Report cards and grading

Setting an item that assesses six different skills

A worked Geography item on Kampala's growth, and how its parts were built to reach every base.

Irumba Sharif · 30 August 2026 · 2 min read

Writing a competency-based item is harder than writing questions. Every base has to be reachable, and no part should assess two bases at once. Here is a Geography Paper 2 item taken apart.

The scenario

Kampala's population at census years: 458,503 in 1980; 774,241 in 1991; 1,208,544 in 2002; 1,507,114 in 2014; and a projection of 1,875,400 for 2024. Alongside it, the growth of informal settlements such as Kisenyi and Bwaise, congestion on Jinja Road and the Northern Bypass, and government plans for satellite towns at Mukono, Wakiso and Entebbe.

One scenario, carrying numbers, places and a policy question. That is deliberate: it has to support six different kinds of thinking.

The six bases, and the part that reaches each

• Data interpretation: describe the trend, and suggest why growth was faster between 1991 and 2002
• Statistical skill: calculate the percentage increase from 1991 to 2002, and the average annual growth rate from 2002 to 2014, showing working
• Explanation of causal factors: four factors behind rural-urban migration into Kampala
• Analysis of consequences: three positive effects on economic development, three negative effects on the urban environment
• Map and spatial reasoning: a labelled sketch map locating Kampala relative to the satellite towns, and how those towns decongest the city
• Evaluation and judgement: weigh the claim that relocating markets and industries to satellite towns is the most effective solution

What makes it work

**The data is real and awkward.** Growth between periods is uneven, so "describe the trend" has something to describe. A neat doubling every decade would have made the first base worthless.

**Calculation is separated from interpretation.** A learner who can read a trend but cannot compute a growth rate scores well on one base and poorly on another, and the rubric records both. If the two had been folded into one part, that learner would get a single muddled score.

**The evaluation is genuinely arguable.** Satellite towns might be the best answer, or they might not, and a learner can score 4 either way if the judgement is reasoned. An item whose evaluation has a correct answer is not assessing judgement.

Setting your own

Work backwards. List the bases you intend to assess, then write a scenario rich enough to support all of them, then write one part per base. If you find a part that would score against two bases, split it. If you find a base with no part reaching it, you have a five-base item, and the conversion table you use has to change accordingly.

A caution about the totals

The minimum on a six-base item is 6, not 0, because a learner who attempts every base earns at least 1 on each. Marking down to zero is not available to you, and a scoring sheet that allows it is producing scores the conversion table cannot read.