The Fertiliser Drops Experiment: Is Marcia's Conclusion Safe?
Marcia is a Form 4 student who lives beside a pond next to a vegetable farm. She has read that fertiliser washed off fields can make pond water turn green, and she decides to test it. She fills five identical glass jars with 400 mL of pond water taken from the same spot and adds liquid fertiliser: 0 drops to Jar A, 1 drop to B, 2 drops to C, 3 drops to D and 4 drops to E. She labels each jar and stands all five on the same windowsill. On Day 4 a teacher needs that bench and moves Jars D and E to a shelf two metres from the window; Marcia leaves them there. After 10 days she lowers a white disc into each jar and records the depth at which it disappears from view: A 28 cm, B 21 cm, C 15 cm, D 9 cm, E 5 cm. She reads each jar only once and writes in her report, 'The more fertiliser, the more algae - my hypothesis is proved.'
1. Identify the independent variable, the dependent variable and three variables Marcia should have kept constant. Then explain how the move of Jars D and E affects what she can conclude.
Read the worked answer
Independent variable: the number of drops of fertiliser added (0 to 4). Dependent variable: the amount of algal growth, measured as the depth at which the white disc disappears - a smaller depth means cloudier water, so more algae.
Constants she should have kept the same: volume of pond water (400 mL), jar size, type and cleanliness of glass, source of the water and its starting population of algae, light received, temperature, and the time allowed for growth (10 days).
The move: light is a raw material for photosynthesis and strongly affects how fast algae grow. Jars D and E had less light for 6 of the 10 days, so on light alone we would expect them to be less cloudy than the others. In fact they were the cloudiest of all. The light difference therefore works against her hypothesis rather than creating it, so the fertiliser trend is probably real. But it still weakens her claim: those two jars were treated differently in a second way at the same time, so the difference between D and E and the rest is confounded. Jar E's single reading should be reported with a note about the move, not quietly deleted, and the whole experiment should be repeated with all jars under the same light.
2. Marcia says her hypothesis is 'proved'. Give two reasons why 'supported by the evidence' is more accurate, and describe one change that would improve reliability and one that would improve validity.
Read the worked answer
Why 'proved' is too strong:
- She used one jar per concentration, with a single reading each. With no repeats she cannot tell whether the pattern comes from the fertiliser or from chance differences between jars - one jar may simply have had more algae in it to start with.
- One experiment, in one pond, under one set of conditions cannot prove a general rule. It supports a hypothesis but does not rule out other explanations, such as the fertiliser colouring the water itself or feeding bacteria as well as algae.
Improving reliability: set up three jars at each concentration (15 jars in total) and/or take three readings from each jar and average them. This reduces the effect of random error such as a stray bubble or a tilted disc.
Improving validity: make sure the disc depth is really measuring algae. Examine a drop under a microscope, or filter a known volume and weigh or observe the residue. If the cloudiness is mostly silt or the yellow-brown colour of the fertiliser, the method is not measuring what she claims, and the conclusion fails even if the numbers look tidy.
3. The farmer wants to know how much fertiliser is safe to apply. Describe a follow-up investigation Marcia could design, naming the independent variable, a controlled variable, and the kind of result that would let the farmer make a decision.
Read the worked answer
A dose–response investigation that copies the farmer's real situation more closely.
Independent variable: fertiliser application rate, for example 0, 5, 10, 20, 40 and 80 kg per hectare applied to six similar plots of the same crop and soil beside the pond.
Controlled variables: plot size, soil type, slope, crop, planting date, rainfall or irrigation, and distance from the pond - or, in a simpler version, six beakers of pond water with a wide range of fertiliser concentrations (0, 1, 2, 4, 8, 16 drops) each set up in triplicate.
Measurements: crop yield from each plot and algal growth or nitrate concentration in water draining from it, sampled after each rain.
What would let the farmer decide: a graph of algal growth (or nitrate in runoff) against application rate. If the curve is flat at low rates and then rises steeply, the 'safe' rate is below the point where the rise begins - say 10 kg per hectare. If crop yield keeps rising up to 40 kg per hectare while algal growth only rises above 20, the farmer faces a real trade-off between yield and pond damage, and the decision must be stated as a trade-off rather than as a single correct number.