Collaborative Assessment Log: Fill & Download for Free

Many softwares are available online and offline to assess tangible results of an organization such as growth, profits, finance etc,. However, for considering human centric elements like team and cultural effectiveness, no software can provide very accurate results. These are factors that require to be assessed through human interactions alone and can result into more actionable outcomes. For more information on culture and team assessments, log on to www.hardindiwali.com

Organization’s collaborative culture is a factor of many parameters such as the vision and mission of the company and how employees relate to it, communication, hierarchical relationships, values of the organization, collaboration within people etc,. It is difficult for a set framework to assess the culture of different organizations considering human behaviours and values are different in every company. Thus a more personal and human interaction and assessment will provide more actionable results and outcomes. For more information on culture assessments log on to www.hardindiwali.com

First off: I don't really know the answer. The theory of these models is a very technical area, and as I understand there is no fleshed-out theory (yet) of "exact" hypothesis tests for survival analysis, because you would need an exact distribution of censoring times.Second: If you are planning to submit to NEJM or JAMA, you are best off working with a real biostatistician or epidemiologist to help you design your study and draft a Statistical Analysis Plan before you go and do it, even if it's a case-control study. You can hire these people (e.g. STATKING Consulting or many other firms) if you don't have access to academic collaborators. This is simply the norm in this area -- unlike in the physical sciences, medical researchers generally do not analyze their own data. They bring in statisticians. And pre-specification of the method is important; unless this is purely "hypothesis-generating" work, you can't just collect all the data and then noodle around with the analysis until you find something significant.Third: In designing the study and choosing the statistical methods, it is about as important to pick something culturally acceptable to the readership as it is to pick a technique with provably nice mathematical properties. The readers and referees at the NEJM are familiar with Kaplan-Meier estimates, Cox proportional-hazards models, and the log-rank and Wilcoxon tests.Fourth: My own understanding is that the Cox proportional hazards model and the log-rank test are complementary and are not going to disagree. You can use a multivariate Cox regression to estimate the hazard ratio (relative risk) for various characteristics, and you can use the log-rank test to test significance at a desired alpha. Indeed, the log-rank test is apparently THE score test for the Cox proportional-hazards model.Many, many NEJM papers (seems like >60%) do a Cox proportional-hazards analysis and present p-values from the log-rank test. So this is not going to surprise anybody if your HRs are from the Cox model and your p-values are from the log-rank test.My understanding is that that is the difference between the techniques: the Cox proportional-hazards model is an estimator for a hazard ratio, and the log-rank test is a hypothesis test attempting to reject the hypothesis that a hazard ratio equals unity. I don't quite know the precise difference between when you use log-rank versus Wilcoxon.Fifth: I see you are at Harvard -- you have a whole biostatistics and epidemiology department at the School of Public Health, filled with some of the world's experts in this stuff. They will probably know a lot more than anybody on Quora. Your colleagues David Harrington, David Hunter, and James Ware are three of the four statistical consultants to the NEJM. Brent Coull is also there and a nice guy.