MAGIC
Five factors that need to be considered for responsible communication of research on sex/gender and the brain
WHY THIS MATTERS
Outdated and exaggerated claims about sex/gender and the brain are far too common
Problems often start with a disconnect between the strength and nature of research findings and the impression created by a narrative.
The risks are growing as a result of mandates to use sex as a biological variable in relevant research and an emphasis on ‘impact’ as a measure of research(er) success.
GIVE US YOUR FEEDBACK
What do you think?
How can we develop the guidelines and ensure they are put into practice?
Notes
1. We use the term ‘Sex’ to refer to a set of biological attributes associated with physical and physiological features including chromosomes, gene expression, hormone function, and reproductive/sexual anatomy; ‘Gender’ to refer to socially constructed roles, behaviours and identities of female, male and gender-diverse people; and ‘Sex/Gender’ to indicate the entanglement of an individual’s biological sex with psychological
and social attributes of their environment.
2. In describing the size and nature of statistical findings authors should:
describe the effect size in a way that allows readers who are unfamiliar with normal distributions to get a sense of the extent of similarity as well as difference. Without this, non expert readers are likely to imagine two non-overlapping groups, or a generalisable difference as big as that seen in height (d~1.7) where men are usually, albeit not always, taller than women. Particular caution is needed where a sex/gender ‘difference’ is closer in size to that seen in the incidence of left-handedness (d~0.1) to ensure readers understand that this means no broad sex/gender generalisations can be drawn.
explain how serious the risk of ‘false positives’ is, given the proportion of the sex/gender comparisons made that passed a test of statistical significance
3. The guidelines are drawn from Robert P. Abelson’s MAGIC framework for organising a principled argument from quantitative evidence
(Abelson, R. P. Statistics as principled argument, 1995, Hillsdale). They are not intended to prevent fraud or disinformation, nor to set blanket
restrictions on how research is communicated.