INNOVATIVE PERSONALITY-BASED DIGITALSERVICES

Ricardo Buettner, Institute of Management & Information Systems, FOM University of AppliedSciences, Munich, Germany, ricardo.buettner@fom.de

AbstractSince the advent of social media, human-internet interaction has changed dramatically towards greaterindividual characteristic-based services. Since personality traits are the most stable behavioral dispo-sitions of an individual, it is surprising that the digital service industry has not focused its attention onpersonality-based services. Consequently the value creation potential of personality-based services iscurrently largely ignored. That is why in this paper I demonstrate that social media data contains fruitfulinformation about a user’s personality which in turn can be used for novel personality-based services, e.g.in marketing and recruiting.

Keywords: Personality Mining Service, Predictive Analytics, Machine Learning, Random Forest, C5.0,Five Factor Model, Big Five, Extraversion, Emotional Stability, Openness to Experience, Conscientious-ness, Agreeableness, XING, Social Media, Online Social Networks.

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

1 INTRODUCTIONThe human personality significantly influences the way people think, feel and, especially, behave (Barrickand Mount, 1991; Judge et al., 1999). Personality traits are defined as “endogenous, stable, hierarchicallystructured basic dispositions governed by biological factors such as genes and brain structures” (Romeroet al., 2009, p. 535). These traits remain quite stable over an entire lifetime and through varying situations(Costa and McCrae, 1992; Romero et al., 2009), and that is why it would potentially be very fruitful toknow a user’s personality in order to provide suitable products and services (Buettner, 2014a).While the digital service industry largely ignores this value potential, scholars have begun to explorethe opportunities concerning mining a personality from social media data. While three research groupshave mainly worked on social media based personality mining (Ortigosa, Quiroga, and Carro (2011) onFacebook, Faliagka, Iliadis, et al. (2014), Faliagka, Ramantas, et al. (2012), and Faliagka, Tsakalidis, andTzimas (2012) on LinkedIn, and Bai, Zhu, and Cheng (2012) on Renren), in this paper I go two stepsfurther (a) by providing empirical evidence for validly mining a comprehensive personality instead ofonly one or two personality traits and (b) by demonstrating two novel digital personality-based services.Consequently I formulate the following research questions:

RQ1: Is it reliably possible to comprehensively determine a user’s personality from social media data?RQ2: What digital personality-based services are possible?

Concerning the first research question (RQ1) I evaluate various data mining approaches using data fromXING. XING is the second most important career-oriented social networking site in Europe with 15 mio.members and 200,000 company profiles. The second research question (RQ2) will be addressed based onthe literature. The most important contributions from this work are:

1. I reliably predict all of the big five personality traits with a predictive gain between 31.4 and 46.2percent using decision trees (C5.0).

2. The decision based approach significantly outperforms prior models in terms of accuracy, specificity,precision and negative predictive value.

3. I demonstrate two novel personality-based digital services, i.e. a personality-based product recom-mender service and a personality-based recruiting matching service.Next I will present the research background including theories on personality and personality-relatedIT/IS and social media research. After that I outline the research methodology, including a description ofthe empirical data set and a short description of the machine learning environment, before presenting thepersonality prediction results and the two novel personality-based digital services. After that, I discuss theresults before concluding with limitations and future research.

2 RESEARCH BACKGROUND2.1 Theories on Personality

Human personality is characterized and measured through personality traits, which are defined as “en-dogenous, stable, hierarchically structured basic dispositions governed by biological factors such asgenes and brain structures” (Romero et al., 2009, p. 535). These traits remain quite stable over an entirelifetime and through varying situations (Costa and McCrae, 1992; Romero et al., 2009). Personalitysignificantly influences the way people think, feel and, especially, behave (e.g. Barrick and Mount (1991)and Judge et al. (1999)). Because of its significant impact on behavior, there are several models forcapturing personality, the most important theories relating to which are the psychoanalytical personalitytheory of Sigmund Freud, the personality theory of C. G. Jung, the personality theory of Carl Rogers andthe Three Factor Theory of Hans J. Eysenck. The most commonly used model to describe personality isthe Five Factor Model (FFM) of Goldberg (1990) and Costa and McCrae (1992), which is also seen as astate-of-the-art measuring model for personality (Gosling, Rentfrow, and Swann Jr., 2003; McCrae andCosta, 1999; Romero et al., 2009). The FFM states and measures human personality as a result of mainlybiological-determined “basic tendencies”: Openness to Experience, Conscientiousness, Extraversion,Agreeableness and Neuroticism, commonly known as the Big Five (Costa and McCrae, 1992). Thecorresponding “Five Factor Theory on Personality” (FFT) uses the Big Five to explain a significant partof human behavior (Costa and McCrae, 1992) and was successfully applied to various research domains,Barrick and Mount (1991), for example in predicting job performance by means of the Big Five andexplaining career success with reference to the Big Five (Judge et al., 1999).

2.2 Personality-related IT/IS Research

Despite psychologists’ insight into the significant impact of personality on behavior (e.g. Barrick andMount (1991)), IT/IS research has for a long time pretty much ignored this factor. However, recent IT/IS

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

research has turned towards personality as a potential predictor of IT usage patterns (Devaraj, Easley,and Crant, 2008; Junglas, N. A. Johnson, and Spitzmüller, 2008; McElroy et al., 2007; Venkatesh andWindeler, 2012). McElroy et al. (2007) directly tested the effect of personality on internet use in general.The results supported the use of personality as an explanatory factor finding that a meaningful part ofthe variance in IS use can be explained by the Big Five personality traits. Devaraj, Easley, and Crant(2008) demonstrated the potential utility of incorporating personality into IT/IS research in the context oftechnology acceptance and use. Junglas, N. A. Johnson, and Spitzmüller (2008) revealed the important roleof personality traits in perceptions of privacy to explain behavioral intentions towards adopting locationbased IT-services. Venkatesh and Windeler (2012) analyzed the impact of the FFM on team technologyuse and found a positive influence of Agreeableness, Conscientiousness, Extraversion, and Openness toExperience on technology use. It was also demonstrated that the Five Factor Model by McCrae and Costa(1999) is the dominant personality model in IT/IS research.

2.3 Personality-related Research on Social Media

Scholars have found relationships between personality traits and social media features (see table 1).

Application Exemplary references

Facebook Amichai-Hamburger and Vinitzky (2010), Moore and McElroy (2012), andMuscanell and Guadagno (2012)

Twitter Golbeck et al. (2011), Hughes et al. (2012), and Quercia et al. (2011)YouTube Aran, Biel, and Gatica-Perez (2014), Biel and Gatica-Perez (2013), and Biel,

Teijeiro-Mosquera, and Gatica-Perez (2012)MySpace Balmaceda, Schiaffino, and Godoy (2014), Muscanell and Guadagno (2012),

and Wilson, Fornasier, and White (2010)LinkedIn Caers and Castelyns (2011), Faliagka, Tsakalidis, and Tzimas (2012), and

Loiacono et al. (2012)RenRen J.-L. Wang et al. (2012a) and Yu and M. Wu (2010)

Table 1: Social media applications containing personality-relevant indicators.

Extraverted people have a higher need for social affiliation/personal communication (Costa and McCrae,1992), for strategic self-presentation (Krämer and Winter, 2008; Seidman, 2013) and as a result they havemore satisfying/stable friendships (McCrae and Costa, 1999) than introverts. Extraverts are more likely touse social media in general (Correa, Hinsley, and Zúñiga, 2010; Gosling, Augustine, et al., 2011; Hugheset al., 2012). Researchers found positive relationships between extraversion and the number of contacts(Aharony, 2013; Amichai-Hamburger and Vinitzky, 2010; Winter et al., 2014), the number of picturesposted (Gosling, Augustine, et al., 2011; Muscanell and Guadagno, 2012), the number of status updates(Garcia and Sikström, 2014), and the usage frequency (Michikyan, Subrahmanyam, and Dennis, 2014).People who have lower Neuroticism values are high in self-esteem and have less pessimistic attitudes thanthose who have higher Neuroticism values (McCrae and Costa, 1999). Because they feel less isolatedand experience less psychological distress (Costa and McCrae, 1992), emotionally stable individuals whohave lower Neuroticism values are less likely to use social media at all (Correa, Hinsley, and Zúñiga,2010; Hughes et al., 2012). The usage intensity is also found to be positively correlated with Neuroticism.Individuals with low Neuroticism values spend less time on social media (Moore and McElroy, 2012; Ryanand Xenos, 2011), update their status less often (J.-L. Wang et al., 2012b), belong to fewer groups (Skues,Williams, and Wise, 2012) and are less addicted to social media usage (Karl, Peluchette, and Schlaegel,2010). People who are high in Openness to Experience have broad interests and seek novelty (McCrae andCosta, 1999). Therefore, Openness to Experience is regarded as correlating positively with social mediause (Amichai-Hamburger and Vinitzky, 2010; Correa, Hinsley, and Zúñiga, 2010; Hughes et al., 2012).Individuals who score high on Openness to Experience also show higher social media usage intensity.They spend more time on social media (Skues, Williams, and Wise, 2012), have more friends (Gosling,Augustine, et al., 2011; Skues, Williams, and Wise, 2012), play more games (J.-L. Wang et al., 2012b)and are more active (Ross et al., 2009) than individuals low on Openness to Experience. Conscientiouspeople make long-term plans, are diligent and have organized support networks (McCrae and Costa, 1999).Social media could be seen as a sort of distraction for conscientious people (Hughes et al., 2012), butthere are contradictory findings on the relationship between Conscientiousness and social media usage.Conscientious individuals are less likely to use social media (Ryan and Xenos, 2011) and also spend lesstime on social media (Gosling, Augustine, et al., 2011; Ryan and Xenos, 2011; Wilson, Fornasier, and

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

White, 2010). Agreeable people are friendly, kind, sympathetic and warm (Costa and McCrae, 1992) andhave a tendency to be trusting, sympathetic, and cooperative (Amichai-Hamburger and Vinitzky, 2010).Individuals high on Agreeableness have more pictures on their social media profile (Ivcevic and Ambady,2012), give more information about their activities and interests (Ivcevic and Ambady, 2012; S. S. Wang,2013), view their own and other’s pages more often (Gosling, Augustine, et al., 2011), have more postsfrom their friends on their wall (Ivcevic and Ambady, 2012) and often comment on social networkingsites (J.-L. Wang et al., 2012b). On the other hand, individuals high on Agreeableness use fewer pagefeatures (Amichai-Hamburger and Vinitzky, 2010), have fewer back-and-forth conversations (Ivcevic andAmbady, 2013) and are less likely to become addicted to social media (Karl, Peluchette, and Schlaegel,2010).

2.4 Personality Mining

Three research groups have mainly worked on social media based personality mining: Ortigosa, Quiroga,and Carro (2011) on Facebook, Faliagka, Iliadis, et al. (2014), Faliagka, Ramantas, et al. (2012), andFaliagka, Tsakalidis, and Tzimas (2012) on LinkedIn, and Bai, Zhu, and Cheng (2012) on Renren. MiningFacebook data, Ortigosa, Quiroga, and Carro (2011) predicted the personality trait “emotional stability” atan accuracy above 63 percent (classification trees, J48, C4.5 algorithm). As a result of the comparison ofdifferent techniques they emphasized that classification trees achieved the best results (Ortigosa, Quiroga,and Carro, 2011, p. 565). Faliagka, Iliadis, et al. (2014), Faliagka, Ramantas, et al. (2012), and Faliagka,Tsakalidis, and Tzimas (2012) also achieved only moderate results through the use of linear regression,regression trees (M5) and support vector machines in order to analyze LinkedIn data. In line with thisresult, Bai, Zhu, and Cheng (2012) also reported that they tested “many classification algorithms such asNaive Bayesion (NB), Support Vector Machine (SVM), Decision Tree and so on, and demonstrated “thatthe C4.5 Decision Tree (Quinlan, 1993) can get the best results” (Bai, Zhu, and Cheng, 2012, p. 5). Byconsidering just the two extreme personality cases (no middle group), within their Renren analysis theyreached a two class classification accuracy of above 69 percent.

3 METHODOLOGY3.1 Empirical XING dataset and sample quality

Based on Correa, Hinsley, and Zúñiga (2010), Jenkins-Guarnieri, Wright, and Hudiburgh (2012), Linet al. (2012) and Ross et al. (2009) I use the XING features shown in table 2. FFM personality traits werecaptured with the Ten Item Personality Inventory (TIPI) from Gosling, Rentfrow, and Swann Jr. (2003)using a 5-point Likert scale (rT IPI = 0.72) and normalized to [0,1]. Finally, demographics (gender and age)were requested. In relation to research question 1 I first collected social media data by asking MBA andBachelor students electronically to take part in a survey concerning social media. The call for participationwas presented as a blackboard entry within the students online learning portal at our university. The callfor participation contained a link to an online questionnaire. Please note that our university specializesin extra-occupational MBA and Bachelor students who all have working experience. As a consequencedata were collected from 917 questionnaires. After removing canceled (86), incomplete (15) and invalid(56) answers, 760 questionnaires (∼ 83%) were finally used within the analysis. The criteria for invalidanswers were (a) time needed to complete the personality test (< 25sec), (b) similar answer patterns,and (c) inconsistent responses. 365 (∼ 48%) of the test persons were female, 395 (∼ 52%) male. Theage pattern was as follows: 22 of the questioned participants (∼ 2.9%) were below 20 years old; 539participants (∼ 70.9%), the majority, between the ages of 21 and 30; 129 participants (∼ 17.0%) between31 and 40; 53 participants (∼ 7.0%) between 41 and 50; 15 participants (∼ 2.0%) between 51 and 60and finally 2 participants (∼ 0.3%) 61 or older. Participants comprised 395 individuals (∼ 52%) with apersonal XING-Profile and 365 (∼ 48%) without any profile or activity on XING. 45 (∼ 11.4%) of the395 XING-Users are active daily-users of the platform. 98 (24.8%) are using it on a weekly basis, 74(∼ 18.7%) use XING several times per month, 154 (∼ 40.0%) at least once a month and 24 (∼ 6.1%)never use this social network.Compared to the personality traits of the general population I observed similar trait patterns by gender(table 3). Nevertheless, this result is also justifiable as a consideration of the recruiting method (seesection 2.3). The participants were mainly MBA and Bachelor students who study on an extra-occupationalbasis, and it must be assumed that most of these people are successful professionals. Judge et al. (1999)found strong relationships between FFM traits and career success as conscientiousness positively predictedcareer success. As shown in table 3, my results confirm these observations by Judge et al. (1999) becausein my sample of mostly successful professionals I found higher values for Conscientiousness.

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

# Item text Scale Mean S.D.

I1 How often do you use XING? [1-never..5-daily] 2.96 1.16I2 How often do you use the XING jobsearch function? [1-never..5-daily] 1.70 0.87I3 How often do you use the XING blogging function? [1-never..5-daily] 1.24 0.51I4 How often do you use the XING messaging function? [1-never..5-daily] 2.36 0.90I5 How often do you use the XING event organization function? [1-never..5-daily] 1.14 0.41I6 How often do you use the XING event participation function? [1-never..5-daily] 1.27 0.48I7 How often do you use the XING advantageous offers function? [1-never..5-daily] 1.29 0.56I8 Have you filled out your educational background on XING? [1-no/2-yes] 1.32 0.47I9 Have you filled out your work experience on XING? [1-no/2-yes] 1.95 0.21I10 Have you filled out your organizations on XING? [1-no/2-yes] 1.68 0.47I11 Have you filled out your interests on XING? [1-no/2-yes] 1.73 0.44I12 Have you filled out your awards on XING? [1-no/2-yes] 1.32 0.47I13 Have you filled out your language skills on XING? [1-no/2-yes] 1.86 0.35I14 Have you filled out your haves on XING? [1-no/2-yes] 1.73 0.44I15 Have you filled out your wants on XING? [1-no/2-yes] 1.69 0.46I16 Have you filled out your about me information on XING? [1-no/2-yes] 1.52 0.50I17 Do you have a XING premium membership? [1-no/2-yes] 1.26 0.44I18 How many XING contacts do you have? [No.] 121 160

Table 2: Measured items for XING usage features.

TIPI (Gosling, Rentfrow, and Swann Jr., 2003) My Samplemale (n = 1,173) female (n = 633) male (n = 395) female (n = 365)

Openness 0.723 0.733 0.702 0.741Conscientiousness 0.698 0.752 0.737 0.795Extraversion 0.542 0.590 0.622 0.678Agreeableness 0.677 0.720 0.611 0.706Neuroticism 0.312 0.390 0.321 0.401

Table 3: Comparison of the [0;1]-normalized TIPI results and our own by gender.

4 RESULTS4.1 Machine Learning Results for Personality Mining (RQ1)

In a first step it is interesting to analyze the relationships between the Big Five personality traits and thespecific XING usage features, which can be found in table 4.

I1 I4 I5 I6 I10 I11 I12 I15 I17 I18

Openness 0.1202 0.1402

Conscientiousness -0.1532

Extraversion 0.1061 0.1121 0.1101 0.0931 0.1141 0.2152

Agreeableness -0.1502 -0.1402 -0.1192 -0.0901

Neuroticism -0.0881 -0.1131 -0.0871 -0.1041 0.1712

Table 4: Significant Spearman-Rho correlations betweenBig Five traits and XING features (1p<0.05, 2p<0.01).

As shown in table 4 only a few significant correlations between specific XING features and the personalitytraits can be found, cf. Cohen (1988). All of them are weak. However, a critical mass of weak relationshipscould have a good level of predictive power. That is why I applied machine learning algorithms forpersonality trait prediction. In order to compare 192 various machine learning algorithms I used theR x64 3.2.2 environment (R Core Team, 2015) combined with the caret package by Max Kuhn for all

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

analyses. Based on the TIPI results I built two mean-balanced classes for each personality trait. Formachine learning and evaluation purposes I split the n=395 sample in a training partition (nT=261) andan evaluation partition (nE=134). I systematically compared the machine learning outputs in terms ofaccuracy (ACC), sensitivity (true positive rate, TPR), specificity (SPC), precision (positive predictivevalue, PPV) and negative predictive value (NPV) as quality criteria and found that the C5.0 trees approach(cf. Kuhn and K. Johnson (2013, pp. 394)) delivers the best output. The results are shown in table 5.

ACC TPR SPC PPV NPVOpenness 0.731 0.487 0.832 0.543 0.798Conscientiousness 0.657 0.759 0.548 0.659 0.654Extraversion 0.672 0.641 0.700 0.661 0.681Agreeableness 0.664 0.609 0.714 0.661 0.667Neuroticism 0.694 0.625 0.757 0.702 0.757

∅ 0.684 0.624 0.710 0.645 0.711

Table 5: C5.0 quality criteria.

C5.0 is the improved successor of C4.5 which is a classification tree algorithm originally proposed byQuinlan (1993) which is in turn an extension of Quinlan’s earlier ID3 algorithm using the concept ofinformation entropy (Loh, 2008, pp. 5). C5.0’s predecessor C4.5 belongs to the most influential datamining algorithms in the research community (X. Wu et al., 2008). With the C5.0 algorithm I reach animpressive predictive gain between 31.4 and 46.2 percent – which means that in fact the social mediaplatform XING does contain fruitful data for personality mining.

4.2 Novel Personality-based Digital Services (RQ2)

Based on the automatic prediction of a user’s personality many novel digital services can be created, sincepersonality significantly influences the way people think, feel and, especially, behave, e.g. (Barrick andMount, 1991; Judge et al., 1999). Next, I will focus on two very interesting novel digital services.

4.2.1 Towards a Personality-based Recruiting Matching Service

Maurer and Cook (2011) emphasized the key role of the Person-Organization Environment (P-OE) fitin order to recruit high quality job applicants. Meta-analysis investigations showed significant positivecorrelations between the P-OE fit and job performance, job satisfaction, organizational commitment andemployee turnover (Kristof, 1996; A. L. Kristof-Brown, Zimmerman, and E. C. Johnson, 2005; Verquer,Beehr, and Wagner, 2003). The P-OE fit can be broken-down into sub-fits, with the most appropriateconsisting of three all but disjointed sub-fits, which together almost cover the whole notion of the P-OE fit(A. L. Kristof-Brown, Zimmerman, and E. C. Johnson, 2005). These sub-fits are the Person-Organization(P-O) fit (between candidate personality and organizational culture) (Kristof, 1996), the Person-Group(P-G) fit (matching of individual and group roles and interactions) (Werbel and D. J. Johnson, 2001)and the Person-Job (P-J) fit (between a candidate’s skills, knowledge, and abilities and job demands)(Edwards, 1991). An overview of investigations into these and further P-OE sub-fits can be found in(A. L. Kristof-Brown, Zimmerman, and E. C. Johnson, 2005). Buettner (2014a) conceptualized thesesub-fits in a framework for online social network recruiting (see figure 1).The relative importance of these sub-fits depends on the concrete vacancy. A. L. Kristof-Brown (2000)empirically showed that recruiters actually distinguish between these sub-fits. Sekiguchi and Huber (2011)investigated the weighting of the P-O fit and the P-J fit when hiring decision-makers to evaluate jobcandidates. The longer a recruiter expects a candidate to stay within the organization the more importantthe P-O fit becomes. The shorter the envisioned stay the more important the P-J fit becomes, since skilland knowledge acquisition on the job become inefficient for relatively short stays within the organization.The very interesting point from a digital service economy perspective is the fact that the P-O fit and the P-Gfit can be assessed if the individual personalities are known. The P-O fit describes the macro-perspectivecompatibility between the employee’s personalities and the organization’s culture they work in (A. L.Kristof-Brown, Zimmerman, and E. C. Johnson, 2005, pp. 285). To measure the P-O fit several tests havebeen developed, e.g. (Kristof, 1996). The importance of such tests for hiring organizations has been notedin Cable and Parsons (2001, p. 21), further research has been surveyed in A. L. Kristof-Brown (2000),A. L. Kristof-Brown, Zimmerman, and E. C. Johnson (2005), and Verquer, Beehr, and Wagner (2003).While early studies of the P-O fit focused on the personality-climate congruence, current investigations

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

Social context

Integrated Perspective

Micro levelMeso levelMacro level

CandidatePersonality*Poten-

tial Roles*

Organizational Culture*

Group Roles*

P-O fit P-G fit

Group Com-muni-cation styles*

Com-muni-cation style*

Skills, Knowledge& Abilities*

Job Description

P-J fit

Hiring Organization

P-OE fit

* Data inferable from OSN

Figure 1: An interdisciplinary P-OE fit based frame-work for OSN-Recruiting (Buettner, 2014a, p. 1421).

emphasize the value congruence (Edwards and Cable, 2009; Kristof, 1996; Verquer, Beehr, and Wagner,2003). According to these and further studies (Anderson, Spataro, and Flynn, 2008; Tom, 1971) employeesare most successful in organizations with a culture which is compatible with their personalities. Henceassessing the P-O fit during recruiting decisions requires (a) an analysis of the candidate’s personalityand (b) an investigation of the organizational culture. (a) can be derived directly as shown in the MachineLearning Results for Personality Mining (RQ1) section. (b) can be assessed by knowing a critical mass oforganization members or alternatively by organizational culture instruments (Jung et al., 2009) such as theOrganizational Culture Profile by O’Reilly III, Chatman, and Caldwell (1991).But it is not just the P-O fit that can be evaluated by knowing the personality. The P-G fit can also be partlyassessed by personality information. Measuring the P-G fit requires (a) an analysis of the candidate’ssocial interaction characteristics and (b) an investigation of the group’s social fabric. Individual interactioncharacteristics are partly determined by personality traits (Mount, Barrick, and Stewart, 1998). Personalitytraits have been used to forecast team performance directly, cf. Bell (2007) and Mount, Barrick, and Stewart(1998). Mount, Barrick, and Stewart (1998) showed by conducting a meta-analysis that agreeableness,emotional stability, and conscientiousness are positively correlated with team performance. A. Kristof-Brown, Barrick, and Stevens (2005) revealed that teams are more attracted to an individual with anextraversion score that is complementary to the groups’ extraversion score (i.e. high individual-low teamor low individual-high team levels).

4.2.2 Towards a Personality-based Product Recommender Service

Recommender services have gained a lot of attention since the advent of the internet. Prior designs forrecommender systems have mainly focused on user preference information (e.g. user rating), content-based information (e.g. item prices) and collaborative information (e.g. recommendation of friends).Personality as a main driver of buying behavior has been largely neglected. However, very recent researchon recommender services is interested in personality-based approaches. For example, Rana and Jain (2015)emphasized this potential in their current overview (“personality attributes ... could then be implementedin recommender system[s]” (Rana and Jain, 2015, p. 143)). Concerning the use of personality informationin recommender systems, Cantador and Fernández-Tobías (2014) states that “there is plenty of room foralternative, more sophisticated methods” (Cantador and Fernández-Tobías, 2014, p. 43).In fact, a few researchers initially sketched personality-based approaches: For instance, Hu and Pu (2010)proposed a general method that infers a user’s music preferences in terms of their personalities. W. Wu,Chen, and He (2013) presented a strategy that explicitly embeds a users’ personality – as a moderatingfactor – to adjust the item’s degree of diversity within multiple recommendations. Fernández-Tobíasand Cantador (2015) presented a study comparing collaborative filtering methods enhanced with userpersonality traits and showed that incorporating personality information facilitates improvement in theaccuracy of recommendations. Hu and Pu (2011) aimed to address the cold-start problem by incorporatinga user’s personality into the collaborative filtering framework.The relationship between personality and consumer behavior is not new. Many decades ago, marketingscholars found substantial correlations between personality traits and preferred products such as mouth-wash, alcoholic drinks, automobiles, etc. (Kassarjian, 1971). But nowadays it is possible to predict a user’spersonality from social media data. That is why mining a user’s personality is very fruitful for designingfuture recommender systems. Consequently new business opportunities towards personality-based recom-

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

mender services when analyzing social media footprints are possible. The principle of such services issketched in figure 2.

Recommended products

Recommender system

(e.g. similarity)

Userpersonality

(e.g. Big Five)

Personality prediction(e.g. C5.0)

Social media data

(e.g. XING)

Figure 2: Personality-based Product Recommender Services for analyzing social media footprints.

5 DISCUSSIONIn line with prior research I found a few significant correlations between specific social media usagefeatures and users’ personality traits (see table 4). It is also in line with prior research that all of thesesignificant correlations are small. However, despite these small correlations I could predict all of the fivepersonality traits with an impressive predictive gain between 31.4 and 46.2 percent – which means that infact the social media platform XING contains fruitful data for personality mining. Concerning researchquestion 1 I can conclude that it is reliably possible to comprehensively determine a user’s personalityfrom social media data.While scholars have mainly used simple linear regression models or in the best case the C4.5 J48 approachfor personality mining issues, in this paper I evaluated the C5.0 algorithm. As a result I found that theC5.0 algorithm substantially outperforms existing models in terms of accuracy, specificity, precisionand negative predictive value on an average over all of the big five personality traits (see table 5). Forinstance, Kosinski, Bachrach, et al. (2014) and Kosinski, Stillwell, and Graepel (2013) reached only weakpersonality prediction results based on linear regression models. In contrast, I reached a strong predictivegain based on decision trees (C5.0). Since I reached predictive gains up to 46.2 percent (see table 5), Ifound empirical evidence that the C5.0 approach can be used for validly mining social media for a user’spersonality which is in turn very interesting for new marketing/personalization and recruiting services.I consequently sketched two novel digital services, namely a personality-based recruiting matching serviceand a personality-based product recommender service which in turn addresses research question 2.With this research I contributed to digital services research and personality mining research.

6 LIMITATIONS AND FUTURE RESEARCHIn order to increase the external validity of the approach presented here, future work will apply personality-based digital services within a queue of recruiting projects, funded by the German Federal Ministryof Education and Research (BMBF) under contracts 17103X10 and 03FH055PX2 to make employeecontracting in Germany more sophisticated through automated negotiation (Buettner, 2006a,b, 2007a,b,2009; Buettner and Kirn, 2008; Buettner and Landes, 2012). In the next step this work will be extensivelyevaluated in our laboratory (Buettner, 2013a,b, 2014b, 2015c, 2016b; Buettner, Daxenberger, Eckhardt,et al., 2013; Buettner, Daxenberger, and Woesle, 2013; Buettner, Sauer, et al., 2015), before beingimplemented in external recruiting software, i.e., career-oriented social networking sites (Buettner, 2015b,2016a) and crowdsourcing platforms (Buettner, 2014c, 2015a).A limitation of this work comprises the methodological problems of measuring the personality (traits)and the XING features. The personality model used, FFM, was measured with the Ten Item PersonalityInventory from Gosling, Rentfrow, and Swann Jr. (2003). Despite the proven value of this personality in-ventory in various studies and an acceptable reliability (rT IPI = 0.72), more extensive multi-item measuresof the Big Five exist. That is why alternative personality measures (e.g. BFI-S from Hahn, Gottschling,and Spinath (2012)) should be applied for personality–social media usage investigations. However, Iused the inventory from Gosling, Rentfrow, and Swann Jr. (2003) in order to reduce participant’s effortand thereby the exit rate. The operationalization of the XING features is also problematic. Nevertheless,I conceptionalized the XING features to the best of my knowledge based on prior research (Correa,Hinsley, and Zúñiga, 2010; Jenkins-Guarnieri, Wright, and Hudiburgh, 2012; Lin et al., 2012; Ross et al.,2009). Future research could extract theses features directly from the specific platform. I did not grab thisinformation directly because of privacy issues. Future research should address this problem and shoulduse privacy preserving data mining techniques such as Bae et al. (2014) or Kim et al. (2008) in order toretrieve personality relevant social media features without violating privacy.

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

ReferencesAharony, N. (2013). “Facebook use by Library and Information Science students.” Aslib Proceedings

65 (1), 19–39.Amichai-Hamburger, Y. and G. Vinitzky (2010). “Social network use and personality.” Computers in

Human Behavior 26 (6), 1289–1295.Anderson, C., S. E. Spataro, and F. J. Flynn (2008). “Personality and Organizational Culture as Determi-

nants of Influence.” Journal of Applied Psychology 93 (3), 702–710.Aran, O., J.-I. Biel, and D. Gatica-Perez (2014). “Broadcasting oneself: Visual Discovery of Vlogging

Styles.” Multimedia, IEEE Transactions 16 (1), 201–215.Bae, D.-H., J.-M. Lee, S.-W. Kim, Y. Won, and Y. S. Park (2014). “Analyzing Network Privacy Preserving

Methods: A Perspective of Social Network Characteristics.” IEICE Transactions on Information andSystems E97-D (6), 1664–1667.

Bai, S., T. Zhu, and L. Cheng (2012). “Big-Five Personality Prediction Based on User Behaviors at SocialNetwork Sites.” arXiv:1204.4809.

Balmaceda, J. M., S. Schiaffino, and D. Godoy (2014). “How do personality traits affect communicationamong users in online social networks?” Online Information Review 38 (1), 136–153.

Barrick, M. R. and M. K. Mount (1991). “The Big Five Personality Dimensions and Job Performance: AMeta-Analysis.” Personnel Psychology 44 (1), 1–26.

Bell, S. T. (2007). “Deep-level composition variables as predictors of team performance: A meta-analysis.”Journal of Applied Psychology 92 (3), 595–615.

Biel, J.-I. and D. Gatica-Perez (2013). “The YouTube Lens: Crowdsourced Personality Impressions andAudiovisual Analysis of Vlogs.” IEEE Transactions on Multimedia 15 (1), 41–55.

Biel, J.-I., L. Teijeiro-Mosquera, and D. Gatica-Perez (2012). “FaceTube: predicting personality fromfacial expressions of emotion in online conversational video.” In: Proceedings of the 14th InternationalConference on Multimodal Interaction. ICMI ’12. Santa Monica, California, USA: ACM, pp. 53–56.

Buettner, R. (2006a). “A Classification Structure for Automated Negotiations.” In: IEEE/WIC/ACMWI-IAT 2006 Proc. Pp. 523–530.

Buettner, R. (2006b). “The State of the Art in Automated Negotiation Models of the Behavior andInformation Perspective.” ITSSA 1 (4), 351–356.

Buettner, R. (2007a). “Electronic Negotiations of the Transactional Costs Perspective.” In: IADIS’07WWW/Internet Proc., Vol. 2, pp. 99–105. ISBN: 9789728924447.

Buettner, R. (2007b). “Imperfect Information in Electronic Negotiations: An Empirical Study.” In:IADIS’07 WWW/Internet Proc., Vol. 2, pp. 116–121. ISBN: 9789728924447.

Buettner, R. (2009). “Cooperation in Hunting and Food-sharing: A Two-Player Bio-inspired Trust Model.”In: BIONETICS ’09: Proceedings of the Fourth International Conference on Bio-Inspired Models ofNetwork, Information, and Computing Systems (BIONETICS ’09), Avignon, France, December 9-11,2009. Vol. 39. LNICST. Springer-Verlag, pp. 1–10. ISBN: 9783642128073.

Buettner, R. (2013a). “Cognitive Workload of Humans Using Artificial Intelligence Systems: TowardsObjective Measurement Applying Eye-Tracking Technology.” In: KI 2013 Proc. Vol. 8077. LNAI,pp. 37–48.

Buettner, R. (2013b). “Social inclusion in eParticipation and eGovernment solutions: A systematiclaboratory-experimental approach using objective psychophysiological measures.” In: EGOV/ePart2013: Proceedings of the Joint Conference of IFIP EGOV 2013 & IFIP ePart 2013, September 16-19,Koblenz, Germany, 2013. Vol. P-221. Lecture Notes in Informatics (LNI). Gesellschaft für Informatik(GI), pp. 260–261.

Buettner, R. (2014a). “A Framework for Recommender Systems in Online Social Network Recruiting.”In: HICSS 2014 Proceedings: 47th Hawaii International Conference on System Sciences (HICSS-47),Januar 6-9, 2014, Big Island, Hawaii, pp. 1415–1424.

Buettner, R. (2014b). “Analyzing Mental Workload States on the Basis of the Pupillary Hippus.” In:NeuroIS ’14 Proc. P. 52.

Buettner, R. (2014c). “Crowdsourcing of a Human Resource Management Perspective: State of the Art,Challenges & Future Need for Research.” Presentation at VHB ’14 Conference, Leipzig, Germany,June 11-13, unpublished.

Buettner, R. (2015a). “A Systematic Literature Review of Crowdsourcing Research from a HumanResource Management Perspective.” In: HICSS-48 Proc. Pp. 4609–4618.

Buettner, R. (2015b). “Analyzing the Problem of Employee Internal Social Network Site Avoidance: AreUsers Resistant due to their Privacy Concerns?” In: HICSS-48 Proc. Pp. 1819–1828.

Buettner, R. (2015c). “Investigation of the Relationship Between Visual Website Complexity and Users’Mental Workload: A NeuroIS Perspective.” In: Information Systems and Neuro Science: GmundenRetreat on NeuroIS 2015. Vol. 10. LNISO, pp. 123–128.

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

Buettner, R. (2016a). “Getting a Job via Career-oriented Social Networking Sites: The Weakness of Ties.”In: HICSS-49 Proc. Pp. 2156–2165.

Buettner, R. (2016b). “The relationship between visual website complexity and a user’s mental workload:A NeuroIS perspective.” In: Information Systems and Neuro Science: Gmunden Retreat on NeuroIS2016, June 6-8, 2016, Gmunden, Austria. in press.

Buettner, R., B. Daxenberger, A. Eckhardt, and C. Maier (2013). “Cognitive Workload Induced byInformation Systems: Introducing an Objective Way of Measuring based on Pupillary DiameterResponses.” In: Pre-ICIS HCI/MIS 2013 Proc. Paper 20.

Buettner, R., B. Daxenberger, and C. Woesle (2013). “User acceptance in different electronic negotiationsystems - a comparative approach.” In: In ICEBE 2013: Proceedings of the 10th IEEE InternationalConference on e-Business Engineering, September 11 - 13, Coventry, UK. IEEE CS Press, pp. 1–8.

Buettner, R. and S. Kirn (2008). “Bargaining Power in Electronic Negotiations: A Bilateral NegotiationMechanism.” In: EC-Web ’08 Proceedings. Vol. 5183. LNCS, pp. 92–101.

Buettner, R. and J. Landes (2012). “Web Service-based Applications for Electronic Labor Markets: AMulti-dimensional Price VCG Auction with Individual Utilities.” In: ICIW 2012 Proc. Pp. 168–177.

Buettner, R., S. Sauer, C. Maier, and A. Eckhardt (2015). “Towards ex ante Prediction of User Performance:A novel NeuroIS Methodology based on Real-Time Measurement of Mental Effort.” In: HICSS-48Proc. Pp. 533–542.

Cable, D. M. and C. K. Parsons (2001). “Socialization tactics and person-organization fit.” PersonnelPsychology 54 (1), 1–23.

Caers, R. and V. Castelyns (2011). “LinkedIn and Facebook in Belgium: The Influences and Biases ofSocial Network Sites in Recruitment and Selection Procedures.” Social Science Computer Review29 (4), 437–448.

Cantador, I. and I. Fernández-Tobías (2014). “On the Exploitation of User Personality in RecommenderSystems.” In: DMRS ’14 Proc.: Proceedings of the International Workshop on Decision Making andRecommender Systems. CEUR Workshop Proceedings 1278, pp. 42–45.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. 2. Hillsdale, NJ, USA: LawrenceErlbaum.

Correa, T., A. W. Hinsley, and H. G. de Zúñiga (2010). “Who interacts on the Web?: The intersectionof users’ personality and social media use.” Computers in Human Behavior 26 (2), 247–253. ISSN:0747-5632.

Costa, P. T. and R. R. McCrae (1992). Revised NEO personality inventory (NEO-PI-R) and the NEOFive-Factor inventory (NEO-FFI): Professional manual. Odessa, FL, USA: PAR.

Devaraj, S., R. F. Easley, and J. M. Crant (2008). “How Does Personality Matter? Relating the Five-FactorModel to Technology Acceptance and Use.” Information Systems Research 19 (1), 93–105.

Edwards, J. R. (1991). “Person-job fit: A conceptual integration, literature review, and methodologicalcritique. International review of industrial and organizational psychology.” In: International review ofindustrial and organizational psychology. Ed. by C. L. Cooper and I. T. Robertson. Vol. 6. Oxford:John Wiley & Sons, pp. 283–357.

Edwards, J. R. and D. M. Cable (2009). “The Value of Value Congruence.” Journal of Applied Psychology94 (3), 654–677. ISSN: 0021-9010.

Faliagka, E., L. Iliadis, I. Karydis, M. Rigou, S. Sioutas, A. Tsakalidis, and G. Tzimas (2014). “On-lineconsistent ranking on e-recruitment: seeking the truth behind a well-formed CV.” Artificial IntelligenceReview 42 (3), 515–528.

Faliagka, E., K. Ramantas, A. Tsakalidis, and G. Tzimas (2012). “Application of Machine LearningAlgorithms to an online Recruitment System.” In: ICIW ’12 Proc.

Faliagka, E., A. Tsakalidis, and G. Tzimas (2012). “An integrated e-recruitment system for automatedpersonality mining and applicant ranking.” Internet Research 22 (5), 551–568.

Fernández-Tobías, I. and I. Cantador (2015). “On the Use of Cross-Domain User Preferences andPersonality Traits in Collaborative Filtering.” In: UMAP ’15 Proc. LNCS 9146, pp. 343–349.

Garcia, D. and S. Sikström (2014). “The dark side of Facebook: Semantic representations of status updatespredict the Dark Triad of personality.” Personality and Individual Differences 67, 92–96.

Golbeck, J., C. Robles, M. Edmondson, and K. Turner (2011). “Predicting Personality from Twitter.” In:Proceedings of the Third International Conference on Privacy, Security, Risk and Trust (passat) andof the Third International Conference on Social Computing (socialcom), pp. 149–156.

Goldberg, L. R. (1990). “An Alternative “Description of Personality”: The Big-Five Factor Structure.”Journal of Personality and Social Psychology 59 (6), 1216–1229.

Gosling, S. D., A. A. Augustine, S. Vazire, N. Holtzman, and S. Gaddis (2011). “Manifestations ofPersonality in Online Social Networks: Self-Reported Facebook-Related Behaviors and ObservableProfile Information.” Cyberpsychology, Behavior, and Social Networking 14 (9), 483–488.

Gosling, S. D., P. J. Rentfrow, and W. B. Swann Jr. (2003). “A very brief measure of the Big-Fivepersonality domains.” Journal of Research in Personality 37 (6), 504–528.

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

Hahn, E., J. Gottschling, and F. M. Spinath (2012). “Short measurements of personality – Validity andreliability of the GSOEP Big Five Inventory (BFI-S).” Journal of Research in Personality 46 (3),355–359.

Hu, R. and P. Pu (2010). “A Study on User Perception of Personality-Based Recommender Systems.” In:User Modeling, Adaptation, and Personalization. Vol. 6075. LNCS, pp. 291–302.

Hu, R. and P. Pu (2011). “Enhancing Collaborative Filtering Systems with Personality Information.” In:RecSys ’11: Proceedings of the 5th ACM conference on Recommender systems.

Hughes, D. J., M. Rowe, M. Batey, and A. Lee (2012). “A tale of two sites: Twitter vs. Facebook and thepersonality predictors of social media usage.” Computers in Human Behavior 28 (2), 561–569. ISSN:0747-5632.

Ivcevic, Z. and N. Ambady (2012). “Personality impressions from identity claims on Facebook.” Psychol-ogy of Popular Media Culture 1 (1), 38–45.

Ivcevic, Z. and N. Ambady (2013). “Face to (Face)Book: The Two Faces of Social Behavior?” Journal ofPersonality 3 (3), 290–301.

Jenkins-Guarnieri, M. A., S. L. Wright, and L. M. Hudiburgh (2012). “The relationships among at-tachment style, personality traits, interpersonal competency, and Facebook use.” Journal of AppliedDevelopmental Psychology 33 (6), 294–301.

Judge, T. A., C. A. Higgins, C. J. Thoresen, and M. R. Barrick (1999). “The Big Five Personality Traits,General Mental Ability, and Career Success across the Life Span.” Personnel Psychology 52 (3),621–652.

Jung, T., T. Scott, H. T. O. Davies, P. Bower, D. Whalley, R. McNally, and R. Mannion (2009). “In-struments for Exploring Organizational Culture: A Review of the Literature.” Public AdministrationReview 69 (6), 1087–1096. ISSN: 1540-6210.

Junglas, I. A., N. A. Johnson, and C. Spitzmüller (2008). “Personality traits and concern for privacy: anempirical study in the context of location-based services.” European Journal of Information Systems17, 387–402.

Karl, K., J. Peluchette, and C. Schlaegel (2010). “Who’s Posting Facebook Faux Pas? A Cross-CulturalExamination of Personality Differences.” International Journal of Selection and Assessment 18 (2),174–186.

Kassarjian, H. H. (1971). “Personality and Consumer Behavior: A Review.” Journal of Marketing Research8 (4), 409–418.

Kim, S.-W., S. Park, J.-I. Won, and S.-W. Kim (2008). “Privacy preserving data mining of sequentialpatterns for network traffic data.” Information Sciences 178 (3), 694–713.

Kosinski, M., Y. Bachrach, P. Kohli, D. Stillwell, and T. Graepel (2014). “Manifestations of user personalityin website choice and behaviour on online social networks.” Machine Learning 95 (3), 357–380.

Kosinski, M., D. Stillwell, and T. Graepel (2013). “Private traits and attributes are predictable from digitalrecords of human behavior.” Proceedings of the National Academy of Sciences 110 (15), 5802–5805.

Krämer, N. and S. Winter (2008). “Impression Management 2.0: The Relationship of Self-Esteem,Extraversion, Self-Efficacy, and Self-Presentation Within Social Networking.” Journal of MediaPsychology: Theories, Methods, and Applications 20 (3), 106–116.

Kristof, A. L. (1996). “Person-organization fit: An integrative review of its conceptualizations, measure-ment, and implications.” Personnel Psychology 49 (1), 1–49.

Kristof-Brown, A. L. (2000). “Perceived applicant fit: Distinguishing between recruiters’ perceptions ofperson-job and person-organization fit.” Personnel Psychology 53 (3), 643–671. ISSN: 1744-6570.

Kristof-Brown, A. L., R. D. Zimmerman, and E. C. Johnson (2005). “Consequences of individuals’ fit atwork: A meta-analysis of person-job, person-organization, person-group, and person-supervisor fit.”Personnel Psychology 58 (2), 281–342. ISSN: 1744-6570.

Kristof-Brown, A., M. R. Barrick, and C. K. Stevens (2005). “When Opposites Attract: A Multi-SampleDemonstration of Complementary Person-Team Fit on Extraversion.” Journal of Personality 73 (4),935–958. ISSN: 1467-6494.

Kuhn, M. and K. Johnson (2013). Applied Predictive Modeling. New York: Springer.Lin, J.-H., W. Peng, M. Kim, S. Y. Kim, and R. LaRose (2012). “Social networking and adjustments

among international students.” New Media & Society 14 (3), 421–440.Loh, W.-Y. (2008). “Classification and Regression Tree Methods.” In: Encyclopedia of Statistics in Quality

and Reliability. In Encyclopedia of Statistics in Quality and Reliability, Ruggeri, Kenett and Faltin(eds.) New York: Wiley, pp. 315–323. ISBN: 9780470061572.

Loiacono, E., D. Carey, A. Misch, A. Spencer, and R. Speranza (2012). “Personality Impacts on Self-disclosure Behavior on Social Networking Sites.” In: AMCIS 2012 Proceedings. Vol. 6.

Maurer, S. D. and D. P. Cook (2011). “Using company web sites to e-recruit qualified applicants: A jobmarketing based review of theory-based research.” Computers in Human Behavior 27 (1), 106–117.ISSN: 0747-5632.

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

McCrae, R. R. and P. T. Costa (1999). “A five-factor theory of personality.” In: Handbook of personality:Theory and research. NewYork: Guilford: Pervin, Lawrence A. and John, Oliver P., pp. 139–152.

McElroy, J. C., A. R. Hendrickson, A. M. Townsend, and S. M. DeMarie (2007). “Dispositional Factorsin Internet Use: Personality versus Cognitive Style.” MIS Quarterly 31 (4), 809–820.

Michikyan, M., K. Subrahmanyam, and J. Dennis (2014). “Can you tell who I am? Neuroticism, ex-traversion, and online self-presentation among young adults.” Computers in Human Behavior 33,179–183.

Moore, K. and J. C. McElroy (2012). “The influence of personality on Facebook usage, wall postings, andregret.” Computers in Human Behavior 28 (1), 267–274.

Mount, M. K., M. R. Barrick, and G. L. Stewart (1998). “Five-Factor Model of personality and Perfor-mance in Jobs Involving Interpersonal Interactions.” Human Performance 11 (2-3), 145–165.

Muscanell, N. L. and R. E. Guadagno (2012). “Make new friends or keep the old: Gender and personalitydifferences in social networking use.” Computers in Human Behavior 28 (1), 107–112.

O’Reilly III, C. A., J. Chatman, and D. F. Caldwell (1991). “People and Organizational Culture: A ProfileComparison Approach to Assessing Person-Organization Fit.” The Academy of Management Journal34 (3), 487–516.

Ortigosa, A., J. I. Quiroga, and R. M. Carro (2011). “Inferring User Personality in Social Networks: ACase Study in Facebook.” In: ISDA ’11 Proc. Pp. 563–568.

Quercia, D., M. Kosinski, D. Stillwell, and J. Crowcroft (2011). “Our Twitter Profiles, Our Selves:Predicting Personality with Twitter.” In: Proceedings of the Third International Conference on Privacy,Security, Risk and Trust (passat) and of the Third International Conference on Social Computing(socialcom), pp. 307–314.

Quinlan, J. R. (1993). C4.5: Programs for Machine Learning. San Mateo, CA: Morgan Kaufmann.R Core Team (2015). R: A Language and Environment for Statistical Computing. R Foundation for

Statistical Computing. Vienna, Austria.Rana, C. and S. K. Jain (2015). “A study of the dynamic features of recommender systems.” Artificial

Intelligence Review 43 (1), 141–153.Romero, E., P. Villar, M. Á. Luengo, and J. A. Gómez-Fraguela (2009). “Traits, personal strivings and

well-being.” Journal of Research in Personality 43 (4), 535–546.Ross, C., E. S. Orr, M. Sisic, J. M. Arseneault, M. G. Simmering, and R. R. Orr (2009). “Personality and

motivations associated with Facebook use.” Computers in Human Behavior 25 (2), 578–586. ISSN:0747-5632.

Ryan, T. and S. Xenos (2011). “Who uses Facebook? An investigation into the relationship between theBig Five, shyness, narcissism, loneliness, and Facebook usage.” Computers in Human Behavior 27 (5),1658–1664. ISSN: 0747-5632.

Seidman, G. (2013). “Self-presentation and belonging on Facebook: How personality influences socialmedia use and motivations.” Pers Individ Dif 54 (3), 402–407.

Sekiguchi, T. and V. L. Huber (2011). “The use of person-organization fit and person-job fit informationin making selection decisions.” Organizational Behavior and Human Decision Processes 116 (2),203–216. ISSN: 0749-5978.

Skues, J. L., B. Williams, and L. Wise (2012). “The effects of personality traits, self-esteem, loneliness,and narcissism on Facebook use among university students.” Computers in Human Behavior 28 (6),2414–2419.

Tom, V. R. (1971). “The Role of Personality and Organizational Images in the Recruiting Process.”Organizational Behavior and Human Performance 6 (5), 573–592. ISSN: 0030-5073.

Venkatesh, V. and J. B. Windeler (2012). “Hype or Help? A Longitudinal Field Study of Virtual WorldUse for Team Collaboration.” Journal of the Association for Information Systems 13 (10), 735–771.

Verquer, M. L., T. A. Beehr, and S. H. Wagner (2003). “A meta-analysis of relations between person-organization fit and work attitudes.” Journal of Vocational Behavior 63 (3), 473–489. ISSN: 0001-8791.

Wang, J.-L., L. A. Jackson, D.-J. Zhang, and Z.-Q. Su (2012a). “The relationships among the Big FivePersonality factors, self-esteem, narcissism, and sensation-seeking to Chinese University students’uses of social networking sites (SNSs).” Computers in Human Behavior 28 (6), 2313–2319.

Wang, J.-L., L. A. Jackson, D.-J. Zhang, and Z.-Q. Su (2012b). “The relationships among the Big FivePersonality factors, self-esteem, narcissism, and sensation-seeking to Chinese University students’uses of social networking sites (SNSs).” Computers in Human Behavior 28 (6), 2313–2319.

Wang, S. S. (2013). “’I Share, Therefore I Am’: Personality Traits, Life Satisfaction, and FacebookCheck-Ins.” Cyberpsychology, Behavior, and Social Networking 16 (12), 870–877.

Werbel, J. D. and D. J. Johnson (2001). “The Use of Person-Group Fit for Employment Selection: AMissing Link in Person-Environment Fit.” Human Resource Management 40 (3), 227–240.

Wilson, K., S. Fornasier, and K. M. White (2010). “Psychological Predictors of Young Adults’ Use ofSocial Networking Sites.” Cyberpsychology, Behavior, and Social Networking 13 (2), 173–177.

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.

Winter, S., G. Neubaum, S. C. Eimler, V. Gordon, J. Theil, J. Herrmann, J. Meinert, and N. C. Krämer(2014). “Another brick in the Facebook wall - How personality traits relate to the content of statusupdates.” Computers in Human Behavior 34, 194–202.

Wu, W., L. Chen, and L. He (2013). “Using Personality to Adjust Diversity in Recommender Systems.”In: HT ’13: Proceedings of the 24th ACM Conference on Hypertext and Social Media. Paris, France:ACM, pp. 225–229. ISBN: 9781450319676.

Wu, X., V. Kumar, J. Ross Quinlan, J. Ghosh, Q. Yang, H. Motoda, G. J. McLachlan, A. Ng, B. Liu,P. S. Yu, Z.-H. Zhou, M. Steinbach, D. J. Hand, and D. Steinberg (2008). “Top 10 algorithms in datamining.” English. Knowledge and Information Systems 14 (1), 1–37. ISSN: 0219-1377.

Yu, L. and M. Wu (2010). “The Relation of Personality and Self-disclosure on Renren.” In: 2nd Symposiumon Web Society (SWS), pp. 435–442.

Buettner, R.: Innovative Personality-based Digital Services. In PACIS 2016 Proceedings: 20th Pacific Asia Conference on Information Systems (PACIS), June 27 - July 1, Chiayi, Taiwan.