(m)simtest ?

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

(m)simtest ?

Bianca Vieru-Dimulescu
Hi,.

We have 2 values (first formant F1, second formant F2) for a given
phoneme for six languages. We want to see whether the languages are
significantly different one from another for this given phoneme.

We have done a manova on our data and it works well, but we doesn't
allow us to see which pair of  languages are different.

If we have only one formant for the phoneme, we would use simtest to see
the pairs significantly different... but the simtest doesn't work for 2
responses, so we look for a multivariate simtest.


Thanks for your advice!



So our data:
#read the data file
 >tous_y <- read.table("moyenne.txt", header=TRUE, sep="\t",
na.strings="NA", dec=".", strip.white=TRUE)
 >lg<-as.factor(c(rep("anglais",6),rep("allemand",6),rep("arabe",6),rep("espagnol",6),
rep("italien",6), rep("portugais",6), rep("francais",6)))
 >tousy<-data.frame(tous_y,lg)

#simtest for only one response
 >summary(simtest(F1_nearey~lg, type="Tukey", data=tousy))      

Absolute Error Tolerance:  0.001

Coefficients:
                       Estimate t value Std.Err. p raw p Bonf p adj
lgitalien-lgfrancais     -0.074  -1.898    0.039 0.066      1 0.495
lgitalien-lganglais      -0.064  -1.630    0.039 0.112      1 0.658
lgitalien-lgarabe        -0.055  -1.413    0.039 0.167      1 0.780
lgitalien-lgespagnol     -0.046  -1.166    0.039 0.252      1 0.891
lgportugais-lgfrancais   -0.039  -1.004    0.039 0.322      1 0.938
lgitalien-lgallemand     -0.037  -0.957    0.039 0.345      1 0.948
lgfrancais-lgallemand     0.037  -0.940    0.039 0.353      1 0.948
lgportugais-lgitalien     0.035  -0.894    0.039 0.378      1 0.955
lgportugais-lganglais    -0.029  -0.736    0.039 0.467      1 0.973
lgfrancais-lgespagnol     0.029  -0.732    0.039 0.469      1 0.973
lganglais-lgallemand      0.026  -0.672    0.039 0.506      1 0.978
lgportugais-lgarabe      -0.020  -0.519    0.039 0.607      1 0.992
lgfrancais-lgarabe        0.019  -0.485    0.039 0.631      1 0.994
lgespagnol-lganglais     -0.018  -0.464    0.039 0.646      1 0.994
lgarabe-lgallemand        0.018  -0.455    0.039 0.652      1 0.994
lgportugais-lgespagnol   -0.011  -0.272    0.039 0.787      1 0.999
lgfrancais-lganglais      0.011  -0.268    0.039 0.790      1 0.999
lgespagnol-lgarabe       -0.010  -0.247    0.039 0.807      1 0.999
lgarabe-lganglais        -0.008  -0.217    0.039 0.829      1 0.999
lgespagnol-lgallemand     0.008  -0.208    0.039 0.836      1 0.999
lgportugais-lgallemand   -0.002  -0.064    0.039 0.949      1 0.999


#manova for the two responses
 > fit <- manova(cbind(F1_nearey,F2_nearey)~lg)
 > summary.manova(fit,test = "Pillai")



phoneme F1_nearey F2_nearey
y -0.281 0.206
y -0.333 0.200
y -0.188 0.081
y -0.144 0.153
y -0.265 0.105
y -0.279 0.172
y -0.237 0.147
y -0.249 0.133
y -0.299 0.218
y -0.291 0.139
y -0.340 0.164
y -0.232 0.233
y -0.328 0.184
y -0.250 0.263
y -0.181 0.173
y -0.258 0.152
y -0.261 0.212
y -0.263 0.181
y -0.341 0.154
y -0.141 0.007
y -0.288 0.149
y -0.343 0.063
y -0.151 0.017
y -0.335 0.070
y -0.538 0.148
y -0.291 0.026
y -0.247 0.097
y -0.281 0.146
y -0.246 0.122
y -0.270 0.123
y -0.267 0.135
y -0.248 0.193
y -0.316 0.080
y -0.288 0.117
y -0.288 0.111
y -0.256 0.134
y -0.245 0.173
y -0.206 0.227
y -0.251 0.175
y -0.218 0.167
y -0.295 0.135
y -0.212 0.121

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html