Metanalysis: Effect Sizes Based on Correlations (Soybean / Target spot)
Juan Pablo Edwards
07/10/2014
head (dat)## id cult ctrl_dis dis_pres r pairs z v_z
## 1 2012_1 TMG 803 42.5 h -0.18 36 -0.1820 0.03030
## 2 2012_2 TMG 803 36.3 h -0.53 36 -0.5901 0.03030
## 3 2012_3 TMG 803 41.3 h -0.35 36 -0.3654 0.03030
## 4 2012_4 5G830 30.6 h -0.19 36 -0.1923 0.03030
## 5 2012_5 BMX Potencia RR 27.0 l -0.19 36 -0.1923 0.03030
## 6 2012_6 TMG 803 44.8 h -0.49 45 -0.5361 0.02381Fitted models
Random effect model
(fit1 = rma.uni(z, v_z, method="ML",data= dat,slab=id))##
## Random-Effects Model (k = 27; tau^2 estimator: ML)
##
## tau^2 (estimated amount of total heterogeneity): 0.0487 (SE = 0.0210)
## tau (square root of estimated tau^2 value): 0.2206
## I^2 (total heterogeneity / total variability): 63.16%
## H^2 (total variability / sampling variability): 2.71
##
## Test for Heterogeneity:
## Q(df = 26) = 73.7723, p-val < .0001
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## -0.5735 0.0535 -10.7238 <.0001 -0.6783 -0.4687 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1(r.low= rma.uni(z, v_z, method="ML",data= dat, subset=(dis_pres=="l")))##
## Random-Effects Model (k = 14; tau^2 estimator: ML)
##
## tau^2 (estimated amount of total heterogeneity): 0.0232 (SE = 0.0198)
## tau (square root of estimated tau^2 value): 0.1523
## I^2 (total heterogeneity / total variability): 44.39%
## H^2 (total variability / sampling variability): 1.80
##
## Test for Heterogeneity:
## Q(df = 13) = 25.1080, p-val = 0.0223
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## -0.5926 0.0612 -9.6880 <.0001 -0.7125 -0.4727 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1(r.high= rma.uni(z, v_z, method="ML",data= dat, subset=(dis_pres=="h")))##
## Random-Effects Model (k = 13; tau^2 estimator: ML)
##
## tau^2 (estimated amount of total heterogeneity): 0.0748 (SE = 0.0403)
## tau (square root of estimated tau^2 value): 0.2735
## I^2 (total heterogeneity / total variability): 72.98%
## H^2 (total variability / sampling variability): 3.70
##
## Test for Heterogeneity:
## Q(df = 12) = 48.4383, p-val < .0001
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## -0.5517 0.0889 -6.2074 <.0001 -0.7258 -0.3775 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Model including Cultivar as qualitative moderator
(fit.cult = rma.uni(z, v_z, method="ML", mods = ~ cult-1, data= dat))##
## Mixed-Effects Model (k = 27; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 0.0000 (SE = 0.0077)
## tau (square root of estimated tau^2 value): 0.0013
## I^2 (residual heterogeneity / unaccounted variability): 0.01%
## H^2 (unaccounted variability / sampling variability): 1.00
##
## Test for Residual Heterogeneity:
## QE(df = 9) = 20.4837, p-val = 0.0152
##
## Test of Moderators (coefficient(s) 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18):
## QM(df = 18) = 369.0216, p-val < .0001
##
## Model Results:
##
## se zval pval ci.lb ci.ub
## cult5G830 -0.1923 0.1741 -1.1049 0.2692 -0.5335 0.1489
## cultAtiva RR -0.7414 0.1741 -4.2590 <.0001 -1.0826 -0.4002
## cultBMX Potencia RR -0.5451 0.1163 -4.6894 <.0001 -0.7730 -0.3173
## cultBRSGO 8151 RR -0.6742 0.1231 -5.4768 <.0001 -0.9154 -0.4329
## cultBRSGO 8661 RR -0.5101 0.1741 -2.9301 0.0034 -0.8513 -0.1689
## cultBRSGO 9160 RR -0.7454 0.0933 -7.9935 <.0001 -0.9282 -0.5626
## cultDow 5G830 RR -1.0203 0.1562 -6.5331 <.0001 -1.3264 -0.7142
## cultM 9144 RR -0.8480 0.1741 -4.8710 <.0001 -1.1892 -0.5068
## cultMsoy 8336 -0.3095 0.1741 -1.7780 0.0754 -0.6507 0.0317
## cultMSoy 9144 RR -1.0714 0.1741 -6.1547 <.0001 -1.4126 -0.7302
## cultNA 5909 RR -0.6184 0.1857 -3.3300 0.0009 -0.9823 -0.2544
## cultP98Y30 -0.5230 0.1741 -3.0042 0.0027 -0.8642 -0.1818
## cultST 810 -0.6475 0.1741 -3.7196 0.0002 -0.9887 -0.3063
## cultSyn 1180 -0.8872 0.1562 -5.6806 <.0001 -1.1933 -0.5811
## cultTMG 1179 RR -0.2260 0.1163 -1.9439 0.0519 -0.4538 0.0019
## cultTMG 132 RR -0.7089 0.1741 -4.0723 <.0001 -1.0501 -0.3677
## cultTMG 7188 RR -0.3205 0.1562 -2.0524 0.0401 -0.6267 -0.0144
## cultTMG 803 -0.3936 0.0758 -5.1915 <.0001 -0.5422 -0.2450
##
## cult5G830
## cultAtiva RR ***
## cultBMX Potencia RR ***
## cultBRSGO 8151 RR ***
## cultBRSGO 8661 RR **
## cultBRSGO 9160 RR ***
## cultDow 5G830 RR ***
## cultM 9144 RR ***
## cultMsoy 8336 .
## cultMSoy 9144 RR ***
## cultNA 5909 RR ***
## cultP98Y30 **
## cultST 810 ***
## cultSyn 1180 ***
## cultTMG 1179 RR .
## cultTMG 132 RR ***
## cultTMG 7188 RR *
## cultTMG 803 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1cult_eff = round(fisherz2r(fit.cult$b),2) # cultivar r ยป Represents disease "sensibility"?
rownames(cult_eff)=gsub('cult','',rownames(cult_eff))
cult_effect=as.data.frame(cult_eff)
cult_effect$cultivar=row.names(cult_eff)
cult_effect=cult_effect[,c(2,1)]
row.names(cult_effect)<-NULL; names(cult_effect)[2]="r"
n.cult=table(dat$cult)
n <- c()
for(i in 1:18) {n <- c(n,n.cult[[i]])}
cultivar= names(n.cult)
n.cultivar=data.frame(cultivar,n)
cultivar_effect=merge(cult_effect, n.cultivar, by="cultivar")
cultivar_effect[with(cultivar_effect, order(-r)), ]## cultivar r n
## 1 5G830 -0.19 1
## 15 TMG 1179 RR -0.22 2
## 9 Msoy 8336 -0.30 1
## 17 TMG 7188 RR -0.31 1
## 18 TMG 803 -0.37 5
## 5 BRSGO 8661 RR -0.47 1
## 12 P98Y30 -0.48 1
## 3 BMX Potencia RR -0.50 2
## 11 NA 5909 RR -0.55 1
## 13 ST 810 -0.57 1
## 4 BRSGO 8151 RR -0.59 2
## 16 TMG 132 RR -0.61 1
## 2 Ativa RR -0.63 1
## 6 BRSGO 9160 RR -0.63 3
## 8 M 9144 RR -0.69 1
## 14 Syn 1180 -0.71 1
## 7 Dow 5G830 RR -0.77 1
## 10 MSoy 9144 RR -0.79 1