INVESTIGATING SCIENCE FACT OR FALACY

Sample Bias

When the sample doesn't truly represent the population, skewing results and reducing validity.

Convenience Sampling

Selecting whatever is most readily available, likely causing bias.

Voluntary Response Bias

Samples directed towards people who choose to participate, overrepresenting strong opinions.

Sample Size

Larger sample sizes lead to more reliable results; smaller sizes allow random chance to distort outcomes.

Confidence Level

Usually set at 95%, meaning results would match the real population 95% of the time if repeated.

Stratified Sampling

Dividing the population into groups and sampling each group proportionally.

Cluster Sampling

Dividing the population into clusters (e.g., cities), randomly selecting clusters, then sampling within them.

Simple Random Sampling

Every member of the population has an equal chance of being selected, removing bias but is difficult in practice with large populations.

Emotive Advertising

Advertising that triggers an emotional response to influence consumer behavior rather than using factual evidence.

Innate Bias

Unconscious, instinctive bias we are born with; our brains are wired to react emotionally before thinking logically.

Logical Fallacy

A flaw in reasoning that makes an argument invalid despite appearing convincing.

Efficacy

How well a product actually does what it claims to do.

Conflict of Interest

When a company/industry has financial motivation to suppress or distort scientific findings that threaten their profits.

Double Blind Trials

Trials attempting to test whether a treatment works while removing bias and placebo effects to produce fair and reliable results.

Placebo Effect

A phenomenon where a person's physical or mental health improves after receiving a 'dummy' treatment.

Weak Scientific Evidence

Even if a study sounds scientific, a small or biased sample makes claims unreliable (e.g. diet drink weight-loss studies).

Control Group

A group that does not receive the treatment, used as a comparison to test if the treatment works.

Improving Trial Reliability

Use large sample sizes, random assignment, repeat trials, keep variables constant, include placebo control.

Fossil Fuel Industry

Knew about climate change since 1956 but funded doubt to protect profits.

Tobacco Industry

Funded research to create doubt about smoking causing lung cancer.

Asbestos Industry

Suppressed evidence of lung disease while selling asbestos products until 1987.

Volkswagen Emissions Scandal

Cars were programmed to cheat emissions tests, misleading consumers.

Uncontrolled data sampling

data collected without a planned sampling method or control over who participates.

Lower sample size

larger standard deviation, less reliable results, lower validity