Beyond rating scales: With targeted evaluation, large language models are poised for psychological assessment.

In this narrative review, we survey recent empirical evaluations of AI-based language assessments and present a case for the technology of large language models to be poised for changing standardized psychological assessment. Artificial intelligence has been undergoing a purported "paradigm shift" initiated by new machine learning models, large language models (e.g., BERT, LAMMA, and that behind ChatGPT). These models have led to unprecedented accuracy over most computerized language processing tasks, from web searches to automatic machine translation and question answering, while their dialogue-based forms, like ChatGPT have captured the interest of over a million users. The success of the large language model is mostly attributed to its capability to numerically represent words in their context, long a weakness of previous attempts to automate psychological assessment from language. While potential applications for automated therapy are beginning to be studied on the heels of chatGPT's success, here we present evidence that suggests, with thorough validation of targeted deployment scenarios, that AI's newest technology can move mental health assessment away from rating scales and to instead use how people naturally communicate, in language.

Natural language analyzed with AI-based transformers predict traditional subjective well-being measures approaching the theoretical upper limits in accuracy.

We show that using a recent break-through in artificial intelligence -transformers-, psychological assessments from text-responses can approach theoretical upper limits in accuracy, converging with standard psychological rating scales. Text-responses use people's primary form of communication -natural language- and have been suggested as a more ecologically-valid response format than closed-ended rating scales that dominate social science. However, previous language analysis techniques left a gap between how accurately they converged with standard rating scales and how well ratings scales converge with themselves - a theoretical upper-limit in accuracy. Most recently, AI-based language analysis has gone through a transformation as nearly all of its applications, from Web search to personalized assistants (e.g., Alexa and Siri), have shown unprecedented improvement by using transformers. We evaluate transformers for estimating psychological well-being from questionnaire text- and descriptive word-responses, and find accuracies converging with rating scales that approach the theoretical upper limits (Pearson r = 0.85, p < 0.001, N = 608; in line with most metrics of rating scale reliability). These findings suggest an avenue for modernizing the ubiquitous questionnaire and ultimately opening doors to a greater understanding of the human condition.

Abbreviated Three-Item Versions of the Satisfaction with Life Scale and the Harmony in Life Scale Yield as Strong Psychometric Properties as the Original Scales.

The cognitive components of subjective well-being can be measured with the Satisfaction with life scale (SWLS) and the Harmony in life scale (HILS), which both comprise five items each. The aim of this article is to abbreviate these scales and examine their psychometric properties and validity. Three datasets including test-retest data are used (N = 787; N = 860; N = 343). The two first datasets were already collected, whereas the third dataset included delivering the three-item scales (SWLS-3; HILS-3) together (in random order) with one shared instruction. The last study was pre-registered, including open data and code. The SWLS-3 and the HILS-3 demonstrate good psychometric properties, including very high internal consistency and item total correlations, strong test-retest reliability, where two-factor models of cognitive well-being tend to yield very good fit indices. Further, the scales demonstrate measurement invariance across time and gender. In fact, the three-item scales demonstrate as strong psychometric properties as compared with the five-item scales. Additionally, the scales demonstrate similar validity by yielding similar correlations to assessments of well-being, mental health problems and social desirability. Thus, the SWLS-3 and the HILS-3 can efficiently be used together with one shared instruction, without compromising (and in most aspects even yielding small improvements) the psychometric soundness of the scales.

Sad mood and poor sleep are related to task-unrelated thoughts and experience of diminished cognitive control.

Previous studies have indicated that a sad mood and sleep deprivation increase mind wandering, but it is unclear whether these associations reflect reduced effort in concentrating on the task at hand or diminished cognitive control. In an internet-based experiment, participants completed a sleep disturbance questionnaire followed by a complex span task and a 2-back task with thought-sampling probes. Subsequently, participants underwent a positive, neutral, or negative mood induction prior to repeating the 2-back. The results (N = 504) replicated the finding of increased task-unrelated thoughts following sad mood induction, B = 0.56 (SE = 0.14), p < 0.01, d = 0.31. Unguided thoughts were increased following sad mood induction, B = 0.31 (0.13), p = 0.02, but working memory did not significantly moderate this association (p = 0.31). People reported a lower degree of trying to concentrate on the 2-back after the sad mood induction, B = -0.07 (0.04), p = 0.04, but actual performance was not affected (p = 0.46). Sleep disturbances showed small associations with task-unrelated, B = 0.23 (0.08), p < 0.01, and unguided thoughts, B = 0.32 (0.08), p < 0.01. This study strengthens the evidence that a sad mood and poor sleep relate to mind wandering.

A daily diary study on maladaptive daydreaming, mind wandering, and sleep disturbances: Examining within-person and between-persons relations.

Cross-sectional and experimental research have shown that task-unrelated thoughts (i.e., mind wandering) relate to sleep disturbances, but there is little research on whether this association generalizes to the day-level and other kinds of task-unrelated mentation. We employed a longitudinal daily diary design to examine the within-person and between-person associations between three self-report instruments measuring mind wandering, maladaptive daydreaming (a condition characterized by having elaborate fantasy daydreams so insistent that they interfere with daily functioning) and sleep disturbances. A final sample of 126 participants self-identified as experiencing maladaptive daydreaming completed up to 8 consecutive daily reports (in total 869 daily observations). The scales showed acceptable-to-excellent within-person reliability (i.e., systematic day-to-day change) and excellent between-person reliability. The proportion of between-person variance was 36% for sleep disturbances, 57% for mind wandering, and 75% for maladaptive daydreaming, respectively (the remaining being stochastic and systematic within-person variance). Contrary to our pre-registered hypothesis, maladaptive daydreaming did not significantly predict sleep disturbances the following night, B = -0.00 (SE = 0.04), p = .956. Exploratory analyses indicated that while nightly sleep disturbances predicted mind wandering the following day, B = 0.20 (SE = 0.04), p < .001, it did not significantly predict maladaptive daydreaming the following day, B = -0.04 (SE = 0.05), p = .452. Moreover, daily mind wandering did not significantly predict sleep disturbances the following night, B = 0.02 (SE = 0.05), p = .731. All variables correlated at the between-person level. We discuss the implications concerning the differences between maladaptive daydreaming and mind wandering and the possibility of targeting sleep for mind wandering interventions.

Online validation of combined mood induction procedures.

Film clips, music, and self-referential statements (termed Velten, after their originator) have been successfully used to temporarily induce sadness and happiness. However, there is little research on the effectiveness of these procedures combined, particularly in internet-based settings, and whether Velten statements contribute to alter mood beyond the effect of simple instructions to close one's eyes and enter the targeted mood. In Study 1 (N = 106) we examined the effectiveness 80 Velten statements (positive, negative, neutral-self, neutral-facts) to create brief and effective sets that might be used in future research. In Study 2 (N = 445) we examined the effect size of 8-min combined mood induction procedures, which presented video clips in the first half and music excerpts with Velten statements or closed eyes instructions in the second half. Participants answered questionnaires on social desirability, joviality, and sadness before being randomly assigned to 1 of 7 groups varying in Valence (positive, negative, neutral) and Velten (closed eyes control, self-referential Velten, and, in the case of neutral condition, factual statements). Subsequently, participants completed the joviality and sadness scales a second time. Compared to the neutral conditions, the positive mood inductions increased joviality (Hedges G = 1.35, 95% CI [1.07, 1.63]), whereas the negative mood inductions increased sadness (Hedges G = 1.28, 95% CI [1.01, 1.55]). We did not observe any significant difference between Velten and closed eyes instructions in inducing joviality or sadness, nor did we observe any significant difference between neutral Velten statements referring to self and facts. Although social desirability bias was associated with reports of greater joviality and lower sadness, it could not account for the effects of the positive and negative mood induction procedures. We conclude that these combined mood induction procedures can be used in online research to study happy and sad mood.

Semantic measures: Using natural language processing to measure, differentiate, and describe psychological constructs.

Psychological constructs, such as emotions, thoughts, and attitudes are often measured by asking individuals to reply to questions using closed-ended numerical rating scales. However, when asking people about their state of mind in a natural context ("How are you?"), we receive open-ended answers using words ("Fine and happy!") and not closed-ended answers using numbers ("7") or categories ("A lot"). Nevertheless, to date it has been difficult to objectively quantify responses to open-ended questions. We develop an approach using open-ended questions in which the responses are analyzed using natural language processing (Latent Semantic Analyses). This approach of using open-ended, semantic questions is compared with traditional rating scales in nine studies (N = 92-854), including two different study paradigms. The first paradigm requires participants to describe psychological aspects of external stimuli (facial expressions) and the second paradigm involves asking participants to report their subjective well-being and mental health problems. The results demonstrate that the approach using semantic questions yields good statistical properties with competitive, or higher, validity and reliability compared with corresponding numerical rating scales. As these semantic measures are based on natural language and measure, differentiate, and describe psychological constructs, they have the potential of complementing and extending traditional rating scales. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

The affective profiles, psychological well-being, and harmony: environmental mastery and self-acceptance predict the sense of a harmonious life.

Background. An important outcome from the debate on whether wellness equals happiness, is the need of research focusing on how psychological well-being might influence humans' ability to adapt to the changing environment and live in harmony. To get a detailed picture of the influence of positive and negative affect, the current study employed the affective profiles model in which individuals are categorised into groups based on either high positive and low negative affect (self-fulfilling); high positive and high negative affect (high affective); low positive and low negative affect (low affective); and high negative and low positive affect (self-destructive). The aims were to (1) investigate differences between affective profiles in psychological well-being and harmony and (2) how psychological well-being and its dimensions relate to harmony within the four affective profiles. Method. 500 participants (mean age = 34.14 years, SD. = ±12.75 years; 187 males and 313 females) were recruited online and required to answer three self-report measures: The Positive Affect and Negative Affect Schedule; The Scales of Psychological Well-Being (short version) and The Harmony in Life Scale. We conducted a Multivariate Analysis of Variance where the affective profiles and gender were the independent factors and psychological well-being composite score, its six dimensions as well as the harmony in life score were the dependent factors. In addition, we conducted four multi-group (i.e., the four affective profiles) moderation analyses with the psychological well-being dimensions as predictors and harmony in life as the dependent variables. Results. Individuals categorised as self-fulfilling, as compared to the other profiles, tended to score higher on the psychological well-being dimensions: positive relations, environmental mastery, self-acceptance, autonomy, personal growth, and purpose in life. In addition, 47% to 66% of the variance of the harmony in life was explained by the dimensions of psychological well-being within the four affective profiles. Specifically, harmony in life was significantly predicted by environmental mastery and self-acceptance across all affective profiles. However, for the low affective group high purpose in life predicted low levels of harmony in life. Conclusions. The results demonstrated that affective profiles systematically relate to psychological well-being and harmony in life. Notably, individuals categorised as self-fulfilling tended to report higher levels of both psychological well-being and harmony in life when compared with the other profiles. Meanwhile individuals in the self-destructive group reported the lowest levels of psychological well-being and harmony when compared with the three other profiles. It is proposed that self-acceptance and environmental acceptance might enable individuals to go from self-destructive to a self-fulfilling state that also involves harmony in life.

Iranian and Swedish adolescents: differences in personality traits and well-being.

Introduction. This study addresses the need to further contextualize research on well-being (e.g., Kjell, 2011) in terms of cross-cultural aspects of personality traits among adolescents and by examining two different conceptualizations of well-being: subjective well-being (i.e., life satisfaction, positive and negative affect) and psychological well-being (i.e., positive relations with others, environmental mastery, self-acceptance, autonomy, personal growth, and life purpose). Methods. Iranian (N = 122, mean age 15.23 years) and Swedish (N = 109, mean age 16.69 years) adolescents were asked to fill out a Big Five personality test, as well as questionnaires assessing subjective well-being and psychological well-being. Results. Swedes reported higher subjective and psychological well-being, while Iranians reported higher degree of Agreeableness, Openness and Conscientiousness. Neuroticism and Extraversion did not differ between cultures. Neuroticism was related to well-being within both cultures. Openness was related to well-being only among Iranians, and Extraversion only among Swedes. A mediation analysis within the Swedish sample, the only sample meeting statistical criteria for mediation analysis to be conducted, demonstrated that psychological well-being mediated the relationship between Neuroticism and subjective well-being as well as between Extraversion and subjective well-being. Conclusions. Certain personality traits, such as Extraversion, Openness, and Conscientiousness, relate differently to well-being measures across cultures. Meanwhile, Neuroticism seems to relate similarly across cultures at least with regard to subjective well-being. Furthermore, the results give an indication on how psychological well-being might mediate the relationship between certain personality traits and subjective well-being. Overall, the complexity of the results illustrates the need for more research whilst supporting the importance of contextualizing well-being research.

Exploring the impact of positive and negative emotions on cooperative behaviour in a Prisoner's Dilemma Game.

Objective. To explore the influences of discrete positive and negative emotions on cooperation in the context of a social dilemma game. Design. Two controlled studies were undertaken. In Study 1, 69 participants were randomly assigned to an essay emotion manipulation task designed to induce either guilt, joy or no strong emotion. In Study 2, 95 participants were randomly assigned to one of the same three tasks, and the impact of emotional condition on cooperation was explored using a repeated Prisoner's Dilemma Game. Results. Study 1 established that the manipulation task was successful in inducing the specified emotions. The analysis from Study 2 revealed no significant main effects for emotions, in contrast to previous research. However, there was a significant effect for participants' pre-existing tendency to cooperate (social value orientation; SVO). Conclusion. Methodological explanations for the result are explored, including the possible impact of trial-and-error strategies, different cooperation games and endogenous vs exogenous emotions.