計劃進修PhD/Research討論區 (11) All money is a matter of belief

推得勁快，痴撚線

開埋訓

快過299792458m/s

想問meta analysis唔入non significant findings 去分析會有咩bias? 因為其實入哂non significant findings最後個combined effect可能變返significant result. 咁有咩意思?

呢個問題存在好耐
open database 係解決之道

呢個情況其實似你普通run individual data時
你放兩組15+15人時唔sig, 放30+30時可能sig, 每個人自己baseline posttest比可能唔食

一切係effect size and power既問題

如果你唔放non sig data, 一來係publication bias selection bias,
二來驗證唔到個effect size係因為power定其他

當然，如果你越放得多越heterogeneous, 咁代表不同paper variety 高

講到尾，meta analysis都只不過係一種estimation，我唔明點解咁多人覺得佢好神

sor, 有感而發

入定唔入好 我覺得入左反而會影響左個result⋯

首先個問題係
有冇publication bias, 呢個係現時既meta analysis 答唔到
因為你搵到null result 既好大機會pub 唔到
咁你published 左既好大機會已經biased
第二個問題係convergence
因為我一個study 都有佢地自己既study error
例如因為sample size太細
所以error 大左，false positive and negative 多左
咁每個study 搵到個results 都會唔同
咁meta analysis 做既野就係搵個convergence
有冇個consistent 既finding

點解open database 最好
我一次過可以睇晒上面2個問題

簡單講下sigma algebra是乜唻ok?
Probability space個幾個element又係乜

簡單講下sigma algebra是乜唻ok?
Probability space個幾個element又係乜

sample space element = a possible outcome
event space (or sigma algebra) includes all possible events, where an event is just a subset of the sample space (but not all subsets are events)
for instance, if you roll a die, sample space would be {1,2,3,4,5,6}
an event is something like "getting an even number"; mathematically it is the set {2,4,6}

其實probability space咪係measure space的一個特例黎

簡單講下sigma algebra是乜唻ok?
Probability space個幾個element又係乜

簡單講下sigma algebra是乜唻ok?
Probability space個幾個element又係乜

簡單講下sigma algebra是乜唻ok?
Probability space個幾個element又係乜

簡單講下sigma algebra是乜唻ok?
Probability space個幾個element又係乜

簡單講下sigma algebra是乜唻ok?
Probability space個幾個element又係乜

簡單講下sigma algebra是乜唻ok?
Probability space個幾個element又係乜

想問meta analysis唔入non significant findings 去分析會有咩bias? 因為其實入哂non significant findings最後個combined effect可能變返significant result. 咁有咩意思?

呢個問題存在好耐
open database 係解決之道

呢個情況其實似你普通run individual data時
你放兩組15+15人時唔sig, 放30+30時可能sig, 每個人自己baseline posttest比可能唔食

一切係effect size and power既問題

如果你唔放non sig data, 一來係publication bias selection bias,
二來驗證唔到個effect size係因為power定其他

當然，如果你越放得多越heterogeneous, 咁代表不同paper variety 高

講到尾，meta analysis都只不過係一種estimation，我唔明點解咁多人覺得佢好神

sor, 有感而發

入定唔入好 我覺得入左反而會影響左個result⋯

首先個問題係
有冇publication bias, 呢個係現時既meta analysis 答唔到
因為你搵到null result 既好大機會pub 唔到
咁你published 左既好大機會已經biased
第二個問題係convergence
因為我一個study 都有佢地自己既study error
例如因為sample size太細
所以error 大左，false positive and negative 多左
咁每個study 搵到個results 都會唔同
咁meta analysis 做既野就係搵個convergence
有冇個consistent 既finding

點解open database 最好
我一次過可以睇晒上面2個問題

簡單講下sigma algebra是乜唻ok?
Probability space個幾個element又係乜

Ph.D. students face significant mental health challenges
By Elisabeth PainApr. 4, 2017 , 3:15 PM
Approximately one-third of Ph.D. students are at risk of having or developing a common psychiatric disorder like depression, a recent study reports. Although these results come from a small sample—3659 students at universities in Flanders, Belgium, 90% of whom were studying the sciences and social sciences—they are nonetheless an important addition to the growing literature about the prevalence of mental health issues in academia. One key message for scientific trainees that are struggling with these types of challenges, write co-authors Katia Levecque and Frederik Anseel of Ghent University in an email to Science Careers, is that “you are not alone.” Beyond that, the authors encourage Ph.D. students to appreciate how important it is to take care of themselves. “Mental health problems can develop into serious threats to one’s wellbeing and career, and can have detrimental consequences in the long-term,” they write. So, if you’re struggling, it’s important to “[s]eek professional help or seek help in your personal environment, even if you think it’s probably a temporary thing.”

According to their survey, 51% of respondents had experienced at least two symptoms of poor mental health in recent weeks, indicating psychological distress. Moreover, 32% reported at least four symptoms, indicating a risk for common psychiatric disorders, which was more than twice the prevalence among highly educated comparison groups. The most commonly reported symptoms included feeling under constant strain, being unhappy and depressed, losing sleep because of worry, and not being able to overcome difficulties or enjoy day-to-day activities. The greatest predictor for experiencing mental health challenges was having difficulty taking care of family needs due to conflicting work commitments. High job demands and low job control were also associated with increased symptoms.

On the plus side, having an inspirational supervisor partially offset these risks. So did interest in an academic career, even among students who thought they had little chance of ultimately making it. Seeing a Ph.D. as good preparation for a nonacademic career and an added value for employers was also beneficial. “When people have a clear vision of the future and the path that they are taking, this provides a sense of meaningfulness, progress and control, which should be a protective factor against mental health problems,” the authors explain.

According to Nathan Vanderford, an assistant dean for academic development at the University of Kentucky in Lexington who also studies mental health in academic trainees, “[t]he study underscores what has long been presumed; that work conditions and career outlook plays a key role in the mental state of PhD trainees,” he writes in an email to Science Careers.

“[I]nstitutions, departments and PIs have long ignored the systemic mental health issues among PhD trainees,” Vanderford continues. “Data such as this should make the issues irrefutable and should, for ethical and moral reasons, force the hand of these entities to take on the responsibility of helping to provide PhD trainees with the support they need to navigate the very stressful journey of earning a PhD.” Levecque and Anseel point out that small steps such as facilitating work-life balance or “offering PhD students clear and full information on job expectations and career prospects, both in and outside academia,” could have a significant positive impact.

One message for PIs is that “investing in their own leadership competencies could really make a difference” for Ph.D. students, the authors add. They encourage PIs to take the issue seriously and learn how to spot signs of potential emotional distress. Gail Kinman, a professor at the University of Bedfordshire in the United Kingdom who acted as a reviewer for the paper, agrees. “PIs should look out for students who isolate themselves, who seem anxious and withdrawn, who are not meeting deadlines. Nobody would expect a PI to be able to diagnose mental health problems but they should be able to spot changes in their students and have the knowledge required to refer them for support,” she writes in an email to Science Careers.

Although the survey is specific to Flanders, many of the characteristics of working toward a Ph.D. are similar around the globe, making the findings generalizable, the authors argue. They hope the study, which has generated conversation on Twitter, will help break the silence around mental health issues in academia. “[I]t is a public secret that fear of stigma, retaliation or the expected negative impact on one’s future career often inhibits people suffering from mental health issues to make it public,” they write. This lack of visibility is problematic because feeling isolated can cause students’ mental health to deteriorate even further. It also means that there is less pressure on institutions and people in power to tackle the issue.

Rather than demonizing academia, action should be taken, the authors emphasize. As academics, “[w]e have had our share of struggles and challenges to overcome, but still think this is one of the most rewarding and meaningful careers one can have. So, if there’s a problem, let’s do something about it and make this a great place to work again. For everyone.”

http://www.sciencemag.org/careers/2017/04/phd-students-face-significant-mental-health-challenges

One message for PIs is that “investing in their own leadership competencies could really make a difference” for Ph.D. students, the authors add. They encourage PIs to take the issue seriously and learn how to spot signs of potential emotional distress. Gail Kinman, a professor at the University of Bedfordshire in the United Kingdom who acted as a reviewer for the paper, agrees. “PIs should look out for students who isolate themselves, who seem anxious and withdrawn, who are not meeting deadlines. Nobody would expect a PI to be able to diagnose mental health problems but they should be able to spot changes in their students and have the knowledge required to refer them for support,” she writes in an email to Science Careers.

Although the survey is specific to Flanders, many of the characteristics of working toward a Ph.D. are similar around the globe, making the findings generalizable, the authors argue. They hope the study, which has generated conversation on Twitter, will help break the silence around mental health issues in academia. “[I]t is a public secret that fear of stigma, retaliation or the expected negative impact on one’s future career often inhibits people suffering from mental health issues to make it public,” they write. This lack of visibility is problematic because feeling isolated can cause students’ mental health to deteriorate even further. It also means that there is less pressure on institutions and people in power to tackle the issue.

Rather than demonizing academia, action should be taken, the authors emphasize. As academics, “[w]e have had our share of struggles and challenges to overcome, but still think this is one of the most rewarding and meaningful careers one can have. So, if there’s a problem, let’s do something about it and make this a great place to work again. For everyone.”

http://www.sciencemag.org/careers/2017/04/phd-students-face-significant-mental-health-challenges

物理博士又黎趕客

One message for PIs is that “investing in their own leadership competencies could really make a difference” for Ph.D. students, the authors add. They encourage PIs to take the issue seriously and learn how to spot signs of potential emotional distress. Gail Kinman, a professor at the University of Bedfordshire in the United Kingdom who acted as a reviewer for the paper, agrees. “PIs should look out for students who isolate themselves, who seem anxious and withdrawn, who are not meeting deadlines. Nobody would expect a PI to be able to diagnose mental health problems but they should be able to spot changes in their students and have the knowledge required to refer them for support,” she writes in an email to Science Careers.

Although the survey is specific to Flanders, many of the characteristics of working toward a Ph.D. are similar around the globe, making the findings generalizable, the authors argue. They hope the study, which has generated conversation on Twitter, will help break the silence around mental health issues in academia. “[I]t is a public secret that fear of stigma, retaliation or the expected negative impact on one’s future career often inhibits people suffering from mental health issues to make it public,” they write. This lack of visibility is problematic because feeling isolated can cause students’ mental health to deteriorate even further. It also means that there is less pressure on institutions and people in power to tackle the issue.

Rather than demonizing academia, action should be taken, the authors emphasize. As academics, “[w]e have had our share of struggles and challenges to overcome, but still think this is one of the most rewarding and meaningful careers one can have. So, if there’s a problem, let’s do something about it and make this a great place to work again. For everyone.”

http://www.sciencemag.org/careers/2017/04/phd-students-face-significant-mental-health-challenges

物理博士又黎趕客

我不入火坑，誰入火坑？

One message for PIs is that “investing in their own leadership competencies could really make a difference” for Ph.D. students, the authors add. They encourage PIs to take the issue seriously and learn how to spot signs of potential emotional distress. Gail Kinman, a professor at the University of Bedfordshire in the United Kingdom who acted as a reviewer for the paper, agrees. “PIs should look out for students who isolate themselves, who seem anxious and withdrawn, who are not meeting deadlines. Nobody would expect a PI to be able to diagnose mental health problems but they should be able to spot changes in their students and have the knowledge required to refer them for support,” she writes in an email to Science Careers.

Although the survey is specific to Flanders, many of the characteristics of working toward a Ph.D. are similar around the globe, making the findings generalizable, the authors argue. They hope the study, which has generated conversation on Twitter, will help break the silence around mental health issues in academia. “[I]t is a public secret that fear of stigma, retaliation or the expected negative impact on one’s future career often inhibits people suffering from mental health issues to make it public,” they write. This lack of visibility is problematic because feeling isolated can cause students’ mental health to deteriorate even further. It also means that there is less pressure on institutions and people in power to tackle the issue.

Rather than demonizing academia, action should be taken, the authors emphasize. As academics, “[w]e have had our share of struggles and challenges to overcome, but still think this is one of the most rewarding and meaningful careers one can have. So, if there’s a problem, let’s do something about it and make this a great place to work again. For everyone.”

http://www.sciencemag.org/careers/2017/04/phd-students-face-significant-mental-health-challenges

物理博士又黎趕客

Haha..睇我下個post

送人入火坑系列：
I was rejected by every PhD program I applied to. This year, I got into my top choice. Here’s how

Around this time last year, I was reading my final rejection letter from the University of California-Berkeley. It wasn’t much of a surprise. By the time the email appeared in my inbox, four other universities—Yale, Princeton, Harvard, and the University of Pennsylvania—had already denied me admittance into their anthropology PhD programs. But Berkeley’s letter particularly stung. Their committee claimed that I was “a strong candidate,” and they were certain I had “already received offers from many other strong programs.” I won’t lie: Tears were shed.

Perhaps I’d been overconfident, but the string of rejections caught me off-guard. During the application process, I’d felt as if I had my bases covered: A full scholarship during my undergraduate studies as an honors student, a master’s degree in anthropology, a year of research under my belt while on a prestigious Fulbright fellowship. One of my letters of recommendation had been written by a MacArthur genius. I’d thought I was a shoo-in. Where had I gone wrong?

The simple answer: I’d overestimated the odds of acceptance. In recent years, the number of applications to graduate schools has swelled while the pot of money available to them, especially to applicants outside of the hard sciences, has shrunk. Harvard’s Department of Anthropology, for example, receives over 250 applications each year, but only has spots for fewer than 10 students. That’s around a 4% acceptance rate. Other departments are even more exclusive—not necessarily because they desire to be, but because they lack the resources (including both funding and faculty) to admit more people.

I knew that, with my ultimate goal of becoming a tenured professor, I needed to reapply again next year. So I decided to swallow my pride, reach out to people I knew in PhD programs, tell them about my rejections, and ask for help reapplying. Here are the top five tips that can make all the difference—whether you’re applying to a PhD program, a law school, a medical school, or a terminal master’s program.

1) Current graduate students often know just as much about the application process as professors.

Getting professors’ advice is clearly important—but current graduate students are often better able to give you concrete advice, as they are on similarly low rungs of the academic ladder. That MacArthur genius who wrote my letter of recommendation? He’s over 70, and although he’s a brilliant academic, today’s grad student struggle is completely alien to his experiences half a century ago.

My second time around, I rang up my friend Stephanie, a doctoral student in Duke’s anthropology department. Her advice was invaluable because she’s plugged into what is expected of students today. She advised me about the current dynamics of her department, how much funding was available, and what kinds and quantities of students are accepted into the program each year. She also passed along her own admissions essay, which was an invaluable template. If you don’t know any current graduate students, try contacting students at the universities to which you want to apply (there is usually a list of current students on a department’s webpage).

2) Expand your school pool, but not at the expense of your savings account.

There is a decent chance you will not be accepted into any of your top three programs—not because your work isn’t top-notch, but because there is a flood of applicants for an increasingly limited number of positions. During my first application process, I applied to universities more for their name recognition than for my actual knowledge of their research. My second go-round, I realized that most of the schools to which I applied—Yale, Penn, and Princeton—really didn’t align with my own interests. Other departments were leading research in work I was more interested in, even if they didn’t have the same name recognition.

Deciding exactly how many schools to apply to can be tricky. The consensus I received from current grad students is that you should choose between five and 10 programs, with the hopes you will be accepted into more than one and be able to compare financial offers. I applied to 10 programs to be safe, which was extremely costly—I spent over $1,500 on application fees, transcripts, and sending out my GRE scores. If you haven’t already taken the GRE, tack on an extra $300. Looking back now, applying to 10 programs was overkill. I should have saved myself at least $500 by narrowing my search to my top six schools.

While you shouldn’t pick the grad schools you apply to based on names alone, it’s also important to take program rankings into the equation. Finding a job in academia is harder than ever, and while you can certainly excel at a program with less name recognition, it’s likely that a degree from your 10th-choice school will be less impressive on the job market.

3) Start working on your application even before it’s officially released.

Universities usually open their application portals in September, but you can begin writing drafts of your application essays and contacting professors before then. The earlier you contact a faculty member, the fewer students they have already spoken with, and the more likely it is they will remember your name when admissions time rolls around.
My first time applying, I didn’t buckle down on applications until October, and I even waited to reach out to some professors in the programs I was applying to until December—just weeks before the application was due. Obviously, they didn’t respond.
This time around, I began contacting potential faculty members in June, about six months before the application deadline. Only contact a professor once you have something concrete to say or ask. A vague introduction that asks, “Is this program a good fit for me?” will probably be met with eyerolls—that’s something only you can decide, not a professor you’ve never met before.
Instead, simply write a three to four sentences outlining your work and why you’re interested in working with them in particular. Asking faculty members what kinds of dissertations and projects they have supervised in the past will also help you get a feel for how they mentor graduate students and if you think you’ll click.

4) Speak the language of your department.

Once you have identified your potential programs and the faculty members you want to work with, it’s time to tailor each individual application to the department to which you’re applying. Does the department have a special research center you’re interested in, or have they pioneered a particular form of research? Mention it, and explain how it can help your studies.

Be sure to be specific in discussing why you are choosing the department, and especially which faculty you want to work with and why. Reading faculty’s research—and reading a whole lot of it—is the only way to gain a real sense of which intellectual questions particular departments are trying to tackle, and which professors’ work most closely interlinks with your own academic interests. This kind of reading also helps you understand how the field is interrogating certain problems, and the kinds of terminology you should be using—or at least alluding to—in your essay.

5) Your application essay shouldn’t have answers—it should ask questions.

Admissions committees don’t expect you to have all the answers in a given field. That’s why you’re applying to grad school, after all. But they do expect you to have smart questions. The age-old adage of identifying “a gap in the literature” has become trite in recent years. Instead, ask concise, thoughtful questions about your area of interest, and explain how you hope to go about investigating them. This implicitly demonstrates that not only do you know there is a gap in the literature, but you have a plan about how you will fill that gap.
Admissions is a crapshoot

Even following all of these steps is no guarantee of hitting the grad school jackpot. The reality is that even the best applications are frequently passed over by universities. Acceptance always involves some degree of luck. As one professor at Harvard told me after I had been turned down, “I am not at all sure that the reason for your not getting in was entirely your fault. My view is that it was primarily due to the fact that the applicant pool was very large and the number of places allowed to us very few. To give an example: We had at least seven Africanists who could have expected to be accepted at any first-rate university in the USA. We were only permitted to admit one. As a result of the very low acceptance ratio, the process has a high level of serendipity to it.”

So if you get turned down, don’t get discouraged. On one hand, it may be a blessing in disguise. Grad school certainly isn’t for everyone. As the most educated generation in American history hits the workforce, employers have a glut of qualified millennials to choose from—and a graduate degree doesn’t necessarily have the same luster it once did. Sometimes the best grad school option is foregoing a degree entirely in favor of gaining more practical work experience.

But if you believe you need a graduate degree to achieve your goals, the only option is to keep knocking at the door. Each round of applications provides the opportunity for you to better articulate your goals, qualifications, and why you belong in the program.
This year, I was accepted into three top anthropology PhD programs—including my first choice, UC-Berkeley, where I received a fellowship that fully funds my studies and includes a generous stipend. After attending Berkeley’s prospective students weekend, I happily accepted their offer. While on my visit, one professor said my application had made him exclaim aloud, “We have to have him!” What a difference a year can make.

https://qz.com/950763/how-to-get-into-graduate-school-according-to-current-phd-students/?utm_source=qzfb

送人入火坑系列：
I was rejected by every PhD program I applied to. This year, I got into my top choice. Here’s how

Around this time last year, I was reading my final rejection letter from the University of California-Berkeley. It wasn’t much of a surprise. By the time the email appeared in my inbox, four other universities—Yale, Princeton, Harvard, and the University of Pennsylvania—had already denied me admittance into their anthropology PhD programs. But Berkeley’s letter particularly stung. Their committee claimed that I was “a strong candidate,” and they were certain I had “already received offers from many other strong programs.” I won’t lie: Tears were shed.

Perhaps I’d been overconfident, but the string of rejections caught me off-guard. During the application process, I’d felt as if I had my bases covered: A full scholarship during my undergraduate studies as an honors student, a master’s degree in anthropology, a year of research under my belt while on a prestigious Fulbright fellowship. One of my letters of recommendation had been written by a MacArthur genius. I’d thought I was a shoo-in. Where had I gone wrong?

The simple answer: I’d overestimated the odds of acceptance. In recent years, the number of applications to graduate schools has swelled while the pot of money available to them, especially to applicants outside of the hard sciences, has shrunk. Harvard’s Department of Anthropology, for example, receives over 250 applications each year, but only has spots for fewer than 10 students. That’s around a 4% acceptance rate. Other departments are even more exclusive—not necessarily because they desire to be, but because they lack the resources (including both funding and faculty) to admit more people.

I knew that, with my ultimate goal of becoming a tenured professor, I needed to reapply again next year. So I decided to swallow my pride, reach out to people I knew in PhD programs, tell them about my rejections, and ask for help reapplying. Here are the top five tips that can make all the difference—whether you’re applying to a PhD program, a law school, a medical school, or a terminal master’s program.

1) Current graduate students often know just as much about the application process as professors.

Getting professors’ advice is clearly important—but current graduate students are often better able to give you concrete advice, as they are on similarly low rungs of the academic ladder. That MacArthur genius who wrote my letter of recommendation? He’s over 70, and although he’s a brilliant academic, today’s grad student struggle is completely alien to his experiences half a century ago.

My second time around, I rang up my friend Stephanie, a doctoral student in Duke’s anthropology department. Her advice was invaluable because she’s plugged into what is expected of students today. She advised me about the current dynamics of her department, how much funding was available, and what kinds and quantities of students are accepted into the program each year. She also passed along her own admissions essay, which was an invaluable template. If you don’t know any current graduate students, try contacting students at the universities to which you want to apply (there is usually a list of current students on a department’s webpage).

紅字 屌