A single mood check-in can feel like a verdict. Tap “anxious” before a difficult meeting, “cheerful” after good news, or “distracted” during a noisy afternoon, and the score seems to say something durable about who you are.
A newly published study suggests that conclusion is often too strong. After separating more than 60,000 reports into differences between people, differences between days and changes within a day, researchers found that positive mood was highly fluid. Broader judgments about life, by contrast, looked much more stable. The lesson is not that mood tracking is useless. It is that different questions measure different kinds of truth.
The same person can generate very different scores
The study, published on October 8 in the *British Journal of Psychology*, reanalysed data from 1,299 participants in the Dutch “HowNutsAreTheDutch” project. For up to 30 days, volunteers received three prompts a day—morning, afternoon and evening—and rated statements including “I feel cheerful,” “I feel anxious,” “I worry a lot,” “I am easily distracted” and “I feel my life is worth living.” The final dataset contained 60,189 assessments.
The researchers asked where variation in each answer came from. Some variation reflected lasting differences between participants. Some reflected one person having a better Tuesday than Wednesday. Some followed a recurring time-of-day pattern. The remainder captured unscheduled fluctuations, unmeasured context and measurement noise.
The profiles were strikingly different. About 75% of the variation in “my life is worth living” lay between people, suggesting that this broad evaluation behaved relatively like a stable personal characteristic in this sample. Worry and distractibility sat closer to the middle, with roughly 55% of their variation between people.
Positive feelings moved far more within the same individual. Only 35.2% of the variation in cheerfulness and 33.7% of the variation in relaxation was between people. Put differently, nearly two-thirds of the observed variation in those ratings came from changes within people or small population-level timing effects. A low cheerfulness score may therefore describe a moment more readily than a person.
This distinction matters because psychology has often inferred individual patterns from group averages. A 2018 analysis in *PNAS* warned that relationships found across people do not automatically describe how one person changes over time. The new study gives that abstract problem a practical shape: “Who tends to feel cheerful?” and “When does this person feel cheerful?” are not interchangeable questions.
Clock time explained less than the messiness of life
The most surprising result may be what the clock failed to explain. Systematic morning-to-afternoon-to-evening differences accounted for only 0.3% to 2.8% of variation across the seven main items. Day-to-day changes within a person were larger, ranging from 7% for the life-worth-living item to about 16% for cheerfulness.
Much more variation remained in the residual category. It reached 46% for cheerfulness and 48% for relaxation, compared with 17% for the life evaluation. That residual is not a discovery of a hidden psychological force. It is a mixed box containing genuine reactions to conversations, stressors, environments and events—as well as response error and influences the study never measured.
The implication is subtle. The findings do not show that time of day never matters. Tiredness and energy displayed stronger time-linked patterns in additional analyses. Nor do they prove that a social interaction caused a mood shift, because context was not directly tested. They show that three fixed daily time points captured little of the structure behind most reported emotional movement.
That is precisely why ecological momentary assessment—asking people about experience in ordinary life—can be valuable. It reduces the need to reconstruct an entire week from memory. But immediacy does not eliminate noise, and repeated prompts do not automatically reveal what produced a change.
Mood tracking works better as a pattern than a verdict
For mood apps, research studies and therapy monitoring, the practical message is to match the measurement schedule to the question. A broad life evaluation may not require several prompts each day. A fluctuating state such as cheerfulness may need repeated observations, ideally paired with information about sleep, activity, company or stressful events. A 2025 methodological tutorial similarly argues that researchers should inspect how individual items perform rather than treating every check-in as equally informative.
The new evidence also has important limits. The sample was self-selected, 84% female and relatively highly educated. Although the protocol lasted 30 days, participants completed only about 10 days on average, and response rates fell sharply over time. The original data collection began in 2014. Three prompts spaced six hours apart could miss faster rhythms, while phrases such as “I worry a lot” may invite a general self-assessment even when participants are asked about the present moment. Most importantly, the analysis could not separate true contextual shifts from measurement error.
So a mood log should not be read as a diagnostic instrument or a daily identity test. Its strength lies in accumulation: whether a pattern repeats, what tends to precede it and whether it changes over weeks rather than minutes. One difficult check-in is data. A carefully observed pattern is evidence. Confusing the two turns a useful tool into a label.
