1

Training for Wisdom: The Illeist Diary Method

Author Affiliation:

I. Grossmann 1, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1, e-mail:
igrossma@uwaterloo.ca
Anna Dorfman 1, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1, e-mail:
anna.dorfman@uwaterloo.ca
Harrison Oakes 1, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1, e-mail:
hoakes@uwaterloo.ca
Kathleen D. Vohs 2, 321 19th Avenue South, Minneapolis, MN, USA 55455, e-mail:
kvohs@umn.edu
Henri C. Santos 3, 120 Hamm Drive, Lewisburg, PA, USA 17837, e-mail:
hsantos@geisinger.edu
Abigail Scholer 1, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1, e-mail:
ascholer@uwaterloo.ca

1 University of Waterloo, Canada.

2 Carlson School of Management, University of Minnesota, U.S.A.
3 Autism and Developmental Medicine Institute, Geisinger Health System, U.S.A.

Corresponding Author: Igor Grossmann, 200 University Avenue West, Waterloo, Ontario,
Canada N2L 3G1, Tel. +1 519 888-4567 ext. 31793, e-mail: igrossma@uwaterloo.ca.

Keywords: wisdom, intervention, perspective-taking, intellectual humility, affective forecasting

 2

Abstract

We tested the utility of illeism – a practice of referring to oneself in the third person – for the
trainability of wisdom-related characteristics in everyday life: i) wise reasoning (intellectual
humility, open-mindedness in ways a situation may unfold, perspective-taking, attempts to
integrate different viewpoints) and ii) accuracy in emotional forecasts toward close others. In a
month-long field experiment, people adopted either the third-person training or first-person
control perspective when describing their most significant daily experiences. Assessment of
spontaneous wise reasoning before and after the intervention revealed substantial growth in the
training (vs. control) condition. At the end of the intervention, people forecasted their feelings
toward a close other in challenging situations. A month later, these forecasted feelings were
compared against their experienced feelings. Participants in the training (vs. control) condition
showed greater alignment of forecasts and experiences, largely due to changes in their emotional
experiences. The present research demonstrates a path to evidence-based training of wisdom-
related processes via the practice of illeism.
 3

Training for Wisdom: The Illeist Diary Method

Wisdom is universally valued1. It leads to thoughts, feelings, and behaviors that are both
personally and socially beneficial2,3. Being wise includes epistemic forms of reasoning—
intellectual humility, sensitivity to possible change in social relations, perspective taking, and the
search for ways to integrate different viewpoints4,5. In contrast to general cognitive abilities,
these forms of reasoning predict subjective well-being6 and prosocial behavior7,8. Numerous
traditions in philosophy link wisdom to “knowing thyself,” which includes accurate forecasts
about how one may feel in challenging situations9,10.
The importance of being wise seems straightforward, yet often people fail to exercise
wisdom when reasoning about issues that are important to them11. In particular, personal
conflicts tend to elicit bias and low levels of wisdom11,12. Though emerging scholarship has
started to identify ways to promote wise reasoning in the face of personal challenges, this work
has been either cross-sectional or confined to testing short, artificial scenarios in the laboratory.
A critical question has remained unaddressed: How can one train wise reasoning in daily life?
Here, we report results from a four-week field experiment addressing this question by evaluating
a novel approach to training wise reasoning and testing its impact on accuracy in forecasts of
one’s feelings toward close others 13,14.
Though empirical scholarship has not yet evaluated the effectiveness of wisdom training
in daily life, propositions for wisdom-enhancing training exist 15,16. One dominant proposition
has centered on the use of illeism—the practice of referring to oneself in the third person. It is an
ancient form of writing and speaking, dating as far back as 58 BC in writings by Julius Caesar17.
The idea behind this method is to create psychological distance from one’s immediate self18.
Indeed, third-person self-talk is common among spiritual and political leaders, has been linked to
better self- and emotion-regulation19, and temporarily bolsters wise reasoning11,20. Theoretically,
illeism may foster wisdom by allowing for a big picture perspective on the situation18,21.
Accordingly, we instructed people to repeatedly practice illeism across a range of situations they
encountered daily22. In this way, we aimed to promote a habitual use of wise reasoning.
A community sample (N = 298) participated in a month-long training intervention. At the
end of each day, people reflected on the central event of the day. According to random
assignment, participants in the training condition were instructed to reflect on key daily events
from a third-person perspective (i.e., using 3rd-person pronouns and their name to refer to
themselves) whereas participants in the control condition were told to adopt a first-person
perspective (i.e., using 1st-person singular pronouns; this perspective is habitually used to reflect
on social issues23,24). Via this diary-based intervention, people used the assigned perspective to
reconstruct (20) and reflect on one social experience that occurred that day (e.g., conflict with
another person; celebration with others; bad news in a social setting; for exact wording see SI).
By applying the illeist method to reflections on a wide range of meaningful daily
experiences, we tested whether it could effectively cultivate wise reasoning from before to after
the intervention. Moreover, we examined the downstream consequences of the training, beyond
reasoning on the ‘know thyself’ aspect of wisdom. Namely, we examined post-intervention
accuracy in forecasts of feelings toward close others. We predicted the training group would
show a better alignment of forecasted and experienced emotions toward a close other. Further,
we predicted that greater growth in wise reasoning would foster such alignment between
forecasted and experienced feelings toward close others. Unless noted, we pre-registered our
predictions, methods, and procedures (osf.io/gb7js).
 4

Results
Training-Based Growth in Wise Reasoning
To assess wise reasoning before and after the intervention, participants came to the lab to
recall their most recent social challenge (i.e., social conflict or irritating social interaction).
Participants used the Event Reconstruction Method to minimize recall bias and reinstate the
emotional experience of the recalled event8,25. Participants provided contextual details about the
event and then described the thoughts that came to their mind as they reflected on the event.
Philosophers and cognitive scientists have defined the central aspects of wise reasoning as a set
of interrelated processes that aid working through complex social issues: (i) intellectual humility
(recognition of one’s limited knowledge), (ii) recognition of the world in flux and change, (iii)
acknowledgment of others’ perspectives, (iv) consideration or search for compromise, and (v)
recognition of the importance of conflict resolution4,26. Following prior research6,27, two
independent and hypothesis-blind raters coded participants’ written narratives on the five wise
reasoning dimensions (see Methods). We fit a general linear model with the dependent variable
transformed to account for violation of normality assumption. Our model treated participants and
coders as random factors, time (pre- vs. post-intervention) as a Level 1 predictor, and perspective
condition as a Level 2 predictor. Results were comparable when using a generalized linear mixed
model with gamma distribution as another way to account for skewness in the data.
Our key hypothesis was that the illeist method of daily reflections (third-person training
condition) would result in greater wise reasoning than habitual reflections (first-person control
condition). In line with our prediction, we observed a Time X Condition interaction, t = 2.65, P =
.008 (Figure 1): Participants in the third-person training condition showed more wise reasoning
after (vs. before) the intervention, B = 0.130, SE = 0.028, t = 4.61, P < 0.0001, whereas wise
reasoning in the first-person control condition did not change, B = 0.022, SE = 0.030, t = 0.74, P
= 0.458. The interaction effect held when controlling for the following potential confounds:
number of adverse experiences during the diary-based training, contextual differences between
social conflicts people reasoned about (i.e., time of day event occurred, length of conflict), and
narrative word count, Time X Condition Interaction: t = 3.12, P = 0.002. In a separate analysis,
the interaction effect was also robust when controlling for age, gender, ethnicity, and fluid
intelligence, t = 3.43, P < 0.001.
Supplementary analyses for each dimension of wise reasoning yielded a more nuanced
pattern of results (see Figure 1 and SI). Participants in the third-person training condition showed
increased intellectual humility, acknowledgement of others’ perspectives, and recognition of
importance of conflict resolution from before to after the intervention, whereas the first-person
control condition did not. The other two dimensions (recognition of the world in flux and
change, consideration or search for compromise) did not show significant differences across time
points. Moreover, supplementary analyses showed training-based effects for greater
emodiversity of negative experiences28,29, which in turn statistically mediated the training effect
on wise reasoning (see SI).
 5

Figure 1. Effects of training on wise reasoning. Panel A: mean estimates with 95% confidence
intervals across five aspects of wise reasoning. Panels B-F: Results across individual dimensions
in each condition pre- and post-intervention.

Forecasting Accuracy
We tested the downstream effects of the illeist training on the accuracy of participants’
forecasted feelings involving an event with a close other. At the end of the diary-based training,
participants considered three interpersonal challenges (betrayal of trust, social exclusion,
disagreement) and three positive events (social support, informal gathering, pleasant
conversation) that could occur during the ensuing month between themselves and a close other.
They rated how they would feel toward this person if the event occurred (see SI for exact
instructions). We tested forecasting accuracy for events that did occur by comparing participants’
ratings of forecasted to experienced emotions, the latter of which were collected one month after
the forecasts.
Results of a model with emotions nested in participants as random factors, valence
(positive vs. negative), event type (positive vs. challenge), and time (forecast vs. experience) as
Level 1 predictors, and perspective condition as a Level 2 predictor indicated a significant 4-way
interaction, t = 2.30, χ2 = 5.30, P = 0.021. To unpack this interaction, we examined effects
separately for positive events and interpersonal challenges.
We observed no significant effects of training condition for accuracy about positive
events, P = 0.236, whereas there was a significant effect of training for accuracy about
interpersonal challenges, t =2.77, χ2 = 7.66, P = 0.006. As Figure 2 indicates, control-group
participants’ forecasts toward close others were significantly different from their experienced
feelings. Specifically, they underestimated the intensity of their negative feelings, t = 3.97, P <
0.0001, and overestimated the intensity of their positive feelings, t = 4.23, P < 0.0001. In
contrast, training-group participants’ forecasts for negative and positive feelings did not differ
from their reported experiences, ts < 1.32, Ps > 0.187. Notably, the training (vs. control) group’s
greater forecasting accuracy was driven by differences in the groups’ experienced, rather than
forecasted, emotions (see Figure 2-A).
 6

Figure 2. Alignment of forecasted and experienced feelings toward a close other during
challenges. Panel A: Effects of training. Panel B: Effects of growth in wise reasoning. Mean
estimates of emotional intensity with 95% confidence intervals.

Last, we explored whether training-based growth in wise reasoning fostered greater

forecasting accuracy. To this end we tested a linear mixed model, with valence (positive vs.
negative), event type (positive vs. challenge), and time (forecast vs. experience) as Level 1
predictors and change in wise reasoning (pre- vs. post-intervention) as a Level 2 predictor.
Results indicated a significant four-way interaction, t = 2.40, χ2 = 5.78, P = 0.016. Again, we
examined effects separately for positive events and interpersonal challenges.
Simple slopes analyses indicated that whereas growth in wise reasoning did not improve
forecasting accuracy for positive events, t < 1, ns, it did significantly predict forecasting accuracy
for interpersonal challenges, t = 2.22, χ2 = 4.94, P = 0.026. Specifically, participants whose wise
reasoning changed little ( 1 SD) from before to after the intervention reported highly inaccurate
forecasts vs. experiences. They reported more intense experiences compared to forecasted
feelings toward close others for negative emotions, t = 4.16, P < 0.0001, and less intense
experiences compared to forecasted feelings for positive emotions, t = 3.33, P = 0.001. In
contrast, this lack of match between forecasted and experienced feelings toward close others was
substantially attenuated among participants who showed more growth (+ 1 SD) in wise reasoning
due to the intervention. They showed no significant difference between forecasted and
experienced feelings toward close others for both negative, t = 1.87, P = 0.062, and positive
emotions, t = 1.72, P = 0.085 (see Figure 2-B). In short, analyses with growth in wise reasoning
as a moderator of emotional forecasting accuracy mirrored the pattern of results for condition
effects.

Discussion
The results of a four-week intervention applying an illeist diary method showed that
third-person writing effectively promoted wiser reasoning, relative to a control condition in
which people reflected on their daily experiences from a first-person perspective. The training
also fostered greater alignment of forecasted and experienced feelings toward a close person
during challenging events like disagreements and social conflicts. This intervention is the first
 7

evidence that wisdom-related cognitive and affective processes can be trained in daily life and of
how to do so.
The present work contributes to the literature on training cognition30,31 and debiasing32.
Prior cognitive training efforts have largely focused on working memory or formal reasoning,
which differ from wise reasoning in several aspects. The former is abstract and decontextualized,
whereas the latter is embedded in the social pragmatics and often involves personally relevant
matters33,34. Moreover, past work has highlighted the difficulties in extending domain-general
training effects to novel situations35–37.
The present training introduces the usefulness of situating repeated training sessions in
the ecological context of daily life experiences. Because people trained across several diverse
situations over the course of a month, repeated sampling may be able to facilitate the
development of something akin to a generalized approach across a range of interpersonal
challenges22. In doing so, this method suggests a greater likelihood that the target characteristic is
applied to novel situations as compared to standardized or laboratory training of situation-
specific states. The present approach also makes use of meaningful firsthand experiences people
encounter in their lives, suggesting greater generalizability and durability compared to laboratory
training on hypothetical scenarios.
Moreover, the present work brings together and jointly advances disparate streams of
scholarship on wise reasoning and forecasting accuracy, showcasing how these wisdom-related
processes can be fostered via daily exercises. The present work extends prior work by providing
the first demonstration of how laboratory insights concerning benefits of illeism18 for wisdom-
related processes can be extended to training-based interventions in the field. Practically, the
present intervention provides proof-of-concept for the viability of training programs for wisdom-
related characteristics in everyday life.
The present research found that training-based effects on accuracy between forecasted
and experienced feelings were most evident for negative experiences. It is possible that positive
experiences during the training period were less relevant than were negative experiences for the
social conflicts on which participants reflected. Even so, being able to differentiate negative (vs.
positive) experiences more 38,39 may be useful when navigating the social world38,40. Together,
these streams of work suggest a viable extension of the current research by examining effects of
illeist training on situations in which differentiation of positive experiences has greater
ecological utility.
Last, the present study extends prior work on affective forecasting, which has focused on
general mood or participant-centered emotions 10. Focusing on forecasts of feelings toward close
others, we observed that the training and control groups differed in extremity of subsequently
reported experiences, such that the training group appeared more balanced, with comparable
intensity of experienced to forecasted feelings toward close others. In contrast, the control group
was more extreme, with diverging intensity of experienced and forecasted emotions. This
observation suggests that illeist diary training reduced the empathy gap41 between forecasted and
experienced feelings toward close others by moderating participants’ emotional reactivity to
reported experiences.
Several caveats are in order before concluding. More work is needed to assess the
suitability of the present training for large-scale applications in clinical or professional settings.
The present training-based intervention focused on interpersonal social challenges, raising the
question whether such training can generalize to wise reasoning and affective processes during
 8

intergroup conflicts, political challenges, or challenges of a non-interpersonal nature (e.g.,

financial decision making).
Conclusion
Wise reasoning is a style of thinking that reflects a balanced, open approach that seeks
reconciliation of disparate views8. It predicts a host of desirable outcomes, from harmonious
interpersonal relationships7,8 to personal well-being6. Here, we showed that wise reasoning is not
the purview of just a few fortunate individuals. The current work invoked the ancient practice of
illeism, in which people refer to themselves in the third person, and found that doing so over the
course of a month improved wise reasoning and emotional tone. The results of this field
experiment suggest that illeist training bolsters wise reasoning in everyday life, with potential
benefits for the health and success of individuals and relationships.
Methods
Recruitment
The study was reviewed and received ethics clearance through a University of Waterloo
Research Ethics Committee (ORE # 31889). We recruited local community members from a
midsized city in southwestern Ontario and students from the local university to “participate in
research on social experiences and personal goals in daily life” for the opportunity to earn up to
$110. The present study was part of a large-scale investigation of goals, emotions, and reasoning.
Once recruited, participants (M age = 22.28, SD age = 6.93, Range age = 17-62; 77% female;
60% without a college degree; Mdn household income = $50,000-75,000; 37% white, 30%
Asian/Asian-Canadian, 11% East Indian, 7% African/African-Canadian, 6% Middle Eastern, 9%
other ethnicity) completed the study.
Participants first completed a set of online and laboratory sessions. They then completed
a 4-week daily diary intervention, after which they completed a second set of online and
laboratory sessions, including a follow-up survey one month after the daily diary period. Full
project documentation, including a structural and timeline summary, is on Open Science
Framework (osf.io/6p2us).
Pre-/Post-Diary Measures
A week before and after the training intervention (i.e., 4-week daily diary period),
participants visited the laboratory. An experimenter accompanied participants to a computer,
with further instructions provided via a display interface. First, participants read a definition of a
difficult social event (see SI) and recalled a recent difficult social event (either a social conflict
or irritating interaction). Participants reconstructed the experience via the Event Reconstruction
Method25, providing the first name of the other person involved in the event, contextual
information such as day of the week and location, and a detailed description of the event. Upon
reconstructing the event, participants spent at least 30 seconds reflecting on the difficult social
event, with the screen not advancing before that time. Subsequently, participants wrote at least
one paragraph describing the thoughts that came to mind as they reflected on the event, which
we used to quantify wise reasoning prior to and after the training intervention period.
Participants then completed an exploratory questionnaire of reasoning processes8 (see SI). On a
subsequent computerized task, we obtained an estimate of participant’s non-verbal fluid
intelligence via a version of the Raven's Progressive Matrices42, for control analyses.
Quantifying wise reasoning in laboratory narratives. Two raters (blind to condition and
hypotheses) coded written stream-of-thought narratives, quantifying degree of wise reasoning on
five previously established dimensions of wise reasoning4: intellectual humility, recognition of
world in flux/change, acknowledgement of others’ perspectives, search for a compromise, and
 9

consideration of conflict resolution. For each dimension, coders used a 5-point scale: 1 = nothing
about the theme mentioned, 2 = one instance of the theme is mentioned, but is not described in
depth, 3 = several instances of the theme are mentioned, but none are described in depth; 4 = one
or more instances of the theme are mentioned, and one of them is described in depth; 5 = several
instances of the theme are mentioned, and two or more of them are described in depth. Inter-rater
reliability for each dimension was very good (Kendall’s Ws > 0.84).
Daily Diary Intervention
Perspective manipulation. Prior to beginning the 4-week diary intervention period,
participants were randomly assigned to use one of two perspectives when reflecting on major
social events: first-person (i.e., using 1st-person pronouns like I, me, my, and mine) or third-
person (i.e., using 3rd-person pronouns like s/he, her/him, they and theirs, and using their name
to refer to themselves). Prior research has indicated a first-person perspective is the habitual
baseline in reflections on daily social events18,24,43, so we chose first-person instructions as the
control condition.
Participants were instructed to complete a diary at the end of each day, for the duration of
four weeks. A computer program guided them through the diary reporting. First, they saw events
that constituted a range of social experiences (i.e., conflict/argument with another person;
annoying/irritating social situation; celebration with others; enjoyable social event; sad/bad news
in a social setting; see SI). The content of the survey was then based on the events participants
indicated they had experienced. If participants indicated they had experienced multiple events in
one day, the program selected an event, prioritizing an adverse over pleasurable event, and
randomly selecting one of the events from the same-valence group. If participants indicated that
they had experienced none of the social events listed, they did not partake in the training on that
day. Instead, they provided reports of their general emotions about the day and whether they had
experienced some other personal challenges (collected to ensure comparable length across diary
days).
Participants reconstructed the social event using the Event Reconstruction Method (23).
They rated the intensity of the event (0 = Not at all intense; 6 = Extremely intense). On Days 1-6,
participants next described their current thoughts about the event using their randomly assigned
perspective. We included short writing prompts to ensure that participants understood how to
write using a first- or third-person perspective. Participants in the first- [third-] person condition
were instructed to use pronouns I/me [their name] as much as possible as they were describing
“the event and their stream of thoughts.” We further provided an example. For the first-person
condition, we stated “For example, you might write “I think… I feel…” For the third-person
condition, we stated “For example, if your name were Chris, you might write “Chris thinks…;
Chris feels…” (see SI for exact wording).
Day 7 of each week was a free-writing day. On Day 7 participants were not instructed to
adopt a particular perspective, though they completed the diary measures as on Days 1-6. Day 7
diaries acted as a manipulation check by enabling us to measure the degree to which participants’
writing about that day’s social event corresponded to their assigned perspective (see SI).
Affective Forecasting
On the last day of the diary (which did not include perspective instructions), participants
completed measures about common social interactions with family and friends. First, participants
thought of someone (an adult) they had a close relationships with and with whom they would
likely interact frequently over the next month. Participants indicated the person’s initials, their
relationship to the person (e.g., partner, best friend, sibling), and the degree to which their
 10

relationship might change in the coming month: “Do you think that one month from now, your
relationship with [Person X] will change for the better or for worse?” (1 = Definitely no; 5 =
Definitely yes).
Forecasted emotions. Participants next read descriptions of six common social events
that could happen between them and their selected person over the next month (see SI for
verbatim instructions and descriptions of each event). Participants described how each event
could unfold and then forecasted their future emotions if it were to occur. Participants first rated
their anticipated mood if the event were to occur (1 = Very unhappy; 7 = Very happy; see
relevant analyses in SI). Next, they rated their anticipated feelings toward the other person—i.e.,
the extent to which they would feel warm, upset, trustful, angry, liking, frustrated, and content
toward the person they selected if the event were to occur (1 = Not at all; 7 = Very much).
Lastly, participants indicated the likelihood that, over the next month, the six events would occur
in the context of that relationship (1 = Not at all likely, 5 = Very likely).
One-month follow-up survey. One month from the date of the post-diary laboratory visit
period, participants completed a follow-up survey online. It started by reminding them that a
month ago they considered six events that might happen between them and a close other.
Participants were first reminded of the close other they selected a month ago. Next, participants
were presented with the list of the six events in a randomized order. For each event, participants
first indicated whether the events had indeed occurred. For each event that participants indicated
having occurred, they briefly described it using the Event Reconstruction Method, and rated their
feelings toward the other person after the event took place using items similar to those used in
the forecasting session.
 11

References

1. Dahlsgaard, K., Peterson, C. & Seligman, M. E. P. Shared Virtue: The Convergence of

Valued Human Strengths Across Culture and History. Rev. Gen. Psychol. 9, 203–213
(2005).
2. Grossmann, I. & Brienza, J. The Strengths of Wisdom Provide Unique Contributions to
Improved Leadership, Sustainability, Inequality, Gross National Happiness, and Civic
Discourse in the Face of Contemporary World Problems. J. Intell. 6, 22 (2018).
3. Sternberg, R. J. Building wisdom and character. in Health, happiness, and well-being:
Better living through psychological science (eds. Lynn, S. J., O’Donohue, W. &
Lilienfeld, S.) 296–316 (Sage, 2014).
4. Grossmann, I. Wisdom in Context. Perspect. Psychol. Sci. 12, 233–257 (2017).
5. Bangen, K. J., Meeks, T. W. & Jeste, D. V. Defining and assessing Wisdom: A review of
the literature. Am. J. Geriatr. Psychiatry 21, 1254–1266 (2013).
6. Grossmann, I., Na, J., Varnum, M. E. W., Kitayama, S. & Nisbett, R. E. A route to well-
being: Intelligence versus wise reasoning. J. Exp. Psychol. Gen. 142, 944–953 (2013).
7. Grossmann, I., Brienza, J. P. & Bobocel, D. R. Wise deliberation sustains cooperation.
Nat. Hum. Behav. 1, (2017).
8. Brienza, J. P., Kung, F. Y. H., Santos, H. C., Bobocel, D. R. R. & Grossmann, I. Wisdom,
bias, and balance: Toward a process-sensitive measurement of wisdom-related cognition.
J. Pers. Soc. Psychol. 115, 1093–1126 (2018).
9. Seligman, M. E. P., Railton, P., Baumeister, R. F. & Sripada, C. Homo Prospectus.
(Oxford University Press, 2016).
10. Wilson, T. D. & Gilbert, D. T. Affective forecasting. in Advances in Experimental Social
Psychology (ed. Zanna, M. P.) Volume 35, 345–411 (Academic Press, 2003).
11. Grossmann, I. & Kross, E. Exploring Solomon’s Paradox: Self-distancing eliminates the
self-other asymmetry in wise reasoning about close relationships in younger and older
adults. Psychol. Sci. 25, 1571–1580 (2014).
12. Staudinger, U. M. & Baltes, P. B. Interactive minds: A facilitative setting for wisdom-
related performance? J. Pers. Soc. Psychol. 71, 746–762 (1996).
13. Seligman, M. E. P., Railton, P., Baumeister, R. F. & Sripada, C. Navigating Into the
Future or Driven by the Past. Perspect. Psychol. Sci. 8, 119–141 (2013).
14. Barrett, L. F., Salovey, P. & Mayer, J. D. The Wisdom in Feeling: Psychological
Processes in Emotional Intelligence. (The Guilford Press, 2002).
15. Sternberg, R. J., Jarvin, L. & Reznitskaya, A. Teaching for Wisdom Through History:
Infusing Wise Thinking Skills in the School Curriculum. in Teaching for Wisdom (eds.
Ferrari, M. & Potworowski, G.) 37–57 (Springer Netherlands, 2008). doi:10.1007/978-1-
4020-6532-3_3
 12

16. Ferrari, M. & Potworowski, G. Teaching for Wisdom. (Springer Netherlands, 2008).
doi:10.1007/978-1-4020-6532-3_4
17. Raaflaub, K. A. & Strassler, R. B. The Landmark Julius Caesar: The Complete Works:
Gallic War, Civil War, Alexandrian War, African War, and Spanish War. (Pantheon,
2017).
18. Kross, E. & Ayduk, O. Self-Distancing: Theory, Research, and Current Directions. Adv.
Exp. Soc. Psychol. 55, 81–136 (2017).
19. Kross, E. et al. Self-talk as a regulatory mechanism: How you do it matters. J. Pers. Soc.
Psychol. 106, 304–324 (2014).
20. Grossmann, I., Oakes, H. & Santos, H. C. Wise Reasoning Benefits from Emodiversity,
Irrespective of Emotional Intensity. J. Exp. Psychol. Gen. (2019).
doi:10.1037/xge0000543
21. Gross, J. J. Emotion regulation: Current status and future prospects. Psychol. Inq. 26, 1–26
(2015).
22. Fleeson, W. & Jayawickreme, E. Whole trait theory. J. Res. Pers. 56, 82–92 (2015).
23. McIsaac, H. K. & Eich, E. Vantage point in traumatic memory. Psychol. Sci. 15, 248–53
(2004).
24. Grossmann, I. & Kross, E. The Impact of Culture on Adaptive Versus Maladaptive Self-
Reflection. Psychol. Sci. 21, 1150–1157 (2010).
25. Schwarz, N., Kahneman, D. & Xu, J. Global and episodic reports of hedonic experience.
in Calendar and time diary: Methods in life events research (eds. Belli, R., Alwin, D. &
Stafford, F.) 157–174 (Sage Publications, 2009).
26. Oakes, H., Brienza, J. P., Elnakouri, A. & Grossmann, I. Wise reasoning: Converging
evidence for the psychology of sound judgment. in Cambridge Handbook of Wisdom (eds.
Sternberg, R. J. & Glück, J.) (Cambridge University Press).
27. Grossmann, I. et al. Reasoning about social conflicts improves into old age. Proc. Natl.
Acad. Sci. U. S. A. 107, 7246–7250 (2010).
28. Quoidbach, J. et al. Emodiversity and the emotional ecosystem. J. Exp. Psychol. Gen. 143,
2057–2066 (2014).
29. Kashdan, T. B., Barrett, L. F. & McKnight, P. E. Unpacking emotion differentiation:
Transforming unpleasant experience by perceiving distinctions in negativity. Curr. Dir.
Psychol. Sci. 24, 10–16 (2015).
30. Larrick, R. P., Morgan, J. N. & Nisbett, R. E. Teaching the use of cost-benefit reasoning
in everyday life. Psychol. Sci. 1, 362–370 (1990).
31. Perkins, D. N. & Grotzer, T. A. Teaching intelligence. Am. Psychol. 52, 1125–1133
(1997).
32. Larrick, R. P. Debiasing. in Blackwell Handbook of Judgment and Decision Making 316–
 13

338 (Blackwell Publishing Ltd). doi:10.1002/9780470752937.ch16

33. Sternberg, R. J. The Scientific Study of Personal Wisdom. The Scientific Study of Personal
Wisdom (Springer Netherlands, 2013). doi:10.1007/978-94-007-7987-7
34. Santos, H. C., Huynh, A. C. & Grossmann, I. Wisdom in a complex world: A situated
account of wise reasoning and its development. Soc. Personal. Psychol. Compass 11,
e12341 (2017).
35. Halpern, D. F. Teaching critical thinking for transfer across domains: Disposition, skills,
structure training, and metacognitive monitoring. Am. Psychol. 53, 449–455 (1998).
36. Sala, G. & Gobet, F. Cognitive Training Does Not Enhance General Cognition. Trends
Cogn. Sci. 23, 9–20 (2019).
37. Kassai, R., Futo, J., Demetrovics, Z. & Takacs, Z. K. A meta-analysis of the experimental
evidence on the near- and far-transfer effects among children’s executive function skills.
Psychol. Bull. 145, 165–188 (2019).
38. Alves, H., Koch, A. & Unkelbach, C. Why Good Is More Alike Than Bad: Processing
Implications. Trends Cogn. Sci. 21, 69–79 (2017).
39. Grossmann, I., Huynh, A. C. & Ellsworth, P. C. Emotional complexity: Clarifying
definitions and cultural correlates. J. Pers. Soc. Psychol. 111, 895–916 (2016).
40. Rozin, P. & Royzman, E. B. Negativity bias, negativity dominance, and contagion.
Personal. Soc. Psychol. Rev. 5, 296-320. (2001).
41. Boven, L. Van, Loewenstein, G., Dunning, D. & Nordgren, L. F. Changing Places: A
Dual Judgment Model of Empathy Gaps in Emotional Perspective Taking. in Advances in
Experimental Social Psychology (eds. Zanna, M. P. & Olson, J. M.) 117–171 (Academic
Press, 2013). doi:10.1016/B978-0-12-407188-9.00003-X
42. Raven, J. The Raven Progressive Matrices: A review of national norming studies and
ethnic and socioeconomic variation in intelligence of students. J. Educ. Meas. 26, 1–16
(1989).
43. Grossmann, I., Gerlach, T. M. & Denissen, J. J. A. Wise Reasoning in the Face of
Everyday Life Challenges. Soc. Psychol. Personal. Sci. 7, 611–622 (2016).
Acknowledgments: Takuya Shibayuma and Oliver Twardus provided valuable assistance
qualitative coding, whereas Konstantyn Sharpinsky provided assistance with data integrity
cross-validation. The present research was funded by Social Sciences and Humanities
Research Council of Canada Insight Grant 435-2014-0685 (to I.G.), Templeton Science of
Prospection Award (to I.G. and K.V.) and Templeton Pathways to Character Award (to I.G.).
All data and statistical analyses that support the findings of this study are publicly available
on Open Science Framework website with identifier osf.io/gb7js.
Author contributions: I.G. provided study concept for part A and I.G. and K.V. provided
study concept for Part B. I.G., H.O., and A.S. designed part A, I.G. and H. S. designed part B.
I.G. and A.D. pre-registered studies. I.G. and H.S. collected data. I.G. and A. D. carried out
 14

data analysis. I.G. drafted the initial version of the manuscript. All authors contributed to
revision of the manuscript and approved the final manuscript for submission.
The authors declare no conflict of interest.
 15

Supplementary Information (SI)

for

Training for Wisdom: The Illeist Diary Method

Author Affiliation:

I. Grossmann1, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1, e-mail:
igrossma@uwaterloo.ca
Anna Dorfman1, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1, e-mail:
anna.dorfman@uwaterloo.ca
Harrison Oakes1, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1, e-mail:
hoakes@uwaterloo.ca
Kathleen D. Vohs2, 321 19th Avenue South, Minneapolis, MN, USA 55455, e-mail:
kvohs@umn.edu
Henri C. Santos3, 120 Hamm Drive, Lewisburg, PA, USA 17837, e-mail: hsantos@geisinger.edu
Abigail Scholer1, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1, e-mail:
ascholer@uwaterloo.ca

1 University of Waterloo, Canada.

2 Carlson School of Management, University of Minnesota, U.S.A.
3 Autism and Developmental Medicine Institute, Geisinger Health System, U.S.A.

Corresponding Author: Igor Grossmann, 200 University Avenue West, Waterloo, Ontario,
Canada N2L 3G1, Tel. +1 519 888-4567 ext. 31793, e-mail: igrossma@uwaterloo.ca.
 16

Supplementary Methods and Results

Structure and Timeline of Wisdom Training Project
Stage Content Timeline

● Schedule intake survey

1 Recruitment & Demographics Survey
● Email online survey

● ~90 min
2 Pre-Diary Laboratory Session
● 2-week delay, then diary begins

● 15 min/daily
3 Diary Study + Forecasting (final Day 7)
● 4 weeks

● ~90 min
4 Post-Diary Laboratory Session
● A week after diary completion
● ~20-30 min
5 One-Month Follow-Up Survey ● 1 month post second laboratory
session

General Analytical Notes

Content-analytic and affective measures often violate assumptions of normality.
Therefore, we normalized dependent variables involving wise reasoning and feelings toward
close others that violated the normality assumption in general linear model analyses by
converting the scores to normal scores using the QuantPsyc package in R1, maintaining the
original mean and standard deviation. Additionally, we performed supplementary generalized
linear mixed model with gamma distribution to account for skewness in the data, which yielded
identical results (see R notebook on the OSF page of the research project).
We used lme42 (42) to perform general and generalized multilevel analyses. We used
jtools3 to provide estimates of significance and to perform simple slopes analyses for the
purposes of understanding two- and three-way interactions. Because wise reasoning scores were
highly skewed, we cross-validated the results applying a generalized linear model with gamma
distribution to the non-transformed scores, which yielded largely identical results (see R
Notebook on Open Science Framework page of the project).

Sample Size and Power Considerations

We aimed to recruit at least 50 participants per cell, similar to the laboratory studies
employing third-person perspective manipulations4,5. We anticipated some attrition over the
course of three months (from the initial online session until the final follow-up measures).
Further, we anticipated the attrition among the participants recruited via the University of
Waterloo subject pool, who could decide to opt-in to the month-long training and subsequent in-
lab sessions for extra pay. Due to these considerations and because we anticipated some
participants not complying with the training, we tripled our target sample size for the initial (pre-
diary) laboratory session. Therefore, we recruited 290 participants for the first laboratory session.
 17

After excluding participants based on pre-registered exclusion criteria, 149 participants (77third-
person/ 72first-person) completed at least two training sessions reflecting on social events during the
diary. On average participants completed 9 training sessions (M = 8.74, Mdn = 8, SD = 4.65,
Range = 2-19). Based on initial pilot data estimates about the frequency of the social events, this
estimate appears typical of the typical number of the targeted social events people report over the
course of a month. The total number of completed diary days, including days without training
but not considering free-writing days, was higher (M = 16.35, Mdn = 18, SD = 5.54, Range = 2-
24).
One hundred twenty participants (57third-person/ 63first-person) completed the post-diary in-lab
sessions and 66 participants (30 third-person / 36 first-person) successfully filled out the one-month
follow-up questionnaire. At the end, these participants provided ratings of experiences for 600
forecasted events in total, with most participants providing ratings for at least one event
occurring over the course of the month. We provide attrition-bias-probing analyses below.
Following the pre-registered analytic plan, we used a full information maximum-
likelihood estimator to fit mixed models and parsimoniously account for missing data with
minimal information loss 6. This approach allows to control for attrition between different
measurement points. To enhance power and generalizability, we further used a mixed effects
framework, with responses from respective dimensions (e.g., wise reasoning, event type,
emotion valence type) as random factors nested within participants, rather than averaging
responses in higher-order categories and examining interactions of nested factors or creating
difference scores (e.g., pre-post / positive-negative).
Supplementary GPower (3.1.9) analyses of the conventional α (5%) with β/α ratio of 4,
assuming an effect size of r = 0.21 (the average effect size in social psychology7), indicated that
the observed sample of 149 pre-diary participants for training wise reasoning produced a power
estimate of .76. Simulations of the power for the Time X Condition effect of wise reasoning (n =
20) via the simr package 8 with Kenward-Roger method for estimation of observed effect size of
0.11 suggested that the power of the study was 75% [50.90, 91.34].
Exclusions
Following pre-registered protocol (osf.io/egcn3), we excluded 15 participants (8 = third-
person condition / 7 = first-person condition) who reported fewer than two social events across
four diary weeks and were therefore not subjected to training.
Methods
Laboratory Sessions Before and After the Diary Intervention
Social event description. Upon being welcomed by an experimenter, most of the
instructions were administered via a computer. For descriptions of social events, participants
received a definition of a difficult social event first:
Our social life can be a source of great joy, but it can also produce great distress. For instance,
some people face conflicts or arguments with other people, such as an argument with a family
member, a relationship conflict with one's partner, or a disagreement with colleagues at work.
 18

Other people encounter social interactions with other people involving annoyances or other
things that irritate them. For example, some people are annoyed/irritated by a friend who
constantly monopolizes their conversations, a partner forgetting to do something they promised
to do or having to eat at a restaurant they particularly dislike because a family member insists on
having a birthday celebration there.
Event reconstruction, reasoning, and post-stream-of-thought questionnaire.
Participants’ recall of social events was guided by probes that encouraged detailed reconstructing
of the experience. This procedure has previously been shown to attenuate recall biases9.
Participants recalled “what you did and how you felt during this event.” Subsequently,
participants provided the first name of the other person involved in the event. Next, they
provided details about the event (i.e., day of the week, time of day, location, length), and lastly,
they described the event in writing in as much detail as possible.
Following the narratives (which subsequently were analyzed for wise reasoning),
participants completed the 21-item Situated Wise Reasoning Scale (SWiS10). The items were
presented in random order over three pages. The instructions on each page read:
As you reflected on and wrote about your most recent difficult social event, to what extent did
you engage in the following thoughts? None of the statements listed below are supposed to be
"good" or "bad". We are simply interested in how people approach difficult situations.

Participants reported the extent to which they engaged in four themes of wisdom-related
reasoning while reflecting on their reinstated social conflict: (i) intellectual humility, which
involves recognizing the limits of one’s knowledge (4 items; sample item: “Looked for any
extraordinary circumstances before forming my opinion”; α = 0.76), (ii) recognition of change (4
items; sample item: “Believed the situation could lead to a number of different outcomes”; α =
0.77), (iii) recognition of others’ perspectives (4 items; sample item: “Made an effort to take the
other person's perspective”; α = 0.80), (iv) consideration or search for compromise and conflict
resolution (5 items; sample item: “Tried my best to find a way to accommodate both of us”; α =
0.78”). All items were rated on a 5-point scale (0 = Not at all, 4 = Very much). We excluded the
fifth component of the SWiS, adopting an outsider’s perspective, because it was not assessed in
the performance measure of wise reasoning. it
Forecasting accuracy scenarios and prompts. During the post-intervention laboratory
session, participants read about six common social events that could occur between themselves
and a person to whom they are close. Participants read the following instructions:
Over the next few pages, you will read about 6 common social events that could happen between
you and [Person X] over the next month. For each event, briefly write how you think this event
might unfold between you and [Person X] and answer a few questions about how you might feel
about these events.
Although you may find some of these events unlikely, please try to consider as best as you can
how you would feel about such events if they did happen.
The six common social events were presented in random order. After reading a brief description
of each event, participants wrote 1-2 sentences about how they thought the event with Person X
 19

could unfold, to help participants reflect on the experienced. These responses were short and
were therefore not analyzed for the purposes of this project.
Betrayal of trust. Participants read, “[Person X] does something that makes you lose
some trust in him/her. This can include lying to you, failing to keep a promise, or speaking badly
about you to others.”
Exclusion. Participants read, “[Person X] makes you feel left out. This can include
failing to invite you to an activity, failing to include you in a conversation, or otherwise ignoring
you.”
Disagreement. Participants read, “You disagree about something with [Person X] and
this turns into a heated argument with him/her.”
Support. Participants read, “You are going through a difficult time and [Person X] is
there for you. This can include staying with you when you are sick, being a good listener, or
giving you some help to get through a problem.”
Informal gathering of friends. Participants read, “You are hanging out with [Person X],
and this can include activities like having dinner together, a celebration, or meeting to do an
activity you all enjoy like watching sports or playing games.”
Pleasant conversation. Participants read, “You are talking with [Person X], and you feel
understood and appreciated during this interaction.”
Results
Effects of Training on Forecasting Accuracy for General Mood
Participants’ ratings of forecasted and experienced mood were subjected to a general mixed
effects model, with participants and events as random factors, event valence (positive vs.
negative) and time (forecast vs. experience) as Level 1 predictors, and perspective condition as a
Level 2 predictor. The results indicated a significant effect of event type, B = 2.99, SE = 0.09,
t(653) = 34.02, P < 0.001, showing that people reported greater happiness after positive events
compared to negative events, and no significant 2- or 3-way interactions, ts <1, ns, suggesting
that there was no difference in anticipated vs. experienced mood, and no training effects.
Emodiversity: Differentiation and Balance of Emotions Experienced During the Conflict
We explored training-based shifts in emodiversity, an index combining the breadth (i.e.,
number of negative emotions experienced) and evenness of affect during the conflict11–13. In a
conflict situation, a person with a narrow and less balanced emotional experience may report
being uniformly angry, whereas a person with a broader and more balanced emotional
experience may report being annoyed, worried, and upset as well14,15.
There are reasons to think that emodiversity would result from the training intervention.
Some work has characterized a balanced representation of emotional experiences as an affective
aspect of wisdom11,16. Also, empirical research has shown that emodiversity can promote mental
and physical health 17, as can engaging in wise reasoning 18. Therefore, we tested whether such
emodiversity could be enhanced via illeist training.
 20

Assessing emotions. After reconstructing the conflict event in the lab and describing
their stream of thoughts, participants rated the extent to which they experienced each of 20
emotions during the difficult social event. Participants rated 10 negative emotions (i.e., afraid,
angry, guilty, sad, nervous, annoyed, ashamed, alone, irritable, blue) and 10 positive emotions
(i.e., happy, proud, joyful, strong, delighted, confident, cheerful, calm, relaxed, at ease),
presented at random, (0 = Not at all; 6 = Very much). Because a key focus of the reflection
concerned a negative experience, we performed separate analyses for positive and negative
emodiversity.
Quantifying emodiversity. To calculate emodiversity, we followed established methods
to quantify the breadth and evenness of participants’ emotions based on Shannon’s entropy
𝑠
formula∑𝑖=1(𝑝𝑖 × 𝑙𝑛 𝑝𝑖 ), via the vegan package in R19. Following prior work11,14, we regressed
out average positive and negative intensity and performed subsequent analyses on
unstandardized residuals, thereby limiting confounding by emotional intensity.
Analyses. We fit a series of linear mixed models with participants as a random factor,
time (pre-post) as a Level 1 predictor, and perspective condition as a Level 2 predictor. We
observed no pre-post growth, or condition-dependent-growth in positive emodiversity, all ts < 1,
ns.
In contrast, we observed a significant training effect for negative emodiversity, Time X
Condition interaction: t = 2.03, χ2 = 4.12, P = 0.042: Participants in the third-person condition
showed greater emodiversity after the intervention than before it, B = 0.064, SE = 0.031, t = 2.07,
P = 0.040. Participants in the first-person control condition did not show notable change, B = -
0.027, SE = 0.033, t = 0.84, P = 0.404. This interaction effect was robust when controlling for
demographics and the types of training experiences and events on which people reflected, Time
X Condition interaction: t = 2.20, χ2 = 4.83, P = 0.028.
Replicating prior research (24), negative emodiversity was also associated with wise
reasoning, B = 0.222, SE = 0.067, t = 3.30, P < 0.001. Additional analyses probing indirect
effects indicated that 13% of the variance in wise reasoning growth in the third-person training
condition of the intervention was accounted for by increases in negative emodiversity, B = 0.018,
95% CI [0.004; 0.03], P = 0.008.
Intervention-Based Changes on the Self-Reported Wise Reasoning Questionnaire
In exploratory analyses, we fit a linear mixed model with participants, and scale items as
random factors, time (pre-post) as a Level 1 predictor, and perspective condition as a Level 2
predictor. Results indicated a significant main effect of time, B = -0.070, SE = 0.029, χ2 = 5.62,
P = 0.018, suggesting that participants’ self-reported reasoning scores decreased after the
intervention as compared to before it. There was a non-significant trend for moderation by
training-condition, B = 0.099, SE = 0.059, = 1.69, P = 0.091. We conducted simple slopes
analyses to probe this trend. Participants in the first-person control condition showed a
significant reduction in self-reported wise reasoning after (vs. before) the diary, B = -0.12, SE =
0.04, t = 2.84, P = 0.005, whereas participants in the third-person training condition sustained a
comparable degree of self-reported wise reasoning when comparing scores before and after the
intervention, B = -0.02, SE = 0.04, t = 0.55, ns. Examining the sub-components revealed that this
was trend was driven by changes in the search for conflict resolution component: first-person: B
 21

= -0.17, SE = 0.08, t = 2.19, P = 0.029, third-person: B = -0.01, SE = 0.07, t = 0.15, ns., with no
significant effects on other components, ts < 1.49, ns. These exploratory analyses suggest that the
effect of training is more pronounced in the pre-registered method of coding spontaneous
reasoning in written narratives as compared to the exploratory method of self-report
questionnaire. Though speculative, this discrepancy may be due to distinct levels of analysis
(spontaneous open-ended responses vs. structured responses prompted via a questionnaire), or
the procedural flow (participants first wrote down their narratives and only thereafter filled out a
questionnaire). Future research should aim to replicate and further unpack this exploratory
pattern of results.
Control Analyses
Manipulation checks. As described in the preregistered analysis protocol, we performed
two types of manipulation check analyses. First, we examined whether participants
spontaneously utilized third-person singular pronouns (he/she) and their name or first-person
singular pronouns when describing their thoughts on the free-writing day (Day 7) of each week.
To this end, a hypothesis-blind research associate counted the first-person singular pronouns and
the third-person pronouns and name references to the participant in the study. We examined the
proportions of first- vs. third-person self-references as a degree of spontaneous self-immersion
on days without prompts to apply an assigned perspective. We focused on the relative proportion
to control for general degree of self-references. We fit a linear mixed model with participants as
a random factor and perspective condition as a between-subjects predictor. Results indicated that
the manipulation was successful, with 16% less spontaneous self-immersion in the third-person
condition than in the first-person condition, B = -0.100, SE = 0.028, t = 3.58, P = 0.0003.
Second, we examined training-driven shifts in narratives participants wrote in the
laboratory sessions that took place before and after the training intervention. In the pre-
registration protocol, we stated that we would test whether participants spontaneously utilized
third-person singular pronouns (he/she) and their name rather than first-person singular
pronouns. However, the results indicated that no participant referred to themselves in third-
person in the post-intervention laboratory session, which took place one week after participants
completed the diary. In hindsight, this is not surprising given that spontaneous self-reference in
the third person in a written narrative about a social event is quite rare.
Therefore, deviating from the preregistered plan, we examined another way to measure
spontaneous self-immersion used in prior research20,21: the proportions of independent, self-
immersive references (first-person singular pronoun; me, mine) relative to the interdependent
self-references (first-person plural pronouns; us, our), focusing on the proportions to control for
number of total pronouns. Given the exploratory nature of these analyses, we examined simple
effects despite a non-significant Condition X Time interaction, t = 1.573, P = 0.116. Participants
in the third-person condition demonstrated a significant drop in the degree of self-immersion
from pre- to post-intervention, B = -0.046, SE = 0.022, t = 2.062, P = 0.041. There was no
difference from pre- to post-intervention observed in the first-person control condition, B =
0.005, SE = 0.024, t = 0.212, ns.
 22

Distinctiveness of the post-training events. We hypothesized that our intervention

during the diary would have carry-over effects on reasoning about a novel challenge. However, it
is possible that some participants selected a recurring event in the post-diary laboratory session.
To control for the novelty of the event, we performed analyses with dummy-coded event novelty
as a control variable (coded by a hypothesis- and condition-blind research assistant) and wise
reasoning as the outcome. Results indicated a main effect of novelty, such that wise reasoning
was greater for events on which participants had previously reflected during the diary period, B =
0.157, SE = 0.052, t = 3.03, χ2 = 9.18, P = 0.002. Results also indicated a main effect of training,
such that wise reasoning was greater post- as compared to pre-intervention, B = 0.053, SE =
0.022, t = 2.44, χ2 = 5.93, P = 0.015. Furthermore, controlling for novelty, this growth effect for
wise reasoning continued to be qualified by a significant Time X Condition interaction, t = 2.53,
χ2 = 6.38, P = 0.012. Simple slopes indicated significant growth pre- to post-intervention in the
third-person training condition, B = 0.103, SE = 0.030, t = 3.479, P < 0.001, but not the first-
person control condition, B = .0001, SE = 0.030, t = 0.007, ns.
Probing Attrition-Based Bias
The attrition between lab sessions was not significantly different between conditions, B =
0.18, SE = 2.39, z = 0.07, P = 0.94.
 23

Training-Based Growth in Wise Reasoning: Sub-Dimension Analyses

We observed significant training-contingent growth (pre- to post-intervention) in three of

five dimensions of wise reasoning, with a fourth dimension of search for compromise showing a
weak trend in the same direction (see Figure 1 in the main test and the Table below).

Estimates for each wise reasoning dimension separately for training and control conditions
Est. S.E. t P
Third-person
Intellectual Humility training 0.1574 0.0394 3.9988 0.0001

First-person control 0.0531 0.0412 1.2889 0.1981

Third-person
Recognition of Change training -0.0405 0.069 -0.5864 0.5586

First-person control -0.0335 0.072 -0.4657 0.6421

Third-person
Perspective-taking training 0.2315 0.0607 3.8137 0.0002

First-person control 0.0637 0.0635 1.0042 0.3158

Third-person
Search for Compromise training 0.0424 0.0411 1.0307 0.3033

First-person control -0.0098 0.0429 -0.2285 0.8194

Search for Conflict Third-person
Resolution training 0.2898 0.069 4.2024 0.0001

First-person control 0.0242 0.0721 0.3356 0.7373

 24

Supplementary References

1. Fletcher, T. D. QuantPsyc: Quantitative Psychology Tools. R package version 1.5. (2012).

2. Bates, D., Maechler, M., Bolker, B. & Walker, S. Fitting Linear Mixed-Effects Models
Using lme4. J. Stat. Softw. 67, 1–48 (2015).
3. Long, J. A. jtools: Analysis and Presentation of Social Scientific Data. (2018).
4. Grossmann, I. & Kross, E. Exploring Solomon’s Paradox: Self-distancing eliminates the
self-other asymmetry in wise reasoning about close relationships in younger and older
adults. Psychol. Sci. 25, 1571–1580 (2014).
5. Kross, E. & Grossmann, I. Boosting wisdom: Distance from the self enhances wise
reasoning, attitudes, and behavior. J. Exp. Psychol. Gen. 141, 43–48 (2012).
6. Enders, C. & Bandalos, D. The Relative Performance of Full Information Maximum
Likelihood Estimation for Missing Data in Structural Equation Models. Struct. Equ.
Model. A Multidiscip. J. 8, 430–457 (2001).
7. Richard, F. D., Bond Jr., C. F. & Stokes-Zoota, J. J. One Hundred Years of Social
Psychology Quantitatively Described. Rev. Gen. Psychol. 7, 331–363 (2003).
8. Green, P. & MacLeod, C. J. simr: an R package for power analysis of generalised linear
mixed models by simulation. Methods Ecol. Evol. 7, 493–498 (2016).
9. Schwarz, N., Kahneman, D. & Xu, J. Global and episodic reports of hedonic experience.
in Calendar and time diary: Methods in life events research (eds. Belli, R., Alwin, D. &
Stafford, F.) 157–174 (Sage Publications, 2009).
10. Brienza, J. P., Kung, F. Y. H., Santos, H. C., Bobocel, D. R. R. & Grossmann, I. Wisdom,
bias, and balance: Toward a process-sensitive measurement of wisdom-related cognition.
J. Pers. Soc. Psychol. 115, 1093–1126 (2018).
11. Grossmann, I., Oakes, H. & Santos, H. C. Wise Reasoning Benefits from Emodiversity,
Irrespective of Emotional Intensity. J. Exp. Psychol. Gen. (2019).
doi:10.1037/xge0000543
12. Benson, L., Ram, N., Almeida, D. M., Zautra, A. J. & Ong, A. D. Fusing Biodiversity
Metrics into Investigations of Daily Life: Illustrations and Recommendations With
Emodiversity. Journals Gerontol. Ser. B Psychol. Sci. Soc. Sci. 73, 75–86 (2018).
13. Ong, A. D., Benson, L., Zautra, A. J. & Ram, N. Emodiversity and biomarkers of
inflammation. Emotion 18, 3–14 (2018).
14. Grossmann, I., Huynh, A. C. & Ellsworth, P. C. Emotional complexity: Clarifying
definitions and cultural correlates. J. Pers. Soc. Psychol. 111, 895–916 (2016).
15. Lindquist, K. A. & Barrett, L. F. Emotional complexity. in The Handbook of Emotions
(eds. Lewis, M., Haviland-Jones, J. M. & Barrett, L. F.) 513–530 (Guilford, 2010).
16. Bangen, K. J., Meeks, T. W. & Jeste, D. V. Defining and assessing Wisdom: A review of
the literature. Am. J. Geriatr. Psychiatry 21, 1254–1266 (2013).
17. Quoidbach, J. et al. Emodiversity and the emotional ecosystem. J. Exp. Psychol. Gen. 143,
2057–2066 (2014).
18. Grossmann, I., Na, J., Varnum, M. E. W., Kitayama, S. & Nisbett, R. E. A route to well-
being: Intelligence versus wise reasoning. J. Exp. Psychol. Gen. 142, 944–953 (2013).
19. Oksanen, J. et al. vegan: Community Ecology Package. (2018).
20. Grossmann, I. & Kross, E. The Impact of Culture on Adaptive Versus Maladaptive Self-
Reflection. Psychol. Sci. 21, 1150–1157 (2010).
21. Pennebaker, J. W., Mehl, M. R. & Niederhoffer, K. G. Psychological aspects of natural
 25

language use: Our words, our selves. Annu. Rev. Psychol. 54, 547–577 (2003).