7 Testing For Differences Between Groups Flashcards
Type 1 error
Detecting an effect that isn’t present
Type 2 error
Failing to detect an effect that is present
P>0.05
Stick to null hypothesis (no relationship)
P<=0.05
Accept alternative hypothesis
One sample t test
One variable
Ordinal/interval/ratio
Independent sample t test
Two variables
Nominal by ordinal/interval/ratio
If t value <1.89 = not statistically significant
Lavines test significance <0.05 = not significant
Chi square
Two variables
Nominal by nominal
Adjusted standardised coefficients need to be >= 2 to designate cells with significant differences
Paired sample t test
2 variables
Nominal by ordinal/interval/ratio
Test if difference in similar ratings for same respondent
ANOVA
2+ variables
Nominal by ordinal/interval/ratio
Check ANOVA sig table first
If Lavines >0.05 look at bon Feroni table