Create two observation cards

Card A was captured at 10:10 Eastern and says its data is current through 10:10. It reports 1.20 million ABC shares traded from 9:30 through that time. Its displayed price change uses a $50 opening reference. Record the provider, symbol, capture time, data time, session definition, units, and reference alongside the number.

Card B was also captured at 10:10 Eastern, but its data label says 10:05. It reports 1.70 million ABC shares traded from 4:00 through 10:05. Its price change uses a previous-close reference of $48.50. In this fictional packet, those definitions are documented. In a real exercise, you would need the provider's actual labels or documentation.

Separate capture time from data time

The screenshots were saved together, but their observations end five minutes apart. A common capture time does not make their underlying data simultaneous. Mark the pair as temporally unmatched before calculating any difference. The appropriate follow-up is to obtain comparable records through one chosen endpoint, if the sources make them available.

Do not replace a missing data time with your computer's clock. That would create a fact the source did not provide. If the timestamp's meaning is unclear, record the exact label and leave the timing status unresolved until you can establish whether it refers to a trade, quote, calculation, or page refresh.

Match the measurement window

Card A starts counting at 9:30, while Card B starts at 4:00. Subtracting 1.20 million from 1.70 million gives 500,000, but you cannot label that difference premarket volume. The endpoints differ too, and the packet has not established whether the sources include the same underlying records.

A defensible comparison needs matching start and end times, the same units, and compatible coverage definitions. Write the missing request precisely: “Obtain each source's ABC share volume from 9:30 through 10:05 Eastern, with the documented coverage.” If that information is unavailable, the honest result is that the totals are not yet directly comparable.

Recalculate a percentage on one stated basis

For a separate arithmetic step, suppose you obtain a verified common endpoint price of $51.50 in the fictional packet. Using the $50 opening reference gives ($51.50 minus $50) divided by $50, or 3%. Using the $48.50 previous-close reference gives $3 divided by $48.50, or approximately 6.19%.

Both calculations can be correct because they answer different questions. One measures change from the chosen opening reference; the other measures change from the chosen previous close. The 3.19 percentage-point difference between the rounded returns is not evidence that one provider overstated the same return.

If your question concerns change from the opening reference, calculate both comparisons from that reference and the same endpoint. Label the resulting calculation as your normalized comparison. Do not present it as a corrected vendor value unless you have established that the vendor intended to report that same measure.

Give each question a clear status

Use three statuses in your report. “Comparable” means the important definitions match. “Reconciled” means you can explain a difference from documented inputs. “Unresolved” means a necessary definition or observation is missing. These are labels for your investigation, not ratings of the providers' overall quality.

In this example, the original volume totals remain unmatched in time and session coverage. The percentage difference is reconciled after the common endpoint and two reference prices are specified. You have resolved one question while leaving another open. That is a more precise result than declaring one entire screen trustworthy and the other incorrect.

Finish with a reproducible comparison note

Write: “The screenshots share a capture time but have different data times. Their volume windows also differ, so I cannot attribute the 500,000-share gap to one cause. Their percentage changes use different reference prices; recalculating from the stated inputs explains that difference.” Attach the observation cards and arithmetic.

For your next study session, compare two observations that already share a defined question and measurement window. Keep the source definitions with your notes. The goal is a comparison someone else can reproduce, with uncertainty remaining visible wherever the available evidence does not settle it.