/0022-3514.54.6.1063
1089 Weiss, P. R. (2000). Emotion and learning. Training and Development, 54(11), 4448.
1090 http://eric.ed.gov/?id=EJ615980
1091 Witherington, D. C. (2007). The Dynamic Systems Approach as Metatheory for
1092 Developmental Psychology. Human Development, 50, 127-153.
1093 http://dx.doi.org/10.1159/000100943
1094 Yan, Z., & Fischer, K. (2007). Pattern Emergence and Pattern Transition in
1095 Microdevelopmental Variation: Evidence of Complex Dynamics of Developmental
1096 Processes. Journal of Developmental Processes, 2(2), 3962.
1097

23
1098 Tables
1099
1100 Table 1 Example of the computation of the confidence interval for the scores in the
1101 intensity of distress of one of the participants. Data associated to the exceptional point
1102 (scores in distress, UCL and LCL), located in m.p.5, are boldfaced.
1103
Confidence Interval
Scores in
Regression line Variance (data- LCL (regression line 1,65 )
m.p. distress UCL (regression line + 1,65 )
(intercept + slope) * data regression line) ^ 2
(17)
1 2 1.666 0.111 2.754 0.579
2 1 1.690 0.477 2.778 0.603
3 1 1.714 0.510 2.802 0.627
4 2 1.738 0.068 2.826 0.650
5 3 1.762 1.533 2.849 0.674
6 2 1.787 0.046 2.873 0.698
7 2 1.809 0.0363 2.897 0.722
8 1 1.833 0.694 2.921 0.746
Average variance 0.434
Standard deviation () 0.659
1104

1105
1106

i ew
Table 2 Amount and percentage (rounded) of high and low exceptional points.

v
re
Type of exceptional Exceptional points in positive emotions Exceptional points in negative emotions
points trajectories trajectories

In
Percentage in Percentage in the Percentage in Percentage in the
No. positive emotions total no. of No. negative emotions total no. of
trajectories trajectories trajectories trajectories
High exceptional points 20 38% 14% 67 75% 47%
Low exceptional points 33 62% 23% 22 25% 16%
Total 53 100% 89 100%
1107
1108 Table 3 Amount and percentage (rounded) of pre-decrease and pre-increase turning points.
1109
Type of turning
Turning points in positive emotions trajectories Turning points in negative emotions trajectories
points
Percentage in Percentage in the Percentage in Percentage in the
No. positive emotions total no. of No. negative emotions total no. of
trajectories trajectories trajectories trajectories
Pre-decrease
2 17% 7% 15 83% 50%
turning points
Pre-increase
10 83% 33% 3 17% 10%
turning points
Total 12 100% 18 100%
1110
1111
1112 Table 4 Average and range of complexity scores for each measurement point.
1113
Measurement points
1 2 3 4 5 6 7 8
Average
2.450 2.980 2.996 3.245 3.176 3.086 3.133 3.279
complexity scores
Range complexity
1.5 - 5 1.5 4 2-5 2-5 2.5 5 1.5 - 5 2.5 5 2-5
scores

24
1114
1115 Table 5 Significant differences in complexity scores between m.p. in which turning points
1116 (t.p.) were identified and m.p. in which no turning points were identified. Note: * p < .05. **
1117 p < .01
1118
Types of turning Scores in complexity (scale 1 to 5) Difference between Scores in
points mean scores in complexity higher
M.p. with t.p. M.p. without t.p. complexity of m.p. in m.p. with t.p.
with and without t.p. than in m.p.
Mean Std dev Mean Std dev without t.p. (p)

All 3.731 0.629 3.162 0.78 0.569 .002**

Pre-decrease t.p. in
3.250 0.25 3.231 0.793 0.019 .562
positive emotions

Pre-increase t.p. in
4.250 0.433 3.201 0.776 1.049 .003**
positive emotions

Pre-decrease t.p. in
3.889 0.489 3.191 0.783 0.698 .001**
negative emotions

Pre-increase t.p. in
3.50 0.25 3.231 0.793 0.269 .533

w
negative emotions

e
1119
1120
1121

1122

re vi
In

25
1123 Figures
5

Intensity of the emoiton

4

Distress (17)
3
Regression line

2 UCL
LCL
1
1 2 3 4 5 6 7 8
Measurement points
1124
1125 Fig.1 Example of an exceptional point (m.p.5).
1126 UCL = Upper Control Limit; LCL = Lower Control Limit
1127
5
Intensity of emotion

w
4

i e
Enthusiasm (4)
3

v
Regression line

re
2 UCL

In
LCL
1
1 2 3 4 5 6 7 8
Measurement points
1128
1129 Fig. 2 Example of a pre-decrease turning point (m.p.6) in a positive emotion trajectory
1130 UCL = Upper Control Limit; LCL = Lower Control Limit
1131
5
Intensity of emotion

Joy (14)
3
Regression line

2 UCL
LCL
1
1 2 3 4 5 6 7 8
Measurement points
1132
1133 Fig. 3 Example of a pre-increase turning point (m.p.6) in a positive emotion trajectory
1134 UCL = Upper Control Limit; LCL = Lower Control Limit
1135

26
 5

Intensity of emotion
4

Overload (5)
3
Regression line
UCL
2
LCL

1
1 2 3 4 5 6 7 8
Measurement points
1136
1137 Fig. 4 Example of a pre-decrease turning point (m.p. 4) in a negative emotion trajectory
1138 UCL = Upper Control Limit; LCL = Lower Control Limit
1139
5
Intensity of emotion

vi ew Distress (3)
Regression line

re
2 UCL
LCL

In
1
1 2 3 4 5 6 7 8
Measurement points
1140
1141 Fig. 5 Example of a pre-increase turning point (m.p. 3) in a negative emotion trajectory
1142 UCL = Upper Control Limit; LCL = Lower Control Limit
1143
1144

1145

27
1146
1147
1148 Appendix 1 Overview of the timeframe of the training course. The follow-up questionnaires
1149 are boldfaced.
1150
Day 1
Course introduction
Trainers sets students up
and then presents
instructions for exercise 1
Students do exercise 1
Trainers set-up and
instructions for exercise 2
10.00 Students do exercise 2
11.30
Students sharing and
reflection (set-down,
consolidating learning)
Trainers set-up and
instructions for exercise 3
Students do Exercise 3
Trainers set-up and
Appendix
Day 2
Students revision
Students sharing and
reflection (set-down,
consolidating learning)
Trainers set-up and
instructions for exercise 1
Students do exercise 1
Trainers set-up and
instructions for exercise 2
Students do exercise 2
Trainers set-up and
instructions for exercise 3
Students do exercise 3
Day 3 Day 4 Day 5
Students revision Students revision Students revision
Trainers set-up and Students sharing and Trainers set-up and
instructions for exercise 1 reflection (set-down, instructions for exercise 1
consolidating learning)
Questionnaire 5 Questionnaire 7 Students do exercise 1
Students do exercise 1 Trainers set-up and Trainers set-up and
instructions for exercises1 instructions for exercise 2
and 2
Trainers set-up and Students do exercises 1 Students do exercise 2
instructions for exercise 2 and 2
(for the break)
Trainers set-up and
instructions for exercise 3
Students do exercise 3

w
instructions for exercise 4

e
(for the break)

i
Coffee break

v
Students sharing and Trainers set-up and Trainers set-up and Trainers set-up and Trainers set-up and

re
reflection (set-down, instructions for exercise 4 instructions for exercise 3 instructions for exercise 4 instructions for exercise 3
consolidating learning)
Trainers set-up and

In
Students do exercise 4 Students do exercise 3 Students do exercise 4 Students do exercise 3
instructions for exercise 5
Students do exercise 5 Students sharing and Trainers set-up and Trainers set-up and Trainers set-up and
reflection (set-down, instructions for exercise 4 instructions for exercise 5 instructions for exercise 4
12.00
consolidating learning)

14.00 Trainers set-up and Trainers set-up and Students do exercise 4 Students do exercise 5 Students do exercise 4
instructions for exercise 6 instructions for exercise 5
Students do exercise 6 Students do exercise 5 Students sharing and Students sharing and
reflection (set-down, reflection (set-down,
consolidating learning) consolidating learning)
Students sharing and Students sharing and Trainers set-up and
reflection (set-down, reflection (set-down, instructions for exercise 5
consolidating learning) consolidating learning)
Questionnaire 1 Questionnaire 3 Students do exercise 5
Students sharing and
reflection (set-down,
consolidating learning)
Trainers set-up and
instructions for individual
exercise
Lunch break End of the course
Trainers set-up and Trainers set-up and Trainers set-up and Trainers set-up and
instructions for exercise 7 instructions for exercise 6 instructions for exercise 5 instructions for exercise
5.2
Students do exercise 7 Students do exercise 6 Students do exercise 5 Students do exercise 5.2
Students sharing and Students sharing and Trainers set-up and Trainers set-up and
reflection (set-down, reflection (set-down, instructions for exercise 6 instructions for exercise 6
consolidating learning) consolidating learning)
16.00 Trainers set-up and Trainers set-up and Students do exercise 6 Students do exercise 6
instructions for exercise 8 instructions for exercise 7
18.00 Students do exercise 8 Students do exercise 7 Students sharing and Students sharing and
reflection (set-down, reflection (set-down,
consolidating learning) consolidating learning)
Students sharing and Questionnaire 4
reflection (set-down,
consolidating learning)
Trainers set-up and
instructions for exercise 9
(for the break)
Coffee break

28
 Trainers set-up and Students sharing and
instructions for exercise reflection (set-down,
10 consolidating learning)
Students do exercise 10 Trainers set-up and
18.30 instructions for exercise 8
Questionnaire 2 Students do exercise 8
20.00
Students sharing and
reflection (set-down,
consolidating learning)
Trainers set-up and Students sharing and
instructions for exercise 7 reflection (set-down,
consolidating learning)
Students do exercise 7 Students do exercise 7 in
life
Questionnaire 6 Trainers set-up and
instructions for exercise 8
Students do exercise 8
Questionnaire 8

1151
1152
1153

vi ew
In re

29
 w
Appendix 2 Distribution of turning points (t.p.) across participants and across measurement points (m.p.). Pre-increase turning points in positive

e
emotions are shaded light grey and pre-decrease turning points in negative emotions are shaded dark grey.
M.p. /
Types
t.p. t.p.
M.p. 2
Pre-decrease Pre-increase
t.p.

re vit.p.
M.p. 3
Pre-decrease Pre-increase
t.p. t.p.
M.p. 4
Pre-decrease Pre-increase
t.p. t.p.
M.p. 5
Pre-decrease Pre-increase
t.p.
M.p. 6
Pre-decrease
t.p.
Pre-increase
t.p.
M.p. 7
Pre-decrease
t.p.
Pre-increase
t.p.

In
N
Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. Pos. Neg. Total m.p.
Particip.
Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. Emo. no. t.p. with
t. p.
1
2
3 1 1 2 2

4 1 1 1

5 2 2 4 1

6 1 1 1

7
8 2 2 1

9 1 1 1

10 1 1 1

11 3 3 1

12 2 2 1

13
14 2 1 3 6 3

15 2 3 5 1

16 1 1 1

17 1 1
Total no.
t.p. 1 5 3 1 3 4 2 1 1 3 6 30
No.
participa
nts with
t.p. 1 3 2 1 2 2 1 1 1 2 2

30
 w
Appendix 3 Trajectories of complexity scores for each student over the 8 measurement points. Three additional trajectories are displayed: 1) highest

e
complexity score, 2) average scores of complexity, and 3) lowest complexity score.

5.0

re vi Highest complexity score

In
4.5 Average scores complexity
Lowest complexity score
4.0 Student 1
Student 2
3.5
Student 3
Student 4
Scores of complexity

3.0
Student 5
Student 6
2.5
Student 7
Student 8
2.0
Student 9
Student 10
1.5
Student 11
Student 12
1.0
Student 13

0.5
Student 14
Student 15

0.0 Student 16
1 2 3 4 5 6 7 8 Student 17
Measurement points

31
 Figure 1.JPEG

vi ew
In re
 Figure 2.JPEG

vi ew
In re
 Figure 3.JPEG

vi ew
In re
 Figure 4.JPEG

vi ew
In re
 Figure 5.JPEG

vi ew
In re