[R-meta] Ratio of Means

[Nathan Pace]

Thanks to Gerta, Wolfgang, Michael, and James for comments and pointers to functions and codes.

In my data set the slope is one (0.97 +/- 0.04).

The QE for residual heterogeneity is very large (~30000).

In van Houwelingen (page 605, Figure 5) a coverage ellipse is plotted for vaccination data.

An equation is shown in the figure legend.

I don’t understand the source of the constant (5.99) on the rhs.

Comments?


Nathan



-----Original Message-----
From: "Viechtbauer Wolfgang (SP)" <wolfgang.viechtbauer at maastrichtuniversity.nl>
Date: Friday, July 21, 2017 at 06:12
To: Michael Dewey <lists at dewedy.myzen.co.uk>, Nathan L Pace <n.l.pace at utah.edu>, "r-sig-meta-analysis at r-project.org" <r-sig-meta-analysis at r-project.org>
Subject: RE: [R-meta] Ratio of Means

    Also, I cannot reproduce this problem.
    
    library(metafor)
    
    arm.df <- read.table(header=TRUE, text = "
    id            study group  mean    sd   n
    1        Arici 2009   Exp 35.73  5.24  27
    2        Arici 2009   Ctl 62.93  8.67  27
    3       Arslan 2011   Exp  4.25  2.03  20")
    
    escalc(measure='MNLN', mi=mean, sdi=sd, ni=n) ### this obviously fails since the variables cannot be found
    
    escalc(measure='MNLN', mi=mean, sdi=sd, ni=n, data=arm.df) ### works
    
    attach(arm.df)
    escalc(measure='MNLN', mi=mean, sdi=sd, ni=n) ### also works, but don't use attach (shudder)
    detach(arm.df)
    
    Best,
    Wolfgang
    
    -----Original Message-----
    From: Michael Dewey [mailto:lists at dewey.myzen.co.uk] 
    Sent: Friday, July 21, 2017 13:01
    To: Nathan Pace; Viechtbauer Wolfgang (SP); r-sig-meta-analysis at r-project.org
    Subject: Re: [R-meta] Ratio of Means
    
    Dear Nathan
    
    Try using the data parameter rather than attaching arm.df.
    Alternatively call the standard deviation something other than sd.
    
    On 20/07/2017 22:17, Nathan Pace wrote:
    > Hi,
    >
    > Thanks for the pointers and links.
    >
    > Next problem is using the measure = “MNLN” in escalc.
    >
    > I created the 2k data frame.
    >
    >> arm.df
    >               study group  mean    sd   n
    > 1        Arici 2009   Exp 35.73  5.24  27
    > 2        Arici 2009   Ctl 62.93  8.67  27
    > 3       Arslan 2011   Exp  4.25  2.03  20
    > 4       Arslan 2011   Ctl 11.25  5.03  20
    > 5       Arslan 2013   Exp  5.15  4.29 100
    > 6       Arslan 2013   Ctl   9.1  4.68 100
    > 7      Aydogan 2008   Exp  0.27  0.08  40
    > 8      Aydogan 2008   Ctl  0.35  0.09  40
    > 9        Cakan 2008   Exp    11     2  20
    > 10       Cakan 2008   Ctl    12     2  20
    > 11       Cobby 1999   Exp    35  20.4  24
    > 12       Cobby 1999   Ctl  54.9  28.3  21
    > 13        Dahl 1997   Exp  21.2  7.88  22
    > 14        Dahl 1997   Ctl  18.7   5.9  21
    > 15      Dilmen 2010   Exp  18.1 13.93  20
    > 16      Dilmen 2010   Ctl    30 11.32  19
    > 17      Durmus 2007   Exp  30.5 11.55  25
    > 18      Durmus 2007   Ctl 42.74 12.33  25
    > 19        Emir 2010   Exp  7.65  0.78  30
    > 20        Emir 2010   Ctl  13.8  2.26  30
    > 21       Fayaz 2004   Exp    12     6  17
    > 22       Fayaz 2004   Ctl    22    13  17
    > 23      Jokela 2010   Exp 38.85  16.8  40
    > 24      Jokela 2010   Ctl 45.15  18.9  40
    > 25     Khalili 2013   Exp     3  2.28  25
    > 26     Khalili 2013   Ctl   4.2  1.57  25
    > 27  Kilieaslan 2010   Exp 20.56   0.4  25
    > 28  Kilieaslan 2010   Ctl  28.6  0.35  25
    > 29     Koppert 2006   Exp    17    22  25
    > 30     Koppert 2006   Ctl    38    22  25
    > 31    Kvalsvik 2003   Exp    16     5  30
    > 32    Kvalsvik 2003   Ctl    20     6  30
    > 33  Montgomery 1996   Exp  27.1    27  19
    > 34  Montgomery 1996   Ctl  34.5  15.1  20
    > 35        Moon 2011   Exp   8.7   5.4  36
    > 36        Moon 2011   Ctl  12.6   6.9  35
    > 37 Munishankar 2008   Exp  33.8  23.8  25
    > 38 Munishankar 2008   Ctl  44.1  24.4  25
    > 39     Sinatra 2005   Exp  38.3  35.1  49
    > 40     Sinatra 2005   Ctl  57.4  52.3  52
    > 41        Syal 2010   Exp  9.53  1.88  30
    > 42        Syal 2010   Ctl 10.19  2.08  30
    > 43      Toygar 2008   Exp  16.5  11.1  30
    > 44      Toygar 2008   Ctl  34.3  11.8  30
    > 45      Yalcin 2012   Exp 48.53  12.4  26
    > 46      Yalcin 2012   Ctl 73.03 22.41  27
    >
    > The call to escalc returned an error:
    >
    >> arm.df <- escalc(measure = 'MNLN', mi = mean, sdi = sd, ni = n)
    > Error in sdi < 0 :
    >   comparison (3) is possible only for atomic and list types
    >
    > ?????
    >
    > Nathan