Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Metan and forestplot

    It is a pleasure for me to be alble to ask for you precious help.
    I have a problem with the small stata version when using the function metan (for meta-analysis). since I don't have enough space to save variables and right now not enough money to upgrade (:-)) metan would not return the forest plot because one variable is missing. is there anyone helping me with the graft? just for this time? I am attaching the dataset. very simple. the syntax to obtain the graph would be:
    metan SIZE meanIO stdvIO SIZE meanAH stdvAH
    or

    . forestplot _ES _LCI _UCI
    the package to download is simply "metan" and for the second syntax "metan" and "ipdmetan"
    thanks in advance
    enrica
    Attached Files

  • #2
    Dear Enrica,

    Would it be possible to use -dataex- to enable us to recreate your dataset without having to risk downloading a file? If it's an aggregate-data meta-analysis, there can't be that many observations and so the dataex output shouldn't be too unwieldy. I'd then be happy to help.

    Although ironically, you're saying that in fact your dataset is too large ... at least, too large to be analysed by -metan- in Small Stata? That is amazing. Is that the case even when you first drop all other variables except SIZE, meanIO, stdvlO, meanAH and stdvAH?? I was aware -metan- was inefficient in its use of temporary variables, but I didn't realise it was that inefficient.

    For your information, I am the author of -ipdmetan-, and am currently in the process of updating the package so that it can handle -metan- syntaxes directly; in other words, you will be able to use -admetan- (a command within the ipdmetan package) to analyse your data rather than -metan-, and hopefully it will be more Small Stata-friendly. I'll post on Statalist when that's available.

    Thanks,

    David.

    Comment

    Working...
    X