Blog: Critique of Feltron’s Report and Generating Alternative Measurements of Affect
As observed in Feltron’s 2009 “Report on an Assessment of Demeanor”, Feltron captured a mood report based on two construct (i.e. extroversion and introversion; happy and sad) as measurements. In addition, qualitative data from interviews were provided from Feltron’s peers in contributing to the assessment of Feltron’s mood. Though results were insightful, however, issues of mutually-exclusive binary categorizations may limit the rich and complex expression of mood (affect). This blog report aims to provide a different approach on the measurement and assessment of affect. As indicated, “Affect is the experience of feeling or emotion and is a key process of an organism’s interaction with stimuli.”This analysis will present methods of self-reported measures of mood and observer-rated behavior, in hopes of examining self-reported mood, mood congruence and affect display.
I have generated two alternative measurements in assessing mood. The first measurement is an experience sampling mobile app survey based on a semantic differential response format referenced from the State-Trait Emotion Measure(STEM). The STEM consists of five positive and five negative emotions and was specifically selected because it can effectively capture general categories of affect with enough details to describe the complexity of a specific mood. Furthermore, a semantic differential format is employed so that the respondent can be assessed on a set of bipolar adjective pairs for ease of information input. The metric of this study is through a five-point Likert scale, where the user can only select one option amongst the scale. Also, by using a randomized signal-contingent measure, the user will not be able to anticipate mood assessment. The mood assessment would be requested several times of a day.
Another method that will be used to measure affect would be observer-rated behaviors from close peers and family members (observers). A similar method of an experience sampling mobile app would be triggered by a GPS sensor that connects to the individual’s mobile device. The observer will be prompted to rate the individual’s affect and mood when the sensor detects that there has been more than fifteen minutes of close contact with the individual. This is an event-contingent system, where a preselected sample of the individual is asked to perform this task. This group of individuals is selected based on the individual’s self-identified close relationship and social life prior to the study. The method and metric of the survey will be in a similar fashion like the individual-survey, which is also a semantic differential response format with a five-point Likert scale based on the framework of STEM, with an additional option to input qualitative data. Responses of the mood assessment will be compiled to produce a more robust measure in assessing the individual’s affect display.
STRENGTHS AND LIMITATIONS
One of the strength that these measures offer is to consider mood in terms of how it is represented. One of the biggest challenges with measuring mood is its transparent cognitive processes and intangibility to measure. By measuring actual affect and perceived affect, this comparison will give insights as to how mood is represented from both a personal level and from perceived affection display. Non-congruent mood between the individual and the observer does not suggest inappropriateness of mood, but could provide certain level of insights. One of the limitations of these operations is the issue of how representative and valid is the observer’s reflection. Uncontrolled variables such as biases can be heavily introduced by the observer’s own mood during these assessments. However, since no single-handed observer’s reports is used to attest for the individual’s affect, by taking the mean of a collection of observer report will provide for stronger results, therefore more robust in external validity.
RELIABILITY AND OVERALL VALIDITY OF APPROACH
The design of the methods is highly replicable based on several factors. The survey of the mobile app are all standardized and systematized for all users to ensure interval validity and to avoid complications such as researcher bias, demand characteristics and variability in measurement artifacts. The survey will be constructed based on the consideration of question bias, wording and placement to avoid anchoring effects and to ensure internal validity. As for external validity, there has been a concern as to how representative observer ratings are in relationship to the individual. This concern arises mainly due to the problem with whether the individual will alter affect display base on the relationship he or she has with the observer. This is a major concern, because there are cases where there could be potentially incongruence between actual affect and perceived affect, but is socially acceptable. This could due to hierarchical roles or demand characteristics, such as when the observer is the individual’s supervisor or mentor. A method to reduce this concern is to produce a set of criteria for selecting observers, such as issues of trust, rapport and level of comfort the individual has with the observer.
PRACTICAL USAGE IMPLICATIONS
A potential practical usage for this application is to help psychologists in building better understanding of their patient’s mental status. Current reports include self-reports from patients or diary studies. However, retrospective memory of reporting self-diaries have always been a form of concern. Using mobile application tracking patient’s mood in real time can help reduce this bias. Also, reports of this individual’s mood display from a supportive group of family and peers can help clinicians in assessing different perspectives of the patient’s condition.