Masters: Stats at LSE or Econ at Cambridge?

Watch
Chr0n
Badges: 14
Rep:
?
#1
Report Thread starter 6 years ago
#1
Hey guys,

so, some of you may know that I was pretty set on going to LSE for an M.Sc. Statistics next year. However, I have just received an offer for the M.Phil. in Economics at Cambridge and am really torn between the two. (The date on the offer is the 31, luckily ). Especially because I can take 4 econometrics modules at Cambridge (basically statistics) + some really interesting finance modules..

I am planning to do a PhD in the states in either Statistics or Machine Learning. So, I don't think doing statistics over economics would be a huge advantage (so long as the degree is quantitative it should be fine). If anything, I feel like if I do statistics now, I will be bored in the first year of a statistics programme in the states as I would learn the same material again.

What would you do?
0
reply
Chr0n
Badges: 14
Rep:
?
#2
Report Thread starter 6 years ago
#2
bump
0
reply
poohat
Badges: 18
Rep:
?
#3
Report 6 years ago
#3
Ironically you would probably be better doing a Statistics MSc at Cambridge or an Economics MSc at LSE. However either will look good on your CV and will give you a shot at decent PhD programs.

Especially because I can take 4 econometrics modules at Cambridge (basically statistics) + some really interesting finance modules..
Kind of, but economics is a fairly narrow branch of statistics. You will come out knowing everything there is to know about linear regression,classical time series analysis, and testing for causality, but not much about anything else. That isn't necessarily a bad thing though since as you say, the first year of grad school would fill in the gaps.

I feel like if I do statistics now, I will be bored in the first year of a statistics programme in the states as I would learn the same material again
Given that most 1st year US grad students find the first year incredibly, incredibly stressful and by far the worst part of the program, being 'bored' because its too easy is a good thing, not a bad thing.
0
reply
Chr0n
Badges: 14
Rep:
?
#4
Report Thread starter 6 years ago
#4
(Original post by poohat)
Ironically you would probably be better doing a Statistics MSc at Cambridge or an Economics MSc at LSE. However either will look good on your CV and will give you a shot at decent PhD programs.
Haha I know! Unfortunately, Cambridge only offers statistics as part of Maths 3. Given that I did PPE I thought I would have no chance of getting into that. I was very interested in the LSE Mathematical Economics and Econometrics programme, but the fees put me off. Especially because you can take 90% of the same modules in the statistics programme for half the price.

(Original post by poohat)
Given that most 1st year US grad students find the first year incredibly, incredibly stressful and by far the worst part of the program, being 'bored' because its too easy is a good thing, not a bad thing.
More so than the MPhil in Cambridge? In any case I think Econometrics would still be a good grounding for a statistics degree, as would statistics.
0
reply
Chr0n
Badges: 14
Rep:
?
#5
Report Thread starter 6 years ago
#5
I have now also received an offer from UCL for the M.Sc. Computational Statistics and Machine Learning. So, now I really don't know what to do anymore. Not only is it Cambridge vs. UCL vs. LSE but also Econometrics vs. Statistics vs. Machine Learning.
0
reply
Cal-lum
Badges: 13
Rep:
?
#6
Report 6 years ago
#6
I would personally go for Cambridge.

The courses are going to get you to the same place, and what you will learn will cross over too, so I would go for Cambridge based on the location, experience, etc... remember to take into account the environment and your living as well as the course!
0
reply
Blutooth
Badges: 16
Rep:
?
#7
Report 6 years ago
#7
(Original post by Chr0n)
I have now also received an offer from UCL for the M.Sc. Computational Statistics and Machine Learning. So, now I really don't know what to do anymore. Not only is it Cambridge vs. UCL vs. LSE but also Econometrics vs. Statistics vs. Machine Learning.

If you want to do Machine Learning, I imagine you'd be able to get in through economics, though I doubt your course would be tailored to your interests. You would have to do a lot of Macro & Micro. However, I think getting into a PHD in stats/ ML is definitely possible with an econ background. They care much more about what you know (linear algebra/ calculus / probability/ stats/ programming?) than your degree programme.


Anyway, here are my opinions:

University environment:
Cambridge>>>UCL>=LSE

Renown of the university
Cambridge>LSE>=UCL

Chances of getting into a good Phd program in statistics
LSE Stats = UCL CSML = Cambridge Economics

Chances of getting into a good Phd program in ML
UCL CSML >= LSE Stats> =Cambridge Economics (

Best course- given your interests
UCL = LSE>>> Cambridge

Cambridge might be good if you want to keep the door open and do something in Econ.

Personally, I'd probably either do Econ at Cambridge or CSML at UCL. I think UCL has the best ML course in the country, but going to Cambridge would be a unique experience at any college. I guess you have to deice whether the place and prestige or the degree is more important to you. Having said that, Econ is a bit expensive. I would probably regretfully have to turn down Econ at Cambridge. No point studying something you are not really interested in. But I reckon your chances would be more or less the same wherever you go- there are interviews for PHds and they'd get a chance to see how much you really knew at interview.

Good luck.
0
reply
Chr0n
Badges: 14
Rep:
?
#8
Report Thread starter 6 years ago
#8
(Original post by Blutooth)
If you want to do Machine Learning, I imagine you'd be able to get in through economics, though I doubt your course would be tailored to your interests. You would have to do a lot of Macro & Micro. However, I think getting into a PHD in stats/ ML is definitely possible with an econ background. They care much more about what you know (linear algebra/ calculus / probability/ stats/ programming?) than your degree programme.
That is the thing, in Cambridge I can take 4 out of 8 modules in Econometrics, 2 in Mathematical Finance (e.g. Asset Pricing) and then do only the compulsory Micro and Macro courses. On the other hand, I am very interested in forecasting financial markets and the Micro theory might actually be very useful for portfolio theory. Similarly, I can easily join a stats department and do a thesis in 'Statistical Learning' rather than 'Machine Learning'. Which is exactly the same thing, but in the stats department and not the CS department. Going the Cambridge would also allow me to sit in on the Mathematics III lectures and gain a thorough grounding in Mathematical Statistics.

UCL on the other hand would give me the thorough grounding in both Statistics and ML and I could easily take one or two advanced Econometrics classes on the side.


(Original post by Blutooth)
University environment:
Cambridge>>>UCL>=LSE

Renown of the university
Cambridge>LSE>=UCL

vs.

Best course- given your interests
UCL = LSE>>> Cambridge
This is exactly why I am so torn :P.

All in all I suppose it is a very good dilemma to have though.
0
reply
Blutooth
Badges: 16
Rep:
?
#9
Report 6 years ago
#9
(Original post by Chr0n)
That is the thing, in Cambridge I can take 4 out of 8 modules in Econometrics, 2 in Mathematical Finance (e.g. Asset Pricing) and then do only the compulsory Micro and Macro courses. On the other hand, I am very interested in forecasting financial markets and the Micro theory might actually be very useful for portfolio theory. Similarly, I can easily join a stats department and do a thesis in 'Statistical Learning' rather than 'Machine Learning'. Which is exactly the same thing, but in the stats department and not the CS department. Going the Cambridge would also allow me to sit in on the Mathematics III lectures and gain a thorough grounding in Mathematical Statistics.

UCL on the other hand would give me the thorough grounding in both Statistics and ML and I could easily take one or two advanced Econometrics classes on the side.



This is exactly why I am so torn :P.

All in all I suppose it is a very good dilemma to have though.
If you are doing half the modules in econometrics I guess that would be ok. Still, I think the kind of statistics you'd be doing would be watered down in comparison to that studying on a statistics/ machine learning masters. I was in a similar position, as I was thinking of applying to UCL for CSML. In the end I went for ML at Cambridge, even though there dis a bit less variety in terms of modules offered. Still, I think it will be great next year.

If you do go for Cambridge you could drop in on stats/ ml lectures elsewhere- but then you might find that a bit too much work. Congrats on your offers- great dilemma to have!
0
reply
poohat
Badges: 18
Rep:
?
#10
Report 6 years ago
#10
(Original post by Chr0n)
I have now also received an offer from UCL for the M.Sc. Computational Statistics and Machine Learning. So, now I really don't know what to do anymore. Not only is it Cambridge vs. UCL vs. LSE but also Econometrics vs. Statistics vs. Machine Learning.
That UCL course is arguably better than the other two, but Cambridge has much more name recognition in the US.

I would probably forget about LSE at this point, its a straight choice between course and prestige.
0
reply
Chr0n
Badges: 14
Rep:
?
#11
Report Thread starter 6 years ago
#11
(Original post by Blutooth)
If you are doing half the modules in econometrics I guess that would be ok. Still, I think the kind of statistics you'd be doing would be watered down in comparison to that studying on a statistics/machine learning masters.
Hm I don't think that the Statistics would be watered down as much as you think. Econometrics at graduate level is, usually, very quantitative and the Econometricians at Cambridge are meant to be very theoretically minded. In some courses, like Time Series, you are actually better of in an Econometrics department as much of modern Time Series has been developed by Econometricians (e.g. Hamilton has long been a standard textbook on Time Series - written by an Econometrician. Although Hamilton is of course quite old and in need of revision now). What I fear more is not so much that the stuff will be watered down as that I will miss out on a lot of breadth. I will learn everything there is to know about Time Series, Causality, Instrumental Variables, etc. but will not cover other standard topics such as classification, multilevel modelling, etc. The question of course is, whether this is relevant if I do a PhD in statistics in the states as that would fill the gaps.

The other worry I have is that a terminal masters in Economics is really not that useful and it closes of all the possibilities of doing a statistics / ml PhD in the UK, unless I do another masters first.

(Original post by Blutooth)
I was in a similar position, as I was thinking of applying to UCL for CSML. In the end I went for ML at Cambridge, even though there dis a bit less variety in terms of modules offered. Still, I think it will be great next year.
Haha, yes, I now think I should have applied to that course. I looked at it at the time but didn't like the language processing part. Congrats on the offer!


(Original post by poohat)
That UCL course is arguably better than the other two, but Cambridge has more recognition in the US. I would probably forget about LSE at this point, its a straight choice between course and prestige.
Haha, yes indeed! And of course quality of life. I would much prefer the Cambridge environment to London. I think London is too crowded. However, since it's only one year it might not matter that much.
0
reply
poohat
Badges: 18
Rep:
?
#12
Report 6 years ago
#12
One thing to note is that people outside economics often don't realise how statisticla/mathematical it is, so if you do the econ degree you should try and make sure your transcript has things like 'real analysis' on it, and the names of some statistics/metrics modules that make it clear they arent just linear regression.

I think if you were doing a PhD in the UK then I would say that the UCL course is the best choice, but if you want to go to the US then Cambridge is maybe slightly better, but you might have to convince an admissions committee that you do have a strong maths/stats background. Its not entirely out of the question that (e.g.) a CS adcom would assume that an economics grad didnt have the required quantitative background, even though that is obviously absurd.

I would guess (and this really is a guess) that statisticians would view an econometrics masters more favorable than CS/machine learning people would, since there is a far higher probability that they would know what it is.
0
reply
Chr0n
Badges: 14
Rep:
?
#13
Report Thread starter 6 years ago
#13
I suppose you are right. I won't be able to put stuff like Real Analysis on my transcript, as I can only take courses from the Economics department, which usually don't have mathematical names, but I suppose this can always be rectified in the personal statement or the reference letters. I will also definitely do my dissertation in Econometrics, so that should be good to display my quantitative skills too.

Oh regarding LSE, I think you are right. The LSE is far too applied for my taste anyway, I much prefer hardcore Mathematical Statistics (hence see for example my earlier thread Warwick vs. LSE). The UCL course can be made out to be quite theoretical, which I like.
0
reply
poohat
Badges: 18
Rep:
?
#14
Report 6 years ago
#14
If you want to do hardcore mathematical statistics then why not just do an Econ PhD and specialize in econometrics? I would only advise that if you could get into a top 5-10 US Econ PhD program but if you graduated as one of the top few people in your MSc then that wouldn't be out of the question.

Economics pays a _lot_ more than statistics in the US (starting academic salaries are around $120-140k in econ compared to about $80k in stats), and a bit more than CS too.

The main advantage of statistics and ML is that you get to work on a wide range of diverse applications, but if you just want to do mathematics then econ is the way to go (as long as you can get into a top program)
0
reply
Blutooth
Badges: 16
Rep:
?
#15
Report 6 years ago
#15
(Original post by Chr0n)
Hm I don't think that the Statistics would be watered down as much as you think. Econometrics at graduate level is, usually, very quantitative and the Econometricians at Cambridge are meant to be very theoretically minded. In some courses, like Time Series, you are actually better of in an Econometrics department as much of modern Time Series has been developed by Econometricians (e.g. Hamilton has long been a standard textbook on Time Series - written by an Econometrician. Although Hamilton is of course quite old and in need of revision now). What I fear more is not so much that the stuff will be watered down as that I will miss out on a lot of breadth. I will learn everything there is to know about Time Series, Causality, Instrumental Variables, etc. but will not cover other standard topics such as classification, multilevel modelling, etc. The question of course is, whether this is relevant if I do a PhD in statistics in the states as that would fill the gaps.

The other worry I have is that a terminal masters in Economics is really not that useful and it closes of all the possibilities of doing a statistics / ml PhD in the UK, unless I do another masters first.


Haha, yes, I now think I should have applied to that course. I looked at it at the time but didn't like the language processing part. Congrats on the offer!



Haha, yes indeed! And of course quality of life. I would much prefer the Cambridge environment to London. I think London is too crowded. However, since it's only one year it might not matter that much.
I suppose you might be right. My opinion has been formed by picking up econometrics texts from my library in college and comparing them to the likes of Bishop and Murphy. They might not have been representative though.

You raise another good point about breadth. I suppose you could explore the other topics that interest you in the dissertation. I would like to apply machine learning to finance, so if you do go to Cambridge send me a pm
0
reply
X

Quick Reply

Attached files
Write a reply...
Reply
new posts
Back
to top
Latest
My Feed

See more of what you like on
The Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

Personalise

Poll: What factors affect your mental health most right now? Post-lockdown edition

Anxiousness about restrictions easing (55)
6.16%
Uncertainty around my education (101)
11.31%
Uncertainty around my future career prospects (94)
10.53%
Lack of purpose or motivation (111)
12.43%
Lack of support system (eg. teachers, counsellors, delays in care) (44)
4.93%
Impact lockdown had on physical health (48)
5.38%
Social worries (incl. loneliness/making friends) (99)
11.09%
Financial worries (60)
6.72%
Concern about myself or my loves ones getting/having been ill (41)
4.59%
Exposure to negative news/social media (56)
6.27%
Difficulty accessing real life entertainment (26)
2.91%
Lack of confidence in making big life decisions (86)
9.63%
Worry about missed opportunities during the pandemic (72)
8.06%

Watched Threads

View All