Acts of kindness: one review in detail

Step 1 Warm-up recall

From M4.L1: write the formula for converting r to d, and say why publication bias matters.

Your answer

Private: nothing you type here is sent or stored.

Step 2 Learning goal

By the end of this lesson you can:

• summarise Curry and colleagues (2018) on acts of kindness (Foundation);
• translate d = 0.28 into a plain-language chance of superiority (Core);
• say what an author’s “no indication of publication bias” does and does not guarantee (Core).

Step 3 A familiar example

Someone tells you they bought a stranger a coffee and felt great all afternoon. One happy story is not a trial. A review asks what happens, on average, across many people who did something kind compared with those who did not.

Step 4 Clear explanation

Curry, Rowland and Van Lissa (2018) searched 489 articles and included 24 articles describing 27 studies, with 4,045 participants in total, giving 52 effect sizes. They concluded that the overall effect of performing kind acts on the well-being of the person who acts is small to medium (d = 0.28). They reported no indication of publication bias, and found that the effect was not moderated by sex, age, type of participant, intervention, control condition or outcome measure.

What does d = 0.28 mean in plain language? Imagine picking one treated person and one control person at random. The common-language effect size is the chance that the treated person has the higher score. It is about 58% for d = 0.28, against 50% for no effect. The value comes from the normal curve: it equals the area below d ÷ √2.

The authors’ statement about publication bias is a finding of their checks, not a guarantee. Checks have limited power when there are few studies, and the 2019 reanalysis in M4.L1 shows how contested such conclusions can be for the wider field.

Step 5 Worked example (4 marks)

A kindness review reports d = 0.28 from 4,045 people.

1. Chance of superiority ≈ 58%. (1 mark)
2. That is eight points above a coin toss: a real but modest shift. (1 mark)
3. A person in the control group will often do better than a person who acted kindly, because the groups overlap heavily. (1 mark)
4. A fair statement: acts of kindness are linked to a small-to-medium average improvement in well-being for the person acting, with uncertainty about size and about whom it helps most. (1 mark)

Step 6 Learner practice (instant-marked)

Step 7 Recap

• Curry and colleagues (2018): 27 studies, 4,045 participants, small-to-medium effect (d = 0.28) of kind acts on the actor’s well-being.
• d = 0.28 is a chance of superiority of about 58%.
• “No indication of publication bias” is a result of checks, not a guarantee.

Step 8 Exit ticket (ungraded)

Explain d = 0.28 to a friend in two sentences, without using the letter d.

Your explanation

Private: nothing you type here is sent or stored.

Step 9 Go further (Extension)

Read the abstract of Curry, Rowland and Van Lissa (2018) in the Journal of Experimental Social Psychology. Find the limitations they discuss.