A star becomes a little dimmer, then brightens again. The dip returns several days later. Have we found a planet? We have found a signal worth investigating. The useful next step is to test what else could have produced it.
The gold standard of vetting an exoplanet transit candidate is a clear chain from recorded light to competing explanations, discriminating observations and a conclusion whose confidence matches the evidence. This guide works through that chain with invented data. It is for learners who can read a simple graph and want to understand why discovery requires more than recognising a familiar shape.
THE GOLD STANDARD OF ASTRONOMY
Choose the question you need to answer
Understand the claim · Chapters 1–2. Separate geometry from source identification.
Plan the investigation · Chapters 3–5. Choose checks, preserve data and state limits.
Check and continue · Chapters 6–8. Try changed cases and follow the reading routes.
Open the chapter contents · Try the transfer checks · The Gold Standard of Astronomy
Chapter contents
Understand the claim
Plan the investigation
3. Choose checks that distinguish competing explanations
A light curve records measured brightness over time. A transit interpretation proposes that an orbiting object crosses the visible face of a star and blocks some of its light. NASA’s transit-method explanation provides the basic geometry. The interpretation depends on which star supplies the light, which star dims and whether the measurements preserve the event faithfully.
Keep three statements separate: a dip was recorded; the dip repeats; a planet around a particular star explains it. Each adds a claim. Repetition strengthens the reason to investigate, but an eclipsing pair of stars can repeat too.
Write the proposed system before calculating a planet’s size. Is there one resolved star? Is there another source inside the measurement aperture? Does the archive identify the object as a candidate or as a confirmed planet? Preserve those labels in your own report.
The learner’s first useful sentence is therefore modest: “The data contain a recurring brightness decrease consistent with a transit-like event.” The remaining chapters ask what would justify saying more.
Imagine an aperture containing target star A and an unresolved eclipsing system B. Outside eclipse, A contributes 9,800 arbitrary flux units and B contributes 200. During an eclipse, B loses 50 units. The detector sees a combined baseline of 10,000 and a decrease to 9,950.
The observed fractional depth is 50 ÷ 10,000 = 0.005, or 0.5%. Yet the eclipse removes 50 ÷ 200 = 25% of B’s own light. A relatively deep event in a faint neighbour has become a shallow event in the combined measurement. These numbers illustrate dilution; they are not data for a named star.
If we instead assumed an isolated star and a small opaque planet crossing a uniformly bright disc, the approximation depth ≈ (planet radius ÷ stellar radius)² would give a radius ratio of √0.005 ≈ 0.071. The arithmetic is correct under that model. The model assignment may be wrong.
This is the central lesson: a precise calculation cannot identify the light source by itself. Real transit fitting also accounts for effects such as the star’s non-uniform surface brightness. The simple radius formula is a teaching approximation, not a substitute for source identification.
| Quantity | Calculation | Meaning |
|---|---|---|
| Combined baseline | 9,800 + 200 = 10,000 | Both sources contribute light |
| Observed dip | 50 / 10,000 = 0.5% | Depth in the unresolved aperture |
| Eclipse in B alone | 50 / 200 = 25% | A different physical interpretation |
A useful check asks the alternatives to predict different observations. Batalha and colleagues describe Kepler vetting tests involving alternating event depths, additional eclipses and changes in the measured light position. Those are examples of an evidence strategy, not an automatic classroom certification procedure.
For the blended-light case, ask where the missing light originates. Compare images during and outside the event. A positional change correlated with the dip can point towards another source, but its meaning depends on the image quality, uncertainty and neighbouring objects. An unexplained shift calls for investigation rather than a universal rejection rule.
Consider the event sequence too. If alternate dips have reliably different depths, folding all dips onto one assumed period may have hidden two kinds of eclipse. Inspect the individual events before accepting the combined curve.
Finally, ask whether another instrument or resolved observation can test the specific ambiguity. Additional evidence is valuable when it bears on an alternative; accumulating more copies of the same unresolved measurement may leave the ambiguity intact.
Start with provenance: target identifier, data source, observation dates, brightness units and any quality flags. Record the measurements you excluded and the rule used to exclude them. Keep the original data so the processing can be revisited.
Next, describe the event without naming its cause: approximate depth, timing, duration and the number of usable occurrences. Show individual events as well as a combined view. Report gaps where an expected event could not have been observed; a missing observation is different from a monitored interval with no dip.
Then build a short alternatives table. For each explanation, identify the evidence it fits, the evidence that challenges it and the unresolved question. A table full of “possible” entries is an honest intermediate result when the data cannot discriminate further.
End with the next observation you would request. In our invented case, resolving or localising the varying source is more directly useful than calculating several extra decimal places in the tentative radius. The best next step follows the uncertainty that controls the conclusion.
A smooth curve may be a product of averaging. Repeatedly combining data can reduce random noise while concealing unequal event shapes or an incorrect period. Keep the plot that would reveal the problem, not only the plot that looks most impressive.
Processing choices matter as well. Removing a slow trend can change a shallow event. Check whether plausible baseline choices preserve the feature. If your conclusion changes with a reasonable processing choice, report that sensitivity and investigate it.
Vetting also has limits of access. A school project may lack the imaging or spectroscopy needed to resolve the strongest alternative. That limitation changes the appropriate conclusion; it does not make the learning worthless.
Passing several checks does not establish atmosphere, surface conditions or life. A transit-sized object is not automatically Earth-like. Separate the property measured from the properties that remain unknown, and consult the original discovery study for its validation or confirmation criteria.
Transfer check: the neighbour brightens
Keep the loss of light from B at 50 units, but increase constant light from A so the total baseline becomes 20,000. Predict the observed depth before calculating. It becomes 50 ÷ 20,000 = 0.25%. The event in B has not changed; the fractional signal has been diluted further. A smaller measured dip need not mean a smaller eclipsing body.
Transfer check: more measurements, same ambiguity
Suppose a second observing run reproduces the 0.5% dip with smaller random uncertainty but still cannot separate A from B. What improved? Confidence in the recurring combined-light signal. What remains unresolved? Its source and physical interpretation. Better precision has not by itself selected the planetary model.
A practical evidence check
Give a partner the synthetic flux table, a short description of the aperture and two possible explanations. Ask them to write one observation that would distinguish those explanations. Then change the target from an unresolved pair to a genuinely isolated source. They should explain which concern disappears and which other uncertainties remain.
The progress test is the quality of that explanation. Memorising a list of false positives is less useful than matching a check to the particular ambiguity.
Does every transit require a radial-velocity mass measurement?
No. Different studies establish planetary status through different combinations of evidence, including statistical validation. Read the criteria used in the particular study rather than inventing one universal checklist.
Is a false positive a mistake by the observer?
Not necessarily. A real recurring signal can have a non-planetary cause. A candidate label preserves the distinction while the interpretation is tested.
Can a student confirm a planet from a plotted curve?
A student can analyse the curve and evaluate a claim. Formal confirmation requires the relevant evidence and scientific assessment. A defensible project may conclude that a candidate remains unresolved.
What should I do when two checks disagree?
Preserve the disagreement, review the assumptions and identify the observation that would resolve it. Do not select the check that gives the preferred answer.
Use the mathematical guide for light-curve calculations, the professional-practice guide for scientific reporting, and the subject root for the larger astronomy map. The learning goal is to make the route from measurement to claim visible enough for another person to challenge it.
The subject and its purpose
The Gold Standard of Astronomy
The Gold Standard of Education: learning that transfers into life
How X Works: the connected subject hub
Calculate and interpret
Exoplanet transits, light curves and orbital periods
Astronomers: reliable observations and honest discovery
Scientific detection bias in observations
Authoritative sources
Sources checked on 21 October 2026. Numerical cases and practice tasks are original teaching examples, not observations from a published experiment or official statistics.
NASA: Transit Method — introductory transit geometry.
Batalha et al.: Pre-Spectroscopic False Positive Elimination of Kepler Planet Candidates — a primary research account of Kepler data validation; the synthetic cases here are separate teaching examples.