Marine Science Inquiry Skills
Core concepts
You must design investigations, analyse data and evaluate evidence. Key ideas: experimental design (hypotheses, independent/dependent/controlled variables, replication, randomisation, control groups); sampling (quadrats, transects, catch surveys, stratified sampling, sample size); data analysis (means, variability, error bars, correlation vs causation, graphing conventions); evaluation (reliability, validity, accuracy, precision, sources of error, limitations); and communication (clear tables, labelled axes with units, justified conclusions). Understand how to distinguish systematic from random error and how to assess whether evidence supports a claim.
Traps to watch
- Correlation is not causation. A trend in two variables does not prove one causes the other; confounding factors must be considered.
- Confusing accuracy and precision. Accuracy is closeness to the true value; precision is repeatability. A precise set of readings can be inaccurate (systematic bias).
- Reliability vs validity. Reliability is consistency on repetition; validity is whether the method measures what it claims.
- Missing units and labels. Marks are lost for unlabelled axes, missing units and unexplained error bars.
- Overgeneralising. Small or non-random samples limit conclusions; state limitations explicitly.
Rule of thumb
For every investigation, state what you changed, what you measured, what you kept constant, and how many replicates. When evaluating, separate random error (reduces precision, fixed by replication) from systematic error (reduces accuracy, fixed by calibration or method change).
Worked example
A student samples fish abundance at two sites using one transect each and concludes Site A is richer. Evaluate: single transects give no replication, so variability is unknown (low reliability); site differences (depth, substrate) are confounders (low validity). Improvement: multiple random transects per site, replicate counts, and report means with error bars before drawing conclusions.