College intro · Water & wetlands
Detection Probability in Practice: an eDNA hypothesis test on a local waterway
A single vial becomes a semester-defining exercise in hypothesis testing and statistical inference.
Students formulate falsifiable null and alternative hypotheses about species presence in a local waterway, design a composite sampling protocol for the whole lab section, and interpret returned detection data through the lens of detection probability, primer specificity, and Type I/II error — producing a full lab report.
Grade band
College intro
Time required
4 lab sessions (150 min each)
Class size
16–24 students, one kit
Sample type
Water, one vial
Standards and outcomes
- Vision & Change Core Competency — Ability to apply the process of science: formulate and test hypotheses using appropriate quantitative reasoning
- Vision & Change Core Competency — Ability to use quantitative reasoning to interpret biological data and estimate uncertainty
- Vision & Change Core Concept — Systems: biological systems are interconnected and interact with each other and their environment
- Course Learning Outcome — Design a controlled or field-based investigation, identify sources of error, and communicate results in standard scientific report format
How one kit serves a whole class
The $200 Classroom Starter Kit holds exactly one collection vial and targets up to five species; results return in 7–10 days as detections plus relative signal strength (not counts) in the class Field Journal. Because only one vial exists for the whole lab section, the section functions as a single distributed research team: subgroups collect sub-samples from different points along the waterway and pool them into one composite sample, and every subgroup's raw field data feeds the same statistical analysis and shared lab report.
Choose one
Research tracks
Track A — Null Hypothesis Boot Camp
Formal hypothesis structure applied to presence/absence data
Driving question. Can we detect genetic material from species predicted, on habitat grounds, to be present at this site?
Species to pick. Up to five species chosen to span a gradient of expected detection probability: one near-certain resident, two plausible-but-unconfirmed species, and two long-shot or historically extirpated species.
What students take away. Students learn to write true null/alternative hypothesis pairs for presence data and to distinguish 'failed to reject H0' from 'proved absence.'
Track B — Primer Specificity and False Positives
Molecular methods and error sources
Driving question. How confident can we be that a detection reflects the target species and not a closely related one?
Species to pick. Up to five species chosen so that at least two pairs are taxonomically close relatives (e.g., two congeneric fish, or a native and a related non-native), forcing discussion of primer cross-reactivity.
What students take away. Students evaluate a detection report critically, distinguishing a confirmed hit from a plausible cross-amplification artifact.
Track C — Detection Probability Across a Gradient
Statistical inference under imperfect detection
Driving question. How does distance from a known population center change the probability of a positive detection?
Species to pick. Up to five species with a documented population center upstream or downstream of the sampling site, so the composite sample can be interpreted as a single point on a probability-of-detection gradient.
What students take away. Students connect a single sampling event to the broader statistical concept that detection probability is not fixed — it depends on distance, effort, and environmental degradation of DNA.
The lesson sequence
Step by step, with teacher notes
- 01
Formulating testable hypotheses
50 minutesWorking in the five-species framework of their assigned track, student teams write formal null and alternative hypotheses for each species before any sample is collected. Each hypothesis must be falsifiable and tied to a specific, cited habitat or distributional rationale.
Teacher notes
- Open with: "In this lab, 'no detection' is a result, not a failure — write your hypotheses so that is actually true."
- Require the literal H0/H1 statement in symbolic or plain-language form for every one of the five species, not just a prediction sentence.
- Push back hard on any hypothesis that cannot be falsified by a negative detection — send students back to revise before proceeding.
- Have each team justify at least one citation (peer-reviewed paper, state agency range map, or museum record) per species.
- Remind students the section shares one vial: species selection is a single class-wide vote, so competing hypotheses must be reconciled into one shared list of five.
Materials
- Peer-reviewed range/distribution sources or agency database access
- Hypothesis worksheet with H0/H1 template
- Whiteboard for class-wide species vote
Student prompts
- State H0 and H1 for each candidate species in falsifiable form.
- What published evidence supports your predicted probability of detection?
- If your hypothesis is wrong, what result would you expect to see instead?
- Why must the whole lab section agree on one shared species list before sampling?
- 02
Designing and executing the composite sampling protocol
90 minutesBecause only one vial exists for the entire section, students design a sampling protocol as a statistical instrument: where along the waterway should sub-samples be drawn to make the composite representative, and how should that decision be documented so a future class can replicate it? Subgroups execute the plan under contamination-control discipline.
Teacher notes
- Frame this explicitly as experimental design, not just fieldwork: "You are one N of one — the sampling design is the only thing standing between this data and pure guesswork."
- Have the class debate and vote on a sub-sampling scheme (e.g., systematic transect vs. targeted micro-habitat sampling) and require a written rationale for the chosen method before anyone enters the water.
- Enforce nitrile gloves before any equipment contact, and sampling upstream of foot traffic, as non-negotiable protocol steps.
- Assign a data-integrity subgroup whose only job is to log GPS points, timestamps, and environmental covariates (temperature, flow, turbidity) for every sub-sample pooled into the composite.
- Model the pooling and sealing step once yourself, then have the data-integrity subgroup verify chain-of-custody documentation before the vial is shipped.
Materials
- 1 classroom eDNA collection vial and prepaid mailer
- Nitrile gloves (one pair per student minimum)
- Sterile sub-sample containers
- GPS-enabled device or app, thermometer, turbidity tube
- Chain-of-custody log sheet
Student prompts
- Justify the sub-sampling scheme in terms of statistical representativeness, not convenience.
- Record GPS coordinates, timestamp, and environmental covariates for every sub-sample pooled.
- What is the single greatest contamination risk in this protocol, and how did the team control for it?
- 03
Molecular methods, error types, and primer specificity
50 minutesDuring the 7–10 day turnaround, students study the amplification and sequencing pipeline in depth, with emphasis on statistical error framing: what a false positive and false negative mean in this context, and how primer specificity determines whether a detection can be trusted.
Teacher notes
- Use the wait window as a full lecture/discussion block, not downtime — assign a primer-design or BLAST-search reading in advance.
- Explicitly map 'false positive' to Type I error and 'false negative' to Type II error on the board, and connect each to a concrete lab decision (e.g., primer specificity reduces Type I error; composite sampling reduces Type II error).
- Say: "A detection is a claim about probability, not a certainty — your job is to state how confident that claim is and why."
- Walk through a real or redacted example of primer cross-reactivity between two related species so students see what a borderline call looks like.
- Assign the lab report introduction and methods sections due at the start of the next class, before results are revealed, so the writing is not contaminated by outcome bias.
Materials
- Diagram of eDNA lab pipeline (extraction, PCR, sequencing, reference matching)
- Primer specificity / BLAST search example handout
- Vocabulary sheet: Type I/II error, sensitivity, specificity, false positive/negative
Student prompts
- Define false positive and false negative in the specific context of this survey.
- How does primer specificity reduce the risk of a false positive between related species?
- Write the introduction and methods sections of your lab report before seeing the results.
- 04
Results reveal, statistical interpretation, and report finalization
50 minutesThe class Field Journal is revealed together. Students test each hypothesis against the detection data, discuss relative signal strength as an imperfect proxy for abundance, and complete their lab reports with a discussion section addressing sources of error and statistical limitations.
Teacher notes
- Reveal live; ask students to independently mark each of their five hypotheses as 'reject H0' or 'fail to reject H0' before any group discussion.
- Immediately correct any student who frames a non-detection as 'proof the species is absent' — reinforce fail-to-reject language.
- Ask: "Which of our five results is most likely to be a false positive or false negative, and what evidence would resolve the ambiguity?"
- Require the discussion section of the lab report to explicitly name at least two sources of error: one biological (e.g., DNA degradation, seasonal shedding) and one methodological (e.g., single composite sample, primer specificity).
- Collect draft lab reports for peer review before final submission, using the rubric as the peer-review checklist.
Materials
- Class Field Journal report (projected)
- Lab report template
- Peer-review rubric copies
Student prompts
- For each species, state whether you reject or fail to reject H0, and why.
- Identify the result most vulnerable to false positive or false negative error, and explain why.
- What would increase statistical confidence in this result if the study were repeated with a second kit?
Student handout
For every student
Thinking prompts
- Write formal null and alternative hypotheses for each of the five target species before results are revealed.
- Justify the sub-sampling and composite-pooling scheme in terms of statistical representativeness.
- Explain the difference between a false positive and a false negative in this study's context.
- Describe how primer specificity affects your confidence in any single detection.
- State, for each species, whether the post-reveal result supports rejecting or failing to reject H0.
- Identify one biological and one methodological source of error that limits this study's conclusions.
Data sheet
- Date, time, and personnel for each sub-sampling event:
- GPS coordinates or precise site description for each sub-sample point:
- Water temperature, turbidity, flow rate, and weather at time of collection:
- Sub-sample volumes and number of pooling points contributing to the composite:
- Pre-reveal hypothesis table (H0/H1 for all five species):
- Chain-of-custody log with preservation and shipping timestamp:
- Post-reveal detection and relative signal table for all five species:
Discussion questions
- Why does a single composite sample limit the statistical power of this study compared to five independent samples?
- How would you redesign the sampling scheme with a second kit to directly test detection probability across distance from a known population?
- What ethical or resource-allocation tradeoffs exist in a $200 kit that forces a class to pick only five species?
- How does 'relative signal strength, not counts' change what kinds of quantitative claims your lab report can responsibly make?
Field checklist
- Null and alternative hypotheses written and reviewed for all five species before sampling.
- Class-wide agreement reached on sub-sampling scheme before fieldwork begins.
- Nitrile gloves worn and equipment kept upstream of foot traffic throughout collection.
- GPS, timestamp, and environmental covariates logged for every sub-sample pooled into the composite.
- Chain-of-custody log completed and verified before shipping.
- Lab report introduction and methods drafted before results were revealed.
Analysis
Hypothesis testing and error analysis of composite eDNA detection data
- 1.For each of the five species, restate H0 and H1 and classify the post-reveal outcome as 'reject H0' or 'fail to reject H0.'
- 2.Report the relative signal strength for each detection and explain what it can and cannot indicate about abundance.
- 3.Identify any result plausibly affected by a false positive (primer cross-reactivity) or false negative (degradation, low shedding rate, dilution in the composite).
- 4.Quantify, qualitatively or semi-quantitatively, how the single-vial/composite-sample design limits statistical confidence relative to a replicated-sample design.
- 5.Propose one specific redesign (sampling scheme, species list, or timing) that would increase statistical power if a second kit were available.
Assessment
20-point rubric
Hypothesis formulation
4 ptsFalsifiable, well-cited H0/H1 pairs are written for all five species prior to results.
Experimental design and field protocol
4 ptsSub-sampling scheme is justified statistically and executed with documented contamination control.
Molecular methods and error concepts
4 ptsFalse positive/negative and primer specificity are explained accurately and applied to this study's data.
Data interpretation
4 ptsDetection and relative signal results are correctly linked back to the stated hypotheses.
Lab report quality and discussion of limitations
4 ptsReport follows standard scientific format and names specific biological and methodological sources of error.
Go further
Extensions
- Pool Field Journal data across multiple sections sampling the same waterway to build a semester-long detection dataset and run a chi-square test on presence/absence across sites.
- Have students write a mock grant proposal requesting a second kit, justifying the added species or sites in terms of statistical power.
- Invite a molecular ecology researcher to critique the class's primer-specificity discussion and lab report conclusions.
- Compare the class's eDNA detection results to a traditional survey method (netting, visual transect) conducted the same week, and discuss method agreement and disagreement.
For the teacher
Answer key and misconceptions
- Expected: student H0/H1 pairs should be symmetric and falsifiable; a common error is writing H1 as a certainty statement rather than a probabilistic claim — correct this explicitly.
- Common misconception: 'we didn't detect it, so it's not there.' Correct framing: failing to reject H0 reflects the detection limits of a single composite sample, not proof of absence.
- Common misconception: any positive detection is fully trustworthy. Correct framing: primer specificity and reference database quality both affect the probability that a detection is a true positive, especially among close relatives.
- Expected: strong lab reports name at least one biological source of error (DNA degradation rate, seasonal shedding, dilution effects of pooling) and one methodological source (single vial, limited species slots, subjective sub-sampling scheme).
- Common misconception: relative signal strength is a population estimate. Correct framing: it is a relative proxy for shed DNA quantity, confounded by animal size, recency of activity, and number of individuals.