10,620 reputation
31944
bio website
location England
age
visits member for 1 year, 9 months
seen 7 hours ago

11h
comment Can we skip the lower order terms in interactions?
What are $Y$, $X$, & $Z$?
14h
comment Interpretation of Zero-One inflated Beta Regression with R (GAMLSS)
?BEINF gives the parametrization actually used in any case.
14h
comment Beyond the basics: intermediate medical statistics textbooks suggestions
The more elementary one I'd recommend is Steyerberg (2009), Clinical Prediction Models: A Practical Approach to Development, Validation, and Updating, for which RMS "provided the main inspiration". (I'm not a medical statistician either, but I tell all & sundry to read these books.)
15h
comment Can we skip the lower order terms in interactions?
Considering your second model: if $W$ is a 0/1 dummy variable, then note that for one category there's no interaction between $X$ & $Z$; for the other there's a two-way interaction between $X$ & $Z$. Note also that when $Z=0$, the slope of $Y$ on $X$ is equal for the two categories. Are you really sure these constraints make sense for whatever you're modelling? - that the two categories distinguished by $W$ differ in this way & that zero is such a special value for $Z$?
15h
revised Can we skip the lower order terms in interactions?
formatted, fixed typo
15h
comment Can we skip the lower order terms in interactions?
A source for what, exactly? If you know how to multiply & how to carry out multiple regression then you know how to carry out a multiple regression in which one of the predictors happens to be the product of some others. Why you should or shouldn't want to is quite thoroughly dealt with in the post you linked to. Venables (1998),"Exegeses on Linear Models", S-Plus User’s Conference, Washington DC also discusses the empirical modelling of interactions, & the marginality principle.
1d
comment Two sample $t$ test in R: where did it go wrong?
Read the manual (t.test). By default you're using the Welch-Satterthwaite approximation to the error degrees of freedom.
1d
comment Interpretation of Zero-One inflated Beta Regression with R (GAMLSS)
$\mu$ & $\sigma$ don't appear on the right-hand side of the p.d.f. formula, whereas a previously unmentioned $p$ & $q$ do. Can you clarify?
Aug
29
revised explanatory variables may bias predictions
added 264 characters in body
Aug
29
comment Is there ever a reason not to use orthogonal polynomials when fitting regressions?
Why the down-vote? (Fair enough if it's for the strained humour.)
Aug
29
comment explanatory variables may bias predictions
Why the down-vote? If there's a mistake it'd be helpful to point it out - I can't see one.
Aug
29
comment Can we skip the lower order terms in interactions?
Of course it can be done (or there'd be no point telling people not to do it) - just add a new variable V=XZW to the design matrix. That you have to ask rather suggests you need to think through carefully your reasons for wanting to - the stated one of only being interested in the higher order term doesn't really cut it.
Aug
29
comment strucchange problem with csv files
Any issues have likely arisen by not defining the data you read in as a time series - see ?ts.
Aug
28
comment R lme4 1.1-7: REML=FALSE giving error “extra argument(s) ‘REML’ disregarded ”
@Patrick: On the face of it, but there's an underlying statistical question: How do you (does it make sense to) fit a non-Gaussian generalized linear mixed model using restricted maximum likelihood? IMO it'd be better to address that in the answers rather than close the question.
Aug
28
reviewed Approve suggested edit on How to test for a home bias in a data set?
Aug
27
reviewed Leave Open R lme4 1.1-7: REML=FALSE giving error “extra argument(s) ‘REML’ disregarded ”
Aug
27
reviewed Leave Open Has anyone publicly shared an implementation of RUSBoost in R?
Aug
27
answered Do you know of any public circular/angular dataset?
Aug
27
comment Mann Whitney test with unequal variances
... implications that one might think; in practice I then look at histograms or whatever to see what's going on.
Aug
27
comment Mann Whitney test with unequal variances
@ttnphns: It tests, in general, whether the probability that an observation from one population is greater than an observation from the other population differs from one half. If you assume the cdfs don't cross that implies stochastic dominance of one over the other. It's often a reasonable assumption - a poison, say, retards the growth of some seedlings more than others, but doesn't advance that of any - & can be informally checked by examining the empirical cdfs. Without that assumption, you can say, well, the probability that an observation &c., but that statement doesn't have all the ...