Fertilizer Trial with Confounded Conditions
Four students test whether a new liquid fertilizer increases bean plant height. They place six plants on the south windowsill and give them 200 mL of water with fertilizer daily. They place six plants on the north windowsill and give them 100 mL of water without fertilizer every other day. All plants use the same soil and pot size. After three weeks the fertilized plants measure 31, 17, 16, 15, 14, and 16 cm; the unfertilized plants average 12.4 cm. The group concludes the fertilizer caused a 5.8 cm increase. One student objects that the north shelf was moved to a hallway for two days during the second week because of a window repair. Another points to the 31 cm plant as proof.
1. Identify two variables confounded with the fertilizer treatment and explain, for each, a mechanism by which it could produce the difference without the fertilizer doing anything.
Read the worked answer
Two confounds stand out. 1. Light and location: the fertilized group sits on the south windowsill; the unfertilized group sits on the north windowsill, and the north shelf was moved to a hallway for two days. More light means more photosynthesis, so position alone could produce taller plants. 2. Water: 200 mL daily versus 100 mL every other day is a large difference in water, and water affects growth directly. Either variable could produce several centimeters of difference with the fertilizer doing nothing. Because the variables change together with the treatment, the comparison cannot separate their effects.
2. The 31 cm plant is in the fertilized group. A teammate says it 'proves' the fertilizer works. Using the group's numbers, evaluate that claim and say what a better summary of the fertilized group would show.
Read the worked answer
The 31 cm plant is one measurement, and it is far from the rest of its group (14–17 cm). It raises the fertilized mean from 15.6 cm to 18.2 cm, so most of the reported 5.8 cm gap depends on it. The median, 16 cm, is barely moved by it and is a fairer summary of a typical plant. Even the median shows a gap of about 3.6 cm over the comparison mean, so something may be happening - but one outlier cannot prove causation, especially with light and water confounded. A better report would give the median, the range, and the sample size.
3. Describe the minimum change to the design that would let the class make a fair test, and state one measurement they should add to check their conclusion.
Read the worked answer
Make location and watering identical: all twelve plants on the same shelf, same pot size and soil, watered on the same schedule with the same volume. Assign plants to fertilizer or no fertilizer at random, so no other trait is systematically tied to a group. The only difference should be the fertilizer itself.
Add a measurement that checks the remaining variables - for example, initial plant height at the start, soil moisture, or light intensity at each pot. Initial height matters most: if the fertilized group started taller, the final difference proves little. With one variable changed and starting conditions recorded, a difference in growth can be attributed to the fertilizer.