Content uploaded by Marc Kohaupt
Author content
All content in this area was uploaded by Marc Kohaupt on Jun 27, 2015
Content may be subject to copyright.
48
Comparing Objective and Subjective Methods to Support
Reflective Learning: an Experiment on the Influence on
Affective Aspects
Verónica Rivera-Pelayo and Marc Kohaupt
FZI Research Center for Information Technology, Karlsruhe, Germany
rivera@fzi.de
marc.kohaupt@gmail.com
Abstract. We conducted an experiment to investigate the influence of objective
and subjective methods to capture affective aspects. We used a sensor to measure
participants’ heart rate and a mood map where participants could report their
emotional state. Results showed that the display of the heart rate had more influ-
ence on the self-reported valence dimension than on the arousal dimension. In
this paper, we discuss several ideas why people might act this way and raise those
ideas for discussion. The discussed reasons may change the way how people feel
when they see their heart rate. If some of the reasons apply in reality, this could
change the way we use objective methods to capture and make affective aspects
aware in TEL.
Keywords: objective method, heart rate, sensor, subjective method, self-report,
mood map, arousal, valence.
1 Introduction
In technology enhanced learning, affective aspects are important in several manners:
they can constrain or serve someone while learning [2], they can be cues to remember
events [11, 17], or they can be the subject of learning, as in emotion regulation [10, 12].
Affective aspects like emotional states or physiological reactions can be captured with
self-reporting techniques or with sensors [12, 13,16], but also with a combination of
both. Data captured with self-reported techniques offer a subjective perspective on
emotions, whereas data from sensors can be a reference point that delivers an objective
perspective. Although so far no physiological parameter by itself could represent a spe-
cial emotional or affective state, a vast amount of rese arch has investigated which pa-
rameters may be a cue or indicator of certain emotions [9, 19]. The comparison between
using subjective and objective approaches has also gained attention in scenarios not
related to learning, e.g. to assess physical activity levels [21].
Boud et al. [2] define reflective learning, i.e. learning from own experiences, as
“those intellectual and affective activities in which individuals engage to explore their
experiences in order to lead to new understandings and appreciations”. The learner’s
49
context makes up the objective and environmental experience that can be compared to
the subjective experience of the learner in order to make sense out of it. This compari-
son can lead to discrepancies as subjective and objective experiences differ in their
underlying perception of situations as well as in their interpretation of the gathered data
[15].
To gain insights about the differences between using objective and subjective meth-
ods to capture affective aspects and how they influence the user, we designed and con-
ducted an experiment. In this experiment, we employed a heart rate sensor to measure
participants’ heart rate [3, 16] and a self-reporting tool (mood map [12, 13]) to capture
their emotional state. We chose these two capturing methods in order to have an auto-
matic approach that offers an objective perspective in comparison to a self-reporting
approach which shows the subjective perspective of the user. The experiment aimed at
examining if there are influences on participants when they were shown their current
heart rate and how this influence looks like. After the experiment, they had to remember
certain events which occurred during the experiment. While remembering the events,
the data gathered with both methods were reported to the participants.
2 Capturing Affective Aspects: an Experiment
According to cognitive dissonance theory [4], there is a tendency for individuals to seek
consistency among their cognitions (i.e. beliefs, opinions). When there is a mismatch
or psychological discomfort (dissonance) between attitudes and behaviour, it can lead
to rethinking attitudes and experiences. Cognitive dissonance is also experienced by an
individual when confronted by new information that conflicts with existing beliefs,
ideas or values. This conflict may be originated also by discrepancies between the ob-
jective and the subjective assessments of own emotions. Therefore, our hypotheses are
that the visualization of the individuals’ heart rate (a) could influence them in their self-
report and (b) could act as memory cue (trigger) and make participants remember a
higher number of events.
2.1 Experimental Design
We designed an experiment with two groups (N = 75, Mage = 23.3, SDage = 2.8, 38
female, 37 male, most of them being students). The treatment group (N = 32, Mage =
22.5, SDage = 2.5, 17 female, 15 male) used a self-reporting method to capture their
emotional state (subjective) while they were shown their heart rate measured by the
sensor (objective). Subjects in the control group (N = 43, Mage = 24.0, SDage = 3.0, 21
female, 22 male) had only the subjective perspective, i.e. they used only the self-report-
ing method without seeing their monitored heart rate. For self-reporting, we used an
adaptation of the mood map [12, 13], a tool developed in the MIRROR
8
project to sup-
port reflective learning on mood.
8
http://www.mirror-project.eu
50
The mood map is a two-dimensional affect scale consisting of arousal and valence.
It is based on the Circumplex Model of Affect [23] which defines every affective state
as the linear combination of two dimensions: valence and arousal. The heart rate was
captured by a movisens ekgMove sensor [14] and the history of the last 60 seconds was
displayed to the participants. The heart rate and the mood map were displayed on a
Samsung Galaxy Tab 10.1N tablet computer. The tablet was placed in front of the par-
ticipants so that they could easily input their emotional state without executing consid-
erable movements.
At capturing stage, we showed 12 emotional film clips [6-8, 20, 22, 26] which aimed
at creating several experiences during the experiment, analogue to the experiences that
are the basis for reflective learning. The film clips were extracted from feature films
and contain scenes which target at provoking certain emotions: neutral (All the Presi-
dent's Men, Hannah and her Sisters), fear (Halloween, Silence of Lambs), disgust (Pink
Flamingos, The Godfather), anger (Gandhi, My Bodyguard), sadness (An Officer and
a Gentleman, The Champ) and amusement (When Harry met Sally, An Officer and a
Gentleman). For all film clips, the German dubbing was used. The duration of the film
clips was 2:11 minutes on average and 26:16 minutes in total (with 54 additional sec-
onds after each film clip for mood self-reporting; 10:48 minutes in total). The film clips
were shown on a LCD monitor in front of each participant. To control confounding
variables, the participants were separated by cabins, wore headphones, the ambient light
was switched off and the blinds were lowered.
After each film clip, the participants reported their emotional state using the mood
map (Fig 1). After the 12 film clips were shown, the participants filled in a question-
naire.
Figure 1. App for treatment group at capturing stage with heart rate display (left)
and mood map (right)
Amongst other questions in this questionnaire, the participants had to write down
any events of the film clips they remembered. At this stage, the app changed to the
51
reporting view (Figure 2) and displayed the history of the gathered data, so that partic-
ipants could use it as a cue.
Figure 2. App for treatment group at reporting stage with heart rate history and mood history
from first to last film clip
In the subsequent analysis of the results, the remembered events were checked by
the investigator and a film clip was marked as remembered when at least one event in
the film clip was remembered correctly. The gathered data of the questionnaire and the
data of the mood map for each film clip were analysed with t-tests (normal distributed
values) or Wilcox-Mann-Whitney-Tests (if no normal distribution given).
2.2 Results
No differences regarding the number of remembered film clips were found between the
groups (p = 0.612, control group: M = 7.7, SD = 3.1, treatment group: M = 8.0, SD =
3.2). This suggests that visualizing the heart rate does not increase or decrease the num-
ber of events the participants remember when reflecting on it. Nonetheless, the analysis
of the gathered data revealed a remarkable difference regarding where they placed their
emotional state on the mood map. As the mood map consists of two dimensions (arousal
and valence), we investigated how both dimensions differ between the groups.
The heart rate describes the physiological state of the participants, and the arousal
dimension of the mood map also describes the physiological state (i.e. the self-report
about whether they feel sleepy or excited). Research on the link between physiology
and affect, e.g. in the xDelia project, has indicated that constantly measured heart rate
is a proxy for arousal [18].
Therefore, we also analysed if the heart rate and the value of the arousal dimension
show similar trends and values. The results showed that there were only few significant
52
results with respect to arousal levels (p < 0.1): in 2 out of 12 film clips, the arousal
differed (two-tailed tests, p1… 12 = (0.936, 0.429, 0.557, 0.540, *0.057, 0.340, 0.873,
*0.011, 0.620, 0.167, 0.849, 0.430)). However, the relationship with the valence dimen-
sion presented significant results in several film clips (p < 0.1). In 6 out of 12 film clips
the value of the valence dimension differed between the groups (two-tailed tests, p1…12
= (*0.086, 0.298, 0.456, *0.052, *0.017, *0.018, 0.877, *0.006, *0.083, 0.861, 0.414,
0.446)). This means that participants who saw their heart rate reported different valence
levels compared to those participants who did not see their heart rate. This impact of
visualizing participants' heart rate was an unexpected effect and therefore we believe it
is worth further discussion and investigation.
3 Discussion
After obtaining these results, several aspects had to be considered. Firstly, this result
was not the expected outcome as it was believed that the participants would potentially
map their heart rate to the self-report of their arousal. Instead, the results show that not
the arousal but the valence is influenced by whether they saw their heart rate or not. As
the experiment was not designed to study this fact, no conclusions about possible un-
derlying reasons could be recognized yet. However, we consider several ideas and hy-
potheses.
At first, we take a look on the physiological background of the heart rate: usually an
individual has a certain resting heart rate. Her heart rate will rise if she is stressed or
certain emotions occur. On the other hand, her heart rate will fall if she is relaxed
[24,25]. Consequently, there is a neutral heart rate and a deviation of this neutral heart
rate depending on the situation.
Said what the heart rate is physiologically depending on, we should think about what
people actually associate with (a) a rising/falling heart rate and (b) the mere display of
the heart rate at all.
A possible reason is that we may associate a low/falling or high/rising heart rate with
a negative or positive (or vice versa) feeling. But a problem arises with this assumption:
is it even possible to associate a high heart rate with a negative or positive feeling? An
emotion of anger as well as being amorous will result in a high heart rate – the first one
is a negative feeling while the second one is a positive feeling. There is probably the
same problem for a low heart rate: it can be easily associated with being bored (proba-
bly a negative feeling) or with being deeply relaxed (probably a positive feeling).
Therefore, it is rather unlikely that a subject establishes this wrong association between
heart rate and valence. Nonetheless, further experiments would be needed to investigate
this aspect in detail.
Another possible reason is that people might just feel (un)comfortable about having
a computer surveying and displaying their heart rate. They might feel (un)com-fortable
about the fact that they are aware of their heart rate (e.g. a well-trained person who likes
to see her low heart rate, or a person which has issues about her high heart rate).
The reason might be related to cognitive dissonance too: a person might think that
she is relaxed – or she is just not aware about her current heart rate. At the time when
53
she looks at the visualization of the heart rate, she becomes aware that her heart rate
had actually risen: this might cause a cognitive dissonance, i.e. there are two percep-
tions that do not fit mentally. This might cause feeling uncomfortable.
Eventually, the reason could also be that people refuse or approve the objective
method to gain information about their physiological state. If they are open minded and
interested in the technology, they may have a higher valence in general compared to a
lower valence if they are not (whether it is wittingly or unwittingly).
Finally, the context where the experiment took place should also be considered. Alt-
hough the study was conducted within the EU Project “MIRROR – Reflective Learning
at Work”, the study had no setting related to the workplace, but was conducted in a lab
setting. However, future studies should be conducted also in work related settings to
investigate how we deal with emotions as well as the physiological reactions they lead
to in the workplace. Showing emotions is often not welcome at the workplace, and we
are probably often quite good at fulfilling this expectation [1, 5]. However, most people
will probably fail to control their physiological reactions. This means their heart rate
will still rise and even lead to more stress (and of course, the initial negative emotion
would still exist).
4 Conclusion
We conducted an experiment to investigate the influence of objective (display of the
heart rate) and subjective (mood map) methods to capture affective aspects. Our ex-
periment reveals that there were no differences in the number of remembered experi-
ences but the display of the heart rate had an influence on users’ self-reported affective
states. However, contrary to the expectations, the objective method had more influence
on the valence dimension than on the arousal dimension. In this paper, we stated first
ideas why people might act this way and raised those ideas for discussion.
The discussed possible reasons why displaying the heart rate has a higher influence
on users' valence than on arousal levels include: (a) people may associate a deviation
from the neutral heart rate with a positive or negative feeling; (b) people may feel
(un)comfortable about a computer surveying their heart rate; (c) people may feel
(un)comfortable about being aware of their heart rate; and (d) a wrong estimation of
their heart rate followed by being aware of their real heart rate may lead to a cognitive
dissonance. These reasons, if they apply, have in common that they may change the
way people feel when they see their heart rate.
If some of the stated reasons apply in reality, this could change the way we use ob-
jective methods to capture and raise awareness of affective aspects in TEL. These facts
should be taken into account and avoid that subjective assessments are influenced by
objective measurements if this is undesired. Therefore, we want to raise these reasons
for discussion and open new trends for further investigation.
54
Acknowledgement
This work is co-funded by the project “MIRROR – Reflective learning at work,” funded
under the FP7 of the European Commission (project number 257617). We would like
to thank all participants in the experiment and the reviewers for their valuable feedback.
References
1. Ashforth, B.E.: Emotion in the workplace: A reappraisal. Human Relations 48(2), 97-125
(1995)
2. Boud, D., Keogh, R., Walker, D.: Promoting reflection in learning: a model. In: Boud, D.,
Keogh, R., Walker, D. (eds.) Reflection: Turning Experience into Learning, pp. 18-40.
Routledge Falmer, London (1985)
3. Fessl, A., Rivera Pelayo, V., Müller, L., Pammer, V., Lindstaedt, S.: Motivation and user
acceptance of using physiological data to support individual reflection. In: Schmidt, A.P.,
Kunzmann, C., Braun, S., Holocher-Ertl, T., Cress, U., Mazarakis, A., Müller, L., Rivera
Pelayo, V. (eds.) Proceedings of the 2nd International Workshop on Motivational and A
ective Aspects in Technology Enhanced Learning (MATEL 2011) (2011)
4. Festinger, L.: A Theory of Cognitive Dissonance. Stanford Univ. Press (1957)
5. Grandey, A.A.: Emotional regulation in the workplace: A new way to conceptualize
emotional labor. Journal of Occupational Health Psychology 5(1), 95-110 (2000)
6. Gross, J.J., Levenson, R.W.: Emotion elicitation using films. Cognition & Emotion 9(1), 87-
108 (1995)
7. Hagemann, D., Naumann, E., Maier, S., Becker, G., L•urken, A., Bartussek, D.: The
assessment of reflective reactivity using films: Validity, reliability and sex differences.
Personality and Individual Differences 26(4), 627-639 (1999)
8. Hewig, J., Hagemann, D., Seifert, J., Gollwitzer, M., Naumann, E., Bartussek, D.: A revised
film set for the induction of basic emotions. Cognition & Emotion 19(7), 1095-1109 (2005)
9. Kolodyazhniy, V., Kreibig, S.D., Gross, J.J., Roth, W.T., Wilhelm, F.H.: An affective
computing approach to physiological emotion specificity: Toward subject-independent and
stimulus-independent classification of film-induced emotions. Psychophysiology 48(7), 908-
922 (2011), http://dx.doi.org/10.1111/j.1469-8986.2010.01170.x
10. Krogstie, B.R., Divitini, M.: Reflective learning in the workplace: The role of emotion. In:
Hernández-Leo, D., Ley, T., Klamma, R., Harrer, A. (eds.) Scaling up learning for sustained
impact, Lecture Notes in Computer Science, vol. 8095, pp. 615-616. Springer, Berlin (2013)
11.McDuff, D., Karlson, A., Kapoor, A., Roseway, A., Czerwinski, M.: Affectaura: an intelligent
system for emotional memory. In: Proceedings of the 2012 ACM annual conference on
Human Factors in Computing Systems. pp. 849-858. CHI ’12 (2012),
http://doi.acm.org/10.1145/2207676.2208525
12. Mora, S., Rivera-Pelayo, V., Müller, L.: Supporting Mood Awareness in Collaborative
Settings. In: Proceedings of the 7th International Conference on Collaborative Computing:
Networking, Applications andWorksharing (CollaborateCom ’11). pp. 268-277 (Oct 2011)
13. Morris, M.E., Kathawala, Q., Leen, T.K., Gorenstein, E.E., Guilak, F., Labhard, M., Deleeuw,
W.: Mobile therapy: case study evaluations of a cell phone application for emotional self-
awareness. Journal of Medical Internet Research 12(2), e10 (2010),
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2885784/
14. movisens GmbH: Ekg- und Aktivitätssensor, http://www.movisens.com/de/produkte/ekg-
und-aktivitaetssensor/
55
15. Müller, L., Krogstie, B., Schmidt, A.: Towards capturing learning experiences. In:
Ravenscroft, A., Sharples, M. (eds.) Context and Technology Enhanced Learning (ConTEL):
Theory, methodology and design, EC-TEL 2011, Palermo, Italy (2011)
16. Müller, L., Rivera Pelayo, V., Kunzmann, C., Schmidt, A.P.: From stress awareness to coping
strategies of medical staff: Supporting reflection on physiological data. In: Hutchison, D.,
Kanade, T., Kittler, J., Kleinberg, J.M., Mattern, F., Mitchell, J.C., Naor, M., Nierstrasz, O.,
Pandu Rangan, C., Steffen, B., Sudan, M., Terzopoulos, D., Tygar, D., Vardi, M.Y., Weikum,
G., Salah, A.A., Lepri, B. (eds.) Human Behavior Unterstanding, Lecture Notes in Computer
Science, vol. 7065, pp. 93-103. Springer, Berlin and Heidelberg (2011)
17. Müller, L., Rivera Pelayo, V., Mora, S., Fessl, A.: Prototype for capturing apps - version 2:
Deliverable 3.3 (2013), http://mirror-project.eu/research-results/downloads/finish/5/98
18. Peffer, G., Fenton-O'Creevy, M., Adam, M., Astor, P., Cederholm, H., Clough,G., Conole,
G., Davies, G., Eriksson, J., Gaved, M., Heuer, S., Jercic, P., Lindley, C., Lins, J.T., van
Overveld, M., Scanlon, E., Schaaf, K., Smidts, A.: xdelia internal report: emotion-centred
nancial decision making and learning (September 2012), http://oro.open.ac.uk/34544/, iSBN
978-1-78007-652-2
19.Peira, N., Fredrikson, M., Pourtois, G.: Controlling the emotional heart: heart rate biofeedback
improves cardiac control during emotional reactions. International Journal of
Psychophysiology 91(3), 225-231 (2014), http://dx.doi.org/10.1016/j.ijpsycho.2013.12.008
20.Philippot, P.: Inducing and assessing differentiated emotion-feeling states in the laboratory.
Cognition & Emotion 7(2), 171-193 (1993)
21. Prince, S.A., Adamo, K.B., Hamel, M.E., Hardt, J., Gorber, S.C., Tremblay, M.: A
comparison of direct versus self-report measures for assessing physical activity in adults: a
systematic review. The International Journal of Behavioral Nutrition and Physical Activity
56(5) (2008)
22. Rottenberg, J., Ray, R.D., Gross, J.J.: Emotion elicitation using lms. In: Coan, J.A., Allen,
John J. B. (eds.) Handbook of Emotion Elicitation and Assessment, pp. 9-28. Series in
affective science, Oxford University Press, Oxford (2007)
23.Russell, J.A.: A circumplex model of a affect. Journal of Personality and Social Psychology
39(6), 1161-1178 (1980)
24. Shapiro, D., Jamner, L.D., Goldstein, I.B., Delano, R.J.: Striking a chord: Moods, blood
pressure, and heart rate in everyday life. Psychophysiology 38(2), 197-204 (2001)
25.Stewart, S.H., Buffett-Jerrott, S.E., Kokaram, R.: Heartbeat awareness and heart rate reactivity
in anxiety sensitivity: A further investigation. Journal of Anxiety Disorders 15(6), 535-553
(2001)
26. Tomarken, A.J., Davidson, R.J., Henriques, J.B.: Resting frontal brain asymmetry predicts a
ective responses to films. Journal of Personality and Social Psychology 59(4), 791-801 (1990)
