ENEN90028 Chap.8 Analysis, Reporting and Assessment
Analysis, Reporting and Assessment
The final project stages preserve the design in analysis, graphics and claims. Exploration checks units, missing values, hierarchy and distribution before testing. Type I and Type II errors and statistical power frame the meaning of results. Results describe patterns and uncertainty, while Discussion interprets them against the causal model, prior research, limitations and the assessment endpoint.
What this chapter covers
- 01
Exploratory checks and graphical summaries
- 02
Type I and Type II errors
- 03
Power effect size variability and replication
- 04
Methods results and discussion roles
- 05
Presentation questions and design defence
- 06
Published report criteria and bounded conclusions
Worked example · free
Separate result from interpretation
- 1The contrast is 14.0 minus 18.0, which equals negative 4.0 species.
- 1Results state both means and the four species difference without causal language.
- 1Discussion says the lower richness is consistent with the predicted pathway.
- 1Discussion qualifies attribution using baseline, replication and remaining land-use differences.
Key terms
- Exploratory analysis
- Initial checking and visualisation used to understand data quality, structure and distribution.
- Type I error
- Concluding that an effect exists when the null model is true.
- Type II error
- Failing to detect an effect that exists.
- Statistical power
- The probability of detecting an effect under stated assumptions.
- Results section
- The report section that describes observations, estimates, uncertainty, tables and graphs.
- Discussion section
- The report section that interprets results against mechanisms, research and limitations.
Analysis, Reporting and Assessment FAQ
Does a non-significant result prove no impact?
No. Failure to reject a null model may reflect a small effect, high variability, weak replication, inappropriate scale or low statistical power. Report the estimate and uncertainty, then discuss what effect sizes the design could plausibly detect and which limitations narrow the conclusion.
What belongs in Results rather than Discussion?
Results should present tables, graphs, estimates, uncertainty and descriptive text about observed patterns. Discussion explains what those patterns mean, whether mechanisms were supported, how they compare with research, which alternatives remain and what the design permits the project to conclude.
How should presentation questions be answered?
Lead with the design decision, give the evidence or reason and state the consequence. Be ready to explain endpoint choice, sampling unit, site selection, scale, quality controls, camp changes and limitations. Distinguish what the data show from mechanisms supported only by literature.
Assessment move
Use timed quiz sets for fast conceptual discrimination, then use longer oral prompts to defend design choices. For report preparation, audit the chain from question and conceptual model to sampling unit, quality control, analysis, graph and conclusion. Recompute values shown in figures, state independent sample size, and keep Results descriptive.
End Discussion claims with the comparison that supports them or the limitation that narrows them. Organise revision around the assessment formats actually published for this subject rather than preparing for a nonexistent final examination. For each short quiz, practise ten-item sets at roughly one and a half minutes per item.
Identify the tested distinction before reading every option, eliminate choices that confuse endpoints with indicators, observations with replicates or association with causation, select the best answer and return to difficult items if time remains. Review distractors by explaining the design principle they violate.
For the team planning report, use a one-page audit that begins with the human intervention and environmental value, then traces mechanisms to assessment endpoints, indicators, sampling units, controls or gradients, scale, field methods, equipment, quality controls and risks. A link missing from that chain is a design problem, not a writing problem.
For presentation preparation, rehearse a five-minute explanation and then switch to questions. Answer with decision, evidence and consequence: state what the team chose, why the source or process supports it and how it affects inference or field work. Include likely questions about site selection, independent replication, spacing and support, changes made at camp and the boundary of the conclusion.
For the individual report, begin with an integrity pass over units, missing values, identifiers and hierarchy. Count independent units before calculating summaries. Plot raw or unit-level patterns where possible, inspect variability and potential anomalies, and verify every number transferred to a table or figure. State the analysis used and keep captions descriptive.
Practise separating sections by rewriting causal or interpretive Results sentences: Results should report means, contrasts, spread and uncertainty, while Discussion connects those observations to mechanisms, existing research, alternatives and limitations. Rehearse Type I and Type II errors in plain language. A false positive concludes an effect when the null model is true; a false negative misses an effect that exists.
Power depends on effect size, variability, replication and the decision rule, so a non-significant result does not prove no environmental impact. Finish with a backward audit from the conclusion. Every claim should lead to a graph or estimate, an analysis that respects the hierarchy, a documented field record and a design comparison. Where the chain weakens, narrow the wording explicitly.
Use the published report criteria as a coverage checklist, but do not invent grade-band rubric descriptors that the lecture says are not provided.
