[R-sig-eco] Meta-analysis and funnel plot

[federicotomasetto]

Dear list,
I am running a meta-analysis using correlation coefficient (effect size) and
I am facing a problem when I try to check for publication bias. I am not an
expert in the outcome of the meta-analysis but the funnel plot is not
looking as some example I have found. Why? What is it telling me?
May I ask for some helps? 

Many thanks

##########################################
require(meta)

n <- 40

r <- c(0.590, 0.660, 0.330, -0.840, 0.840, 0.050, 0.020, 0.040, 0.100,
0.050, 0.390, 0.500, 0.610, -0.140, 0.330, 0.130, -0.150, 0.880, 0.780,
0.780, 0.340, 0.190, 0.340, 0.120, 0.830, 0.450, -0.520, 0.490, 0.770,
0.740, 0.440, 0.220, -0.230, 0.670, 0.570, 0.410, 0.660, 0.200, 0.660,
0.920) 

std.r <- sqrt((1 - r^2)^2 / (n - 1))

# Fixed and random effects meta-analysis based on correlation coefficient 
comb.r <- metagen(r, std.r)

#Confidence interval plot with fixed and random effects meta-analysis
forest(comb.r, comb.f = TRUE, comb.r = TRUE, main = "Validity studies", xlab
= "Correlation coefficient")

# Funnel plot and tests for publication bias
funnel(comb.r, main = "Validity studies", xlab = "Correlation coefficient")

##########################################

Federico Tomasetto
PhD Student