A behaviour analysis approach, as described in Part 2 of this series, offers a relatively direct way to assess emotion, personality or skills, and is thus complementary to traditional approaches that rely on indirect measures using verbal self-report.
Traditional psychometrics produce psychological scales aimed to reliably measure variables such as emotional experience, personality, intelligence, attitudes, etc. Published reports in social sciences offer several instruments that have been carefully calibrated on hundreds or thousands of participants. In personality assessment, for example, a scale comprises multiple items probing how the test-taker believes he or she generally thinks, feels or behaves in hypothetical situations. Response formats are mostly standardized (multiple-choice, Likert scale, etc). An individual’s composite score on a psychological scale is interpreted in relation to the distribution of scores of a norm group of people.
However, psychometric questionnaires are challenged by various self-report biases such as social desirability (tendency to report what you think is expected from you). The risk of insincere answers to self-report questionnaires is particularly high in the context of job recruitment (you will bend the truth in order to get hired). The interpretation accuracy is further constrained by the amount of self-knowledge the responder has. We know from research that introspection can be quite inaccurate and that it shows large interindividual differences. This is why researchers often ask close colleagues, friends or family members to also fill in the same questionnaire about the tested individual.
Self-report questionnaires may not always provide the most valid measure of personality or skills. They are likely to draw a distorted picture in contexts where the respondents see a clear benefit of presenting themselves in a certain way and have an inexpensive way of doing so voluntarily. Behaviour assessment provides a complementary alternative because it is more difficult and less obvious how to monitor or control expressive behaviour than it is to adapt verbal responses written on a questionnaire. In sum, behavioural measures may provide a more genuine and direct assessment of someone’s emotion, personality or skills.
The future in person assessment
In the light of these considerations, the future in person assessment is likely to include multimodal behavioural measures to overcome the challenges presented by psychometric questionnaires.
In effect, verbal and nonverbal behaviours are observable, visible manifestations of psychological traits and states. They can be measured objectively using standardized behaviour expert observation (which can be modelled and reproduced using algorithms, see Part 4) or using performance-based tests (such as emotion recognition tests). More subjective measures include intuitive impressions through self-report using non-expert behaviour observations (see Figure 2). However, objective measures can free person assessment from inaccuracies such as biases and everyday beliefs that are not grounded in scientific study. These distort and trouble the lens through which interviewees are observed and judged. Several parameters have already been found to systematically relate to expressions of temporary states such as emotions as well as more stable traits such as personalities. Scientific research indicates that verbal and nonverbal behaviours play a particularly important role in impression formation such as during the job interview. In short: You are what you do, not what you say you do.
Figure 2. Different ways to assess a person (red indicates the current focus of Vima)
By carefully blending state-of-the-art sensing methods for behavioural feature extraction, powerful computational analysis and subject matter expertise from behavioural psychology, Vima effectively takes a wide-angle integrating verbal and nonverbal behaviours to accurately predict personality traits and soft skills.