Second chance at a PhD?

[GradcafeForOlds]

I started a stat PhD a few years ago, and didn't really put in the work I (in retrospect) should've, and ultimately failed quals.  I've been working since then (not in anything where I work with PhD scientists, unfortunately), but part of me really thinks that I want to do a PhD, and that I can actually pull it off now that I know exactly what I need to do.  It's probably too late to start applying this year, so don't worry about immediate actions, but I wanted to check what stuff you guys think I would need to do to actually stand a chance in admissions.

 

Do I need to go back to undergrad to take some more math classes?  I didn't do well in my PhD classes, which is why I ask -- my undergrad grades were fine, but after following them up with some grad school C's I doubt they'll look as enticing.  

 

Should I ask for a letter from my old PhD program?  They probably wouldn't have great things to say about my work ethic or brains, but I thought some schools might want confirmation that I didn't cheat or violate Title 9 or anything disqualifying other than my performance, and that I'm not some eternal pain to have around.  I know this is a thing for Profs at least when they change schools, not sure if it would matter in my case.

 

I can probably guess this one, but should I get a master's first?  I want a PhD pretty surely, but proving myself on my dime is probably the best way to get some trust back.  I'm still in the area, so would my old PhD program have any chance of accepting me for a master's?

 

Finally, I doubt it matters, but should I look at more biostat-y, applied programs?  My grades in grad school were pretty universally bad (including in things like linear models that I like), but being at a pretty theoretical program made it harder to stay on top of everything for me, and maybe programs that aren't as focused on that wouldn't be as concerned about my first time through.  Probably not, I know, but I'm open to biostat too, so I thought I'd test the waters.

[bayessays]

I wouldn't bother going back to take undergrad classes if you already have them. Their major concerns are going to be that you can't handle graduate coursework or that you are lazy or that you got kicked out (but I don't think you really have to worry about them thinking you cheated since you didn't get any Fs and your grades will probably paint the picture).

I'd say a lot probably depends on how long it's been, what you've been up to in the meantime, and what type of program you went to before.  You want to create a compelling narrative.

"I started my PhD st University of Chicago, and because of the extremely theoretical nature of the program and my lack of maturity at the time, I struggled and left after two years.  In the meantime, I've been working the last 3 years at a health organization, and I now have the context and focus to want to pursue applied statistics research and contribute to public health, which is why I would like to get my PhD in biostatistics from the University of Cincinnati" is a relatively compelling narrative.

If you formed any solid relationship in your previous program, such a letter would be very valuable. If you have formed any solid working relationship with a PhD statistician in your job after who can attest to your work ethic, that would be valuable.  If you could retake some of those classes you did poorly in at your old school and form such a relationship to get a letter, that would certainly help but only you can make that financial decision. 

Really, dropping out of a PhD program and coming back to start somewhere else isn't the end of the world.  Your biggest concern though is the Cs, which are usually considered failing grades. If you went to a top program before and are willing to go to a much lower ranked program now, your path will be a lot easier.  A PhD student at a biostat program outside the top 10 is not going to take classes in the same realm of difficulty as a first year U Chicago student, for example.

[GradcafeForOlds]

Thanks for the response.  The only reason I brought up undergrad was to maybe take versions of things I hadn't done well in grad school (like linear models and measure theory, maybe at a grad level of allowed), which might be easier and cheaper than a full master's program.