Research synthesis
methods such as meta-analysis rely on either individual participant data or appropriate
summary statistics (e.g., measurement of precision such as standard deviation
and standard error) of trial data for implementation. A barrier to study
inclusion in research synthesis occurs when no precision metrics are explicitly
included in the primary report. Typically, such otherwise credible studies are
omitted in the research synthesis leading to potential publication bias and loss
of statistical power. We developed a user-friendly R shiny web app to estimate
the mean squared error (MSE) from published reports (e.g., Plant Disease
Management Reports) considering information on treatment means, significance
level, number of replications, and post-hoc test results from balanced and
randomized trials analyzed via ANOVA. To achieve this, users upload a csv file containing
trial information into the app, then specify the experimental design and the
post-hoc test that was applied in the trial analysis. Multiple trials can be
processed by specifying a trial identifier column in the csv file. Validation of
the procedure using 1,000 individual participant simulated datasets with
variable number of treatments and effect sizes demonstrated a strong Lin’s concordance
correlation (0.503 ≤ ρc ≤ 0.994) between values of MSE obtained with ANOVA and MSE FindR. Additionally, values
of the bias correction factor ranged from 0.875 to 0.999 across all simulation scenarios,
indicating an excellent agreement between the best fitted and identity lines. With
MSE FindR, researchers can conveniently obtain an estimate of variability from
published reports lacking measurement of precision, ultimately enabling the inclusion
of such studies in a quantitative review of research data evaluating product
efficacy or genotype performance in coordinated Fusarium head scab wheat and
barley trials in the US. Inclusion of summary
statistics (e.g., means, estimates of precision, significance level, and the number
of replications) should still be considered as a standard approach in summary
trial reports for the US wheat and barley scab initiative community.