Can we believe the majority of biological sciences research?

Isabel Winney
8th December 2016

Can we believe the majority of biological sciences research?

  • Yes?
  • No?

Yes

  • Highly skilled researchers
  • Peer review
  • Robust scientific method

Yes but...

  • Highly skilled researchers
    • Who are self-taught
    • Who love to address new topics and use new techniques
    • Who are under huge pressure to publish, get grants, survive
  • Peer review
    • Is full of bias against certain labs / techniques / new findings / old findings
    • Is slow
    • Publications can be corrected, but most are not (from my friend the meta-analyst)
    • Corrections are viewed very negatively, then forgotten
  • Robust scientific method

Do we spot our mistakes?

Easily?

Not easily?

Publication bias

Competition

Lack of blinding

In the news this week

In the news this week

Misuse of p-values

Misuse of p-values

Cognitive bias

  • Selective memory
  • Too much information
  • Positive perception
  • Limited time
  • The God complex
  • Self-deception

Researcher degrees of freedom

HARKing

Model selection (whole other debate)

Increased false-positive rates

Short detour on computing

Formal teaching in computing? Excel???

http://mikecroucher.github.io/MLPM_talk/

Can we believe the majority of biological sciences research?

  • Yes?
  • No?