Report Your Actual Data

This example underscores a crucial principle in data collection and analysis: report your data as you actually find it, without smoothing or 'fudging.'

The graph shown is a mock-up, but it's based on a real experience. Years ago, my company was contracted to collect, characterize, and analyze customer financial data for a client in the financial services sector. Understanding customer demographics is essential for financial services, especially when identifying where most assets reside. In this case, the financial power was concentrated among baby boomers, with a peak around the birth year of 1955.

When we plotted the collected data, we noticed a sharp spike at 1955 that seemed anomalous. Nevertheless, I adhere to a policy of “playing it as it lies” — reporting findings as they appear, without embellishment. We presented the graph as it was, odd as it looked.

Before finalizing the report, we investigated the anomaly and discovered the cause: data entry staff, when uncertain of a birth year, were instructed to enter '55' in the two-digit year field. By reporting this faithfully, we provided our client with important actionable insights:

1. Their data collection process had gaps, including missing or erroneous entries.
2. By estimating the excess in the spike, we could approximate an error rate.
3. Our honest reporting demonstrated the integrity of our analysis.

For those trained to collect, analyze, and report data correctly, this may seem obvious. However, the practice of reporting raw, unaltered data has declined in recent decades, and, in many cases, raw data is no longer accessible. Reporting faithfully remains essential for transparency and trustworthiness in analysis.