How is trust learned?

How Is Trust Learned?

Master Thesis

Aron Groen

In collaboration with Saskia van der Hoogen

a a o o a o o a

Master thesis Psychology, specialization Economic

&

Consumer Psychology Institute of Psychology

Faculty of Social and Behavioral Sciences

_-

Leiden University Date: June 14 2019

Student number: s22623 12

First examiner of the university: Michael

Giffin

Second examiner ofthe university: (OpL) External

Abstract

Trust is an essential aspect of our economic and social lives. Although trust has been extensively measured in the Trust Game, the difference between social and non-social contexts has not yet been investigated. The current research uses neuroticism, being strongly linked to trust, to explain possible differences in average and learned trust in both contexts. During the study, 101 participants completed the brief version of the Eyseneck Inventory of Personality and took part in both social and non-social conditions of the Trust Game. Results indicated a negative effect

of

neuroticism on average trust, but no effect on learned trust. There was no significant interaction between neuroticism and condition on trust. The findings expand our understanding of the relationship between neuroticism and trust in different contexts.

Introduction

Trust is a concept which is diffrcult to measure and subject to ambiguous interpretation. Yet, it is relevant to understand how trust is learned since it is a basic element in our economic and social lives. The current research aims to clarify how trust is learned and the role

of

neuroticism in learning to trust.

Trust concerns ideas that one has about the behavior of another (Cox,2004). Many definitions put emphasis on this idea, such as'oan action that is trusting of another is one that creates the possibility of mutual benefit,

if

the other person is cooperative, and the risk of loss to oneself if the other person defects" (Cox,2004,p.263) and "positive expectations of the

intentions or behavior of another" (Rousseau et al., 1998, p. 395).

The two-stage Trust Game by Berg, Dickhaut

&

McCabe (1995) is a popular way to measure trust (Johnson

&

Mislin,

20II).

A

short description is as follows: the game starts by the "trustor" receiving an endowment, usually $10 (Camerer,2011). The trustor can invest a portion of this endowment in the opposing player (the "trustee") which is often tripled equaling $30. The amount invested is a measure of trust. The trustee, in turn, can return apart of this endowment to the trustor indicating trustworthiness. The game can consist of multiple rounds. Previous

research showed a strong positive correlation between the amount sent by the trustor and the amount returned by the trustee (Glaeser et a1., 2000). The amount returned by the trustee in turn has a positive influence on how much the trustor invests in subsequent rounds in iterated versions of the Trust Game (Anderhub, Engelmann

&

Güth, 2002).

Earlier research stated that single and multiple-round Trust Games measure different versions of trust: respectively average and relationship-specific trust (Van den Bos, van

Dijk

&

Crone, 2013). Measured in single-round Trust Games, 'average trust' is influenced by personality. Relationship-specific trust is less directly influenced by personality, as inferences concerning the opponents' perceived tnrstworthiness play a role as well(Zhang, 1999). Both types of trust are uncorrelated as shown by analyses of multiple questionnaires measuring the two constructs (Zhang,1999; Johnson-George

&

Swap, 1982). The current research aims to investigate both average and learned trust.

A major part of the Trust Game research so far used computer opponents because it made data collection easier. These experiments aimed to make participants think their counterparts were real. Yet, manipulation checks of this deception were often missing in such experiments

(Johnson

&

Mislin, 2009). This could have confounded experimental results. To illustrate, other research showed higher amounts sent in economic exchanges with persons than with computers (Bottom et a1.,2006). As the experimental method in Bottom et al. their study did not include the Trust game, there is uncertainty about the difference between person- and computer contexts in the current experiment.

Investigating the comparison between person-person and person-computer interaction is also relevant for showing the influence of social factors on trust. Recent experimental research using the Ultimatum Game showed that economically-optimal behavior is learned faster in person-computer (non-social) than person-person (social) situations (Giffrn et al., manuscript

in

preparation). Participants in the social conditions also gave higher offers than those in the non-social condition, in congruence with research described earlier. The difference was ascribed to the so-called fairness nonn, making participants rely on social concerns rather than showing economically optimal behavior. An important fact to note is that the Ultimatum Game is similar to the Trust Game but focuses less on trust by not allowing reciprocal offers (Civai

&

Hawes, 2016). As such, there is still uncertainty regarding what the difference between social and non-social contexts looks like when it comes to (learning) trust. To predict trust, examining

personality traits can be useful as personality and the two types of trust are at least to some extent related. Other research underlines the important role of personality by indicating opposite effects on trust by the personality traits extraversion and neuroticism (Evans

&

Revelle, 2008).

Neuroticism may especially help explain average trust levels and trust learning. Being part of the popular Big Five model of personality (Goldberg, 1990), neuroticism was defined as a

temperamentaltrait of emotionality, having unrealistic ideas or a disposition to experience aversive emotional states (Ormel, Rosmalen

&

Farmer, 2004). The described averseness seems

to contradict the openness to risk described as trust earlier in this paper. Indeed, a strong negative correlation was found between neuroticism and trusting behavior (Dohmen et a1.,2008).

Neuroticism has led to lower amounts invested in the Trust Game in different research contexts (e.g. Ainsworth et al,2014; Evans

&

Revelle, 2008; Müller

&

Schwieren,2012; Ben-Ner

&

Halldorsson,2012). To fuither illustrate, trait anxiety and neuroticism have a strong positive relationship (Schinka et a1.,2004). As neuroticism is a factor with a strong predictive value

of

Neuroticism and trust are linked, but it is not clear to what degree. In order to provide a precise quantification ofneuroticism its influence on average and learned trust, in the current experiment every participant played in both social and non-social conditions. This distinction could influence the negative effect of neuroticism on trust in two ways: leading to relative distrust in computers compared to humans or vice versa. These possible effects result from difference ways

noÍn

adherence and social anxiety could interact. Previous research has shown a strong positive correlation between neuroticism and social anxiety (Boelen

&

Reijntjes,2009) leading to the plausible conclusion that neurotic participants show relatively less trust in the social condition of the Trust Game. However, an alternative account has been demonstrated by research by Heinrichs et al. (2006) indicating social anxiety and social norm adherence are

positively correlated. Participants in the social condition could show more trusting behavior than economically optimal because of so-called 'social sanctioning' (Fehr

&

Gachter, 2000) or 'belrayal aversion' (Bohnet et al., 2008). Similarly, fairness norms were earlier used to explain higher offers in a social relative to a non-social condition. As a result, the prediction that neurotic people show higher trust because they are afraid to violate social noûns seems plausible as well.

With the current design we

will

be able to address the seemingly disparate findings. Together, the above is used to formulate the two hypotheses of the current research:

Hypothesis 1: Neuroticism has a negatÍve effect on average economic

trust

and

trust

learning

in

a computer as well as in people; and

Hypothesis 2: The negative effect of neuroticism is stronger

for

average

trust

and

trust

learning

in

people than computers.

With respect to the two aforementioned hypotheses, the current research tested the effect of neuroticism on average and leamed trust in a social and non-social context. Knowledge of the role of neuroticism could benefit society by knowing better how trust is influenced in cases

of

low economic trust, such as during an economic crisis. We tested our hypotheses by comparing trust in the Trust Game in a computer- or human counterpart condition using a within-subjects design. In each condition participants played multiple rounds against three trustworthiness levels, allowing for manipulation of trust. We measured neuroticism using a questionnaire and analyzed the data together with levels of average and learned trust. We expected our data to show that

neuroticism has a negative influence on both average and learned trust in both social and non-social contexts. We also expected that the negative influence of neuroticism on average and learned trust was stronger in the social context. Using this approach, we aimed to answer the overarching question: what role does neuroticism play in learning to trust in different contexts?

Method Participants

In our pilot study, data of 30 participants was measured online via Mturk. As for the actual experiment, we recruitedI2T participants of an age between 18 and 35 at the social sciences faculty of Leiden University. Each participant in the actual experiment participated in both non-social and social conditions.

All

participants received a show-up fee of €6.50.

Participants in the experiment received an additional sum of between

€l

and €2 depending on their performance in the task.

Materials

The pilot study consisted of a questionnaire asking the participant "How much

will

you return for €0?" for an amount up to €20. For each offer, participants could indicate an amount between €0 and €20. Participants of the Trust Game experiment played against these previously recorded responses. The responses were divided into three trustworthiness levels per condition, meaning a 2x3 within-subjects design. Each trustworthiness level was indicated to the participant by a unique sign.

The Trust Game experiment is based on the core idea by Berg, Dickhaut

&

McCabe (1995). The trustors received an endowment of €20. They could invest a portion of this amount

to their counterpart, the trustee. The indicated investment was the dependent variable trust in the current experiment. The amount was then trþled. The experiment consisted of 72 rounds for each of the two main conditions. Participants randomly played against a different counterpart

each round while the unique sign indicated the opponent.

The independent variable neuroticism was measured using the brief version of the

Eysenck questionnaire of personality (Sato, 2005). This questionnaire measured extraversion and neuroticism

in

12 items.

All

items focused on neuroticism had a factor loading higher than.4. Participants could react on statements concerning neuroticism by indicating

'yes' or'no'.

The minimum neuroticism score of the questionnaire was 0 (extremely non-neurotic). The maximum score was 12 (extremely neurotic).

Procedure

Firstly, we ran a pilot study 30 Mturk participants. After reading the information letter, participants indicated their investments, after which they were debriefed and they received their fee. Recorded investments were split into three different categories of responders: low, high and ascending. The low group consisted of returns below 33o/o of the amount transferred, returns

of

the high group were above 33o/o and in the ascending group retums rose in percentage as a

function of the amount transferred. These trustworthiness categories meant the three sub

conditions for both social and non-social main conditions.

For the main experiment we recruited I27 pafücipants by approaching them on campus. After

filling

out the informed consent form, participants completed the neuroticism questionnaire

and played the Trust Game. They were told that their extra payment was dependent on their investments in the social condition and the choice of the other participant. After reading the instructions, participants completed a quiz in order to check they understood the instructions well. After completing the experiment participants were given a debriefing and their fee. The experiment took roughly one hour.

Variable Unstandardized coeffi cient

p

t

SE B Model 1 Neuroticism

Model2

Neuroticism Model3 Condition Neuroticism Neuroticism

*

Condition -0.25 -0.r2 0.53 -0.20

0.t2

0.t2

1.05

0.t2

l7

-2.04* 046 -0.99 32 0.50 ,67 -t.75 .08 -0.35 -0.2r .84

Model4

Condition -0.28 0.18 -1.s9 11 Neuroticism -1.18 1.55 -0.76 .45 Neuroticism * Condition 0.32 0.25 1.27 .21

Notes.

l)

Models

I

ønd 2 respectively predict mean offer and

dffirence

in rneøn offer using Neuroticism as predictor. Models 3 and 4 measure the interaction between predictors

Neuroticism and Condition on mean offer amount and rnean difference in offers, respectively. 2)

The estimated regression slopes of participants' mean offer with condition as predictor is shown in Figure 1. The relatively flat lines per participant confirm the absent difference in mean offer amount per condition. The graphs representing the interaction effect between neuroticism and condition on average and learned trust are shown rn Figure 2 and Figure 3. As can be inferred from Figure 2, the slope of the regression plot does not clearly differ per condition. Instead the graphs show resemblance to parallelism, underlining the absent interaction effect. As for Figure 3 there is again no clear difference in slope between the graphs as they show parallel tendencies, especially with higher neuroticism scores.

lntcrrcüon Plot a

t

o t I a

.

! a õ E t¡¡ o'-Àbn.aæid Sóc¡il Condlüon

Figure 1. Interaction plot of the estimated mean offers per participant in each condition. Each line represents a separate participant.

lntcnctlon Plot

Condition

--. ilonaûcH

Nrurotlcl¡m 9corr

Figure 2.Interaction plot showing the estimated mean offers per condition for each neuroticism score. Intrrectlon Plot Ê at t

;

a E ù_a

;

E a₅ t¡,¡ Condit¡on -So.fl -- l,lor{cg ?

I

a È o 5 at

{

t

_a a G a Ê ct E I a T Nrurotici¡m Scorr

Figure 3. Interaction plot showing the mean difference in amount offered per condition for each neuroticism score.

To sum up, the results provide some evidence that participants with high neuroticism indicate lower amounts offered on average, but there is no evidence that this effect differs

between conditions. Also, neuroticism does not seem to negatively affect offer differences within participants throughout the Trust Game.

Discussion

In this study, the effect of neuroticism on average and leamed trust was investigated in both social and non-social settings. Neuroticism was found to have a negative influence on average trust: participants who showed more signs of neuroticism exhibited lower levels of tnrst in the Trust Game. Yet, we found no difference in negative effect of neuroticism in social versus non-social contexts. Also, we found no effect of neuroticism on learned trust or a difference in effect in different contexts.

The result that neuroticism negatively predicts average trust is consistent with the theory that neuroticism and trust are strongly linked. It has to be noted that the variance of trust

explained by neuroticism in our study is relatively low: R2

:

.04, but significant:

p :.046.

Low amounts of variance explained are a common sight in social sciences since there are many factors to be accounted for in predicting behavior (Weeda, 2019). Thus, we consider the conclusion that neuroticism decreases average trust plausible.

The finding that neuroticism does not negatively affect leamed trust is not in line with our expectations. We expected this effect based on research showing neuroticism influences learned trust like average trust does, although in a less direct fashion (Zhang,1999). The effect of neuroticism on learned trust could simply be too weak considering neuroticism already explained a marginal amount of variance in average trust. On a theoretical level this could imply that there are more factors to be accounted for, when predicting learned trust using neuroticism, than expected.

The absence of difference in negative effect of neuroticism on average trust between social and non-social contexts was not in line with our expectations either. We expected to see a

difference based on the strong link between neuroticism and social anxiety, making neurotic people less trusting in social situations. There are different possible explanations for the absence of effect. Firstly, decreased trust in social contexts resulting social anxiety may not have

outweighed the tendency for norm adherence. Norm adherence results from the same social anxiety but could lead to increased expressed trust as described in our theory. The absence of an increased negative effect of neuroticism on trust in social contexts could be because of an equilibrium between two processes resulting neuroticism and social anxiety. This equilibrium could be a new theoretical insight concerning the tradeoff between different effects

of

what extent they resonate with different motives for expressing trust, which are certain to lead to a distinct outcome, such as norm adherence leading to more trust.

Another explanation for the absence of different effects of neuroticism on trust in social

and non-social contexts is our experimental method. We may not have induced enough feelings corresponding social situations, relative to the non-social context. The difference in conditions could have been increased

if

participants were to directly interact with their opponents in the social condition. Although we did not do a manipulation check by asking how 'social' this part of the experiment felt for participants, we have reason to believe our conditions were valid. Firstly, we used no deception, giving us more hope that subjects believed that the experiment concerned a social situation. Secondly, we made sure participants realized the social nature of the condition by quizzingthem during the introduction. Future research could compare our current social condition to a condition in which participants briefly interact with each other during the experiment and measure possible differences in trust.

A third explanation focusing on the salience of social processes in the social condition is the fact that the experiment was repetitive and boring.

A

substantial amount of participants reported boredom or annoyance. This could have decreased the salience

ofthe

social nature

of

the task and therefore, to some extent, its external validity. Thus, the lack of difference found in negative effect of neuroticism on trust between contexts needs to be interpreted with caution.

It

could be that non-bored neurotic people show larger differences in trust between contexts. To prevent boredom and annoyance, future research could decrease the number ofrounds per condition in order to measure whether a decreased reliability resulting less rounds comes at an increase in effect size and validity.

The preliminary conclusion can be that in this research a negative effect of neuroticism on average trust was found, but that this effect did not differ between social and non-social contexts. Regarding trust learning, neuroticism alone does not seem to be of substantial impact. This conclusion is relevant to further understand how trust is learned and what role personality plays in trusting behavior. In a society in which trust is such a basic element, our results seem to imply that neurotic people have a harder time functioning optimally in economic interactions.

Results

Analyses were run on neuroticism scores obtained using the Eysenck questionnaire

of

personality (Sato,2005) as well as amount sent during the Trust Game. An amount of 72 offers per condition meant 144 offers

to

analyze per participant. Of the I27 participants,26 showed invalid or missing data and 3 participants reported psychiatric disorders. These data were left out of the analysis, resulting in a participant amount of

N:

I0l

(Nwonen: 87).

As for the data analysis, there were no extreme outliers with a SD more extreme than + 3.

We calculated mean offers over all trials as well as per condition and the mean difference between participants' first and last offer over sub conditions (meaning learned trust), see Table

I

This allowed for examining the effect of neuroticism on average trust and learned trust in both contexts. Also, these data could be used to investigate a possible interaction effect between neuroticism and condition on average and learned trust.

Table I

Mean Amount Offered, Mean

Dffirence

Between First and Last Offer and Standard Deviations per Conditíon

Condition Mean neuroticism score (SD)

Mean offer

(^sD)

Mean difference first and last offer (SD)

Social e.Os

(4.13)

0.42 (6.0s)

Non-social 8.70

(4.01)

-0.06 (s.81)

Both s.26 (3.37) 8.74

(s.s6)

0.18 (4.18)

Firstly, we ran two paired samples /-tests in order to check for differences in mean offer and mean difference between first and last offer between conditions. There was no significant difference in average trust between conditions: I (100)

:

-1.22,

p:

.23. There was also no significant difference in learned trust: / (100)

:

5.7I,

p:

.57. This means that the condition manipulation seemed to have no effect on learned trust.

We tested hypothesis 1, stating neuroticism has a negative effect on average and learned trust as well as in persons, by examining the predictive value of neuroticism on average and learned trust by frtting two regression models. The first model consisted of neuroticism score as predictor and mean offer as outcome. The second model consisted of neuroticism as predictor as well while containing a different outcome variable: first to last offer difference. The model statistics are denoted tn Table 2 under 'Model I

'

and 'Model 2'. Results of regression model I indicated that neuroticism was a significant negative predictor of average trust: B

:

-.25,

p:

.046. Neuroticism also explained a significant proportion of variance in average

trust, R2 _{r¿¡ur¡"¿:} .04,

F(\,99):

4.I0,

p:

.46. Thus, mean offers tended to be lower for participants who scored higher on neuroticism. To test model

2,we

carned out a multilevel regression analysis with intercepts varying between subjects, using the enter method. Results indicated that neuroticism was a non-significant predictor of learned trust: B

:

-0.12,

p

:

.32.

Neuroticism also explained a non-significant proportion of variance in learned trust, ,R 2 _ad¡usted <

.01, ,F(1,

99):0.97,

p:

.33. Thus, neuroticism did not seem to have an effect on learned trust. The results described above suggest that neuroticism is of influence on participants' average trust, but not on learned trust. We tested hypothesis 2, statingthat there is an interaction effect between neuroticism score and condition on average and learned trust, fitting two more models. Model3 was again analyzedusing a multilevel regression, using neuroticism and condition as predictors and mean amount offered as outcome variable. Neuroticism was a non-significant predictor of average trust: ,B

:

-.21,

p:

.08. Condition was also a non-significant predictor of average trust: B

:

0.53,

p

:

.62. To round off, there was a non-significant

interaction between condition and neuroticism score: B

:

-0.35,

p

:

.84. Model 4 was analyzed using the same regression analysis with neuroticism score and condition as predictors, but with mean difference between first and last offers as outcome variable. Neuroticism was a non-significantpredictor of learned trust: -B

:

-0.28,p:.11.

Conditionwas also anon-significant predictor of leamed trust: : B

:

-I.18,

p

:

.45. There was a non-significant interaction between condition and neuroticism on learned trust: B

:

0.32,

p

:

.21.

Table2

References

Ainsworth, S. E., Baumeister, R. F.,

Ariely,D., &Vohs,

K. D. (2014).Ego depletion

decreases trust in economic decision making. Journal of Experimental Social Psychology, 54,40-49.

Anderhub, V., Engelmann, D.,

&

Güth, W. (2002). An experimental study of the repeated trust game with incomplete information. Journal of Economic Behavior &. Organization, 48(2), r97-216.

Ben-Ner,

4., &

Halldorsson, F. (2010). Trusting and trustworthiness: What are they, how to measure them, and what affects them. Journal of Economic

Psychology,3l(I),

64-79. Berg, J., Dickhaut, J.,

&

McCabe, K. (1995). Trust, reciprocity, and social history. Games

and economic behavior,

l0(l),

t22-142.

Boelen, P.

4.,

&

Reijntjes, A. (2009). Intolerance of uncertainty and social anxiety. Journal of anxiety dis orders, 2

3(l\,

1 30- 1 35.

Bohnet, L, Greig, F., Herrmann,B.,

&

Zeckhauser, R. (2008). Betrayal aversion: Evidence from brazil, china, oman, switzerland, turkey, and the united states. American Economic Review, 9

8(l),

29 4-3 10.

Bottom, W.P., Holloway, J., Miller, G. J., Mislin, A.,

&

Whitford, A. (2006). Building a pathway to cooperation: Negotiation and social exchange between principal

and

agent. Adminis trøtive Science _Quarterly, 5/ (1 _), 29 -58.

Camerer, C. F. (2011). Behavioral game theory: Experirnents in strategic interaction. Princeton University Press.Chaudhuri,4., Sopher,8.,

&

Strand,P. (2002). Cooperation in social dilemmas, trust

and

reciprocity. Journal of Economic Psychologt,23(2),231-249. Civai, C.,

&

Hawes, D. R. (2016). Game theory in neuroeconomics. lnNeuroeconomics (pp.

13 -37). Springer, Berlin, Heidelberg.

Cox, J. C. (2004). How to identify trust and reciprocity. Games and economic behavior,46(2), 260-281.

DeYoung, C.G., Peterson, J.

8., &

Higgins, D. M. (2002). Higher-order factors of the Big Five predict conformity: Are there neuroses of health?. Personality and Individual dffirences, 3 3 (4), 533 -5 52.

Di Cagno, D.,

&

Sciubba, E. (2010). Trust, trustworthiness and social networks:

Playing

a trust game when networks are formed in the lab. Journal of Economic Behavior

&

Organization, 75(2), 156-167 .

Dohmen T, Falk A, Huffrnan D, Sunde U. 2008. Representative trust and reciprocity: prevalence and determinants. Econ. Inq. 46:84-90

Evans, A. M.,

&

Revelle, W. (2008). Survey and behavioral measurements of interpersonal trust. Journal of Re s e ar ch in P ers onality, 4 2 (6), 1 5 8 5 - 1 593.

Fehr, E.,

&

Gachter, S. (2000). Cooperation and punishment in public goods experiments. American Economic Review, 90(4), 980-994.

Giffin, M. R., Lebreton, M., Gross, J.,

&

De Dreu, C. (2018, December 1). Do social nonns impede learning? Manuscript in preparation.

Glaeser, E. L., Laibson, D. I., Scheinkman,

J.4.,

&

Soutter, C. L. (2000). Measuringtrast. The

quarterly

j

ournal of economics, I I 5 (3), 8 1 1 -846.

Goldberg, L. R. (1990). An alternative" description of personality": the big-five factor structure. Journal of personality and social psychologlt, 59(6),1216.

Hoff,

K.4.,

&

Bashir, M. (2015). Trust in automation: Integrating empirical evidence on factors that influence trust. Humqn Factors, 57(3),407-434.

Heinrichs, N., Rapee,

R.M.,

Alden,

L.4.,

Bögels, S., Hofmann, S. G., Oh, K. J.,

&

Sakano,

Y.

(2006). Cultural differences in perceived social nonns and social anxiety. Behaviour research ønd therapy, 44(8), ll87-1197 .

Johnson, N. D.,

&

Mislin, A. A. (2011). Trust games: A meta-analysis. Journal

of

Economic Psychologt, 3 2(5), 865-889.

Johnson-George, C.,

&

Swap, W. C. (1982). Measurement of specific interpersonal trust: Construction and validation of a scale to assess trust in a specific other. Journal

of

pers onality and Social Psycholo

gt,

4 3 (6), 1306.

King-Casas,8., Tomlin, D., Anen, C., Camerer, C. F., _Quartz, S. R.,

&

Montague, P.

R.

(2005) Getting to know you: reputation and trust in a two-person economic exchange. Science,

308(57 t8),78-83.

Müller, J.,

&

Schwieren, C. (2012\. What can the Big Five personality factors contribute to explain small-scale economic behavior?

Ormel, J., Rosmalen,J.,

&

Farmer, A. (2004). Neuroticism: a non-informative marker

of

vulnerability to psychopathology. So cial Psychiatry and P sychiatric Epidemiolo

gt,

39(1 1), 906-912

Rousseau,

D.M.,

Sitkin, S. 8., Burt, R. S.,

&

Camerer, C. (1998). Not so different after all:

A

cross-discipline view of trust. Academy of management review,23(3),393-404. Sato, T. (2005). The Eysenck personality questionnaire brief version: Factor structure and

reliability. The Journal of psychologlt, 139(6), 545-552.

Schinka,

J.4.,

Busch,

R.M.,&

Robichaux-Keene, N. (2004). A meta-analysis of the

association between the serotonin transporter gene polyrnorphism (5-HTTLPR) and trait anxiety. Molecular psychiatry, 9(2), 197.

Song, F. (2008). Trust and reciprocity behavior and behavioral forecasts: Individuals versus group-representatives. Garnes and Ec onomic Be hav ior, 6 2(2), 67 5 -69 6.

Van den Bos, 'W., van Dijk, E.,

&

Crone, E. A. (2012). Learning whom to trust in repeated social interactions: a developmental perspective. Group Processes & Intergroup

Relations, I 5(2), 243 -256.

Weeda, _{W. (2019). Applied Data Analysis: lecture 2 [Powerpoint Slides]. Retrieved from} blackboard. leidenuniv. nl

Zak,P. J. (2008). The neurobiology of trust. Scientffic American,298(6),88-95.

Zhang, J. (1999). Distinction between general trust and specific trust: Their unique patterns with personality trait domains, distinct roles in interpersonal relationships, and different functions in path models of trusting behavior.

Appendix A - Assumptions

Linearity

Checking the assumption of linearity, we created a graph of the standardized predicted value and residual as shown rn Figure 3. As can be observed, there are no clear signs

of

curvilinearity. Thus, we believe there is no systematic deviation, considering the assumption

of

linearity met.

Standerdizcd Prcdictcd

Mean OlTcr

HcightAnd

Rc¡idurl

PGr

Participant

öo

0 0 00 00

0

0 o(D

o

m oÕoo

o o

(}@o

o o

ooo

_o(Ð

o o

o

ôo

ooæ

æ

o

æ

o o o Õ o o -1 o

E

E

a, +r .c

E

_J

L

o-E

_J

.tr

E

É

E

Ê rt fl

o

o

o o

oo

(DÔ

OÕ

oco

00

oüE

_o

ooo

o

oæ

fD

o

æoo

o o

ooo

{)

oo

m

{o o 1 o o 0 o o o

o

o

0

o o

o

o

o0

o

ô

oo

o

Õo

o

öÕ

o

_oo

Õo

o

_oo

o

00

o o

o

mo

oo

o (D o o

o

eco

oo

o 0 o o

æ

ooo

o o oooo00

ÔÕô

(DÕ

o

ooo

o

Standardizrd

Residual

Figure 4. Standardized Residual versus Predicted Value plot for offer amount scores.

Normality

The assumption of normality is not relevant since for our analysis F is robust against violations. o o o 3 ?

Homoscedasticity

Checking for homoscedasticity, we again checked the plot rn Figure 3. There were no clear signs of diablo shapes. Thus, we consider the assumption not violated.

Multicollinearity

In order to check for multicollinearity, we inspected the 'coefficients' table of the regression ouþut.

If

the VIF value, indicated in this table, exceeds 10, there is reason to worry (Field, 20Í3). Yet, there is no variable value considerably higher than 1, providing us

with

enough evidence to consider this assumption met.

Independent

Errors

We made sure the assumption of independent errors was not violated by running a

universiteit

Leiden

Assessment Form

Master

Thesis

PsychologyA

FifSt

exúminef

Name

Student:

Aron Groen

Student

lD:

s2262312

Master

programme:

Ous.

Psychology

/

O

MSc Psychology (research)

Master

Specialisation:

Economic and Consumer

Psychology

g

Course code (click

to

auto-fill):

6464.'.

number

of

ECs:

20 Date of receipt Master Thesis at

OSC :

A. Name

first

examiner

lnstitute of

Psychology, Unit B. Name second examiner

lnstitute of

Psychology, Unit C. Name external supervisor (irappticable)i

Affiliation

Assessment

form

C attached

MichaelGiffin

Social and Organisational

Psychology

g

Jörg Gross

Social and Organisational

Psychology

O

Ot"t

/

no@

How ls Trust Learned? Title thesis:

Checked

for

plagiarism (Turnitin)

Approval by

the

Committee Ethics Psychology or

the Committee

Medical Ethics (LUMC)

Number:

Date of approval:

Final grade in consultation

with

second examiner:

written:

E,)\^l

Date:

Name

first

examiner Michael Giffin

Name second examiner

Jörg

Gross

g

Process

o

Practical input

To what extent hæ the student invested time / effort and showed research skills in the various stages ofthe research?

lntellectual input

To what extent did the student contribute independently to the final product?

o

Ethical aspects

To what extent has the student paid sufficient attention to e.g. professional treatment of

participants and conscientious use ofdata and scientific sources?

Evaluation of process (U, S or G):

Content: Thesis

o

Abstract

To what extent is the Abstract prepared in accordance with the requirements ofrelevant joumals? o o

tol

Unsatisfactory Satisfactory Good Unsatisfactory Satisfactory Good

o

lntroduction

To what extent does the Introduction reflect a good range ofrelevant literature, a good conceptual flow and relevant aims and hypotheses?

o

Method

To what extent does the Method section include all major elements, such as description of the

design, sample, meariures, procedures, and statistical choices in such a way that the study could

be readily replicated?

o

Results

To what extent does the Results section report on analyses appropriate to the research questions?

To what extent has the student paid attention to accuracy?

o

Discussion

To what extent does the Discussion section include all major elements?

o

References

APA rules in text and literature lists.

o Overall presentat¡on

Is the presentation ofthe thesis adequate with regards to overall lay-out, structure, title, tables, statistics, appendices, APA rules etc.?

o

Style of writing

To what extent does the student's writing style reflect effort / capacþ to write clearly /

engagingly?

Evaluation of content (lO-point scale):

Additional remarks U

S

G

U

S

G

t!

IA

ø

ú

Universiteit Leiden

Assessment

form

Master

Thesis Psychology

I

Name Student:

Student lD:

Master Programme: Master Specialisation :

A. Name

first

examiner

lnstitute of

Psychology, Unit B. Name second examiner

lnstitute of

Psychology,

Unit

C. Name external supervisor (¡f appricabre)l

Affiliation

How ls Trust Learned? Title thesis:

s2262312

@tvtsc

Psychology

I

OMSc

Psychology (research)

Economic and Consumer

Psychology

g

MichaelGiffin

Social and Organisational

Psychology

O Jörg Gross

Social and Organisational

Psychology

g

Second

exam¡ner

Advice on

finalgrade

_/

grade range

7-8

25.06.2019

Date:

Name second examiner Jörg Gross

Signature:

Content: Thesis

o

Abstract

To what extent is the Absûact prepared in accordance with the requirements ofrelevant joumals?

Unsatisfactory Satisfactory Good

o

lntroduction

To what extent does the Introduction reflect a good range of relevant lit¡rature, a good conceptual flow and relevant aims and hypotheses?

o

Method

To what extent does the Method section include all major elements, such as description ofthe

design, sample, measures, procedures, and statistical choices in such a way that the study could

be readily replicated?

o

Results

To what extent does the Results section report on analyses appropriate to the research questions?

To what extent has the student paid attention to accuracf

o

Discussion

To what extent does the Discussion section include all major elements?

o

References

APA rules in text and literature lists.

o

Overall presentation

Is the presentation ofthe thesis adequate with regards to overall lay-out, structure, title, tables, statistics, appendices, APA rules etc.?

o

Style of wr¡t¡ng

To what extent does the student's writing style reflect effort / capacity to write clearly I engagingly?

Evaluation of content (10-point scale)

Additional remarks U

S

G

(

r'

r'

r'

7-8

Universiteit læiden

lnternship

Assessment

Form

Nameofstudent:

_Afttl¡l

6R0¿/V

student

ro:

gll

_b

"J.'1 t

2

Maste/s programme Maste/s specialisation Course code (see appendix)

Date of receipt by SSC:

MSc in Psychology

Economic and Consumer Psychology

6464EC110Y(ECP)

NumberofEC: 10

Name of examiner at Leiden University: lnst¡tute of Psychology, Unit:

¿hr

c

_H4E

b-

GI

F

trM

Nameof supervisoratinternshiporganisation (if applicable):

j¡.ej1C.rr¿

_Atltog

,

bx"eille

le

îl"oorS rnternshiporganisation,

_,î"+[

*i,

(on-iurtrq,^

4

An^l¿gl1

_(llt^)

|

Social and Organisational Psychology

confirm that the internship report has been checked for plagiarism (Turnitin) and that the check did not g¡ve rise to any suspicion of plagiarism.

!

(l

Grade: Written

Ei3\!

q

o\,r\

lirrt

Signature of examiner at Leiden University:

_\)"\\À'

1.

Assessment

of

proctice

(attitude,

skillsl

by

the examiner at

Leiden

University

very good

Ptease place a

/

,n

rn

box YES to confirm that, in the case of an external internship, the supervisor at the internship organisation was consulted before the final assessment. Please place a check mark in the NA (Not box if the internship was an internal one.

YES NA

2.

Assessment

of

lnternship

Report by

the exam¡ner at

Leiden

University

Aspects oddressed in occordance with the guidelines Description ofcontext of

internship

very good Reflection on learning

goals

very good

Reflection on

activities

very good

Link between theory and

practice

very good

References to literature/writing

style

very good

Additional remarks (mandatory short explanation of why the

grade

was

g¡ven)

Accoe,tg

+"

Aru^¡

,-l**.l^

^FtJf

òtk'.,.'\4,

lq

w'.,

þ-r,l;rt,

tw\b,p.¡int\í,

,*l

_t/\$'^1.

_{t

_wnr

_qrti.

_ì,

_\rqn

q^)

_tò

òr\ñ{"

1!ù

vo">

-¡d

P

\^f\^^

\\-

[rJ r-ro\

[r,

g**lL;,^l|È

Ll

leqon5\rÌn"

N,

ç

_ç'rn

_{6,^a¡f1ìr*\}

_i\,,N\

,

Ulr

**

nt

_/u,^

rft{*y

_çtr^"1

r*,\\\.

_\",r..

_,^F

\*,

Consent

form for

publication of the thesis

in

the Student Repository

Student Name Student lD. Master Programme Master specialisation Thesis title: Aron Groen s2262372

MSc Psychology

/

MSc Psychology (research) Economic & Consumer Psychology

How ls Trust Learned?

Every thesis is to be stored in the Leiden University Student Repository. All parties involved in this thesis indicate their opinion regarding its availability for the public below. Only with consent by all under mentioned parties involved will publication take place. Open to the public means that the thesis is retrievable by search

engines such as Google. Full embargo implies storage in the Student Repos¡tory but only accessible by staff for quality assessment purposes. Note: lf there is a full embargo on the thesis the student him of herself is

also prohibited to make the thes¡s publicly available.

Student: lNameI ISignature] lDatel First Examiner: IName] ISignature] IDate]

Aron Groen YES Open to the public

29-70-20L9

t("^to

the pubtic

tr

Fullembargo

External supervisor (if applicable): IName]

[Agency]

ISignature]

lDatel

B

Open to the public

tr

Fullembargo