HSC Science Extension flashcards

31 cards. Read the term and the definition here. The deck at the top of this page is optional.

Deductive reasoning

Reasoning from general premises to a specific conclusion. If premises are true and the logical form is valid, the conclusion necessarily follows, so deduction cannot yield new content beyond what premises contain.

Inductive reasoning

Reasoning from specific observations to a broader generalisation. It expands knowledge but never guarantees truth, so inductive conclusions remain probabilistic and are always open to revision by new evidence.

Abductive reasoning

Inference to the best explanation: choosing the hypothesis that most plausibly accounts for the observed evidence. It generates testable candidates rather than proving them, so abductive conclusions are provisional and require confirmation.

Falsifiability

Popper's criterion that a scientific claim must specify observations that would refute it. A hypothesis compatible with every possible observation makes no risky prediction and therefore lies outside empirical science.

Kuhnian paradigm shift

Kuhn's model in which normal science accumulates anomalies until a crisis produces a new paradigm. Paradigms are not abandoned because of single anomalies; a rival framework must explain both old and new evidence.

Confirmation bias

The tendency to seek, notice, or weight evidence that supports existing beliefs while ignoring contradictory data. Pre-registered predictions and deliberate disconfirming tests are the standard methodological remedies.

Correlation versus causation

A correlation shows variables vary together; causation requires temporal precedence, a plausible mechanism, and control of confounders. Observational association alone cannot establish that one variable produces the other.

Research ethics principles

Core principles of merit and integrity, justice, beneficence, and respect for persons. Human research requires informed consent, voluntary participation, confidentiality, and prior ethics committee approval.

Operational definition

The specific procedure or indicator used to measure a variable, such as defining growth as height in millimetres measured weekly. Operationalising variables makes hypotheses testable and methods reproducible by other researchers.

Null hypothesis

A statement that no effect or no relationship exists between the variables studied. Statistical tests assess how compatible observed data are with this null, before any alternative hypothesis is considered.

Independent and dependent variables

The independent variable is manipulated or selected by the researcher; the dependent variable is measured as the outcome. Controlled variables are held constant so they cannot confound the observed relationship.

Internal validity

The extent to which a study supports a causal inference, requiring control of confounders, appropriate comparison groups, and minimal measurement bias. Threats include selection, attrition, and instrumentation effects.

External validity

The extent to which findings generalise beyond the sample, setting, and conditions studied. Representative random sampling and realistic conditions strengthen it; narrow convenience samples weaken it.

Reliability versus validity

Reliability is consistency of repeated measurements; validity is whether a measure captures the intended construct. A scale can be reliably wrong, so reliability is necessary but not sufficient for validity.

Stratified random sampling

Dividing the population into subgroups and randomly sampling within each, preserving subgroup proportions. It reduces selection bias and improves precision when strata differ markedly on the outcome.

Literature review purpose

To map existing knowledge, define key terms, identify the gap the study addresses, and justify the chosen method. It must cite sources accurately and avoid unsupported or uncited claims.

Risk assessment in proposals

Identifying hazards, estimating likelihood and severity, and specifying mitigation for each. Human or animal participants additionally require documented ethics approval before any data collection begins.

Categorical versus numerical data

Categorical data are nominal or ordinal labels; numerical data are discrete counts or continuous measurements. Data type determines permissible statistics, since means of nominal categories are meaningless.

Mean, median and mode

Measures of central tendency: arithmetic average, middle value, and most frequent value. The median resists outliers, while the mean is sensitive to extreme scores in skewed distributions.

Standard deviation

A measure of spread describing typical distance of scores from the mean. A large standard deviation indicates high variability, so group means may differ while distributions overlap substantially.

p-value definition

The probability of obtaining data at least as extreme as observed, assuming the null hypothesis is true. It is not the probability the null is true nor the probability of replication.

Statistical versus practical significance

A statistically significant result (typically p < 0.05) may be trivially small. Report effect size and confidence intervals to judge whether the difference matters in real contexts.

95% confidence interval

A range constructed so that 95% of intervals from repeated sampling contain the true parameter. It is not a 95% probability that the parameter lies within one particular interval.

Type I and Type II errors

Type I error rejects a true null (false positive); Type II error fails to reject a false null (false negative). Statistical power equals one minus the Type II error rate.

Measurement uncertainty

The range within which a measured value plausibly lies, arising from instrument resolution, calibration, and random variation. Report it with appropriate significant figures and propagate it through calculations.

Evidence-limited conclusions

Conclusions restricted to what the data support, distinguishing demonstrated inference from speculation. They acknowledge limitations and state what additional evidence would test the claim further.

Report structure

Title, abstract, introduction, methods, results, discussion, conclusion, references, and appendices. Each section has a distinct function, and raw data belong in appendices for verification.

Abstract content

A self-contained summary stating aim, method, key results, and conclusion, usually 150-250 words. It contains no citations or new information and must be intelligible without the full report.

Reproducible methods section

A methods section reporting materials, apparatus, procedure, variables, and analysis in enough detail for another researcher to repeat the study and obtain comparable results.

Results versus discussion

Results present findings, tables, and figures without interpretation; the discussion interprets them against the hypothesis and literature, explains limitations, and states implications without overclaiming.

Citation and reference matching

Every in-text citation must appear in the reference list and vice versa, in one consistent style. All borrowed ideas require acknowledgement to avoid plagiarism and to allow verification.