Блог

Я изучаю метаанализ в R. Я хотел бы вычислить регтест из библиотеки metafor. Смотрите пример моих данных:

 library("metafor")  value lt;- c(0.0001,0.0003,0.0002,0.0009,0.01,0.002) se lt;- c(1,3,2,4,0,3) data lt;- data.frame(value, se)  regtest(x = data$value, sei = data$se, model="rma")  

Но я получил ошибку Error in rma.uni(yi, vi, mods = X, intercept = FALSE, ...) : Division by zero when computing the inverse variance weights. In addition: Warning message: There are outcomes with non-positive sampling variances.

Кажется, что значения слишком малы, но как я могу это исправить, пожалуйста?

Больше не делите на ноль 😉

 value lt;- c(0.0001,0.0003,0.0002,0.0009,0.01,0.002) se lt;- c(1,3,2,4,1,3) data lt;- data.frame(value, se)  regtest(x = data$value, sei = data$se, model="rma")  

 Regression Test for Funnel Plot Asymmetry  Model: mixed-effects meta-regression model Predictor: standard error  Test for Funnel Plot Asymmetry: z = -0.0025, p = 0.9980 Limit Estimate (as sei -gt; 0): b = 0.0069 (CI: -2.5466, 2.5605)