Three Sources, One Strike: Judging Scale and Reliability
A student is writing a 1,200-word essay on 'the scale of the 1937 strike at the Kansanshi copper mine'. She has three sources. Source A, the company's annual report, states that 'approximately 300 men absented themselves for two days, causing a loss of 1.5% of production'. Source B, the local newspaper owned by a mine shareholder, reports 'a mob of 2,000 natives terrorising the town'. Source C, an oral interview recorded in 1975 with a retired worker, says 'the whole compound emptied - nobody went underground that day'. She cannot obtain police records before the deadline. Her teacher has told her to use all three sources and to avoid simply choosing the 'best' one.
1. Which source is most useful for the essay question, and which is most reliable? Explain why these are two different judgements.
Read the worked answer
Usefulness depends on the question; reliability depends on the source's relationship to what actually happened.
For 'scale', Source A is the most useful: it offers a figure (300) and a measurable consequence (1.5% of production) that could be checked against output records. That does not make it reliable. An annual report is written to reassure shareholders, so absences may be minimised, and 'absent' may mean only men who failed to go underground, excluding those who stayed in the compound in support.
Source B is the least reliable for scale: 'a mob of 2,000' is a rhetorical figure, and the paper's owner had a financial stake in presenting strikers as a threat. But B is highly useful for a different question - how the strike was represented to the white reading public.
Source C is useful for meaning and experience ('the whole compound emptied') rather than number, and it was recorded 38 years later, so it is shaped by memory and by the politics of 1975.
A strong answer separates three questions: how many, how it was seen, and what it meant.
2. How should she handle the conflict between the figures in the essay itself?
Read the worked answer
Do not average the figures and do not silently pick one.
Explain the range and its causes. 300 is best treated as a minimum, coming from an employer with an interest in minimising disruption. 2,000 is best treated as a maximum, coming from a hostile observer with an interest in exaggeration. Then note that the two numbers may not measure the same thing: adult male employees who stopped work, everyone present in the compound (including families), and people in the town are three different populations.
A defensible sentence would be: 'Between 300 and 2,000 people took part; the company figure almost certainly counts only adult male employees, while the newspaper figure probably includes women and children who gathered in support.'
Then use Source C to establish that the stoppage was remembered as total, which matters for its significance even if it cannot fix a number. This demonstrates reasoning about provenance and purpose rather than listing sources.
3. She has space for one more piece of research. What should she look for, and what would it allow her to claim?
Read the worked answer
The most valuable addition would be a source that fixes the denominator: a payroll or compound register for 1937.
If the mine employed, say, 900 African workers, then '300 absent' means one third and '2,000' becomes impossible for the workforce alone. A compound register, a government labour report, or a missionary diary recording the number of men who attended a meeting would each let her convert percentages into people.
A second-best option is a source from a different perspective with no stake in the dispute - for example, a store ledger showing unusually heavy purchases of maize meal during the strike days, which would indicate how many people were being fed outside the mine kitchens.
What she should avoid is another opinion source. Adding a fourth commentator does not resolve a numerical question; adding an administrative record does. The choice of research should be driven by the type of uncertainty, not by the desire for more evidence.