Home › Exam notes › CSIR NET › Scientific Method and Evidence

Scientific Method and Evidence

Scientific Reasoning

Useful for: CSIR NET

Overview

Correlation shows that two variables move together; causation means one variable produces change in the other — correlation alone does not prove causation.

Key points

Correlation coefficients range from −1 (perfect negative) to +1 (perfect positive); 0 means no linear relationship.
A third (confounding) variable may cause both.
Reverse causation: the effect may actually be the cause.
Spurious correlations occur by coincidence.
Randomised experiments are the strongest way to show causation.
Correlation is useful for prediction even without causation.

Explanation

Scatter plots show the form of a relationship.
Strong correlation is not the same as large effect.
Longitudinal data help establish time order.
Controlling variables reduces confounding.
Be cautious with headlines claiming causes from surveys.

Step by step

Identify the concept, method or policy asked.
Recall the definition, thinker, year or rule.
Apply it to the situation and eliminate extreme options.

Examples

Ice-cream sales and drowning both rise in summer → confounding by temperature.
Students who read more score higher → correlation; an experiment is needed to show reading causes the gain.
R = −0.8 between absence and marks → strong negative correlation.

Traps and advanced points

Trap: Inferring causation from correlation.
Trap: Thinking r = 0 means no relationship of any kind.
Trap: Ignoring confounders.

Quick revision

−1 to +1.
Confounding, reverse causation.
Experiments for causation.

Practice MCQs

A good scientific hypothesis must be
1. testable and capable of being shown false
2. always true
3. based only on opinion
4. impossible to test
Read chapter notes · Similar questions
Show answer
Answer: (a) testable and capable of being shown false. Karl Popper called this falsifiability.
In an experiment, the group that does not receive the treatment is the
1. control group
2. test group
3. focus group
4. reference sample only
Read chapter notes · Similar questions
Show answer
Answer: (a) control group. It provides a baseline for comparison.
Repeating an experiment by other researchers to check results is called
1. abstraction
2. plagiarism
3. hypothesis
4. replication
Read chapter notes · Similar questions
Show answer
Answer: (d) replication. Replication builds confidence in findings.
Reasoning from specific observations to a general rule is
1. circular reasoning
2. deductive reasoning
3. inductive reasoning
4. analogy only
Read chapter notes · Similar questions
Show answer
Answer: (c) inductive reasoning. Deduction goes from general to specific.
Evaluation of a research paper by experts before publication is
1. copyright
2. citation
3. indexing
4. peer review
Read chapter notes · Similar questions
Show answer
Answer: (d) peer review. It helps maintain quality.
The variable that the researcher changes on purpose is the
1. constant
2. dependent variable
3. confounding variable
4. independent variable
Read chapter notes · Similar questions
Show answer
Answer: (d) independent variable. Its effect on the dependent variable is measured.