In total, 52 uniform field experiments were conducted in Illinois, Indiana, Iowa, Michigan, South Dakota, and Wisconsin in the United States and Ontario, Canada from 2013 to 2017 comparing crop protection products against sudden death syndrome (SDS) of soybean. Data were analyzed using meta-analytic models to summarize the relationship between foliar disease index (FDX) and yield. For each study, correlation and regression analyses were performed separately to determine three effect sizes: Fisher's transformation of correlation coefficients (Z r ), intercept (β0), and slope (β1). Random- and mixed-effect meta-analyses were used to summarize the effect sizes. Study- and location-specific moderator variables FDX (low < 10% and high ≥ 10%), date of planting (early = prior to 7 May, conventional = 7 to 21 May, and late = after 21 May) cultivar (susceptible and partially resistant to SDS), study location, and growing season were used as fixed effects. The overall mean effect sizes of transformed correlation coefficient [Formula: see text] r was -0.41 and different from zero (P < 0.001), indicating that yield was negatively correlated with FDX. The [Formula: see text] r was affected by disease level (P < 0.01) and cultivar (P = 0.02), with a greater effect at higher disease levels and with susceptible cultivars. The mean [Formula: see text] 0 was 4,121 kg/ha and mean [Formula: see text] 1 was -21 kg/ha/% FDX and were different from zero (P < 0.01). Results from these data indicate that, for every unit of FDX increase, yield was decreased by 0.5%. Study locations and year affected the [Formula: see text] 0 , whereas none of the moderator variables significantly affected [Formula: see text] 1.
Keywords: correlation; field crops; fungi; oilseeds and legumes; regression; sudden death syndrome; yield loss and economic impacts.