Sensitive Experiments

Question 1

What is a sensitive experiment?

Answer 1

one that is likely to detect the effect of an independent variable on the dependent variable when an effect actually exists.

Question 2

What is power in a sensitive experiment?

Answer 2

• an index of the sensitivity of an experiment, in terms of the statistical test used to assess the effect of IV on the DV
• probability of detecting a statistically significant effect when an effect actually exists.

Question 3

How do you increase sensitivity?

Answer 3

• maximising the effect of IV
• reducing error (unexplained) variance due to measurement error
• uncontrolled variability in the experiment
• individual differences

Question 4

How is sensitivity increased by maximising the effect of the IV?

Answer 4

• use a powerful and reliable IV manipulation
• simple behaviour through an adequate range on DV

Question 5

How is sensitivity increased by reducing error?

Answer 5

• use precise measuring tools
• use repeated measures design (same participants are exposed to all levels of the IV so that consistent individual differences that impact the DV can be removed by error term)
• use blocking designs (an individual differences variable known to influence scores on the DV is included as a factor in the research design)
• use analysis of covariance (ANCOVA) (participants are measured on individual differences variable then randomly assigned to the experimental conditions without regards of these score to control the covariate)

Question 6

What occurs when the indivdual differences covariate is an additional IV?

Answer 6

the sum of squares does nt change but the error term does

Question 7

What does an ANCOVA conceptually do?

Answer 7

error term is reduced because some of the variability in DV can be explained by the covariate

Question 8

What are the assumptions for an ANCOVA?

Answer 8

same as ANOVA plus Homogeneity of Regression

Question 9

What is the assumption of homogeneity of regression for an ANCOVA?

Answer 9

if the relationship between the DV and the covariate is different for different levels of the IV and if there is an interaction between covariate and the iv - the assumption is violated and ANCOVA is not appropriate

Question 10

How is the assumption of homogeneity of regression tested?

Answer 10

Test of between subject effects table of spss output -> look at IV*IV row and whether the interaction is significant. If not significant the homogeneity of regression is breached and can proceed with an ANCOVA

Question 11

When reporting results of a pairwise comparison, what statistics are reported?

Answer 11

Mdiff
CI
t
p
d