Message-ID: <5c54172c-d037-8d31-f84e-866a79ac9d52@mpi.nl>
Date: 2019-10-22T02:04:26Z
From: Phillip Alday
Subject: Logit model in R
In-Reply-To: <05b53c66-daf5-8963-8a29-48add54649e7@gmail.com>
Also, since you emphasize that you're in cognitive science, it might
make sense to take a look at the following papers, which would bring
this closer to the topic of mixed models:
Jaeger, T. F. (2008). Categorical Data Analysis: Away from ANOVAs
(Transformation or Not) and Towards Logit Mixed Models. Journal of
Memory and Language , 59 (4), 434?446. doi:10.1016/j.jml.2007.11.007
Davidson, D. J., & Martin, A. E. (2013). Modeling accuracy as a function
of response time with the generalized linear mixed effects model. Acta
Psychologica , 144 , 83?96.
Best,
Phillip
On 22/10/2019 03:17, landon hurley wrote:
> Chiara,
>
>> I would like to ask which code i have to write in R to calculate the
>> percentage of categorial responses "Yes" or "Not" delivered for each of my
>> 15 perceptual stimuli.
> Typically the mean of a sequence of binary yes/no questions would be
> sufficient to answer this question. Take the m x n data set matrix D
> with n> 15 and apply the code
>
> colMeans(D[,1:15])
>
> to compute the mean of each column vector. The sequence 1:15 denotes the
> list sequence from the number 1 to the number 15, increasing by 1 at
> each step. If the 15 stimuli are not in sequential order, then they must
> be identified by the index sequence c(a,b,...,o) for which each letter
> is replaced by the respective column number of matrix D. Alternatively,
> the indices can be column names instead of numbers, for which each
> number must be enclosed in a separate " " quote string.
>
> colMeans(D[,c(a,b,...,o)])
>
> As a side note, you may wish to consider that since this is a mailing
> list for mixed models, it would be perhaps advisable to perhaps consider
> Stack Exchange or some other mailing list or other forum strictly
> devoted to performing basic operations in R. Also, since your email
> message has nothing to do with the implementation of a logit model in R,
> perhaps a better choice of email subject header would benefit in
> directing individuals to addressing your question.
>
> If you are interested in ultimately performing a regression upon a
> categorical unordered outcome measure, then I would recommend
> investigating the glm function in R, with the family operation set to
> 'binomial'.
>
> best,
>> Many thanks,
>> Chiara
>>
>> [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-mixed-models at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>>
>