[R-meta] escalc error interpretation

[Michael Dewey]

Dear Philippe

Looking at out1 and out2 there seems to be a pair which are both zero. 
Perhaps that is the issue.

Michael

On 26/08/2021 17:58, Philippe Tadger wrote:
> Dear all,
> 
> I have an error with the function escalc,
> 
> Error in escalc(measure = "PLO", xi = out1, mi = out2, data = ddd, add =
> 1/2, : One or more group sizes are < 1.
> 
> that I don't know how to interpret what is wrong with my data. As follow
> a reproducible example:
> 
> library(metafor)
> ddd<-data.frame(
>     tn = c(5535L,
>            5535L,1222L,1222L,
>            1032L,1032L,1392L,
>            1392L,990L,990L,
>            182L,182L,1406L,
>            1406L,21043L,21043L,
>            100L,100L,9647L,
>            9647L),
>     tp = c(15L,15L,
>            2L,2L,2L,2L,
>            10L,10L,2L,2L,1L,
>            1L,9L,9L,28L,28L,
>            0L,0L,33L,33L),
>     np = c(15L,15L,
>            2L,2L,2L,2L,
>            10L,10L,2L,2L,1L,
>            1L,9L,9L,28L,28L,
>            0L,0L,36L,36L),
>     nn = c(5802L,
>            5802L,1234L,1234L,
>            1071L,1071L,1438L,
>            1438L,1005L,1005L,
>            194L,194L,1416L,
>            1416L,21900L,
>            21900L,105L,105L,
>            10121L,10121L),
>     MA_num = c(10L,10L,
>                10L,10L,10L,10L,
>                10L,10L,10L,10L,
>                10L,10L,10L,10L,
>                10L,10L,10L,10L,
>                10L,10L),
>     SUBMA_num = c(45L,45L,
>                   45L,45L,45L,45L,
>                   45L,45L,45L,45L,
>                   45L,45L,45L,45L,
>                   45L,45L,45L,45L,
>                   45L,45L),
>     study = c(735L,
>               735L,738L,738L,739L,
>               739L,740L,740L,
>               744L,744L,745L,
>               745L,747L,747L,750L,
>               750L,753L,753L,
>               755L,755L),
>     out1 = c(15L,
>              5535L,2L,1222L,2L,
>              1032L,10L,1392L,2L,
>              990L,1L,182L,9L,
>              1406L,28L,21043L,
>              0L,100L,33L,
>              9647L),
>     out2 = c(0L,267L,
>              0L,12L,0L,39L,
>              0L,46L,0L,15L,0L,
>              12L,0L,10L,0L,
>              857L,0L,5L,3L,
>              474L),
>     study = as.factor(c("1","1","2","2",
>                         "3","3","4","4",
>                         "5","5","6","6",
>                         "7","7","8","8",
>                         "9","9","10",
>                         "10")),
>     group = as.factor(c("1","2","1","2",
>                         "1","2","1","2",
>                         "1","2","1","2",
>                         "1","2","1","2",
>                         "1","2","1","2"))
> )
> levels(ddd$group) <- c("sensitivity", "specificity")
> ddd <- escalc(measure="PLO", xi=out1, mi=out2, data=ddd, add=1/2, to="all",
>                 include=group=="sensitivity")
> 
> The error is triggered for the following conditional in the escalc function
> 
> if (any <https://rdrr.io/r/base/any.html>(ni < 1, na.rm=TRUE
> <https://rdrr.io/r/base/logical.html>)) stop
> <https://rdrr.io/r/base/stop.html>(mstyle$stop
> <https://rdrr.io/r/base/stop.html>("One or more group sizes are < 1."))
> 
> Thanks in advance for your time
> 

-- 
Michael
http://www.dewey.myzen.co.uk/home.html