A person wakes on a day with no urgent appointment, then rises on another day before a demanding journey. Cortisol concentration in saliva may rise during the first part of both mornings, but the size and shape of that rise need not match. The cortisol awakening response, or CAR, names a particular change in cortisol around the transition from sleep to wakefulness. It is not a synonym for having cortisol, for the entire daily hormone rhythm, or for how strongly a person reacts to a stressful conversation. Its meaning depends on precisely when waking occurred, when samples were taken, and what question the investigator is asking.
A response tied to waking
Cortisol secretion has a daily rhythm, often reaching relatively high levels in the morning and declining across much of the day. Superimposed on that broad rhythm is an increase after waking that can be characterized by sampling at awakening and over the following interval. The CAR is this change, conventionally assessed through several early-morning samples. A single high morning value cannot show whether cortisol rose after waking. Nor can a late sample accurately stand in for a sample taken at the moment of waking. These simple distinctions are essential because a plausible biological story can otherwise be built on a measurement artifact.
The endocrine process involves the hypothalamic-pituitary-adrenal axis and circadian regulation, but researchers do not treat it as a single reflex with one trigger. Stalder and colleagues’ later review describes coordination among the sleep-wake transition, circadian timing, anticipatory factors, and mechanisms affecting cortisol release and action. Awakening exposes a person to light, posture change, movement, and the prospect of the day. These events are related but not identical. A person can wake while remaining in bed, get up immediately, or be woken earlier than planned. Interpreting the response requires attention to these conditions rather than assuming that a single external stimulus caused it.
Cortisol is a glucocorticoid hormone with effects across metabolic, immune, and nervous-system processes. Calling the CAR a morning preparation signal is a useful hypothesis about function, not proof that every observed rise optimizes performance. A functional interpretation must connect hormone dynamics to a specified outcome under a specified schedule. It must also consider that secretion, transport, receptor sensitivity, and tissue response are not the same thing. Saliva records one accessible aspect of the system, not the entirety of glucocorticoid action throughout the body.
What an early-morning curve can measure
One approach takes a sample immediately on waking and further samples during roughly the next half hour or longer. Researchers may report the increase from the first value, an area-under-the-curve measure, or a more detailed time course. These summaries answer different questions. The rise from baseline emphasizes change after waking. A total-output measure is influenced by the starting concentration as well as the later values. Two participants can therefore have the same increase but different total cortisol exposure, or the same total output with different curve shapes. Comparing studies without checking which summary was used can create an apparent disagreement that is really a difference in the measure.
Timing is especially consequential. Smyth and colleagues studied the early time course using electronic monitoring and excluded samples that were delayed beyond a narrow window after awakening. Their work illustrates why an unobserved diary entry saying “at waking” may not be enough. If the first sample is collected after cortisol has already begun rising, an estimated baseline is too high and the calculated change can be distorted; the direction of bias depends on the exact summary and delay. The important point is that waking and collection times are part of the biological measure, not administrative details that can be repaired by a polished statistical model.
Expert consensus guidelines recommend recording awakening objectively where feasible, keeping sample times accurate, and reporting the sampling protocol transparently. They also emphasize potential confounds such as sleep timing, medication, and daily context. A researcher may ask for repeated mornings because an individual morning is affected by the circumstances of that day. Even repeated samples do not turn the CAR into an immutable personal trait. They estimate a pattern under specified conditions. The degree of consistency across days is an empirical question, not an assumption licensed by taking more tubes of saliva.
Salivary cortisol can be collected in ordinary life, which improves ecological relevance compared with measuring only in a laboratory. That advantage comes with weaker direct oversight. Electronic time stamps or actigraphy can improve confidence that the protocol was followed, but devices also have limitations. An actigraph estimates sleep and waking from movement; it does not directly observe endocrine events. A time-stamped tube cap shows when it opened, not whether collection technique was perfect. Good measurement combines several imperfect observations and reports residual uncertainty.
Anticipation without a simple stress score
A demanding day may be associated with a different morning response from an undemanding day, and some theories propose that anticipation helps mobilize resources for expected demands. Yet it is too simple to read a large CAR as “more stress” and a small one as “less stress.” Demands differ in predictability, controllability, emotional meaning, and timing. The person’s sleep, wake time, current illness, medication, and sampling adherence may also differ. An association between a stated expectation and a hormone curve does not prove which of these pathways produced the pattern.
Stress-reactivity experiments ask how cortisol changes after a defined challenge such as a social-evaluative task. The CAR asks what happens around waking. Both involve the same hormone, but they are different phenomena with different time anchors. Someone may show a pronounced response to a laboratory challenge and an unremarkable morning rise, or the reverse. The system is context-sensitive. Treating the two curves as interchangeable conceals the mechanisms researchers are trying to distinguish.
Likewise, the CAR is not simply the diurnal slope. The slope describes how cortisol changes across hours from morning toward evening, often using samples over the day. A person could have a normal-looking early rise and a relatively flat later decline. Another could show a modest post-waking rise and a marked daylong fall. Both patterns might matter for a study, but each needs its own measurement. The related Learning Hub entry on the CAR gives a broader orientation; this page concentrates on the mechanism and on the conditions under which its measurement can bear an interpretation.
Anticipation is also not the same as conscious anxiety. A person may know an early meeting is scheduled without feeling distressed, and the body may respond to expected activity or altered waking time. In another case, worries may impair sleep before the samples begin. A study that records only a morning questionnaire cannot cleanly separate these routes. Stronger designs compare well-characterized mornings within people, measure sleep and timing, and state which part of the anticipatory account they actually tested.
Biological timing changes the question
“Morning” is a clock label, while circadian phase is an internal biological state. A shift worker who wakes at midday and a daytime worker who wakes at seven are both at their personal awakening point, yet they may differ in light exposure, accumulated sleep, and circadian alignment. Their CAR measurements cannot be interpreted simply by matching minutes after waking. The same issue arises after travel, irregular sleep schedules, and an alarm that interrupts sleep earlier than usual. A protocol that ignores such differences may wrongly attribute a timing effect to personality or stress.
Stalder and colleagues’ integrative review treats the CAR as a coordinated transition rather than an isolated spike. That framing is valuable because it makes distinct hypotheses possible: whether a change in wake time, circadian phase, light, or anticipated demands alters the response; whether this alteration affects later function; and whether the same result appears in people with different schedules. These hypotheses are not settled merely because a curve exists. They require designs able to separate correlated morning events.
The body also responds to waking through pathways other than cortisol. Alertness, autonomic activity, temperature, and movement change over the same period. A convenient narrative can turn all these changes into one “activation” process, but biological measures need not move together. If an intervention alters heart rate after waking, that does not establish that it changed the CAR. If cortisol changes, it does not establish improved alertness. The value of measuring several systems is to test their relations rather than assume they tell the same story.
From a group pattern to an individual interpretation
Group-level associations can be informative without providing an individual diagnostic test. A study may find that a particular group has, on average, a different CAR from a comparison group. Distributions can overlap substantially, and the difference may depend on the measurement protocol or selected sample. A single person’s curve therefore should not be labeled healthy, burned out, resilient, or ill from a generic internet threshold. Clinical endocrine assessment uses its own indications and validated procedures; a research CAR curve is not a substitute.
Even when two laboratories agree on sampling times, they may differ in whether people are employed, caring for children, studying, or recovering from illness. Such social contexts shape sleep and the meaning of tomorrow’s demands. They are not nuisance variables to erase reflexively. The goal is to say which population and setting support the claim. An observed association in regular-schedule adults may not generalize to adolescents, shift workers, or people with disrupted sleep. Biological specificity includes the life circumstances under which biology was observed.
Repeated-measures designs can help distinguish within-person variation from between-person differences. For example, a person might have a larger rise on their own early-work days than on their own rest days, while people who generally work early may not differ from others on average. These are different questions. A cross-sectional result cannot be repurposed as evidence that changing one person’s expectations will change their hormone response. Causal experiments would require a carefully controlled change, appropriate timing, and enough repeated observations to separate an effect from ordinary day-to-day fluctuation.
Measurement itself can also change a morning. Waking to take a sample immediately may make a participant more alert or alter their routine. This does not invalidate the research, but it reminds us that ecologically sampled physiology is collected inside human behavior. A good protocol makes the burden manageable and records deviations. An article that promises perfect capture of natural life from salivary sampling overstates what even a careful field study can achieve.
How to use the idea responsibly
For a researcher, the first question is not whether cortisol is interesting, but whether the CAR is the right measure for the proposed hypothesis. If the question concerns a response to interpersonal challenge, a challenge protocol is more direct. If it concerns all-day regulation, a daylong sampling schedule is needed. If it concerns waking transition, the CAR is relevant, provided awakening and sample adherence are recorded. Matching construct to protocol prevents a wide-ranging interpretation from being attached to a narrow curve.
For a practitioner or designer, the CAR can suggest that the beginning of the day is biologically distinct, but it does not prescribe a universal morning routine. Claims that a breath exercise, cold exposure, or productivity habit “optimizes cortisol” require direct testing with appropriate comparison groups and meaningful outcomes. Even a demonstrated change in the curve would not automatically show a benefit. The practical question would be whether the person sleeps, functions, and feels better over time, and whether any change outweighs effort or harm.
For readers interpreting their own laboratory results, context matters. Cortisol testing ordered for medical reasons may use different samples and diagnostic logic from CAR research. Questions about symptoms or medication belong with a qualified clinician who knows the indication. The research literature can explain why timing matters, but it cannot interpret an individual’s endocrine status from an isolated graph shared without protocol details.
What the evidence does not show
The evidence does not establish that a larger CAR is always healthier, more productive, or more stressed. It does not make a one-morning saliva curve a diagnosis, a stable personality marker, or a direct measure of resilience. Current models integrate circadian timing, awakening, anticipation, and downstream hormone action, but the contribution of each varies with the design and remains an active research problem. Measurement adherence can materially alter the estimate. Any claim should name the sampling schedule, population, comparator, and outcome before giving the rise a psychological meaning.
The cortisol awakening response is valuable precisely because it asks a narrow, time-sensitive question. It invites investigators to observe a transition that daily averages can miss. Its limits are equally instructive: a hormone curve is evidence about sampled hormone dynamics, not a verdict on the person who produced it. Careful interpretation begins with the moment of waking and moves outward only as far as the data allow.
Research references
- Stalder, T., et al. (2016). Assessment of the cortisol awakening response: Expert consensus guidelines. Psychoneuroendocrinology, 63, 414–432. View source
- Stalder, T., et al. (2025). The cortisol awakening response: Regulation and functional significance. Endocrine Reviews, 46, 43–59. View source
- Smyth, N., Thorn, L., Hucklebridge, F., Evans, P., & Clow, A. (2015). Detailed time course of the cortisol awakening response in healthy participants. Psychoneuroendocrinology, 62, 200–203. View source
- Clow, A., Hucklebridge, F., Stalder, T., Evans, P., & Thorn, L. (2010). The cortisol awakening response: More than a measure of HPA axis function. Neuroscience & Biobehavioral Reviews, 35, 97–103. View source
- Wüst, S., Federenko, I., Hellhammer, D. H., & Kirschbaum, C. (2000). Genetic factors, perceived chronic stress, and the free cortisol response to awakening. Psychoneuroendocrinology, 25(7), 707–720. View source