6.2 Experimental Design, Variables, and Controls

The Core Design Question

Life Science: Biology does not treat laboratory reasoning as a separate end-of-year unit. NYSED's current exam design draws on science and engineering practices across clusters, so experimental design can appear in genetics, homeostasis, ecology, evolution, enzymes, or human impact. A cluster may describe a student investigation, a field study, a simulation, or an engineering test. Your job is to decide whether the data can support the proposed claim.

The first question is always: What comparison is being made? If the comparison is unclear, the conclusion is usually weak. A fair experiment changes one main factor, measures a response, and keeps other important conditions the same.

Variables and Controls

A control group is not the same thing as a controlled variable. In a fertilizer experiment, plants with no fertilizer may be the control group. Light, water, soil type, pot size, plant species, starting height, and temperature are controlled variables. This distinction appears constantly in Regents-style questions.

Worked Biology Example

A student tests whether a soil additive improves bean plant growth. Twenty bean plants receive the additive, and twenty bean plants do not. All plants are the same species, start at similar height, receive the same water, grow in the same light, and are measured after 21 days.

The independent variable is the presence or absence of the soil additive. The dependent variable is plant growth, such as change in height or dry biomass. The no-additive group is the control group. Light, water, plant species, container size, starting height, and growth time are controlled variables. If the additive group averages 18 cm of growth and the control group averages 12 cm, the data support the claim that the additive increased growth under those conditions.

Now change the design: the additive group is placed near a sunny window, while the control group is placed in shade. The experiment no longer isolates the additive. Light is a confounding variable because it could explain the growth difference. More trials would not fix the problem unless light is controlled or included as another intentional variable.

Observations, Experiments, and Models

Not every data source is an experiment. A field study might compare bird nesting success in several habitats. That can reveal patterns, but many factors differ at once: predators, food supply, noise, trees, human disturbance, weather, and competition. A simulation can test a model of population growth, but its conclusion depends on assumptions in the model. A controlled lab experiment can better support cause and effect, but it may simplify real ecosystems.

On the Regents, match the strength of the claim to the strength of the design. "The treatment was associated with higher growth in this trial" is safer than "the treatment will always increase growth in every plant species." Strong answers often include a limitation.

Improving an Investigation

When asked how to improve a design, pick changes that make the evidence more reliable or the comparison fairer:

• Increase sample size so one organism or one trial does not dominate the result.
• Repeat the investigation to check whether the pattern is consistent.
• Add a control group if the design lacks a baseline.
• Keep major conditions constant across groups.
• Randomly assign organisms or samples to groups when practical.
• Use calibrated tools and clearly defined measurement methods.
• Blind the observer when subjective scoring could introduce bias.

Do not choose changes that make the investigation more complicated without improving validity. Adding ten unrelated species, changing several factors at once, or measuring a new outcome can make the conclusion harder to defend.

Regents Traps

Trap 1: Naming the dependent variable as the changed factor. If the question asks whether temperature affects enzyme activity, temperature is the independent variable. Enzyme activity is the dependent variable because it is measured.

Trap 2: Treating a constant as a control group. Same amount of water is a controlled variable. Plants without fertilizer are the control group.

Trap 3: Ignoring starting conditions. If one group begins with larger plants, more bacteria, or older subjects, final values may reflect starting differences. Change from baseline may be more meaningful than final value alone.

Trap 4: Confusing sample size with repeated trials. Testing 30 seeds once gives a larger sample. Repeating the same setup three separate times checks consistency. Strong designs often use both.

Trap 5: Overstating causation from a survey. If students who sleep more also score higher, sleep may be related to score, but other variables such as study time, health, stress, or schedule could matter.

Engineering Design Questions

The engineering part of the Life Science: Biology blueprint can appear in ecosystem and human-impact contexts. An engineering question may ask which solution best meets criteria and constraints. Criteria are success goals, such as reducing bird collisions, lowering runoff, preserving biodiversity, or maintaining crop yield. Constraints are limits, such as cost, safety, time, materials, land use, and side effects. A strong answer compares trade-offs using evidence instead of choosing the option that sounds nicest.

For final review, practice turning any lab description into a design map: changed factor, measured response, baseline, constants, sample size, repeated trials, limitation, and justified conclusion. If you can map the design, you can usually spot the trap.