Skip to content
Tutorials Center
My learning

Topics

Lesson 2 of 37 ยท Getting Started

Thinking Like an Environmental Scientist

Learn hypotheses, variables, controls and sample size, then run a simulated plant experiment and a t-test to see how scientists judge environmental evidence.

Environmental scientists cannot put a whole lake or forest in a laboratory. Instead they combine careful observation, field studies, small controlled experiments and computer models. This lesson shows the habits of mind that make their conclusions trustworthy, so that you can judge claims in the news and design your own simple investigations.

The cycle of inquiry

The "scientific method" is not a fixed recipe. It is a cycle that scientists repeat many times.

flowchart LR
    A["Observe a pattern"] --> B["Ask a question"]
    B --> C["Form a hypothesis"]
    C --> D["Test it: experiment or survey"]
    D --> E["Analyze the data"]
    E --> F["Conclude and share"]
    F --> A

Each result raises new questions, so the cycle starts again.

A hypothesis is a testable, specific explanation, such as: "Plants watered with salty water grow more slowly than plants watered with fresh water." A good hypothesis can be proven wrong by evidence. "Nature is mysterious" cannot be tested, so it is not a hypothesis.

Variables and controls

In an experiment you change one thing and watch what happens to another.

  • The independent variable is the one you deliberately change (salt level in the water).
  • The dependent variable is the one you measure (plant height after 14 days).
  • Controlled variables are kept the same for all groups (light, soil, temperature, pot size).
  • The control group receives the normal condition (fresh water) so you have something to compare with.

If two things change at once you cannot tell which one caused the result. This is why controlled variables matter.

Write down the independent variable, the dependent variable and at least three controlled variables before you start any experiment. If you cannot name them, the design is not ready.

Field studies and correlation

Many environmental questions cannot be tested by changing one thing, because you cannot assign pollution to a river on purpose. Scientists use observational studies (also called field studies): they measure what already happens at many sites and look for patterns.

A pattern in the data is a correlation. Correlation does not prove causation. Ice cream sales and drowning both rise in summer, but ice cream does not cause drowning: a third factor, hot weather, drives both. Scientists build stronger arguments for causation by combining several kinds of evidence: a plausible mechanism, repeated results, results at different places, and experiments where possible.

Sample size and variation

Living things vary. If you grow only one plant in each condition, a difference might be luck. A larger sample size (more plants) smooths out random variation and makes a real difference easier to see. Repeating the measurement is called replication.

Scientists also report the mean (average) and the standard deviation, which measures how spread out the values are. A big gap between means, compared with the spread, is stronger evidence.

A simulated experiment

The program below simulates plant heights for two groups. The data are made up by a random number generator with a fixed seed, so the result is repeatable. It is a simulation to practice the analysis, not a real experiment.

Python Simulated salt experiment

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

With this seed and only 10 plants per group, the test does not reach the usual cutoff even though the true means differ. That is not a mistake: small samples often miss real effects, as the next example shows.

The p-value is the probability of seeing a difference at least this large if there were really no effect. A small p-value (a common cutoff is 0.05) means chance alone is an unlikely explanation. It does not tell you how big or important the effect is, and it does not prove the hypothesis.

Why sample size matters

Try changing the sample size in a second simulation. With the same true difference, small samples often give a confusing answer, and larger samples usually give a clear one.

Python How often do we detect a real difference?

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

You should see the detection rate climb as the sample grows. This is one reason a single small study is rarely decisive.

Models

When an experiment is impossible, scientists build models: simplified descriptions of a system, often as equations run on a computer. Climate models, population models and water-flow models all work this way. A model is only as good as its assumptions, so scientists test it against real observations. You will build small models of your own in later lessons, always labeled as simplified.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. In an experiment on how fertilizer affects crop yield, what is the dependent variable?

    The dependent variable is what you measure in response. The amount of fertilizer is the independent variable you change.
  2. 2. Why do scientists use a control group?

    Without a control group you cannot tell whether the change you made caused the result.
  3. 3. Two measurements rise together over many years. What can you safely conclude?

    Correlation is a pattern. A third factor or coincidence could explain it, so scientists look for mechanisms and repeated results.
  4. 4. What is the main benefit of a larger sample size?

    More samples average out chance differences, so a real effect is easier to detect. It never guarantees a hypothesis.

Summary

  • Science works in a cycle: observe, question, hypothesize, test, analyze, share, repeat.
  • A hypothesis must be testable and could be proven wrong.
  • Change one independent variable, measure a dependent variable, and control everything else.
  • Correlation is not causation; stronger conclusions need several kinds of evidence.
  • Larger samples and replication reduce the influence of chance; a p-value measures how surprising the data would be if there were no effect.
  • Models help when experiments are impossible, but they rest on assumptions.

Next: Earth's Systems and Feedback Loops