EnviroCoder logoEnviroCoder
Classroom project library

College intro · Soil & schoolyard

Testing Soil Management Claims: eDNA, statistical power, and the limits of a single sample

One composite soil sample, five target taxa, and a class-wide test of a real land-management hypothesis.

Students formulate a hypothesis about how a specific land-management practice (tillage, compaction, cover cropping) affects belowground biodiversity, collect a class-wide composite soil sample under a documented protocol, and use returned detection data to evaluate statistical power, sampling bias, and the limits of inference from a single sample — culminating in a formal lab report.

Start a Starter Kit quote

Grade band

College intro

Time required

4 lab sessions (150 min each)

Class size

16–24 students, one kit

Sample type

Soil, one vial

Standards and outcomes

  • Vision & Change Core Competency — Ability to apply the process of science: design an investigation with a clearly falsifiable hypothesis and identify its limitations
  • Vision & Change Core Competency — Ability to use quantitative reasoning to evaluate biological claims and communicate uncertainty
  • Vision & Change Core Concept — Systems: matter and energy flow through soil systems and are influenced by human management
  • Course Learning Outcome — Critically evaluate the statistical power and sampling design of a real or simulated dataset

How one kit serves a whole class

The $200 Classroom Starter Kit provides one soil collection vial, targets up to five species or taxonomic categories, and returns detections plus relative signal in 7–10 days via the class Field Journal — no counts. With only one vial for the entire lab section, the class treats the composite sample as a single statistical observation representing one management condition: soil cores are pulled from multiple points within one plot type and pooled, with every design decision documented for the shared lab report.

Choose one

Research tracks

Track A — Compaction and Detection Probability

Land-use pressure and statistical inference from small N

Driving question. Does a compacted, high-traffic soil plot show reduced detection of soil invertebrates compared to what published literature predicts for undisturbed soil?

Species to pick. Up to five soil invertebrate or microbial categories known to be compaction-sensitive versus compaction-tolerant, drawn from local extension or agricultural science sources.

What students take away. Students compare a single composite sample's results against a published baseline and discuss why N=1 limits the strength of any comparative claim.

Track B — Cover Crop and Decomposer Diversity

Nutrient cycling and management practice

Driving question. Is decomposer diversity detectable at levels consistent with active organic matter cycling under the site's current cover-crop or bare-soil management?

Species to pick. Up to five decomposer taxa (earthworm species, soil fungi, decomposer beetles, nematode categories) selected for known association with organic matter turnover.

What students take away. Students connect detection results to a specific, testable claim about nutrient cycling health drawn from the primary literature.

Track C — False Negatives in Cryptic Soil Taxa

Detection limits and Type II error in belowground sampling

Driving question. How does soil eDNA degradation and patchy distribution affect our ability to detect taxa known to be present from other survey methods?

Species to pick. Up to five taxa for which independent evidence (pitfall traps, visual survey, prior course data) already suggests presence, allowing detection results to be checked against a known ground truth.

What students take away. Students directly quantify a real or apparent false-negative rate by comparing eDNA results to independent survey evidence.

The lesson sequence

Step by step, with teacher notes

  1. 01

    Hypothesis formulation and literature grounding

    50 minutes

    Teams select a track and write formal, falsifiable hypotheses for five soil taxa, each grounded in a cited primary or extension-service source describing that taxon's known response to the management practice under study.

    Teacher notes

    • Open with: "A hypothesis without a citation is a guess — every H1 in this lab needs a source."
    • Require literal H0/H1 statements, not just predictions, and check that each is falsifiable by a null result.
    • Push students to be specific about the comparison being made (e.g., versus a published baseline, versus another known site, versus independent survey data) since there is no true experimental control plot with a second kit.
    • Facilitate the class-wide vote reconciling all team proposals into one shared five-taxon list, since only one vial is available.
    • Flag speculative taxa that are unlikely to have validated primers available and redirect toward well-documented categories.

    Materials

    • Primary literature or extension-service sources on soil taxa and management
    • Hypothesis worksheet with H0/H1 template
    • Whiteboard for class vote

    Student prompts

    • State H0 and H1 for each of the five candidate taxa, citing a specific source.
    • What management practice is currently applied at this site, and what specific effect does the literature predict?
    • Why is this comparison necessarily weaker than a true controlled experiment?
  2. 02

    Composite soil sampling protocol design and execution

    90 minutes

    With one vial for the section, students design a soil-coring scheme (grid, transect, or random-point) to represent the chosen plot fairly, then execute the coring, pooling, and sealing as one documented composite sample. A data-integrity subgroup tracks every design decision for the lab report.

    Teacher notes

    • Frame the coring scheme itself as the experiment's key design variable: "The map of where you core is the design of this whole study — defend it like you would defend a hypothesis."
    • Require gloves before any soil or equipment contact, and a clean coring tool between each core to reduce cross-contamination.
    • Have the class vote on and justify a specific number and spacing of soil cores pooled into the composite, tied explicitly to plot size and expected patchiness.
    • Assign the data-integrity subgroup to log GPS points, depth, and any visible surface cover (leaf litter, bare soil, vegetation type) for every core.
    • Model the pooling and mixing step yourself before the chain-of-custody subgroup seals and labels the sample for shipping.

    Materials

    • 1 classroom soil eDNA collection vial and prepaid mailer
    • Nitrile gloves (one pair per student minimum)
    • Sterile soil corer or trowel, cleaned between cores
    • GPS-enabled device or app, soil thermometer
    • Chain-of-custody log sheet

    Student prompts

    • Justify the number, spacing, and depth of soil cores pooled into the composite sample.
    • Record GPS location, depth, and surface cover for every core contributing to the composite.
    • What single design choice, if changed, would most improve the representativeness of this sample?
  3. 03

    Molecular methods, statistical power, and error types

    50 minutes

    During the turnaround, students study the soil eDNA extraction and amplification pipeline and connect it to statistical power: why a single composite sample has limited power to detect real effects, and how false positives and false negatives specifically arise in soil systems (PCR inhibitors from humic acids, patchy microbial distribution).

    Teacher notes

    • Use this window as a full content block: assign a short reading on PCR inhibition by soil humic substances in advance.
    • Diagram the pipeline (extraction → inhibitor removal → PCR amplification → sequencing → reference matching) and highlight where soil samples specifically risk failure compared to water samples.
    • Say: "Low statistical power doesn't mean bad science — it means you must be honest about what conclusions this single sample can and cannot support."
    • Explicitly connect false positive to Type I error (e.g., contamination or primer cross-reactivity) and false negative to Type II error (e.g., inhibitor suppression, patchy taxon distribution, degradation).
    • Assign lab report introduction and methods sections due before the results reveal to avoid outcome bias in the writing.

    Materials

    • Reading on PCR inhibitors in soil eDNA extraction
    • Diagram of soil eDNA lab pipeline
    • Vocabulary sheet: statistical power, Type I/II error, sensitivity, specificity

    Student prompts

    • Explain how humic acids or other soil inhibitors could produce a false negative.
    • Why does a single composite sample have lower statistical power than five independent samples?
    • Draft your lab report's introduction and methods sections before seeing the results.
  4. 04

    Results reveal and statistical evaluation

    50 minutes

    The class Field Journal is revealed together. Students evaluate each hypothesis, compare results against the cited literature baseline or independent survey evidence, and complete lab reports that explicitly discuss the statistical limitations of an N=1 composite design.

    Teacher notes

    • Reveal results live; have students mark reject/fail-to-reject for each of their five hypotheses independently before group discussion.
    • Correct any student framing a non-detection as definitive absence — reinforce that this reflects statistical power, not biological truth.
    • Ask: "If we had a second kit, which single design change would most increase our confidence in this result?"
    • Require the discussion section to name a specific, quantified (or semi-quantified) limitation tied to the composite sample design.
    • Collect draft reports for structured peer review using the rubric as the review checklist before final submission.

    Materials

    • Class Field Journal report (projected)
    • Lab report template
    • Peer-review rubric copies

    Student prompts

    • For each taxon, state whether you reject or fail to reject H0, and why.
    • Compare your result to the cited baseline or independent survey evidence — do they agree or conflict?
    • What redesign would most increase statistical power if repeated with a second kit?

Student handout

For every student

Thinking prompts

  • Write formal null and alternative hypotheses, with citations, for each of the five target taxa.
  • Justify the soil-coring scheme (number, spacing, depth) in terms of representativeness of the plot.
  • Explain how PCR inhibitors in soil can produce a false negative result.
  • Describe why a single composite sample has lower statistical power than replicated independent samples.
  • State, for each taxon, whether the post-reveal result supports rejecting or failing to reject H0.
  • Compare your eDNA results to at least one independent line of evidence (literature baseline or prior survey data).

Data sheet

  • Date, time, and personnel for each soil-coring event:
  • GPS coordinates and depth for each soil core:
  • Surface cover, soil moisture, and recent management activity (mowing, tillage, irrigation) at time of collection:
  • Number of cores pooled and total composite sample volume:
  • Pre-reveal hypothesis table (H0/H1 for all five taxa, with citations):
  • Chain-of-custody log with preservation and shipping timestamp:
  • Post-reveal detection and relative signal table for all five taxa:

Discussion questions

  • Why does soil, compared to water, pose additional risks for false negatives in eDNA testing?
  • How would you redesign this study with a second kit to directly compare two management plots rather than one plot against a literature baseline?
  • What are the limits of generalizing from one composite sample to a claim about the whole site's soil health?
  • How does 'relative signal strength, not counts' shape what your lab report can responsibly claim about decomposer activity?

Field checklist

  • Null and alternative hypotheses, with citations, written for all five taxa before sampling.
  • Class-wide agreement reached on coring scheme (number, spacing, depth) before fieldwork begins.
  • Nitrile gloves worn and coring tool cleaned between every core.
  • GPS, depth, and surface cover logged for every core pooled into the composite.
  • Chain-of-custody log completed and verified before shipping.
  • Lab report introduction and methods drafted before results were revealed.

Analysis

Statistical power and error analysis of a single composite soil eDNA sample

  1. 1.For each of the five taxa, restate H0 and H1 and classify the post-reveal outcome as 'reject H0' or 'fail to reject H0.'
  2. 2.Compare each result against the cited literature baseline or independent survey evidence, noting agreement or conflict.
  3. 3.Identify any result plausibly affected by a false positive (contamination, primer cross-reactivity) or false negative (PCR inhibition, patchy distribution, degradation).
  4. 4.Discuss, qualitatively, how the single-composite design limits statistical power relative to a replicated-plot design.
  5. 5.Propose one specific redesign (coring scheme, taxon list, or comparison plot) that would increase statistical power with a second kit.

Assessment

20-point rubric

Hypothesis formulation with literature grounding

4 pts

Falsifiable, cited H0/H1 pairs are written for all five taxa prior to results.

Sampling design and field protocol

4 pts

Coring scheme is justified statistically and executed with documented contamination control.

Molecular methods and error concepts

4 pts

PCR inhibition, false positives/negatives, and statistical power are explained accurately.

Data interpretation

4 pts

Detection and relative signal results are correctly linked back to hypotheses and independent evidence.

Lab report quality and discussion of limitations

4 pts

Report follows standard scientific format and explicitly discusses statistical power limitations.

Go further

Extensions

  • Pool Field Journal data with a second section sampling a contrasting plot to build a two-condition comparison and run an appropriate statistical test.
  • Have students write a mock grant proposal requesting a second kit to sample a true control plot, justifying the design in terms of statistical power.
  • Invite a soil scientist or agricultural extension agent to review the class's management-practice conclusions.
  • Compare eDNA detection results to a traditional soil survey method (Berlese funnel extraction, visual sorting) conducted the same week.

For the teacher

Answer key and misconceptions

  • Expected: student H0/H1 pairs should cite a specific management-effect claim from the literature; a common error is writing an untestable, vague prediction — correct this explicitly.
  • Common misconception: 'we didn't detect it, so it's absent from the soil.' Correct framing: a single composite sample has limited statistical power, so failing to reject H0 reflects detection limits, not proof of absence.
  • Common misconception: soil eDNA works exactly like water eDNA. Correct framing: soil samples face additional inhibitor-related false-negative risk from humic acids and other compounds that can suppress PCR amplification.
  • Expected: strong lab reports explicitly quantify or describe the limitation of N=1 composite sampling and propose a specific higher-power redesign.
  • Common misconception: relative signal strength reflects total decomposer biomass. Correct framing: it is a proxy for shed/environmental DNA quantity, confounded by soil chemistry, moisture, and patchy microbial distribution.

Ready to run it with your class?

Design your Starter Kit Back to the library