LogFAQs > #910788566

OK so there is a wealth of studies that have found that there is a significant bias of choosing the first option (primacy) and last options (recency) in a long list. Worth googling if interested.

Though the conditions tested in these studies are not exactly the same as in this example, funamentally they are similar. So when talking about a series of A/B choices in a horizontal line, under certain conditions there may be as much as 25% bias (primacy) and 20% (recency) when the cohort is choosing between similar* options. Ie if 10% of people are indifferent it could have as much as a 2.5-3% impact.

* Similar being cases where the responder is indifferent between the choices, regardless of how different they may be, or the choice itself is similar

I believe this could tip the balance in close battles when they are 1st or 4th in this 4 a-day format, and would also impact the 3 and 2 a day formats in lesser degrees.

To eliminate this bias it would actually be well worth having two versions of the voting grid and either randomising which version is hit or switching them after 12 hours.

Looking through the potentially close matches of 1v16, 2v15, 3v14, and 4v13 (which will be first or last of the four of the day), in R1 it may help:

Alucard>Peach
Bomberman>Kazuma
Bowser>Charizard
Aerith > Waluigi
Riku > Bayonetta
Claire Redfield>Ren
Tails>Drake
Quiet>Aqua

Might be unrealistic to expect a significant number of people to be indifferent about Aerith vs Waluigi, but I think you'd get a fair few on some of the others.