load("~/Dropbox/Projeto_tese/C4 Doenca e dano/META/data/master_file.RData")
load("~/Dropbox/Projeto_tese/C4 Doenca e dano/META/data/summary_dat.RData")
load("~/Dropbox/Projeto_tese/C4 Doenca e dano/META/data/4_correlations.RData")
Dataset
head(long)
## study year state location cultivar fungic chemG bk
## 1 1 2012 MT Campo Verde TMG803 CHECK 0 4
## 2 1 2012 MT Campo Verde TMG803 CZM MBC 4
## 3 1 2012 MT Campo Verde TMG803 CZM_CM_TEBU DMIs_MBC_QoI 4
## 4 1 2012 MT Campo Verde TMG803 CZM_LS MBC_LignoS 4
## 5 1 2012 MT Campo Verde TMG803 EPO_FLUX_PYRA_0.8 DMIs_QoI_SDHI 4
## 6 1 2012 MT Campo Verde TMG803 EPO_FLUX_PYRA_1 DMIs_QoI_SDHI 4
## rep sev yield
## 1 1 45 2642
## 2 1 55 2918
## 3 1 35 2414
## 4 1 40 2708
## 5 1 25 2684
## 6 1 20 2967
Linear model regression - parameteres b0 and b1
long_gp <- groupedData(yield ~ sev | study, data = na.omit(long), order = FALSE)
fit <- lmList(yield ~ sev | study, data=long_gp)
# First parameter: est + sd
b0 = as.data.frame(summary(fit)$coefficients[,,1][,1:2]) %>%
add_rownames(var = "id")%>%
transmute(`study` = as.integer(id), Coef_b0=Estimate, SE_b0=`Std. Error`)
# Second parameter: est + sd
b1 = as.data.frame(summary(fit)$coefficients[,,2][,1:2])%>%
add_rownames(var = "id")%>%
transmute(`study`= as.integer(id), Coef_b1=Estimate, SE_b1=`Std. Error`)
reg = left_join(b0, b1, by = 'study') %>%
right_join(by_study, by = c("study")) %>%
mutate(study = factor(study)) %>%
select(study, year, state, cultivar, Coef_b0:SE_b1)
# Discarding studies with sev ranges < = 10% severity
(reg_red = subset(reg, !study %in% as.numeric(to_discard$study)))
## Source: local data frame [34 x 8]
##
## study year state cultivar Coef_b0 SE_b0 Coef_b1
## (fctr) (int) (fctr) (fctr) (dbl) (dbl) (dbl)
## 1 1 2012 MT TMG803 2898.754 138.20354 -7.584689
## 2 2 2012 MT TMG803 3927.619 141.38724 -24.469702
## 3 3 2012 MT TMG803 3251.294 168.21468 -16.095129
## 4 4 2012 MS 5G830_RR 3445.393 67.35849 -4.360786
## 5 5 2012 MS BMX_Potencia_RR 4218.295 86.50103 -9.249662
## 6 6 2012 MT TMG803 3610.224 83.71836 -7.227121
## 7 7 2012 GO M8336_RR 3849.029 197.06981 -12.180597
## 8 8 2013 MT TMG1179_RR 3877.356 110.35541 -3.623984
## 9 9 2013 MT 5G830_RR 3735.816 117.29455 -22.882334
## 10 10 2013 GO BRSGO_9160_RR 4341.308 176.65041 -37.492322
## .. ... ... ... ... ... ... ...
## Variables not shown: SE_b1 (dbl)
Effect size of reggression coeficcients B0 e B1
B0
b0.rand = rma.uni(Coef_b0, sei = SE_b0, method="ML", data= reg_red)
b0.rand
##
## Random-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of total heterogeneity): 457469.4430 (SE = 114643.2355)
## tau (square root of estimated tau^2 value): 676.3649
## I^2 (total heterogeneity / total variability): 98.27%
## H^2 (total variability / sampling variability): 57.93
##
## Test for Heterogeneity:
## Q(df = 33) = 2197.9272, p-val < .0001
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## 3557.3172 118.0013 30.1464 <.0001 3326.0389 3788.5954 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
confint(b0.rand , digits = 2)
##
## estimate ci.lb ci.ub
## tau^2 457469.44 300449.42 821573.74
## tau 676.36 548.13 906.41
## I^2(%) 98.27 97.40 99.03
## H^2 57.93 38.39 103.24
#forest (b0.rand)
B1
b1.rand = rma.uni(Coef_b1, sei = SE_b1, method="ML", data= reg_red)
b1.rand
##
## Random-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of total heterogeneity): 144.5508 (SE = 43.0165)
## tau (square root of estimated tau^2 value): 12.0229
## I^2 (total heterogeneity / total variability): 87.05%
## H^2 (total variability / sampling variability): 7.72
##
## Test for Heterogeneity:
## Q(df = 33) = 194.4615, p-val < .0001
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## -17.8517 2.3011 -7.7579 <.0001 -22.3617 -13.3416 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
confint(b1.rand , digits = 2)
##
## estimate ci.lb ci.ub
## tau^2 144.55 95.85 356.62
## tau 12.02 9.79 18.88
## I^2(%) 87.05 81.67 94.31
## H^2 7.72 5.46 17.58
#forest (b1.rand)
Cultivar Moderator- to classify tolerance levels
## Warning in rma.uni(Coef_b1, sei = SE_b1, mods = ~cultivar - 1, method =
## "ML", : Redundant predictors dropped from the model.
## est low up cultivar Toler
## 1 -13.426458 -23.47573 -3.3771905 5G830_RR 2_Medium
## 2 -32.195475 -59.90792 -4.4830260 AS_3730_IPRO 3_Low
## 3 -28.389593 -45.59419 -11.1849920 BMX_Ativa_RR 3_Low
## 4 -8.916045 -16.89049 -0.9415982 BMX_Potencia_RR 1_High
## 5 -28.631467 -54.31383 -2.9491079 BRSGO_8151_RR 3_Low
## 6 -30.702977 -41.34798 -20.0579758 BRSGO_9160_RR 3_Low
## 7 -12.180597 -29.27227 4.9110733 M8336_RR 2_Medium
## 8 -28.182773 -36.15697 -20.2085775 M9144_RR 3_Low
## 9 -19.139126 -31.40411 -6.8741380 NA_5909_RG 2_Medium
## 10 -37.341122 -55.21267 -19.4695700 Syn1180_RR 3_Low
## 11 -29.150430 -48.04355 -10.2573068 TMG1175_RR 3_Low
## 12 -4.731433 -14.91747 5.4546023 TMG1179_RR 1_High
## 13 -1.806598 -11.28768 7.6744828 TMG1180_RR 1_High
## 14 -14.215373 -30.77060 2.3398531 TMG7188_RR 2_Medium
## 15 -13.739801 -19.93469 -7.5449079 TMG803 2_Medium
## Warning in left_join_impl(x, y, by$x, by$y): joining factors with different
## levels, coercing to character vector
State as qualitative moderator
B0
( rma.uni(Coef_b0, sei = SE_b0, mods = ~ factor(state), method = "ML", data = reg1) )
##
## Mixed-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 405180.3940 (SE = 101908.6174)
## tau (square root of estimated tau^2 value): 636.5378
## I^2 (residual heterogeneity / unaccounted variability): 98.02%
## H^2 (unaccounted variability / sampling variability): 50.52
## R^2 (amount of heterogeneity accounted for): 11.43%
##
## Test for Residual Heterogeneity:
## QE(df = 29) = 1922.1960, p-val < .0001
##
## Test of Moderators (coefficient(s) 2,3,4,5):
## QM(df = 4) = 4.4486, p-val = 0.3487
##
## Model Results:
##
## estimate se zval pval ci.lb
## intrcpt 3920.9004 399.6010 9.8120 <.0001 3137.6968
## factor(state)GO -283.6681 460.6897 -0.6157 0.5381 -1186.6033
## factor(state)MS -29.8738 477.6968 -0.0625 0.9501 -966.1423
## factor(state)MT -571.9353 432.9380 -1.3211 0.1865 -1420.4782
## factor(state)PR -616.2523 604.8217 -1.0189 0.3083 -1801.6811
## ci.ub
## intrcpt 4704.1040 ***
## factor(state)GO 619.2671
## factor(state)MS 906.3947
## factor(state)MT 276.6077
## factor(state)PR 569.1765
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
b0.state = rma.uni(Coef_b0, sei = SE_b0, mods = ~ factor(state)-1, method = "ML", data = reg1)
##tab_b0.state = data.frame (est = b0.state$b,
# low = b0.state$ci.lb,
# up = b0.state$ci.ub)
#tab_b0.state
B1
( rma.uni(Coef_b1, sei = SE_b1, mods = ~ factor(state), method = "ML", data = reg1) )
##
## Mixed-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 62.9366 (SE = 22.1577)
## tau (square root of estimated tau^2 value): 7.9333
## I^2 (residual heterogeneity / unaccounted variability): 73.74%
## H^2 (unaccounted variability / sampling variability): 3.81
## R^2 (amount of heterogeneity accounted for): 56.46%
##
## Test for Residual Heterogeneity:
## QE(df = 29) = 114.4418, p-val < .0001
##
## Test of Moderators (coefficient(s) 2,3,4,5):
## QM(df = 4) = 25.1822, p-val < .0001
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## intrcpt -33.5566 6.0045 -5.5886 <.0001 -45.3251 -21.7880
## factor(state)GO 4.8279 7.1632 0.6740 0.5003 -9.2117 18.8674
## factor(state)MS 23.9385 7.1641 3.3414 0.0008 9.8971 37.9799
## factor(state)MT 21.6364 6.4590 3.3498 0.0008 8.9770 34.2958
## factor(state)PR 12.2505 8.8117 1.3903 0.1644 -5.0200 29.5210
##
## intrcpt ***
## factor(state)GO
## factor(state)MS ***
## factor(state)MT ***
## factor(state)PR
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
b1.state = rma.uni(Coef_b1, sei = SE_b1, mods = ~ factor(state)-1, method = "ML", data = reg1)
tab_b1.state = data.frame (est = b1.state$b,
low = b1.state$ci.lb,
up = b1.state$ci.ub)
tab_b1.state
## est low up
## factor(state)TO -33.556574 -45.32512 -21.788027
## factor(state)GO -28.728711 -36.38441 -21.073008
## factor(state)MS -9.618057 -17.27719 -1.958928
## factor(state)MT -11.920143 -16.58513 -7.255156
## factor(state)PR -21.306058 -33.94616 -8.665955
Relative yield
(coef(b1.state)/coef(b0.state))*100
## factor(state)TO factor(state)GO factor(state)MS factor(state)MT
## -0.8558385 -0.7898509 -0.2471856 -0.3559351
## factor(state)PR
## -0.6447300
Cultivar Tolerance as qualitative moderator
B0
( rma.uni(Coef_b0, sei = SE_b0, mods = ~ factor(Toler), method = "ML", data = reg1) )
##
## Mixed-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 451157.7755 (SE = 113106.4858)
## tau (square root of estimated tau^2 value): 671.6828
## I^2 (residual heterogeneity / unaccounted variability): 98.21%
## H^2 (unaccounted variability / sampling variability): 55.79
## R^2 (amount of heterogeneity accounted for): 1.38%
##
## Test for Residual Heterogeneity:
## QE(df = 31) = 1990.5169, p-val < .0001
##
## Test of Moderators (coefficient(s) 2,3):
## QM(df = 2) = 0.3748, p-val = 0.8291
##
## Model Results:
##
## estimate se zval pval ci.lb
## intrcpt 3658.4923 225.7345 16.2071 <.0001 3216.0608
## factor(Toler)2_Medium -94.2000 299.4422 -0.3146 0.7531 -681.0960
## factor(Toler)3_Low -180.6478 295.8765 -0.6106 0.5415 -760.5551
## ci.ub
## intrcpt 4100.9238 ***
## factor(Toler)2_Medium 492.6960
## factor(Toler)3_Low 399.2594
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
b0.Tol = rma.uni(Coef_b0, sei = SE_b0, mods = ~ factor(Toler)-1, method = "ML", data = reg1)
#tab_b0.Tol = data.frame (est = b0.Tol$b,
# low = b0.Tol$ci.lb,
# up = b0.Tol$ci.ub)
#tab_b0.Tol
B1
( rma.uni(Coef_b1, sei = SE_b1, mods = ~ Toler, method = "ML", data = reg1) )
##
## Mixed-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 48.9769 (SE = 18.4687)
## tau (square root of estimated tau^2 value): 6.9983
## I^2 (residual heterogeneity / unaccounted variability): 68.89%
## H^2 (unaccounted variability / sampling variability): 3.21
## R^2 (amount of heterogeneity accounted for): 66.12%
##
## Test for Residual Heterogeneity:
## QE(df = 31) = 110.2189, p-val < .0001
##
## Test of Moderators (coefficient(s) 2,3):
## QM(df = 2) = 39.5913, p-val < .0001
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## intrcpt -5.5927 2.8147 -1.9870 0.0469 -11.1093 -0.0761 *
## Toler2_Medium -8.8279 3.7081 -2.3807 0.0173 -16.0956 -1.5601 *
## Toler3_Low -24.4633 3.9538 -6.1873 <.0001 -32.2126 -16.7140 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
b1.Tol = rma.uni(Coef_b1, sei = SE_b1, mods = ~ factor(Toler)-1, method = "ML", data = reg1)
tab_b1.Tol = data.frame (est = b1.Tol$b,
low = b1.Tol$ci.lb,
up = b1.Tol$ci.ub)
tab_b1.Tol
## est low up
## factor(Toler)1_High -5.592727 -11.10934 -0.07611351
## factor(Toler)2_Medium -14.420597 -19.15206 -9.68913163
## factor(Toler)3_Low -30.056018 -35.49834 -24.61369673
Relative yield
(coef(b1.Tol)/coef(b0.Tol))*100
## factor(Toler)1_High factor(Toler)2_Medium factor(Toler)3_Low
## -0.1528697 -0.4045852 -0.8642140
Dataset for correlations
r_study1 = left_join(r_study, select(tab_cv, cultivar, Toler), by = "cultivar")
## Warning in left_join_impl(x, y, by$x, by$y): joining factors with different
## levels, coercing to character vector
r_red = subset(r_study1,!study %in% to_discard$study) #%>% filter(!r>0)
levels(r_study1$state)[c(1,6)] <- "TO"
print(tbl_df(r_red), n=40)
## Source: local data frame [34 x 12]
##
## study year r pairs z Vz state cultivar Dis_level
## (dbl) (dbl) (dbl) (int) (dbl) (dbl) (fctr) (chr) (fctr)
## 1 1 2012 -0.18 36 -0.182 0.030 MT TMG803 3High
## 2 2 2012 -0.52 36 -0.576 0.030 MT TMG803 2Medium
## 3 3 2012 -0.35 36 -0.365 0.030 MT TMG803 3High
## 4 4 2012 -0.19 36 -0.192 0.030 MS 5G830_RR 2Medium
## 5 5 2012 -0.19 45 -0.192 0.024 MS BMX_Potencia_RR 2Medium
## 6 6 2012 -0.49 36 -0.536 0.030 MT TMG803 3High
## 7 7 2012 -0.30 36 -0.310 0.030 GO M8336_RR 3High
## 8 8 2013 -0.13 40 -0.131 0.027 MT TMG1179_RR 2Medium
## 9 9 2013 -0.78 40 -1.045 0.027 MT 5G830_RR 3High
## 10 10 2013 -0.78 40 -1.045 0.027 GO BRSGO_9160_RR 3High
## 11 11 2013 -0.63 40 -0.741 0.027 PR BMX_Potencia_RR 3High
## 12 12 2013 -0.39 40 -0.412 0.027 GO BRSGO_9160_RR 2Medium
## 13 13 2013 -0.74 40 -0.950 0.027 GO Syn1180_RR 2Medium
## 14 14 2013 -0.34 40 -0.354 0.027 MT TMG7188_RR 1Low
## 15 15 2014 -0.55 32 -0.618 0.034 MS NA_5909_RG 2Medium
## 16 16 2014 -0.32 36 -0.332 0.030 MT TMG1179_RR 3High
## 17 17 2014 -0.53 36 -0.590 0.030 GO BRSGO_8151_RR 1Low
## 18 18 2014 -0.25 36 -0.255 0.030 MT TMG803 2Medium
## 19 20 2014 -0.75 36 -0.973 0.030 GO BRSGO_9160_RR 1Low
## 20 21 2014 -0.79 36 -1.071 0.030 TO M9144_RR 3High
## 21 23 2014 -0.69 36 -0.848 0.030 MT M9144_RR 1Low
## 22 24 2014 -0.63 36 -0.741 0.030 PR BMX_Ativa_RR 2Medium
## 23 25 2015 -0.59 40 -0.678 0.027 MS NA_5909_RG 1Low
## 24 26 2015 -0.67 40 -0.811 0.027 MT M9144_RR 2Medium
## 25 27 2015 -0.11 40 -0.110 0.027 MS BMX_Potencia_RR 1Low
## 26 28 2015 -0.18 40 -0.182 0.027 MS BMX_Potencia_RR 2Medium
## 27 29 2015 -0.59 40 -0.678 0.027 GO AS_3730_IPRO 1Low
## 28 30 2015 -0.67 40 -0.811 0.027 GO TMG1175_RR 1Low
## 29 32 2015 -0.53 40 -0.590 0.027 MT TMG803 2Medium
## 30 33 2015 -0.68 40 -0.829 0.027 TO M9144_RR 3High
## 31 34 2015 0.15 40 0.151 0.027 MT TMG1180_RR 1Low
## 32 35 2015 -0.27 40 -0.277 0.027 MT TMG1180_RR 3High
## 33 36 2015 -0.29 40 -0.299 0.027 MT TMG1180_RR 2Medium
## 34 37 2015 -0.85 40 -1.256 0.027 TO M9144_RR 3High
## Variables not shown: yield (dbl), Y_lev (fctr), Toler (chr)
Plots
par(mfrow=c(1,2))
plot(density(r_red$r, col="lightgray", xlim = c(-1,0.2)));rug(r_red$r)
## Warning: In density.default(r_red$r, col = "lightgray", xlim = c(-1, 0.2)) :
## extra arguments 'col', 'xlim' will be disregarded
plot(density(r_red$z, col="lightgray", xlim = c(-1,0.2)));rug(r_red$z)
## Warning: In density.default(r_red$z, col = "lightgray", xlim = c(-1, 0.2)) :
## extra arguments 'col', 'xlim' will be disregarded
par(mfrow=c(1,2))
hist(r_red$r,col="lightgray", xlim = c(-1,0))
hist(r_red$z,col="lightgray", xlim = c(-1.5,0.3))
layout(1)
Exploratory Graphs ~ possible moderators
ggplot(r_red, aes(x=state, y=r, label = r_red$study)) + geom_text() # -PR
ggplot(r_red, aes(year, r, label = r_red$study)) + geom_text()
ggplot(r_red, aes(x=cultivar, y=r, label = r_red$study)) + geom_text() + coord_flip()
ggplot(r_red, aes(x=Dis_level, y=r, label = r_red$study)) + geom_text()
ggplot(r_red, aes(x=Y_lev, y=r, label = r_red$study)) + geom_text()
Random effect model - Overall
(z.random = rma.uni(z, Vz, method="ML", data = r_red))
##
## Random-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of total heterogeneity): 0.0853 (SE = 0.0275)
## tau (square root of estimated tau^2 value): 0.2921
## I^2 (total heterogeneity / total variability): 75.19%
## H^2 (total variability / sampling variability): 4.03
##
## Test for Heterogeneity:
## Q(df = 33) = 138.2701, p-val < .0001
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## -0.5536 0.0578 -9.5796 <.0001 -0.6668 -0.4403 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
forest(z.random)
confint(z.random, digits = 2)
##
## estimate ci.lb ci.ub
## tau^2 0.09 0.05 0.17
## tau 0.29 0.22 0.42
## I^2(%) 75.19 62.88 86.01
## H^2 4.03 2.69 7.15
fisherz2r(c(est = z.random$b, low = z.random$ci.lb, up = z.random$ci.ub))
## est low up
## -0.5031862 -0.5828850 -0.4138978
Year as moderator
(year_mod = rma.uni(z, Vz, method="ML", mods = ~ factor(year), data= r_red))
##
## Mixed-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 0.0692 (SE = 0.0236)
## tau (square root of estimated tau^2 value): 0.2630
## I^2 (residual heterogeneity / unaccounted variability): 71.09%
## H^2 (unaccounted variability / sampling variability): 3.46
## R^2 (amount of heterogeneity accounted for): 18.91%
##
## Test for Residual Heterogeneity:
## QE(df = 30) = 118.5162, p-val < .0001
##
## Test of Moderators (coefficient(s) 2,3,4):
## QM(df = 3) = 5.6900, p-val = 0.1277
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## intrcpt -0.3348 0.1185 -2.8261 0.0047 -0.5670 -0.1026 **
## factor(year)2013 -0.3335 0.1667 -2.0008 0.0454 -0.6601 -0.0068 *
## factor(year)2014 -0.3440 0.1628 -2.1132 0.0346 -0.6630 -0.0249 *
## factor(year)2015 -0.1960 0.1485 -1.3199 0.1869 -0.4871 0.0950
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
confint(year_mod, digits = 2)
##
## estimate ci.lb ci.ub
## tau^2 0.07 0.04 0.17
## tau 0.26 0.20 0.41
## I^2(%) 71.09 59.88 85.59
## H^2 3.46 2.49 6.94
year_mod_1 = rma.uni(z, Vz, method="ML", mods = ~ factor(year)-1, data= r_red)
tab_year = data.frame (est = fisherz2r(year_mod_1$b),
low =fisherz2r(year_mod_1$ci.lb),
up = fisherz2r(year_mod_1$ci.ub))
tab_year
## est low up
## factor(year)2012 -0.3228530 -0.5131848 -0.1022553
## factor(year)2013 -0.5838511 -0.7153278 -0.4124509
## factor(year)2014 -0.5907352 -0.7150976 -0.4301249
## factor(year)2015 -0.4860178 -0.6083442 -0.3411365
Yield level
( rma.uni(z, Vz, method="ML", mods = ~ Y_lev, data= r_red))
##
## Mixed-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 0.0727 (SE = 0.0245)
## tau (square root of estimated tau^2 value): 0.2697
## I^2 (residual heterogeneity / unaccounted variability): 72.09%
## H^2 (unaccounted variability / sampling variability): 3.58
## R^2 (amount of heterogeneity accounted for): 14.72%
##
## Test for Residual Heterogeneity:
## QE(df = 31) = 123.0381, p-val < .0001
##
## Test of Moderators (coefficient(s) 2,3):
## QM(df = 2) = 4.2919, p-val = 0.1170
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## intrcpt -0.7904 0.1296 -6.1001 <.0001 -1.0443 -0.5364 ***
## Y_lev2Medium 0.2375 0.1767 1.3444 0.1788 -0.1088 0.5838
## Y_lev3High 0.3045 0.1470 2.0717 0.0383 0.0164 0.5925 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#Yield_mod1 = rma.uni(z, Vz, method="ML", mods = ~ Y_lev -1, data= r_red)
Disease level
(Dis_mod = rma.uni(z, Vz, method="ML", mods = ~ Dis_level, data= r_red))
##
## Mixed-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 0.0767 (SE = 0.0255)
## tau (square root of estimated tau^2 value): 0.2770
## I^2 (residual heterogeneity / unaccounted variability): 73.16%
## H^2 (unaccounted variability / sampling variability): 3.73
## R^2 (amount of heterogeneity accounted for): 10.04%
##
## Test for Residual Heterogeneity:
## QE(df = 31) = 127.0944, p-val < .0001
##
## Test of Moderators (coefficient(s) 2,3):
## QM(df = 2) = 2.6937, p-val = 0.2601
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## intrcpt -0.5410 0.1079 -5.0153 <.0001 -0.7524 -0.3296 ***
## Dis_level2Medium 0.0846 0.1404 0.6029 0.5466 -0.1905 0.3598
## Dis_level3High -0.1276 0.1428 -0.8934 0.3716 -0.4076 0.1523
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#Dis_mod1 = rma.uni(z, Vz, method="ML", mods = ~ Dis_level -1, data= r_red)
#tab_dis = data.frame (est = fisherz2r(Dis_mod1$b),
# low =fisherz2r(Dis_mod1$ci.lb),
# up = fisherz2r(Dis_mod1$ci.ub))
#tab_dis
State as moderator
( rma.uni(z, Vz, method="ML", mods = ~ state, data= r_red) )
##
## Mixed-Effects Model (k = 34; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 0.0386 (SE = 0.0162)
## tau (square root of estimated tau^2 value): 0.1963
## I^2 (residual heterogeneity / unaccounted variability): 57.78%
## H^2 (unaccounted variability / sampling variability): 2.37
## R^2 (amount of heterogeneity accounted for): 54.80%
##
## Test for Residual Heterogeneity:
## QE(df = 29) = 80.6743, p-val < .0001
##
## Test of Moderators (coefficient(s) 2,3,4,5):
## QM(df = 4) = 23.8676, p-val < .0001
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## intrcpt -1.0517 0.1489 -7.0628 <.0001 -1.3436 -0.7599 ***
## stateGO 0.3290 0.1747 1.8836 0.0596 -0.0133 0.6713 .
## stateMS 0.7279 0.1824 3.9905 <.0001 0.3704 1.0854 ***
## stateMT 0.6220 0.1632 3.8110 0.0001 0.3021 0.9418 ***
## statePR 0.3107 0.2360 1.3168 0.1879 -0.1518 0.7732
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
state_mod = rma.uni(z, Vz, method="ML", mods = ~ state-1, data= r_red)
tab_state = data.frame (est = fisherz2r(state_mod$b),
low =fisherz2r(state_mod$ci.lb),
up = fisherz2r(state_mod$ci.ub))
tab_state
## est low up
## stateTO -0.7824737 -0.8725284 -0.6409956
## stateGO -0.6186036 -0.7170910 -0.4958961
## stateMS -0.3129788 -0.4856114 -0.1168509
## stateMT -0.4051122 -0.5084672 -0.2902500
## statePR -0.6297490 -0.8004202 -0.3646334
Cultivar (Tolerance) as moderator
(cv_mod = rma.uni(z, Vz, method="ML", mods = ~ factor(Toler), data= r_study1))
## Warning in rma.uni(z, Vz, method = "ML", mods = ~factor(Toler), data =
## r_study1): Studies with NAs omitted from model fitting.
##
## Mixed-Effects Model (k = 35; tau^2 estimator: ML)
##
## tau^2 (estimated amount of residual heterogeneity): 0.0219 (SE = 0.0120)
## tau (square root of estimated tau^2 value): 0.1480
## I^2 (residual heterogeneity / unaccounted variability): 43.70%
## H^2 (unaccounted variability / sampling variability): 1.78
## R^2 (amount of heterogeneity accounted for): 73.69%
##
## Test for Residual Heterogeneity:
## QE(df = 32) = 62.0317, p-val = 0.0011
##
## Test of Moderators (coefficient(s) 2,3):
## QM(df = 2) = 43.2437, p-val < .0001
##
## Model Results:
##
## estimate se zval pval ci.lb ci.ub
## intrcpt -0.2338 0.0737 -3.1737 0.0015 -0.3783 -0.0894
## factor(Toler)2_Medium -0.2443 0.0985 -2.4811 0.0131 -0.4372 -0.0513
## factor(Toler)3_Low -0.6074 0.0949 -6.3983 <.0001 -0.7934 -0.4213
##
## intrcpt **
## factor(Toler)2_Medium *
## factor(Toler)3_Low ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
confint(cv_mod, digits = 2)
##
## estimate ci.lb ci.ub
## tau^2 0.02 0.01 0.07
## tau 0.15 0.08 0.26
## I^2(%) 43.70 19.96 70.15
## H^2 1.78 1.25 3.35
cv_mod_1 = rma.uni(z, Vz, method="ML", mods = ~ factor(Toler)-1, data= r_study1)
## Warning in rma.uni(z, Vz, method = "ML", mods = ~factor(Toler) - 1, data =
## r_study1): Studies with NAs omitted from model fitting.
tab_cv_tol = data.frame (est = fisherz2r(cv_mod_1$b),
low =fisherz2r(cv_mod_1$ci.lb),
up = fisherz2r(cv_mod_1$ci.ub))
tab_cv_tol
## est low up
## factor(Toler)1_High -0.2296723 -0.3611930 -0.08919221
## factor(Toler)2_Medium -0.4447261 -0.5413634 -0.33649196
## factor(Toler)3_Low -0.6864452 -0.7436075 -0.61932100