Andreia Fonseca
2013-Oct-10 10:44 UTC
[R] help for compare regression coefficients across groups
Dear all
I have data related to cell count across time in 2 different types of
cells. I have transformed the count data using a log
and I want to test the H0: B cell_ttype1=Bcell_type2 across time
for that I am fitting the following model
fit_all<-lm(data$count~data$cell_type+data$time+data$cell_type*data$time)
the output is
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.0450021 0.0286824 36.434 <2e-16 ***
data$cell_typeOV -0.0456669 0.0405631 -1.126 0.271
data$time 0.0115620 0.0004815 24.015 <2e-16 ***
data$cell_typeOV:data$time -0.0009764 0.0006809 -1.434 0.164
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05
'.' 0.1 ' ' 1
Residual standard error: 0.06318 on 26 degrees of freedom
Multiple R-squared: 0.9764, Adjusted R-squared: 0.9737
F-statistic: 358.8 on 3 and 26 DF, p-value: < 2.2e-16
inspite the fact that the p-value of he interaction is >0.05 may I still
conclude that B cell_ttype1 is different from Bcell_type2 because the
p-value of the fit is lower<0.05?
Thanks in advance for your help.
With kind regards,
Andreia
--
---------------------------------------------------------------------------------------------
Andreia J. Amaral, PhD
BioFIG - Center for Biodiversity, Functional and Integrative Genomics
Instituto de Medicina Molecular
University of Lisbon
Tel: +352 217500000 (ext. office: 28253)
email:andreiaamaral@fm.ul.pt ; andreiaamaral@fc.ul.pt
[[alternative HTML version deleted]]
David Winsemius
2013-Oct-11 01:38 UTC
[R] help for compare regression coefficients across groups
On Oct 10, 2013, at 3:44 AM, Andreia Fonseca wrote:> Dear all > > I have data related to cell count across time in 2 different types of > cells. I have transformed the count data using a log > and I want to test the H0: B cell_ttype1=Bcell_type2 across time > > for that I am fitting the following model > > fit_all<-lm(data$count~data$cell_type+data$time+data$cell_type*data$time) > > the output is > > Coefficients: > Estimate Std. Error t value Pr(>|t|) > (Intercept) 1.0450021 0.0286824 36.434 <2e-16 *** > data$cell_typeOV -0.0456669 0.0405631 -1.126 0.271 > data$time 0.0115620 0.0004815 24.015 <2e-16 *** > data$cell_typeOV:data$time -0.0009764 0.0006809 -1.434 0.164 > --- > Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 > > Residual standard error: 0.06318 on 26 degrees of freedom > Multiple R-squared: 0.9764, Adjusted R-squared: 0.9737 > F-statistic: 358.8 on 3 and 26 DF, p-value: < 2.2e-16 > > > inspite the fact that the p-value of he interaction is >0.05 may I still > conclude that B cell_ttype1 is different from Bcell_type2 because the > p-value of the fit is lower<0.05?You have offered the output of an interaction model and only provided the Wald statistics on terms, which is a situation where those values are usually meaningless. Furthermore, you have not described the study or the data in any detail. You should seek competent local consultation at your institution. If you get an answer based only on this information, that should result in a lower opinion (in the Bayesian sense) of the competence of the responder. -- David.> > Thanks in advance for your help. > > With kind regards, > > Andreia > > -- > --------------------------------------------------------------------------------------------- > Andreia J. Amaral, PhD > BioFIG - Center for Biodiversity, Functional and Integrative Genomics > Instituto de Medicina Molecular > University of Lisbon > Tel: +352 217500000 (ext. office: 28253) > email:andreiaamaral at fm.ul.pt ; andreiaamaral at fc.ul.pt > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at r-project.org mailing list > 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.David Winsemius Alameda, CA, USA