QCAA Agricultural Science EA cheat sheet
QCAA Agricultural Science EA · Keep the key ideas within reach.
Unit 3 - Data Test (IA1): Interpreting Agricultural Data
Descriptive statistics: mean = sum of values ÷ n; median = middle value of ordered data; mode = most frequent value. The median resists outlier skew better than the mean in yield or weight data.
Standard deviation quantifies spread about the mean; a large SD relative to the mean (high coefficient of variation, CV = SD ÷ mean × 100%) signals variable paddock or herd performance.
Correlation coefficient r ranges −1 to +1; r near 0 means no linear relationship. Correlation does not establish causation - a third variable (e.g. rainfall) may drive both.
Linear regression: y = mx + c, where m is the slope (change in y per unit x) and c the intercept. R² is the proportion of variation in y explained by x.
Significant figures and units must match instrument precision; report means to the same decimal place as raw data and always state units (kg/ha, mm, %, t DM/ha).
Graph selection rule: bar charts for categorical comparisons, line graphs for continuous trends over time, scatter plots for relationships between two continuous variables.
Trial design vocabulary: replication (repeated plots), randomisation (unbiased allocation), control (untreated comparison) and blocking (grouping similar units) reduce confounding and bias.
Water-use efficiency = biomass or yield produced ÷ water used (kg/mm or kg/ha per mm); feed conversion ratio = feed consumed ÷ liveweight gain.
When comparing treatments, state the direction and magnitude of the difference plus whether variation overlaps - overlapping error bars weaken claims of a real effect.
Unit 3 - Student Experiment (IA2): Design, Analysis and Evaluation
Research question must identify the independent variable (what is manipulated), dependent variable (what is measured) and at least one controlled variable held constant.
Validity = the experiment measures what it claims; reliability = repeated trials give consistent results; accuracy = closeness to the true value. These are distinct claims requiring distinct evidence.
Replication rule: at least three replicates per treatment allows a mean and a measure of variation; n = 1 cannot support an uncertainty claim.
Randomisation and a control group reduce systematic error; calibration of instruments (e.g. moisture meter, scales) reduces systematic error but not random error.
Percentage error = |measured − accepted| ÷ accepted × 100%; percentage difference compares two experimental values without an accepted standard.
Uncertainty propagation: for sums or differences add absolute uncertainties; for products or quotients add percentage uncertainties.
A valid conclusion states whether the hypothesis is supported, cites the data trend with values, and acknowledges uncertainty; it never overstates beyond the range tested.
Improvements must be specific and linked to an identified limitation (e.g. 'increase replicates from 3 to 6 to reduce the effect of soil variability on mean yield').
Ethical and safety considerations for agricultural trials include animal welfare approval, chemical handling (SDS), and biosecurity controls when moving soil or plant material.
Unit 4 - Research Investigation (IA3): Evidence, Claims and Evaluation
Research question must be answerable from secondary sources and specify a measurable outcome, a system (e.g. a production system) and a context.
Source hierarchy: peer-reviewed journals and government agency reports outrank industry magazines, which outrank general websites; primary data outranks secondary summaries.
A claim is a statement asserted as true; evidence is the data supporting it. Evaluate each source for currency, authority, accuracy, purpose and bias (the CRAAP-style criteria).
Synthesis means combining multiple sources to identify agreement, contradiction and gaps - a list of source summaries is not synthesis.
Quantitative evidence should be quoted with units and context (e.g. 'yield rose 12% over five seasons in a 300 mm rainfall zone') rather than vague statements.
Confounding variables and extrapolation limits must be discussed: results from one soil type, climate or breed may not transfer to another system.
Economic and sustainability reasoning should reference gross margin (revenue − variable costs), break-even yield, and indicators such as soil organic carbon or water-use efficiency.
A quality conclusion answers the research question directly, is proportional to the strength of evidence, and states residual uncertainty and further research needed.