Mean and median
Mean = sum of values ÷ number of values. Median = middle value after sorting; with an even number, average the two middle values. For incomes 10, 12, 13, 15 and 50 (thousand dollars), the mean is 20 and the median 13. A high outlier pulls up the mean, so compare both when judging typical living standards.
Quantiles
Quantiles divide an ordered distribution into equal-sized groups. Quartiles form four groups, quintiles five and deciles ten. The bottom income quintile is the lowest-income 20% of people or households; its share of total income need not be 20%. State whether data refer to people or households and follow the source’s grouping convention.
Weighted index numbers
Price relative = current price ÷ base-period price × 100. Weighted index = sum of (weight × price relative) ÷ sum of weights. If food has weight 60 and index 110, and transport weight 40 and index 120, the combined index is (60 × 110 + 40 × 120) ÷ 100 = 114. The basket costs 14% more than in the base period. Use expenditure weights, not a simple average unless weights are equal.
Inflation from an index
Inflation = (new index − old index) ÷ old index × 100. If an index rises from 114 to 119.7, inflation is 5%, not 5.7%. Index-point change and percentage change are different; compare consistent baskets and time periods.
Seasonally adjusted data
Seasonal adjustment removes estimated recurring calendar-related patterns, such as holiday shopping, to make underlying changes easier to compare. It does not remove the business cycle or every irregular shock. Compare like with like: seasonally adjusted quarter-on-quarter growth should not be compared directly with an unadjusted annual change.
Reading data accurately
Check the unit, period, source and whether values are nominal, real, total or per person. A rate rising from 4% to 6% increases by 2 percentage points, or 50%. For grouped income data, distinguish a group’s share of people from its share of income. Correlation alone does not identify causation.