You will assess a real causal claim from news, a report or your workplace.
The HR newsletter at Daniel's company leads with a cheerful headline: "Mentoring works: mentees promoted 40 percent faster." The article says everyone should sign up for the next intake. Two of Daniel's colleagues have already forwarded it with "worth doing" attached. Daniel is not against mentoring. He just wants to know whether the programme makes people get promoted faster, or whether people who were going to be promoted anyway are the ones who sign up.
This exercise takes about thirty minutes and pulls together everything in this module. You will need one causal claim, a sheet of paper or a spreadsheet, and the four alternatives from lesson 4.1, Three other reasons two things move together.
Choose a real claim that says, or implies, that one thing causes another. Health articles, management books, company reports, internal newsletters and policy announcements are full of them. Words like "boosts", "leads to", "improves", "linked to" and "the secret of" are signals. "Linked to" is worth special attention, because writers often use it to suggest a cause while staying technically safe.
Write the claim in the form "X causes Y", as specifically as you can. "Mentoring helps careers" is too vague to judge. "Joining the company mentoring programme causes employees to be promoted sooner than they otherwise would have been" can be checked.
Set up your sheet with five rows: the claim, the evidence, alternative explanations, what would settle it, and the verdict.
In the evidence row, write down what the claim rests on and answer one question: is it observational or experimental? Observational evidence watches what people chose to do. Experimental evidence comes from a study where someone else, ideally chance, decided who got the treatment, as lesson 4.2, Why experiments beat observation, explained.
Note the size of the group, how the comparison was made and what was left out. If it is a percentage, check whether it is relative or absolute using lesson 3.3, Percentages, risk and the base rate.
Here is Daniel's evidence row, with figures from the made-up newsletter. Mentees in the last three intakes were promoted after an average of 2.1 years, compared with 3.5 years for other staff. That is the "40 percent faster" in the headline: 1.4 years shorter, divided by 3.5. The evidence is observational, because staff chose whether to apply, and managers chose who got in. The article does not say how many mentees there were or whether the two groups were in similar roles and levels.
Now go through the four alternatives one at a time and write at least one specific candidate for each.
For confounders, ask what kind of person would be more likely both to join and to be promoted. Daniel writes: ambition, since ambitious staff are more likely to apply and more likely to push for promotion; manager support, since managers who back their staff may both nominate them and promote them; and role, since some departments promote faster and may also encourage mentoring.
For reverse causation, ask whether the outcome could have led to the supposed cause. Daniel writes: staff already being lined up for promotion may be encouraged to join the programme to prepare.
For selection, ask how the groups were chosen. Daniel writes: applicants were selected by managers, who may have picked people they already rated highly. And the comparison group "other staff" includes people on contracts or in roles with no promotion path.
For chance, ask how big the groups are and how many things were compared. Daniel writes: size of intakes unknown, and if they are small, a few fast promotions could move the average a lot.
In the next row, write what evidence would make the claim more or less likely. Daniel's list: compare mentees with similar staff in the same roles and grades who applied but were not taken, or run the next intake by lottery among eligible applicants. Even a simple check, comparing mentees with colleagues of the same grade and manager rating, would help.
Then give your verdict, using one of three labels.
Likely causal means there is experimental evidence, or strong observational evidence where the main alternatives have been checked and ruled out. Possible means the link is real and a causal story is plausible, but the alternatives have not been ruled out. Not shown means the evidence cannot tell causation apart from the alternatives, or the link itself is doubtful.
Daniel's verdict: "Possible. The link is probably real, but ambition, manager support and selection by managers could explain all of it. A lottery for the next intake, or a comparison with applicants who were not taken, would tell us much more." He adds that this does not mean mentoring is useless, only that the newsletter has not shown how much it helps promotion.
A finished sheet has the claim in "X causes Y" form, the evidence described as observational or experimental with its size, at least one specific candidate for each of the four alternatives, what would settle it, and a verdict with two or three sentences of reasons.
Choose your claim now, preferably one people around you already believe. Other people tend to spot confounders you missed, which is why the activity below ends with a colleague reading your verdict.
Complete the causal claim worksheet for one real claim and share your verdict with a colleague for a second opinion.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).