W9: Performing meta-analysis in R

Question 1

Interpret effect size of random-effects model with 9 studies investigating association between amniotic testosterone and sexually differentiated play play preferences (2)

Answer 1

• ## The effect size estimate is 0.0818 with a p value of 0.2737
• therefore shows that the association between amniotic testosterone and sexually differentiated play preferences is not significant at any level.

Question 2

Interpret effect size CI random-effects model with 9 studies investigating association between amniotic testosterone and sexually differentiated play play preferences (2)

Answer 2

• It has confidence intervals with lower bound of -0.0647 and upper bound of 0.2283.
• This contains 0 and therefore the effect size is not significant

Question 3

Interpret heterogenity with 9 studies investigating association between amniotic testosterone and sexually differentiated play play preferences (3)

Answer 3

• The test for heterogeneity is significant at the 5% level (Q = 18.25, p = 0.0194)
• From the I^2 statistic there is moderate heterogeneity since its between 30 to 60% (I^2 = 56.30%).
• This shows that all the studies included in the meta-analysis showing that not all studies show the same effect with amniotic testosterone and sexually differentiated preferences.

Question 4

What does leave1out (ma) do? - (2)

Answer 4

• This recalculates the meta-analysis whilst excluding each of the sample
• These identifies if any sample is exerting a large influence of meta-analysis:

Question 5

Egger’s test used to assess

Answer 5

possible potential publication biases in meta-analysis via funnel plot asymmetry

Question 6

In this Egger’s test it shows that

Answer 6

• The p value is 0.1848 which is not significant and therefore no evidence of publication bias (less than 0.05 implicates publication bias)

Question 7

Moderator Analysis
Checking if sex is a moderator between correlation of aminotic T and sex-typical play preferences:

Answer 7

• The p value is 0.617 and therefore sex is not a moderator of amniotic testosterone levels and sex-typical play references.