Publications by Keith Goldfeld
Analysing an open cohort stepped-wedge clustered trial with repeated individual binary outcomes
I am currently wrestling with how to analyze data from a stepped-wedge designed cluster randomized trial. A few factors make this analysis particularly interesting. First, we want to allow for the possibility that between-period site-level correlation will decrease (or decay) over time. Second, there is possibly additional clustering at the patie...
8860 sym R (4977 sym/8 pcs) 6 img
Clustered randomized trials and the design effect
I am always saying that simulation can help illuminate interesting statistical concepts or ideas. The design effect that underlies much of clustered analysis is could benefit from a little exploration through simulation. I’ve written about clustered-related methods so much on this blog that I won’t provide links – just peruse the list of en...
11152 sym R (3366 sym/19 pcs) 6 img
Alternatives to reporting a p-value: the case of a contingency table
I frequently find myself in discussions with collaborators about the merits of reporting p-values, particularly in the context of pilot studies or exploratory analysis. Over the past several years, the American Statistical Association has made several strong statements about the need to consider approaches that measure the strength of evidence or...
7828 sym R (7192 sym/8 pcs) 10 img
When you want more than a chi-squared test, consider a measure of association for contingency tables
In my last post, I made the point that p-values should not necessarily be considered sufficient evidence (or evidence at all) in drawing conclusions about associations we are interested in exploring. When it comes to contingency tables that represent the outcomes for two categorical variables, it isn’t so obvious what measure of association sho...
8770 sym R (1620 sym/7 pcs) 16 img
Can unbalanced randomization improve power?
Of course, we’re all thinking about one thing these days, so it seems particularly inconsequential to be writing about anything that doesn’t contribute to solving or addressing in some meaningful way this pandemic crisis. But, I find that working provides a balm from reading and hearing all day about the events swirling around us, both here a...
5319 sym R (2424 sym/7 pcs) 2 img
Yes, unbalanced randomization can improve power, in some situations
Last time I provided some simulations that suggested that there might not be any efficiency-related benefits to using unbalanced randomization when the outcome is binary. This is a quick follow-up to provide a counter-example where the outcome in a two-group comparison is continuous. If the groups have different amounts of variability, intuitivel...
4018 sym R (1839 sym/7 pcs) 4 img
Simulation for power in designing cluster randomized trials
As a biostatistician, I like to be involved in the design of a study as early as possible. I always like to say that I hope one of the first conversations an investigator has is with me, so that I can help clarify the research questions before getting into the design questions related to measurement, unit of randomization, and sample size. In the...
7707 sym R (2950 sym/7 pcs) 8 img
To stratify or not? It might not actually matter…
Continuing with the theme of exploring small issues that come up in trial design, I recently used simulation to assess the impact of stratifying (or not) in the context of a multi-site Covid-19 trial with a binary outcome. The investigators are concerned that baseline health status will affect the probability of an outcome event, and are interest...
5528 sym R (3965 sym/5 pcs) 8 img
Considering the number of categories in an ordinal outcome
In two Covid-19-related trials I’m involved with, the primary or key secondary outcome is the status of a patient at 14 days based on a World Health Organization ordered rating scale. In this particular ordinal scale, there are 11 categories ranging from 0 (uninfected) to 10 (death). In between, a patient can be infected but well enough to rema...
6053 sym R (2033 sym/8 pcs) 4 img
When proportional odds is a poor assumption, collapsing categories is probably not going to save you
Continuing the discussion on cumulative odds models I started last time, I want to investigate a solution I always assumed would help mitigate a failure to meet the proportional odds assumption. I’ve believed if there is a large number of categories and the relative cumulative odds between two groups don’t appear proportional across all categ...
6153 sym R (3892 sym/14 pcs) 8 img