The Fertilizer Trial That Changed Two Things at Once

The comparison is confounded. Light and temperature differ between the south windowsill and the north shelf near a door, and fertilizer is only one of at least three variables that changed. Because the fertilized plants also had more light, the 4.6 cm difference cannot be assigned to fertilizer. Interestingly, low light usually causes stretched, spindly growth, so the north-shelf plants were probably handicapped; that makes the direction of the confound uncertain rather than automatically favoring the fertilizer, which is the honest point to make. The yellow, curled tips point to fertilizer injury - salt accumulation or nutrient burn - so the fertilizer may have helped at this concentration or harmed some plants, and the class cannot tell which. "More is better" is therefore unsupported twice over: the design cannot isolate the cause, and the visible symptoms suggest the treatment was not uniformly beneficial. The conclusion slide must be rewritten as a limitation, not a finding.

Repot all 16 seedlings in identical pots with the new bag of potting mix and place them together on the same shelf, rotating pot positions every other day to even out light. Randomly assign eight pots to fertilizer and eight to plain water - random assignment, not table groups, so prior plant size and shelf position cannot sort systematically. Give all pots the same volume on the same schedule, and measure the light meter reading at the start and end to document that light is comparable. Measure height from the soil line to the growing tip, same person, same ruler, same time of day, with the fertilizer label concealed so the measurer is not biased. One week is short, so the class should report growth since repotting, not total height, and state that as a limitation. The variable is fertilizer concentration; the control is the water-only group under otherwise identical conditions.

Present the original data as a preliminary observation with an identified confound, plus the revised design and its predictions. A defensible poster states: fertilized plants averaged 14.2 cm and water-only plants 9.6 cm, but light and temperature differed, so no causal claim is made; two fertilized plants showed yellow tips consistent with nutrient burn; the revised trial randomizes pots in one location and is in progress. This is stronger than the original conclusion because it is actually supported by the evidence, it demonstrates the practice of identifying variables and limitations, and it gives the audience something to evaluate. The teacher should be explicit that a well-argued "we cannot conclude yet" scores higher than a confident wrong answer, and should have students write the specific prediction - that if fertilizer is responsible, the fertilized group will outgrow the control even with light held constant - so the revised design is testable rather than decorative.