I won't say that I wish I did, because I did my grad work in Spatial Informatics, but I can speak from experience with my peers.  I often have to consult for my colleagues at various locations and in various disciplines.  That says the skill is incredibly applicable and necessary.  I also have taught research statistics for secondary students with all of them serviced by graduate or PhD mentors.  Those grad students and PhD recipients all commented on how the curriculum that I was presenting was new even to them (basically statistical interpretation).  I think that speaks volumes about the regret and the lack in current curricula.

I personally believe that it needs to be a requirement to graduate high school, then to get an undergraduate degree, and then to get an advanced degree.  It is literally that important, even if you never go on to do research.  It is part of being a literate citizen given mass communication practices.

Thank you for your feedback, Matthew. The traditional statistics graduate programs are most likely "evolving" towards Data Science, which can be a good thing in which we witness a better match between academic training and the job market and also grants opportunities for the faculty, and a decline of traditional "deep" statistics courses that challenge the brain and that help create future statistics professors with a deep knowledge of statistics. Maybe it is a matter of having a healthy balance.

Dear Raid

Very pertinent issue. Being a data analyst and a teacher for applied statistics since last more than twenty years and looking researches very closely, I also of the view that it has the incredible potential for quality research, planning and policy making. Currently this can be described in one word as you rightly mentioned "data science', However, the issue is how to integrate the thinking, associated data and right analysis. Courses, training and practicing these issues is one of the option. therefore, I would be also of the view that the basic and advance level statistics course would be always beneficial for all. I would rather add that it should be for common mass also. 

Regards

Dear Rajiv,

The Data Science wave is new. and many things are still left unclear and undefined. Companies and universities seem to be rushing to create data science units quickly. The one obvious good outcome has been a significant demand in statistics and a great increase in starting salaries. My department may get 200 Ph.D.(in Mathematics)  applicants for an opening in Mathematics and only 35 Ph.D. (in statistics) applicants for an opening in Statistics. The job market "is too good".

To recent Ph.D. recipients or Ph.D. students: Do you wish you had taken more statistics courses?

I wish I had taken more statistics courses based on certain mini-research data set (without data set to analyze I might forget the usefulness of the statistics I'd learned).  

Statistics are too wide & we only focus on certain statistics relevant to our thesis during the PhD study.  In order to address above, some universities started to conduct their statistics classes repetitively every semester so that students can re-attend when they have missed / forgotten certain statistics.  Above is a good approach but for those statistics that you didn't use in the actual thesis, you might also forget eventually.  

What I'm doing now is leveraging on the statistics that I'd learned for my research.  In the event on certain statistics that I'm not familiar with, I'll do my own research or attend certain courses organized by universities (but I will bring along some data set for analysis to gain better understanding).

Thank you for your feedback, Han Ping Fung.

The need for continuous education exists, and having course offerings each semester should be helpful in this matter. I try to engage recent graduates in ongoing research projects with me to keep the past students supported and to keep their interest in using statistics high. There are challenges that then must be overcome, such as allowing alumni access to university facilities and computers. I am working things out with Instructional Technology (IT) so that past students receive extended access to software.

Raid Amin, I think for new researcher the statistics courses should be related or blended with data mainpulation software SAS etc. In fact apart from basic statistics, it will be more practical, relevant and future oriented to give extensive experience of handling the data on computer softwares rather than courses on pure statistics itself.

Thank you S. Sharma for your feedback.

I always build my statistics courses around SAS and around many real life data sets. I have to infuse the statistics foundations and then explain the concepts as I analyze data for the class.