Erosion Inquiry With Confounded Variables
At a suburban middle school, two seventh-grade teachers pilot a three-day inquiry on erosion control. Each class receives four identical trays of sieved loam, three watering cans calibrated to deliver 500 mL in 20 s, and a stopwatch. The research question is: 'Does a 2 cm bark-mulch cover reduce soil loss from a 15-degree slope?' One group mulches one tray but tilts it to 25 degrees because the higher angle 'looks more realistic.' A second group collects runoff for 30 s on the first trial and 2 min on the second, then compares total sediment mass. A third group dries sediment overnight before massing it, while another group masses it wet. At share-out, a student concludes that mulch 'definitely works' because the mulched tray lost less soil. The teachers must decide whether to grade the conclusion as written or intervene before the class adopts the claim.
1. Identify the controlled-variable and measurement problems in these student designs, and explain how each problem weakens the inference that mulch caused the difference in soil loss.
Read the worked answer
The core problem is that more than one variable differs between compared trays, so the cause of any difference is ambiguous.
- Slope angle: The mulched tray is at 25 degrees while the comparison tray is at 15 degrees. Steeper slopes generally increase runoff energy and sediment transport, so the mulch effect is confounded with slope. A difference could be due to mulch, slope, or both.
- Runoff duration: Collecting for 30 s versus 2 min changes total sediment mass but not necessarily the erosion rate. The groups are comparing different quantities unless they normalize mass per unit time or collect for the same interval.
- Wet versus dry mass: Water content adds variable mass. A wet sample can outweigh a dry one even if it contains less soil. The class needs an operational definition: dry to constant mass at a stated temperature, or measure sediment volume after settling.
- Replication: One tray per treatment cannot separate treatment effect from tray-to-tray variability.
A defensible inference requires holding slope, water volume, application rate, collection time, and drying protocol constant; randomizing tray assignment; and repeating trials. Otherwise the student's conclusion is an unsupported causal claim.
2. The student's claim that mulch 'definitely works' is based on one comparison. Using ideas about evidence and uncertainty, explain what would be needed before the class could accept that claim, and describe one instructional move that would make the uncertainty visible.
Read the worked answer
A single comparison cannot establish a reliable effect because it provides no estimate of variability. The class would need repeated trials for each treatment, ideally with trays assigned at random, and a consistent protocol. They should report a mean and a measure of spread, such as the range or standard deviation, and compare whether the difference between treatments is larger than the variability among replicates.
Instructionally, the teacher can have groups pool their data on a class graph with individual trial points or error bars. If the mulched and bare ranges overlap substantially, students can see that the evidence is inconclusive; if they separate, the claim is better supported. The teacher should also distinguish 'our data support the claim under these conditions' from 'mulch is proven to work everywhere.' That distinction targets the nature-of-science idea that scientific claims are provisional, bounded by the design, and strengthened by replication rather than by a single vivid result.
3. Design a revised procedure the teachers could use on day 2. Specify at least three variables held constant, the dependent variable and how it will be measured, and one way to reduce measurement error.
Read the worked answer
A revised procedure could use identical trays filled to the same depth with the same sieved loam, each compacted with the same mass or template. All trays are set at 15 degrees. Each receives 500 mL of water delivered from the same height at the same rate. Runoff is collected for a fixed 90 s after water first reaches the tray outlet.
Variables held constant include slope angle, soil type and depth, compaction, water volume and delivery rate, and collection duration. The independent variable is mulch treatment: 0 cm versus 2 cm of the same bark mulch. The dependent variable is dry sediment mass per tray, measured by filtering runoff, drying the residue to constant mass at a stated temperature, and massing it on a balance with a known precision.
To reduce measurement error, use the same balance and drying protocol for all samples, collect three replicates per treatment, and have one trained student mass all samples while another records. Report the mean and spread rather than a single value. This design makes the comparison attributable to mulch, within the stated conditions.