A new study is taking aim at a deceptively difficult problem in psychological science: how to compare people’s momentary feelings when those feelings are measured repeatedly, using questions that may not behave in the same way for everyone. Published in Communications Psychology, the article “Improving the comparability of momentary affect assessments and accounting for item-specific effects in within-person affect dynamics” examines how researchers can make intensive, real-time measurements of emotion more precise and more meaningful.
Momentary affect refers to a person’s emotional state at a particular point in time. Instead of asking someone to summarize how they generally feel, researchers may ask them several times a day whether they feel happy, anxious, calm, irritated, energetic, or sad. These assessments are often delivered through smartphones or other digital devices, allowing scientists to observe emotional changes as they unfold in everyday life. The approach has transformed psychological research, but it also introduces a measurement challenge: a score recorded at 10 a.m. may not be perfectly comparable with a score recorded at 4 p.m., even when the same person answers the same question.
The problem begins with the items used to measure affect. A questionnaire may contain several words or statements intended to capture a broader emotional dimension, such as positive or negative affect. Yet each item can carry its own psychological and statistical characteristics. One person may interpret “excited” as pleasant and energizing, while another may experience it as uncomfortable or overwhelming. “Calm” may signal well-being for one individual but emotional numbness for another. If these item-specific differences are ignored, researchers may mistake the behavior of a question for a genuine change in a person’s emotional state.
Koch, Jähne, Koslowski and colleagues focus on improving the comparability of these repeated assessments while accounting for the distinctive effects of individual items. Their work addresses a central issue in within-person research: the goal is not primarily to compare one person’s average mood with another person’s average mood, but to understand how an individual’s affect changes over time. This distinction matters because psychological processes often operate inside people rather than simply between them. The factors associated with feeling more anxious than usual may be different from the factors associated with one person being more anxious than another.
Within-person affect dynamics can be understood as the patterns linking emotional states across moments. For example, a researcher might ask whether a rise in stress is followed by a decline in positive affect, whether emotions return to baseline after a difficult event, or whether one feeling tends to persist into the next measurement. These patterns are commonly studied with intensive longitudinal data, in which each participant contributes many observations. Statistical models can then estimate temporal associations, emotional persistence, fluctuations, and possible feedback loops.
However, repeated measurements are not automatically equivalent simply because they use the same wording. A particular item may be more sensitive to rapid changes than another. Some emotional terms may produce more variable responses, while others may remain relatively stable. Certain items may also show stronger links to the previous moment, creating the appearance of emotional carryover. If those differences are treated as evidence about affect itself, the resulting model may describe the questionnaire as much as it describes the person.
The study’s methodological emphasis is therefore highly relevant to the expanding field of experience sampling and ecological momentary assessment. These approaches generate rich streams of data, but richness can create false precision if measurement properties are not carefully considered. By modeling item-specific effects, researchers can separate at least two sources of variation: changes in the underlying affective state and systematic differences in how particular questions capture that state. This separation can make estimates of emotional dynamics more comparable across measurement occasions and potentially across studies.
The distinction is technical, but its consequences are easy to understand. Imagine tracking someone’s mood with three items: cheerful, relaxed, and enthusiastic. If enthusiasm naturally rises and falls more sharply than relaxation, a simple average may suggest that the person’s overall positive affect is changing dramatically. An item-sensitive model could reveal that much of the apparent movement comes from the enthusiasm item itself. The person may still be experiencing meaningful emotional variation, but the size and timing of that variation could look different once the measurement behavior of each item is taken into account.
This issue also affects the reproducibility of psychological findings. Researchers may use different sets of affect items, different response scales, or different assessment schedules. Even when studies appear to measure the same construct, their results may not be directly comparable if the items respond differently to momentary changes. Improving comparability could help scientists determine whether apparently conflicting findings reflect genuine differences in emotional processes or merely differences in measurement design. It may also support more reliable synthesis of results across datasets and populations.
The importance of this work extends beyond academic statistics. Momentary affect research is increasingly connected to mental-health monitoring, personalized interventions, and digital behavioral science. If a system is designed to detect when a person’s emotional state is deteriorating, it must distinguish a real change from noise caused by a particular question. More careful measurement could eventually improve the interpretation of mood diaries, clinical assessments, and smartphone-based interventions. Such applications require caution, however, because a statistical model can clarify patterns in reported affect without proving why those patterns occur or guaranteeing an accurate diagnosis.
The paper arrives at a moment when psychology is moving toward more dynamic explanations of human experience. Emotions are not fixed traits that remain constant from one questionnaire to the next; they fluctuate in relation to circumstances, thoughts, bodily states, and social interactions. Capturing that movement is one of the major promises of intensive longitudinal research. Koch, Jähne, Koslowski and colleagues’ contribution centers on making that promise more dependable by recognizing that the measuring instruments themselves can influence the patterns researchers observe.
The broader message is that better emotional science may depend not only on collecting more data, but on understanding each data point more carefully. Thousands of smartphone responses cannot automatically reveal the architecture of a person’s emotional life. To identify meaningful within-person dynamics, researchers must account for the wording, sensitivity, and behavior of the items used to ask about feelings. By bringing item-specific effects into the analysis, the study offers a framework for treating momentary affect as both a psychological phenomenon and a measurement problem—an essential step toward faster, clearer, and more reproducible science of emotion.
Subject of Research: Improving the comparability of momentary affect assessments and accounting for item-specific effects in within-person affect dynamics.
Article Title: Improving the comparability of momentary affect assessments and accounting for item-specific effects in within-person affect dynamics.
Article References: Koch, T., Jähne, M.F., Koslowski, K. et al. Improving the comparability of momentary affect assessments and accounting for item-specific effects in within-person affect dynamics. Commun Psychol (2026). https://doi.org/10.1038/s44271-026-00473-0
Image Credits: AI Generated
DOI: 10.1038/s44271-026-00473-0
Keywords: momentary affect, within-person dynamics, ecological momentary assessment, experience sampling, psychological measurement, item-specific effects, emotional variability, intensive longitudinal data, affect assessment, reproducibility in psychology
