Annuaire IPHC

DEPE | Séminaires BEEPSS / BEEPSS Seminars » CANCELED | How research on mate choice in zebra finches reveals weaknesses (...)

CANCELED | How research on mate choice in zebra finches reveals weaknesses in our scientific method

Dernière mise à jour : jeudi 12 mars 2020, par Catherine Berger

Presented by : Wolfgang Forstmeier, MPI Seewiesen
Date : Thursday 26th March 2020, 13:00-14:30
Place : Amphithéàtre Grünewald, Bâtiment 25

The zebra finch is one of the most frequently studied organisms in terms of its mate choice, which is reflected in more than 150 published empirical studies. However, after studying zebra finch mate choice in several captive populations (both domesticated and recently wild-derived) for 15 years, I here argue that most of the published literature is misleading because of psychological biases (such as confirmation and publication bias). I believe that this is because we have started our research under the plausible assumption that both sexes are choosy and that both sexes recognize and prefer high-quality individuals over low-quality individuals. Yet this assumption has never been tested in a rigorous and unbiased way, but rather we have naively interpreted chance findings of p<0.05 that arise in nearly every exploratory study of multiple traits in a way that is consistent with our assumption (confirmation bias). More rigorous tests, that give the null hypothesis a fair chance, show surprisingly little evidence that zebra finches can recognize and prefer high-quality individuals. Moreover, zebra finches show remarkably little consensus among individuals on who is the most attractive. I argue that such consensus may not have evolved in this lifetime monogamous species, maybe because the sum of all costs (including costs of competition for the most attractive individual) were larger than the sum of all possible benefits (good genes, attractive offspring, and good parent benefits). This example shows that our science can be highly inefficient in cases where the null hypothesis is actually true, mostly because researchers are intrinsically biased against the null hypothesis due to psychological and monetary rewards. For a more efficient science we need to change both our methods and the incentives given to researchers.

www.orn.mpg.de/forstmeier