2.1 Scientific Method and Experimental Design
Key Takeaways
- A scientific hypothesis is a testable, falsifiable prediction; a theory is a well-substantiated explanation; a law describes a recurring pattern without explaining mechanism.
- Every controlled experiment has exactly one independent variable (manipulated), one dependent variable (measured), and several controlled variables (held constant).
- Replicates (repeated trials) and adequate sample size reduce random error; a control group establishes the baseline against which experimental groups are compared.
- Blinding (single-blind hides treatment from subjects; double-blind hides it from both subjects and researchers) eliminates placebo and observer bias.
- Peer review and independent replication are the two filters that move a finding from hypothesis to durable scientific knowledge.
Why This Section Matters
The Praxis Biology (5236) Subarea I, Nature of Science, accounts for roughly 9% of scored questions. The Educational Testing Service (ETS) heavily weights items that ask you to identify variables, distinguish levels of scientific knowledge, and critique a study design from a short paragraph. You do not need to memorize obscure statistics here — but you must recognize the standard scientific-method vocabulary instantly.
Hypothesis, Theory, and Law
These three terms are routinely confused in everyday speech. On the Praxis, they have precise meanings:
| Term | What It Is | Example |
|---|---|---|
| Hypothesis | A testable, falsifiable prediction about how a variable will behave. | "Increasing light intensity will increase the rate of photosynthesis in elodea." |
| Theory | A broad, well-substantiated explanation of a natural phenomenon, supported by extensive evidence. | Cell theory, evolutionary theory, germ theory, plate tectonics. |
| Law | A concise description of an observed pattern, often mathematical, without explaining why it happens. | Mendel's Law of Segregation; the Hardy-Weinberg principle. |
A common misconception your students will bring into class is that a theory is "just a guess." In science, a theory is the strongest form of explanatory knowledge — promotion from hypothesis to theory requires repeated independent confirmation across many lines of evidence.
Falsifiability (Karl Popper)
A statement is scientific only if it could, in principle, be proven wrong by an observation. The claim "all living cells contain DNA or RNA" is falsifiable — finding one cell without either would refute it. The claim "there is an invisible, undetectable force guiding evolution" is not falsifiable and therefore not a scientific hypothesis.
Variables in a Controlled Experiment
A classic Praxis stem describes a short experiment and asks you to identify the variables.
- Independent variable (IV) — the one factor the researcher intentionally changes. Goes on the x-axis.
- Dependent variable (DV) — what is measured in response. Goes on the y-axis.
- Controlled variables (constants) — factors held the same across all groups (temperature, pH, light cycle, species, age).
- Confounding variable — an uncontrolled factor that varies with the IV and offers an alternative explanation for the result. Confounders ruin causal inference.
Worked Example
A teacher grows three groups of bean seedlings: Group A in plain water, Group B in 1% fertilizer solution, Group C in 5% fertilizer solution. All groups receive the same light, temperature, and pot size. After 14 days, seedling heights are measured.
- IV: fertilizer concentration (0%, 1%, 5%).
- DV: seedling height after 14 days.
- Controlled variables: light, temperature, pot size, water volume, species, starting seed mass.
- Control group: Group A (0% fertilizer) — receives no treatment of the IV.
- Experimental groups: Groups B and C.
Control vs. Experimental Groups
The control group receives no treatment of the independent variable (or receives a known-neutral placebo). Its function is to establish the baseline outcome without the manipulation, so that any difference observed in the experimental group can be attributed to the IV rather than to extraneous factors.
Many Praxis stems give a three-condition experiment (e.g., bright / dim / no light) and ask you to identify the control. The answer is always the condition that removes the variable being tested, not the "middle" or "normal" condition.
Replicates and Sample Size
- Replicates are repeated independent trials. Within a single experiment, multiple replicates reduce random error and let you calculate variability (standard deviation, standard error).
- Sample size (n) is the number of independent units (organisms, dishes, patients) per group. Larger n increases statistical power — the probability of detecting a real effect.
- Replicates ≠ resampling the same organism multiple times. Measuring the same plant 10 times gives you 10 measurements but a sample size of 1.
A reasonable rule of thumb for classroom inquiry: at least 3 replicates per condition; ideally 5-10 when feasible.
Blinding
Blinding controls bias generated by expectation, not by the treatment itself.
- Single-blind: subjects do not know which group they are in. Eliminates the placebo effect in human trials.
- Double-blind: neither subjects nor the researchers handling them know group assignment until the data are unblinded. Eliminates observer bias (researchers unconsciously favoring expected outcomes when measuring or coding data).
- Open-label: everyone knows. Acceptable for some pilot work but weakest design.
Double-blind randomized controlled trials (RCTs) are the gold standard for evaluating new drugs and clinical interventions.
Peer Review and Reproducibility
After a study is complete, the findings move through two filters before becoming durable knowledge:
- Peer review — qualified experts in the field evaluate the manuscript's methods, statistics, and conclusions before publication. Peer review screens for clear errors but does not, by itself, prove the result is true.
- Independent replication — other research groups reproduce the experiment with their own materials. Reproducibility is the strongest validation; results that fail to replicate are downgraded or retracted.
Common Praxis Traps
- Confusing "control group" (no-treatment baseline) with "controlled variable" (factor held constant).
- Calling a well-tested theory "just a theory" — never the right answer choice.
- Treating large sample size as a fix for a confounded design. N cannot rescue a missing control group.
- Equating peer review with replication. They are distinct safeguards.
A biology teacher tests whether adding caffeine to pond water affects Daphnia heart rate. She prepares four beakers: 0 mg/L (plain water), 5 mg/L, 10 mg/L, and 20 mg/L caffeine, with 6 Daphnia in each beaker at 22 degrees C. Which is the independent variable?
A graduate student claims her new herbal extract "boosts vitality." When asked how she would test the claim, she answers that 'vitality' can mean anything from energy to mood to immunity, so any positive change counts. Why is this claim NOT a scientific hypothesis?