Match the technique to the data type
A presentation technique displays fieldwork data so a pattern can be seen. Choose by data type: category totals (counts per land-use type) suit a bar or pie chart; data that must show WHERE suits a located bar or pie chart on a sketch map; two continuous variables (distance vs pebble size) suit a scatter graph with a best-fit line. A misfit technique -- a plain bar chart for spatial data -- loses the locational information needed.
Mean, median and range summarise data
Descriptive statistics summarise and compare data sets. The mean is the sum of values divided by their count. The median is the middle value once the data is in order (the mean of the two middle values if the count is even). The range is the highest value minus the lowest, showing how spread out the data is. Comparing the mean at one site with another -- not eyeballing raw readings -- gives evidence that two places genuinely differ.
Conclusion: decide, cite figures, anomaly
A hypothesis is a testable statement written before data collection. A strong conclusion states the decision early -- "IS", "IS NOT" or "IS PARTIALLY supported" -- then backs it with 2-3 figures from your OWN results (means, percentages, a trend), not a restatement of the hypothesis. An anomaly (a point breaking the pattern) should be acknowledged, not ignored, and does not force rejection: a clear trend with one explained anomaly can still support the hypothesis.
Drawn from real examiner reports.
Graphs drawn without axis units
Graphs are frequently drawn without labelled axis UNITS, making an otherwise accurate plot impossible to interpret -- a value of "40" means nothing unless the axis says mm, m, %, or count. Every graph needs each axis labelled with its variable AND unit, plus an even, appropriate scale. An accurate plot still loses marks if the reader cannot tell what the numbers measure.
Digest (## Skills, June 2023 Paper 3 coursework report): graphs are often ineffective through missing axis units, a weakness that mirrors the Paper 4 skill directly (s23 Paper 3 coursework report).
Evaluation lists only negatives
Evaluations often list only what went wrong (a limitation) without pairing each with a specific improvement -- and the two are not the same thing. State both: name the weakness, then the exact change that fixes it (e.g. "only one reading per site" -> "take three repeat readings and average them"). A list of problems alone caps the marks.
Digest (## Skills, June 2023 Paper 3 coursework report): evaluations focus almost solely on negatives without matching improvements, mirroring the Paper 4 skill directly (s23 Paper 3 coursework report).
Stating figures without comparing them
Quoting figures side by side is not a description or comparison. "Site A scored 8, Site B scored 4" earns little until comparison and a quantified difference are added -- "Site A scored HIGHER, by 4 points". A scatter trend needs "pebble size DECREASES as distance INCREASES", not "it goes down". For describe or compare, use comparative words with the figures.
Digest (## Cross-cutting exam technique): quoting comparative statistics side-by-side is not itself a compare/describe answer -- comparative words (higher/lower) must be used (all four sittings).
A vague verdict, not a clear decision
The topic's most repeated Paper 4 error. Some never STATE a decision -- "the hypothesis might be true", or re-describing results with no verdict -- forfeiting the conclusion marks. Others disagree with a decision the question told them to accept. Open with "supported / not supported / partially supported", then justify -- and follow the decision the question gives.
Digest (## Skills, Paper 4 investigation skills): the hypothesis-decision technique is the single most repeated Paper 4 message -- candidates must open with an explicit Yes/No/Partially decision before supporting evidence; vague verdicts and disagreeing with a given decision both forfeit credit (s23 Paper 41/42/43 key messages).
Leaving a completion task blank
Practical completion tasks -- plotting a missing point, completing a part-finished histogram, or adding a line of best fit -- have a high omission rate, skipped on many scripts. They are straightforward marks needing no written explanation, just accurate use of the scale and axes. Never leave one blank: even an approximate point or line earns credit a blank cannot.
Digest (## Skills, Paper 4 investigation skills): practical completion tasks (plotting, histograms, best-fit lines) have a persistently high omission rate (5-20%+ per question) despite being flagged as easier marks not to skip (w22 Paper 41 Q1(a)(i)).
Speculative analysis, not the data
Analysis is often held back by speculative, hedging language -- "the reason might be...", "this could be because..." -- instead of using the collected figures. Credit comes from analysing what the data shows: quote the means, percentages or trend, then interpret them. Tentative guessing with no reference to your own numbers stays at a low level, however plausible.
Digest (## Skills, June 2023 Paper 3 coursework report): analysis is held back by speculative language ("the reason might be..."), a weakness that mirrors the Paper 4 skill directly (s23 Paper 3 coursework report).
Fieldwork data cannot "prove" a hypothesis
A testable wrong belief. Some write that results "prove" the hypothesis, treating one small study as certain fact. A fieldwork enquiry uses a limited sample, sites and days -- so the credited language is "support" or "suggest", never "prove". The same applies to correlation: a strong correlation supports a link, not that one variable CAUSES the other.
Digest (## Skills, Paper 4 investigation skills): the hypothesis-decision technique message emphasises a graded Yes/No/Partially decision rather than certainty, and vague/unqualified verdicts recur as a credit-blocker (s23 Paper 41/42/43 key messages).
One anomaly does not sink the enquiry
A common overreaction. Finding one anomaly -- a point breaking a consistent pattern -- does not mean the investigation failed or the hypothesis is rejected. A clear trend across most sites, with one explained anomaly (a sheltered spot, a measurement error), can still support it. The credited approach: name it, suggest a reason, then state how it affects the conclusion.
Digest (## Cross-cutting exam technique): vague/unqualified verdicts and listing-without-developing recur as credit-blockers across extended-response questions in every sitting, closely related to how anomalies are handled in a fieldwork conclusion.
Limitation, reason, improvement
Structure an evaluation, do not list. For each point: name a SPECIFIC limitation ("only one reading per site"); explain HOW it cut reliability ("no reading to average against"); give a SPECIFIC improvement ("three readings, averaged"). Develop 2-3 points; depth beats a list.
Match the presentation to the data
Keep the information the data needs. Categorical data by place -> a located bar/pie chart on a base map. Two continuous variables -> a scatter graph with a best-fit line. A plain bar chart gives only totals, losing WHERE; a scatter is wrong for one category.
Write the conclusion: decide then cite
A conclusion is not a summary. State the decision FIRST -- the hypothesis IS, IS NOT or is PARTIALLY supported -- then justify with 2-3 figures from your OWN results (a mean, a percentage, a trend). Acknowledge any anomaly briefly with a reason. Say "supports", never "proves".
Full notes, flashcards, Q&A and the topic quiz for every premium subject.
Premium plans are US$8.99/month or US$49.99/year — first month free.
Studying with a parent's blessing? Show them this.
