I am conducting a multilevel regression analysis on the effect of an
intervention on student test results, and am not sure how to implement the
necessary R code to correctly capture the nested structure.
The outcome measure for the study is whether a student passed or failed a
final exam. The structure of the data is students nested within schools,
and then schools nested within random assignment blocks. Treatment (i.e.,
the intervention) was implemented at the school-level. The covariates that
I am planning to use are prior year test scores (this is also a binary
variable for pass or fail), race, and gender.
My ideal output would show the impact of the treatment for each of the
random assignment blocks, and then the weighted average of the impact
across all of the random assignment blocks.
Based on my research thus far, it seems like the **lmer** function from the
**lme4** package would be the best route to go.
This is the code that I have tried:
# Fit multilevel regression with random assignment blocks
glmer2 <- glmer(Post_Test_Score ~ Treatment +
Pre_Test_Score +
(1 | School) +
(1 | Random_Assignment_Block),
data = StudyData,
family = binomial("logit"))
My two questions are the following:
1.) Given the nested structure of my data, would the above regression
output the correct coefficient for the impact of treatment across all
random assignment blocks?
2.) How would I code the interaction effect between Treatment and
Random_Assignment_Block in order to generate separate impact estimates for
each of the random assignment blocks?
[[alternative HTML version deleted]]
1. Please post in plain text, not HTML, which can get garbled. 2. I believe your syntax is incorrect, but I haven't used lmer in a while, and so what I believe should be ignored anyway. HOWEVER, there is a SIG (special interest group) for mixed models, and you have a much better chance of getting reliable advice on such matters there. So you should sign up and post to R-sig-mixed-models on these topics rather than here. Cheers, Bert Bert Gunter "The trouble with having an open mind is that people keep coming along and sticking things into it." -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) On Thu, Jan 28, 2016 at 10:10 PM, David Roy <dmr02004 at gmail.com> wrote:> I am conducting a multilevel regression analysis on the effect of an > intervention on student test results, and am not sure how to implement the > necessary R code to correctly capture the nested structure. > > > > The outcome measure for the study is whether a student passed or failed a > final exam. The structure of the data is students nested within schools, > and then schools nested within random assignment blocks. Treatment (i.e., > the intervention) was implemented at the school-level. The covariates that > I am planning to use are prior year test scores (this is also a binary > variable for pass or fail), race, and gender. > > > > My ideal output would show the impact of the treatment for each of the > random assignment blocks, and then the weighted average of the impact > across all of the random assignment blocks. > > > > Based on my research thus far, it seems like the **lmer** function from the > **lme4** package would be the best route to go. > > > > This is the code that I have tried: > > > > # Fit multilevel regression with random assignment blocks > > glmer2 <- glmer(Post_Test_Score ~ Treatment + > > Pre_Test_Score + > > (1 | School) + > > (1 | Random_Assignment_Block), > > data = StudyData, > > family = binomial("logit")) > > > > My two questions are the following: > > > > 1.) Given the nested structure of my data, would the above regression > output the correct coefficient for the impact of treatment across all > random assignment blocks? > > > > 2.) How would I code the interaction effect between Treatment and > Random_Assignment_Block in order to generate separate impact estimates for > each of the random assignment blocks? > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.
Dear David, R-sig-mixedmodels is a better mailing list for this kind of question. 1) yes 2) use (Treatment | Random_Assignment_Block) instead of (1 | Random_Assignment_Block) Best regards, ir. Thierry Onkelinx Instituut voor natuur- en bosonderzoek / Research Institute for Nature and Forest team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance Kliniekstraat 25 1070 Anderlecht Belgium To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of. ~ Sir Ronald Aylmer Fisher The plural of anecdote is not data. ~ Roger Brinner The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data. ~ John Tukey 2016-01-29 7:10 GMT+01:00 David Roy <dmr02004 at gmail.com>:> I am conducting a multilevel regression analysis on the effect of an > intervention on student test results, and am not sure how to implement the > necessary R code to correctly capture the nested structure. > > > > The outcome measure for the study is whether a student passed or failed a > final exam. The structure of the data is students nested within schools, > and then schools nested within random assignment blocks. Treatment (i.e., > the intervention) was implemented at the school-level. The covariates that > I am planning to use are prior year test scores (this is also a binary > variable for pass or fail), race, and gender. > > > > My ideal output would show the impact of the treatment for each of the > random assignment blocks, and then the weighted average of the impact > across all of the random assignment blocks. > > > > Based on my research thus far, it seems like the **lmer** function from the > **lme4** package would be the best route to go. > > > > This is the code that I have tried: > > > > # Fit multilevel regression with random assignment blocks > > glmer2 <- glmer(Post_Test_Score ~ Treatment + > > Pre_Test_Score + > > (1 | School) + > > (1 | Random_Assignment_Block), > > data = StudyData, > > family = binomial("logit")) > > > > My two questions are the following: > > > > 1.) Given the nested structure of my data, would the above regression > output the correct coefficient for the impact of treatment across all > random assignment blocks? > > > > 2.) How would I code the interaction effect between Treatment and > Random_Assignment_Block in order to generate separate impact estimates for > each of the random assignment blocks? > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide > http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. >[[alternative HTML version deleted]]