AI Analysis: Usual hours worked by occupation, monthly, unadjusted for seasonality

Executive Summary

Statistics Canada's Table 14100422 provides nearly 40 years of monthly data (1987–2026) on usual hours worked across 59 occupational categories, 11 geographic regions, and 3 gender groups, capturing the breadth of Canadian labour market structure. The dataset is heavily right-skewed, with a mean VALUE of 502.7 versus a median of just 11.8, driven by large aggregate categories that dwarf specific sub-occupational records. Despite this wide variation, average hours worked have remained remarkably stable over time, with a near-zero correlation (r = 0.002) between date and value.

Key Findings

• The dataset contains 158,795 valid records spanning January 1987 to February 2026, with approximately 36.5% of VALUE entries missing — a pattern typical of detailed cross-tabulated labour force survey data.
• VALUES range from 0.2 to 496,686.5 thousand persons, with a standard deviation of 6,322, reflecting the mixing of two distinct units of measure: 'Hours' (mean 1,642, median 43.5) and 'Persons in thousands' (mean 35.38, median 4.5).
• 12.8% of records (20,265 out of 158,795) were flagged as statistical outliers using a 3×IQR threshold, with outlier values exceeding 158.1 thousand persons, concentrated in broad national-level and aggregate occupational categories.
• Gender-based analysis shows Men+ average 486.91 versus Women+ at 331.18, with Women+ also recording a slightly lower median (10.6) compared to Men+ (12.3), suggesting persistently lower hours or employment counts for women across occupational categories.
• 59 unique NOC occupational categories are represented, and when aggregate 'Total' categories are excluded, the ranking reveals stark contrasts between the largest employing occupations and the smallest or most specialized sub-groups.
• Correlation analysis found virtually no relationship between date and VALUE (r = 0.002), indicating that average usual hours worked have been structurally stable across the nearly four decades covered by the dataset.
• Full-time workers in the 35+ hours categories exhibit the widest distributional spread and the most extreme outliers, as shown in box plot analysis, while narrow sub-occupations or province-level records consistently produce the smallest non-zero values.