STATS TESTS- KEY NOTES

• • Are results significant? 
  • If not- results may be due to chance
  • Use of a probability level- e.g. (0.05- less than 5% chance that results are due to chance, result is likely to be due to the IV so statistically significant)
  • Drug trials -> stricter probability level -> 1% /0.01
  • Statistical test calculation done- (calculated / observed value) this is compared with a table of critical values (must be greater than / lower than critical value- significant result)
• KEY POINT:

  • ”R” in name
  • CALCULATED value must be GREATER THAN / EQUAL TO the CRITICAL value to be SIGNIFICANT
  • No “R” - CALCULATED value must be EQUAL TO / LESS THAN the CRITICAL value for SIGNIFICANCE
• HOW TO WORK OUT CRITICAL VALUES:

  • Is the hypothesis directional, non-directional?
  • Number of ppts in the study (N) / degrees of freedom (df)
  • Level of significance (normally 0.05 unless stated otherwise)
• PROBABILITY / SIGNIFICANCE:


  • Stats tests can be used to accept the experimental OR null hypothesis 
  • 0.05 / 5% significance level used in psychology (95% chance of significance- acceptable level)
• TYPE 1 ERROR:


  • Mistakenly reject null & accept alternative hypothesis
  • Significance levels too lenient (too high- e.g 10% rather than 5%)
  • ”False hope”- think there is significant difference / correlation, but there isn’t one
• TYPE 2 ERROR:


  • Null hypothesis accepted, alternative rejected
  • Significance level too strict e.g. 0.01 instead of 0.05