Employment for Information: Newspaper, Periodical, Book, and Directory Publishers (NAICS 5111) in the United States (IPUJN5111W010000000) | FRED | St. Louis Fed
- Rows
- 39
- Columns
- 2
- Data range
- 1987-01-01 to 2025-01-01
Graph and download economic data for Employment for Information: Newspaper, Periodical, Book, and Directory Publishers (NAICS 5111) in the United States (IPUJN5111W010000000) from 1987 to 2025 about periodicals, book, printing, information, NAICS, IP, employment, and USA.
AI analysis
Dataset Analysis: Employment in Newspaper, Periodical, Book, and Directory Publishing (NAICS 5111)
1. Dataset Overview and Context
This series, sourced from the Federal Reserve Bank of St. Louis (FRED) via the St. Louis Fed's public API (fredgraph.csv endpoint), tracks employment levels in the traditional print publishing sector — newspapers, periodicals, books, and directories — from 1987 through 2025. As an index-based series (with an implied base period of 100), it captures nearly four decades of structural change in this industry, including the pre-internet peak, the digital disruption of the 2000s–2010s, and post-pandemic conditions. Despite being tagged under "Healthcare Prices" in this catalog (likely a topic-tagging artifact rather than a substantive link), the actual content is an industrial employment index, making it more naturally suited to labor economics, media industry, or macroeconomic structural-change studies. Its long time span and consistent annual/monthly cadence from a trusted government-aggregated source give it solid value as a covariate in correlation studies examining employment trends, automation, or industry decline patterns.
2. Data Quality Observations
The dataset is clean on the surface: zero null cells across both columns, and the join key (observationdate) has zero nulls with 39 distinct values — one per row, confirming no duplicate timestamps. The value column has 38 distinct values against 39 rows, meaning exactly one repeated value, which is immaterial and easily explained by a plateau period or rounding rather than a data quality defect. The duplicate-row check itself is flagged as pending ("Phase D — recomputed against Parquet"), so full deduplication confidence isn't yet certified — this should be verified before treating the dataset as authoritative for merges. Overall, given the small size (39 rows) and FRED's curation standards, this is a high-quality, analysis-ready series once duplicate verification is finalized.
3. Key Column Distributions
The value column (IPUJN5111W010000000) ranges from a minimum of 60.618 to a maximum of 222.478 — a nearly 4x spread — reflecting the dramatic contraction (or earlier expansion, depending on direction) of employment in this sector over time. The mean (161.011) sits below the median (184.284), and the distribution shows negative skew (-0.461), indicating a longer left tail: a cluster of higher historical values with a drag toward lower recent values, consistent with the well-documented decline of print publishing employment in recent decades. The interquartile range (Q1=110.46, Q3=217.089) spans over 100 points, underscoring high variance relative to a 39-observation sample — this is a series with a strong secular trend rather than stationary noise, which matters for correlation/regression design (differencing or detrending will likely be necessary to avoid spurious correlation).
4. Recommended Join Keys
The observationdate column is the clear and only sensible join key — it's a clean date field with zero nulls and full distinctness. For cross-dataset correlation, align on year (or year-month, if higher-frequency data is available) and consider resampling/interpolating this series if paired with monthly or quarterly datasets, since the observed cadence and exact frequency (annual vs. monthly) should be confirmed from the raw timestamps.
5. Suggested Pairings for Correlation Analysis
Strong candidates include: (a) other FRED industry employment series (e.g., NAICS codes for broadcasting, internet publishing/web search NAICS 519130, or advertising) to study labor substitution across media formats; (b) digital advertising revenue or internet usage/broadband adoption indices to test the "digital disruption" hypothesis; (c) CPI for newspapers/periodicals or producer price indices for publishing to link employment to pricing power; (d) macro indicators like real GDP, PPI, or overall nonfarm payrolls as a baseline for detrending; and (e) since this is oddly tagged under "Healthcare Prices," it would also be worth testing against healthcare CPI or PCE health series purely to confirm whether that tag reflects a genuine, non-obvious macro correlation or is simply a misclassification.
Columns
- observation_date (date)
- IPUJN5111W010000000 (decimal)