Lesson 6 of 6 · evaluating a possible discovery responsibly
Build a discovery claim from tests that can fail
10 minute lesson · 3 checks along the way
Your goal for this lessonUse controls, base rates and independent predictions to communicate a possible discovery without converting a test statistic into certainty.
Follow a linked name to explore its archive card. “Back to your lesson” brings you to the same part of the reading.
Words with a dotted underline open a short definition right here in the lesson.
State the claim before testing its confidence
A possible discovery contains several distinct propositions: the detector recorded a feature, the feature came from the target, a proposed substance or process explains it, and a biological origin fits better than alternatives. Each proposition requires evidence. Mars 2020 Perseverance provides a useful reminder: confidence in a chemical map does not by itself decide the origin of the chemistry represented in that map.
Begin with a clear measurement description, including uncertainty and calibration, then identify what each interpretation adds. Preserve the sample history or observing conditions. State which alternatives were tested, which remain feasible, and which cannot yet be evaluated. A phrase such as should be accompanied by those specifics so that readers understand what makes the result promising and what prevents a stronger conclusion.
The same July 2024 Cheyava Falls observation can support different levels of claim. Its visible textures are measured features; mineral identity and formation history need additional analyses. The September 2025 interpretation remained a potential biosignature.
Look closely
Separate one directly visible observation from one origin claim that this frame alone cannot establish. What follow-up would challenge the stronger claim?
See what to notice
Pale patches with dark margins and adjacent white veins are visible. A biological cause is not directly visible; spatial chemistry, age relationships, contamination controls and realistic abiotic experiments can test that interpretation.
Check 1 of 3
10 pointsUse the reading above. You can try again without losing points.
Work through a decision
Worked screening: the starting population matters
Consider 10,000 simulated cases with known labels: 100 biological and 9,900 . This is an invented validation exercise, not an estimate of how often alien life exists. A fixed screening test detects 90 percent of biological cases and incorrectly flags one percent of nonbiological cases. It therefore yields 90 true positives and 99 false positives, for 189 positive results in total.
Among the positives, 90 divided by 189, about 47.6 percent, are biological in this simulation. High did not make every positive a discovery, because the much larger nonbiological group supplied many false alarms. Change the starting mix, detection efficiency, or and the fraction changes. Real life searches lack a complete labeled extraterrestrial training population, so these toy rates must not be silently treated as measured cosmic probabilities.
Count both kinds of positive result
Original instructional confusion matrix for the lesson’s invented validation population. Sensitivity is 90 percent and the abiotic false-positive rate is 1 percent. Counts and denominators are explicit; the starting one-percent biological fraction is not a measured prevalence of alien life.
Open full-size diagram in a new tab: Count both kinds of positive resultFollow up with a different vulnerability in mind
Repeating a measurement can reveal random fluctuations, but a shared contaminant or calibration mistake can repeat too. Strong follow-up targets a vulnerability: a separate preparation tests , another instrument tests a detector artifact, and an additional predicted chemical relationship tests the origin model. Independence is a physical and statistical claim to justify, not a reward earned merely by increasing the number of observations or authors.
If two tests share a , multiplying their quoted probabilities as though independent exaggerates the evidence. Likewise, searching many features and reporting only the most unusual one can make a chance fluctuation appear more surprising than it is. Recording planned comparisons and reporting the search scope helps others assess that problem. Unexpected patterns can still motivate discovery, but their follow-up should distinguish exploration from an independently tested prediction.
Check 2 of 3
10 pointsUse the reading above. You can try again without losing points.
Communicate what would change the conclusion
NASA’s Ladder of Life Detection and discussions of a confidence scale offer ways to organize measurements and communication. They are frameworks for reasoning, not universally binding certificates or automatic numerical probabilities. Their usefulness depends on matching tests to an environment and an explicit claim. A ranked label cannot replace explaining why an instrument responds to the target, how contamination was constrained, or how a nonbiological model was evaluated.
A responsible report gives the dated result, accessible evidence, alternative explanations, and a next observation with outcomes that could strengthen or weaken the interpretation. It can distinguish an exciting candidate from a confirmed explanation without minimizing the discovery’s scientific value. Revising the conclusion after new data is part of the method. The strongest claim is the one whose supporting steps are visible enough for others to challenge, reproduce, and improve.
Check 3 of 3
10 pointsUse the reading above. You can try again without losing points.
Put it together
Your lesson progress
Each check connects to a part of the story. Complete all 3 to earn your 25-point lesson bonus.
Go to the source
Explore the evidence behind this lesson. These optional readings and videos open in a new tab.