Produces a comprehensive data quality report covering digit preference, completeness, plausibility, and sampling weight diagnostics.
steps_data_quality(raw, cleaned, cols)The raw (pre-cleaning) data frame, typically from
import_steps_data().
The cleaned data frame from clean_steps_data().
Column mapping list from detect_steps_columns().
A list of class "steps_quality" with elements:
Terminal-digit tables and heaping indices for physical measurements (SBP, DBP, height, weight, waist).
Per-variable missingness counts and percentages, grouped by STEPS domain.
Summary of values outside plausible ranges.
Sampling weight distribution statistics.
Digit preference / heaping is assessed using the Whipple-style heaping index: the ratio of observed frequency at a digit (0 or 5) to the expected frequency under uniform distribution. An index of 1.0 = no preference; >1.5 = moderate heaping; >2.0 = severe.
Completeness reports missing values for key STEPS variables grouped by Step (behavioural, physical, biochemical).
Plausibility counts values outside WHO-recommended ranges (e.g. height 100–250 cm, weight 20–300 kg, SBP 60–300 mmHg).
Weight diagnostics summarise the distribution of sampling weights and flag potential issues (high CV, zero/NA weights).