Unit 06.04: Measure, size, comparison, period

The number is the easy part. The sentence around it is where analyses go wrong.

Name the measure, the size, the comparison and the period

An honest sentence about a result contains four things: what was measured, how big it was, what it is being compared with, and over what period. Drop any one and the sentence starts claiming more than the data supports.

The most common overreach is causal. 'Highest demand' asserts a cause for the highest recorded rate; the data shows the rate, not the demand behind it — and certainly not what should be done about it.

The second most common is omitting the period. A rate over two months and a rate over two years read identically in a sentence and mean very different things.

This block writes the same finding twice.

suppressPackageStartupMessages(library(dplyr))

by_ward <- data.frame(ward = c("North", "South", "East"),
                      months = c(3L, 2L, 1L),
                      per_1000 = c(34.9, 40.3, 33.2))

top <- by_ward |> filter(per_1000 == max(per_1000))
gap <- round(max(by_ward$per_1000) - min(by_ward$per_1000), 1)

cat("The numbers:\n")
print(by_ward)

cat("\nOverclaiming:\n")
cat("  \"", top$ward, " has the highest demand and needs more staff.\"\n", sep = "")

cat("\nWhat the data supports:\n")
cat("  \"", top$ward, " recorded the highest visit rate at ", top$per_1000,
    " per 1,000 residents,\n", sep = "")
cat("  ", gap, " above the lowest ward, over ", top$months,
    " months of returns.\"\n", sep = "")

cat("\nThe difference: the second names the measure, the size, the comparison\n")
cat("and the period, and it does not claim a cause the data cannot show.\n")

The overclaiming version — 'South has the highest demand and needs more staff' — makes two leaps: from recorded visits to demand, and from demand to a staffing decision. The supported version names the measure (visits per 1,000 residents), the size (40.3), the comparison (7.1 above the lowest ward) and the period (2 months of returns). It is longer, and everything in it can be checked against the table above it.

The mistake this prevents

The mistake is writing the conclusion you expected and attaching the number to it. Read your own sentence back and ask which word the data does not support.

Takeaway

Write findings with all four elements present. Reserve causal language for when you have evidence of a cause, and state the period every time.

Mark complete