Reputation
6,348
Top tag
Next privilege 10,000 Rep.
Access moderator tools
Badges
1 16 37
Newest
 Benefactor
Impact
~429k people reached

Jun
2
comment What's an intuitive way to explain the different types of validity?
There's not going to be one correct answer that will be memorable and intuitive to you, I'm afraid. But there are innumerable book chapters, articles, and websites on this topic. Good luck.
Apr
23
awarded  Benefactor
Apr
23
accepted Can a 1-D risk score (binary outcome) be sensibly used to create multiple treatment groups?
Apr
23
comment Can a 1-D risk score (binary outcome) be sensibly used to create multiple treatment groups?
Thank you, this was very helpful. Assuming that treatments carry their own risks, multiple cut points may well be justified. I'm trying to carry your work forward to simulate events and answer the question of how justified those multiple cutpoints are in the absence of such treatment risks.
Apr
21
revised Can a 1-D risk score (binary outcome) be sensibly used to create multiple treatment groups?
added 13 characters in body; edited title
Apr
21
comment Can a 1-D risk score (binary outcome) be sensibly used to create multiple treatment groups?
Looking back at your answer, it brings up points I didn't address but it seems to affirm the idea that whatever treatment is selected, it should be given to everyone who is treated.
Apr
21
comment Can a 1-D risk score (binary outcome) be sensibly used to create multiple treatment groups?
It makes a big difference that the outcome is binary. Take pregnancy: it is impossible to halfway prevent it. Any attempt to do so would result in wasted resources. It also makes a big difference that the risk score is continuous.
Apr
21
comment Can a 1-D risk score (binary outcome) be sensibly used to create multiple treatment groups?
Your assumptions remove all uncertainty from the situation. Of course if we could be certain who was bound to catch the virus then we would absolutely know what action to take for each person.
Apr
8
answered How best to communicate uncertainty due to data quality and measurement issues
Apr
8
revised How best to communicate uncertainty due to data quality and measurement issues
Please roll back my edits if they don't fit.
Apr
1
comment Do random forest variable importance measures take into account the interactions?
Two thumbs up! Nice work!
Apr
1
comment How to determine how many simulations to run, in order to illustrate “extreme-valued statistics”?
Welcome. Your criteria would need to be more specific than to have "confidence in the conclusions" and to obtain a histogram that is a "sufficiently detailed." (What constitutes "confidence" or "sufficient" for you, or for your audience?) I also think to get a helpful answer you would need to share more info about your simulation and a plot of your distribution so far.
Mar
27
comment How can I test for changes in the distribution of categorical data over time?
@gung - Ahh, but those 2 are not mutually exclusive [Commence throwing rotten fruit] :-)
Mar
27
comment Comparison of observations in subgroups?
From your tag, you're familiar with multi-level modeling. Can you say why you are leaning toward or away from such an approach? (If my edits are not correctly expressing your meaning, please roll them back.)
Mar
27
revised Comparison of observations in subgroups?
more precise?
Mar
26
comment How to specify reference category for binary independent variables in multinomial logistic regression in SPSS
It seems 'NomReg' doesn't allow the specification you are looking to use. With a 'do repeat' command set, though, you could recode many such predictors in one step.
Mar
26
comment Distribution of number of files in folders on personal computers?
Interesting question, but without any empirical data, how would anyone judge the realisticness of a particular answer?
Mar
26
comment Analyzing ordered factor vs continuous variable
Not to overdo it with semantics, but what does "best" mean to you? If you're dissatisfied with your regression approach maybe you could say why.
Mar
26
answered Doesn't Factor Analysis always overfit on a theoretical basis
Mar
26
comment How to quantify the Relative Variable Importance in Logistic Regression in terms of p?
See mwsug.org/proceedings/2009/stats/MWSUG-2009-D10.pdf or jstor.org/discover/10.2307/… . But your question about P seems ultimately inapplicable, as I think @Matt.135 suggested. Rel. imp. has to do with strength of connections between variables, while P involves the probability of a certain outcome given a certain value on 1 or more variables. Ex.: What does "age" have to do with "being married"? Unanswerable: it depends on the value of age. But the variable age can relate to the variable marital status.