PROBABILITY AND SIGNIFICANCE

• WHAT IS A LEVEL OF SIGNIFICANCE?

  • value psychologists are happy to accept
  • also known as the P value
  • probability that their result is due to chance and not manipulation of variables
  • shows likelihood that the observed difference would occur if there was no real difference between the conditions/ relationship between variables
• MEANING OF DIFFERENT P VALUES?

  0.1

  • 10% due to chance
  • 90% confident result is a real effect

  0.05

  • 5% due to chance
  • 95% confident result is a real effect

  0.01

  • 1% due to chance
  • 99% confident result is a real effect

  0.001

  • 0.1% due to chance
  • 99.9% confident result is due to chance
• WHY DO PSYCHOLOGISTS USUALLY USE 0.05?

  strikes balance between likelihood of a type one error or type two error
• WHAT CAN BE USED INSTEAD OF 0.05?

  0.01

  -> when more certainty required (e.g. when testing life-altering drugs)
• WHAT ARE THE TWO TYPES OF ERROR?

  one and two
• WHAT IS A TYPE 1 ERROR?

  • false positive
  • wrongly reject null hypothesis
  • found a difference/ relationship when there was not one
• WHEN ARE TYPE 1 ERRORS LIKELY TO OCCUR?

  • in levels of significance that are too lenient, e.g. 0.1
  • as confidence is so low (90%) that random effects may be perceived as real ones
• WHAT ARE TYPE 2 ERRORS?

  • false negatives
  • wrongly accept null hypothesis
  • found no difference/ relationship when there was one
• WHEN ARE TYPE 2 ERRORS LIKELY TO OCCUR?

  • in levels of significance that are too stringent, e.g. 0.001
  • as confidence is so high (99.9%) that real effects may be missed