A large steel corporation conducted an experiment to compare the average iron contents of two consignments of lumpy iron ore. In accordance with industrial standards, n increments of iron ore were randomly selected from each consignment and measured for iron content. From previous experiments, it is known that iron contents vary over a range of roughly 3%. How large should n be if the steel company wants to estimate the difference in mean iron contents of the two consignments correct to within .05% with 95% confidence? To obtain an approximate value for σ₁and σ₂, set σ₁ = σ₂ = σ,  and set Range = 4σ. Then 3 ≈ 4σ and σ ≈ 3/4.