Compare the Stories

Establish the shared record

The employees and managers agree that these documents are accurate:

Each period covers six months. The count of people who quit doubled from four to eight. A resignation rate would compare that count with the number of employees. We do not know the workforce size or how it changed, so we cannot calculate the rate.

Several facts are missing. How many people quit in earlier years? When did each person decide to leave? Did the office rule apply to them? Where did they go, and why? What else changed at the company? The two sides agree on the supplied facts. They disagree about the cause.

State the explanations fairly

Story A: The office policy contributed to the increase.

Perhaps some employees left because they disliked the office rule. The timing fits: more people quit after the rule was announced and took effect. But this story assumes that the rule applied to some of those people and affected their choices. The records do not show either point.

Story B: The competitor's recruitment contributed to the increase.

Perhaps better job offers drew people away. The other company advertised similar jobs before the period when more people quit. But this story assumes those employees learned about the jobs and acted on them. An advertisement alone does not show that anyone was contacted or hired.

Check: what would count against each account?

For Story A, ask when the employees decided to leave. Suppose dated records showed that all eight had firmly decided before they could have known about the office rule. That would count against the claim that the rule caused those choices. Check whether anyone heard about it before the public announcement. June 1 alone cannot settle that question.

Other evidence could weaken the story without settling it. Reliable accounts from departing workers might point to earlier, unrelated reasons. But silence about the policy in an exit interview is less helpful. An interview may not capture every reason someone left.

For Story B, check where the eight people went. If none joined the competitor, we could reject the claim that this company hired them away. Other job opportunities might still have mattered. That broader claim would need separate evidence.

Also check when they learned about the hiring campaign. If all eight had already decided to leave, the campaign would not explain those decisions. Here too, check for private contact before the public advertisement.

We do not have these records. The exercise shows what to look for. It does not claim we have found it.

Compare: can both stories fit?

An employee might dislike the office rule and get a good job offer. Both reasons could matter. Another person might leave for reasons linked to neither event. These are two possible explanations, not the only two.

Next, look for earlier departure records and workforce counts. Find out who had to return, when people gave notice, where they went, and why they say they left. Compare those covered by the rule with those who were not. Check whether their jobs or other conditions also differed. Those differences could affect the result.

Keep each finding within its limits. One confirmed hire tells us where that person went. A record of contact and the employee's own account could help explain why. They cannot explain every departure. The same applies to someone who names the office rule: their account is evidence about their choice, not everyone's.

Calibrate: what can we say now?

Four employees quit in the first six months; eight quit in the next six. The office rule and the competitor's job advertisement both came before the second period. We do not know why more people quit or how much either event mattered. Without workforce counts, we also cannot tell whether the share of employees who quit increased.

Next useful question: Did more people quit among those covered by the office rule? Check when they decided, why they say they left, and where they went. Look for records of contact with the other company too.

We now know more clearly what the disagreement is. We have not yet settled it.