Matthias R. Mehl

Using linguistic cues for the automatic recognition of personality in conversation and text

It is well known that utterances convey a great deal of information about the speaker in addition to their semantic content. One such type of information consists of cues to the speakers personality traits, the most fundamental dimension of variation between humans. Recent work explores the automatic detection of other types of pragmatic variation in text and conversation, such as emotion, deception, speaker charisma, dominance, point of view, subjectivity, opinion and sentiment. Personality affects these other aspects of linguistic production, and thus personality recognition may be useful for these tasks, in addition to many other potential applications. However, to date, there is little work on the automatic recognition of personality traits. This article reports experimental results for recognition of all Big Five personality traits, in both conversation and text, utilising both self and observer ratings of personality. While other work reports classification results, we experiment with classification, regression and ranking models. For each model, we analyse the effect of different feature sets on accuracy. Results show that for some traits, any type of statistical model performs significantly better than the baseline, but ranking models perform best overall. We also present an experiment suggesting that ranking models are more accurate than multi-class classifiers for modelling personality. In addition, recognition models trained on observed personality perform better than models trained using self-reports, and the optimal feature set depends on the personality trait. A qualitative analysis of the learned models confirms previous findings linking language and personality, while revealing many new linguistic markers.

Linguistic Markers of Psychological Change Surrounding September 11, 2001

The diaries of 1,084 U.S. users of an on-line journaling service were downloaded for a period of 4 months spanning the 2 months prior to and after the September 11 attacks. Linguistic analyses of the journal entries revealed pronounced psychological changes in response to the attacks. In the short term, participants expressed more negative emotions, were more cognitively and socially engaged, and wrote with greater psychological distance. After 2 weeks, their moods and social referencing returned to baseline, and their use of cognitive-analytic words dropped below baseline. Over the next 6 weeks, social referencing decreased, and psychological distancing remained elevated relative to baseline. Although the effects were generally stronger for individuals highly preoccupied with September 11, even participants who hardly wrote about the events showed comparable language changes. This study bypasses many of the methodological obstacles of trauma research and provides a finegrained analysis of the time line of human coping with upheaval.

The Electronically Activated Recorder (EAR): A device for sampling naturalistic daily activities and conversations

A recording device called the Electronically Activated Recorder (EAR) is described. The EAR tape-records for 30 sec once every 12 min for 2–4 days. It is lightweight and portable, and it can be worn comfortably by participants in their natural environment. The acoustic data samples provide a nonobtrusive record of the language used and settings entered by the participant. Preliminary psychometric findings suggest that the EAR data accurately reflect individuals’ natural social, linguistic, and psychological lives. The data presented in this article were collected with a first-generation EAR system based on analog tape recording technology, but a second generation digital EAR is now available.

Knowing me, knowing you: the accuracy and unique predictive validity of self-ratings and other-ratings of daily behavior.

Many people assume that they know themselves better than anyone else knows them. Recent research on inaccuracies in self-perception, however, suggests that self-knowledge may be more limited than people typically assume. In this article, the authors examine the possibility that people may know a person as well as (or better than) that person knows himself or herself. In Study 1, the authors document the strength of laypeoples beliefs that the self is the best expert. In Study 2, the authors provide a direct test of self- and other-accuracy using an objective and representative behavioral criterion. To do this, the authors compared self- and other-ratings of daily behavior to real-life measures of act frequencies assessed unobtrusively over 4 days. Our results show that close others are as accurate as the self in predicting daily behavior. Furthermore, accuracy varies across behaviors for both the self and for others, and the two perspectives often independently predict behavior. These findings suggest that there is no single perspective from which a person is known best and that both the self and others possess unique insight into how a person typically behaves.

Does Posting Facebook Status Updates Increase or Decrease Loneliness? An Online Social Networking Experiment:

Online social networking is a pervasive but empirically understudied phenomenon. Strong public opinions on its consequences exist but are backed up by little empirical evidence and almost no causally conclusive, experimental research. The current study tested the psychological effects of posting status updates on Facebook using an experimental design. For 1 week, participants in the experimental condition were asked to post more than they usually do, whereas participants in the control condition received no instructions. Participants added a lab “Research Profile” as a Facebook friend allowing for the objective documentation of protocol compliance, participants’ status updates, and friends’ responses. Results revealed (1) that the experimentally induced increase in status updating activity reduced loneliness, (2) that the decrease in loneliness was due to participants feeling more connected to their friends on a daily basis, and (3) that the effect of posting on loneliness was independent of direct social feedback (i.e., responses) by friends.

Eavesdropping on Happiness Well-Being Is Related to Having Less Small Talk and More Substantive Conversations

Is the happy life characterized by shallow, happy-go-lucky moments and trivial small talk, or by reflection and profound social encounters? Both notions—the happy ignoramus and the fulfilled deep thinker—exist, but little is known about which interaction style is actually associated with greater happiness (King & Napa, 1998). In this article, we report findings from a naturalistic observation study that investigated whether happy and unhappy people differ in the amount of small talk and substantive conversations they have. Although the macrolevel and long-term implications of happiness have been studied extensively (Eid & Larsen, 2008; Howell & Howell, 2008), little is known about the daily social behavior of happy people, primarily because of the difficulty of objectively measuring everyday behavior. Many behavioral measures (e.g., experience sampling, day-reconstruction method) rely on self-reports and thus cannot be used to disentangle true associations between happiness and behavior from biases or idealized self-views. This is especially true for evaluatively loaded behaviors, such as the substance (or lack thereof) of one’s conversations. To address this difficulty, we used the Electronically Activated Recorder (EAR; Mehl, Pennebaker, Crow, Dabbs, & Price, 2001), a digital audio recorder that unobtrusively tracks real-world behavior by periodically recording snippets of ambient sounds while participants go about their daily lives.

The Social Dynamics of a Cultural Upheaval Social Interactions Surrounding September 11, 2001

Using the Electronically Activated Recorder (EAR), a new methodology for sampling behavioral data in naturalistic settings, we tracked the social lives of 11 people by recording 30-s snippets of ambient sounds in their environment approximately every 12 min. Participants wore the EAR continuously for 10 days from September 11, 2001. Pre-September 11 baseline data were available for all participants. Analyses of the coded sound information showed that although participants did not change in their overall amount of interactions, they gradually shifted from group conversations to dyadic interactions. Exploratory analyses revealed that a relative increase in dyadic interactions over the first 10 days after September 11 was marginally related to better psychological adjustment at follow-up. The findings have relevance for the understanding of stress and affiliation and normal psychological reactions to emergencies.

Computergestützte quantitative Textanalyse. Äquivalenz und Robustheit der deutschen Version des Linguistic Inquiry and Word Count

Zusammenfassung. Im Beitrag wird die deutsche Adaptation des Linguistic Inquiry and Word Count (LIWC) vorgestellt. Das LIWC ist ein computergestutztes Textanalyseprogramm mit integriertem Worterbuch, das von Pennebaker und Kollegen (2001) entwickelt wurde, um Essays aus Experimenten zum expressiven Schreiben zu untersuchen. Es werden zwei Studien zur Gute und Praktikabilitat des deutschen LIWC berichtet: (1) An einer Stichprobe zweisprachig vorliegender Texte (N = 122) wird die Aquivalenz der deutschen Ubersetzung mit dem Original uberpruft. (2) An einer Stichprobe von N = 104 E-Mails wird die Robustheit des Verfahrens gegenuber Schreibfehlern untersucht. Fur die meisten LIWC-Kategorien zeigt sich eine gute Aquivalenz. Des Weiteren erweist sich das LIWC als robust gegenuber Schreibfehlern. Diese Befunde sprechen fur die Brauchbarkeit des LIWC fur die Analyse deutscher Texte. Die beobachtete Robustheit ist von erheblicher praktischer Bedeutung fur die Analyse naturlicher geschriebener Sprache aus der compu...

Naturalistic Observation of Health-Relevant Social Processes: The Electronically Activated Recorder Methodology in Psychosomatics

&NA; This article introduces a novel observational ambulatory monitoring method called the electronically activated recorder (EAR). The EAR is a digital audio recorder that runs on a handheld computer and periodically and unobtrusively records snippets of ambient sounds from participants’ momentary environments. In tracking moment-to-moment ambient sounds, it yields acoustic logs of people’s days as they naturally unfold. In sampling only a fraction of the time, it protects participants’ privacy and makes large observational studies feasible. As a naturalistic observation method, it provides an observer’s account of daily life and is optimized for the objective assessment of audible aspects of social environments, behaviors, and interactions (e.g., habitual preferences for social settings, idiosyncratic interaction styles, subtle emotional expressions). This article discusses the EAR method conceptually and methodologically, reviews prior research with it, and identifies three concrete ways in which it can enrich psychosomatic research. Specifically, it can a) calibrate psychosocial effects on health against frequencies of real-world behavior; b) provide ecological observational measures of health-related social processes that are independent of self-report; and c) help with the assessment of subtle and habitual social behaviors that evade self-report but have important health implications. An important avenue for future research lies in merging traditional self-report–based ambulatory monitoring methods with observational approaches such as the EAR to allow for the simultaneous yet methodologically independent assessment of inner, experiential aspects (e.g., loneliness) and outer, observable aspects (e.g., social isolation) of real-world social processes to reveal their unique effects on health. Abbreviations EAR = electronically activated recorder SECSI = Social Environment Coding of Sound Inventory

When Leaving Your Ex, Love Yourself Observational Ratings of Self-Compassion Predict the Course of Emotional Recovery Following Marital Separation

Divorce is a highly stressful event, and much remains to be learned about the factors that promote psychological resilience when marriages come to an end. In this study, divorcing adults (N = 109) completed a 4-min stream-of-consciousness recording about their marital separation at an initial laboratory visit. Four judges rated the degree to which participants exhibited self-compassion (defined by self-kindness, an awareness of one’s place in shared humanity, and emotional equanimity) in their recordings. Judges evidenced considerable agreement in their ratings of participants’ self-compassion, and these ratings demonstrated strong predictive utility: Higher levels of self-compassion at the initial visit were associated with less divorce-related emotional intrusion into daily life at the start of the study, and this effect persisted up to 9 months later. These effects held when we accounted for a number of competing predictors. Self-compassion is a modifiable variable, and if our findings can be replicated, they may have implications for improving the lives of divorcing adults.