Fertilizer, Windowsills, and a Confounded Experiment
A student tests whether a new liquid fertilizer increases bean plant growth. She fills 10 identical small pots with the same soil and plants one bean seed in each. She places five pots on a sunny windowsill and waters them with fertilizer solution. She places the other five pots on a shaded shelf and waters them with plain water. All pots receive the same amount of liquid each day. After three weeks, the plants on the sunny windowsill are taller. The student concludes that the fertilizer caused the extra growth and writes that the experiment proves the fertilizer works.
1. Identify the variables that differ between the two groups and explain why the conclusion is not supported.
Read the worked answer
At least two variables differ: light exposure (sunny windowsill versus shaded shelf) and water type (fertilizer solution versus plain water). Because both changed at the same time, the experiment cannot separate the effect of fertilizer from the effect of light. The taller plants could be taller because they received more light, because they received fertilizer, or because of both. The conclusion that fertilizer caused the growth is therefore not supported. This is a confounding-variable problem: when two factors change together, the outcome cannot be attributed to one of them. The student's claim that the experiment "proves" the fertilizer works is too strong even if the plants did grow, because proof requires ruling out alternative explanations.
2. Describe a revised design that would test the fertilizer more fairly.
Read the worked answer
A fair design would keep light constant and vary only the fertilizer. For example, place all 10 pots on the same sunny windowsill. Water five pots with plain water and five with the same amount of fertilizer solution. Keep pot size, soil, seed type, watering volume, and temperature the same. If the student also wants to test light, she would need a second experiment with a 2 × 2 design: fertilizer versus no fertilizer, crossed with sunny versus shaded, with several plants in each condition. The key principle is to change one variable at a time, or to use a design that allows each variable's effect to be separated. Randomly assigning pots to conditions also helps reduce accidental bias.
3. Suppose the revised experiment shows the fertilized plants are taller. What additional evidence would make the conclusion stronger, and what should the student avoid claiming?
Read the worked answer
Additional evidence would include more plants per group (to reduce the chance that one unusually tall plant drives the result), repeated trials, measurements taken the same way for every plant, and a report of average height with the variation between plants. The student should also check that the fertilizer did not harm other plant features, such as leaf color or root health. She should avoid claiming that the fertilizer is "best" or "safe for all plants," because the experiment tested one type of plant, one fertilizer dose, and one growing period. A stronger conclusion is limited: "Under these conditions, the fertilizer solution was associated with greater average height in bean plants." That wording matches the evidence without overgeneralizing.