Experimental Design and Hypothesis Concepts

Hypothesis: Definition and Requirements

• A hypothesis is your tentative answer to a research question and should make predictions that could be tested.

• It must have two essential parameters: it has to be testable and it must yield two possible outcomes.

• Words like “prettier” or “smells better” are not testable in a scientific sense, so they do not fit a scientific hypothesis.

• A hypothesis is part of a circular, iterative view of scientific design: observations lead to questions, which lead to hypotheses, which are tested, and data may prompt reformulation of the hypothesis.

• When forming hypotheses, remember they should be testable and falsifiable; falsifiability means there must be a possible outcome that would disprove the hypothesis.

• In many courses, you’ll write null and alternative hypotheses (Ho and Ha). While you may not always be explicitly asked to write Ho in class labs, you should understand what they mean.

Null and Alternative Hypotheses (Ho vs Ha)

• Null hypothesis (Ho): predicts no difference between test groups; any observed difference arises from random variation.

• Alternative/experimental hypothesis (Ha): predicts a difference (or a directional difference) between test groups.

• Example 1 (shoe sizes):

  • $H<em>0:\\mu</em>{\text{males, 10-15}} = \mu_{\text{females, 10-15}}$

  • Ha:\ \mu{\text{males, 10-15}} > \mu_{\text{females, 10-15}}

• Example 2 (chickens):

  • $H<em>0:\ \mu</em>{\text{purple}} = \mu_{\text{brown}}$

  • Ha:\\mu{\text{purple}} > \mu_{\text{brown}}

• Outcomes when testing:

  • Either the average (e.g., shoe size) is greater than the other group, or it is not greater than the other group.

  • Representations can be adjusted (e.g., use “greater than or equal to” depending on the test setup), and the choice of comparison determines the statistical model.

• If data show a difference beyond what would be expected by random variation, you reject Ho and focus on Ha.

• Important nuance: nothing in science is ever proven; conclusions are based on the best available evidence and can be revised with new data or methods.

• A common exercise: write some simple pairs of hypotheses (e.g., about flower color and bee attraction) to practice forming Ho/Ha.

The Difference Between Theories and Laws

• Theories and laws serve different purposes and are not simply “stronger” or “more tested” versions of each other.

• Theories explain why phenomena occur; laws describe relationships or patterns often with a mathematical form.

• Theories and laws are both well-supported by evidence, but they address different questions.

• Examples:

  • Theory: Darwin’s theory of evolution by natural selection explains how fitness differences influence survival and reproduction, leading to evolutionary change.

  • Law: Boyle’s law describes the relationship between pressure and volume in a gas, given by a mathematical relationship (for an ideal gas at constant temperature and amount): $P V = n R T$, implying $P \propto \frac{1}{V}$ when $n, T$ are fixed.

• Limits:

  • Scientific theories explain phenomena but do not claim to address supernatural or metaphysical questions.

  • Observations and interpretations can be biased or misinterpreted; science relies on repeatable results and objective testing.

• The example of stomach ulcers shows how initial explanations (stress, spicy foods) can be revised as evidence accumulates; science cannot address supernatural causes, and hypotheses must be testable and falsifiable.

• Misinterpretation vs falsification: results can be misinterpreted due to cognitive biases; falsification occurs when data directly contradict a hypothesis.

• Historical note: vaccination predates the formal identification of a virus (e.g., Jenner’s smallpox vaccine); technological advances (e.g., understanding of viruses, electron microscopy) later clarified underlying mechanisms.

The Role of Technology, Models, and Real-World Tools

• Technology is the practical application of scientific knowledge and provides tools for inquiry (e.g., vaccines, pasteurization).

• Vaccination history: Edward Jenner developed the smallpox vaccine before viruses were scientifically identified; vaccines can precede full mechanistic knowledge.

• Pasteurization example: ultra-high-temperature (UHT) pasteurization treats milk at high temperature for a short time to inactivate pathogens.

• Models as a scientific tool: models help us understand complex phenomena and make predictions when direct testing is difficult; e.g., a model of blood flow through the heart aids understanding of heart function and predictions of responses to changes.

Variables in a Scientific Investigation

• Independent variable (the manipulated or experimental variable): the factor you intentionally change.

• Dependent variable (the responding variable): the outcome that changes in response to the independent variable.

• Principles for good experiments:

  • Change only one variable at a time to isolate effects.

  • A hypothesis should address one independent and one dependent variable at a time.

  • You can formulate multiple hypotheses for different aspects of an experiment, but test one pair at a time.

• Examples:

  • If you change the amount of sunlight for plants, you might measure plant growth as the dependent variable.

  • Ensure that other factors like watering and temperature remain constant while you vary sunlight.

Recap: The Scientific Process in Everyday Thinking

• The scientific process is a general way of evaluating data and making decisions in daily life (e.g., deciding what to wear based on weather data and observations).

• A hypothesis must satisfy two criteria: it is testable and it yields two possible outcomes.

• A scientific theory explains phenomena; a scientific law describes relationships and often has a mathematical form.

• The independent variable is manipulated by the investigator; the dependent variable changes in response.

• Only a single variable is tested at a time in a given investigation; multiple hypotheses can be formed for different parts of a study.

• To test your understanding, practice writing scientific hypotheses; consider simple, observable questions and work through Ho/Ha formulations.

Quick Practice Prompts (from the transcript)

• Observe purple vs yellow flowers with bee attraction differences; formulate a hypothesis about what makes bees prefer one over the other.

• Consider a test of whether a treatment affects plant growth with sunlight as the manipulated variable; write Ho and Ha and indicate possible outcomes.

Significance for Exams

• Be able to articulate: what a hypothesis is, how it is testable and falsifiable, how Ho and Ha are structured, and how to interpret a rejection of Ho.

• Distinguish clearly between theories (explanations) and laws (descriptions with mathematical relationships).

• Identify independent and dependent variables in a described experiment and justify changing only one variable at a time.

• Recognize that science progresses by testing, observing results, and revising hypotheses; never claiming absolute proof.