← AcademiaOne PhD Hub · 返回首页
AcademiaOneEDUCATION
PhD Personal Statement Guide · US · Science & Engineering 美国理工科个人陈述指南
AcademiaOneEDUCATION
Your PhD application, from idea to offer你的博士申请,从构想到录取
DOCTORAL APPLICATIONS · UNITED STATES · SCIENCE & ENGINEERING   美国理工科博士申请

PhD Personal Statement
& Personal History Guide · Science & Engineering

美国理工科博士个人陈述写作指南

A paragraph-by-paragraph model of the personal statement that goes with a science or engineering PhD application in the US, for self-proposed projects and for advertised studentships where the statement is the main document. Each part shows the structure, a fully worked example from electrical power engineering, the difference between a weak and a stronger line, and the current statement rules of US universities taken from their official pages.

逐段拆解美国理工科博士申请中的个人陈述,既适用于自主选题,也适用于以陈述为主要材料的已发布奖学金项目。每一部分都包含结构说明、一个完整的电力工程范例、“较弱”与“更好”写法的对比,并附美国各校官网上的最新陈述要求。

Start here 开始之前

What US programmes ask for美国各项目要求

Most US applications have two essays. The statement of purpose carries your research plan and faculty fit; the personal statement is the second, shorter essay about your path, your background and what you would bring to the graduate community. Universities name it differently: a personal statement at Cornell, Northwestern, UCLA and Michigan (up to 500 words at the last three), a Personal History Statement at UC Berkeley (about 1.5 to 2 pages), Purdue and UC Santa Barbara, a personal history and diversity statement at UC Davis (4,000 characters) and a Life Experiences Statement at Duke. At Brown, Boston University and USC the personal statement is the main essay and does the work of the statement of purpose, so read the third column of the table for each target.

美国多数申请有两篇文书。目的陈述(SOP)承载研究计划与导师匹配;个人陈述是第二篇、篇幅较短,讲你的经历、背景,以及你能为研究生群体带来什么。各校叫法不同:康奈尔、西北大学、UCLA 与密歇根称个人陈述(personal statement,后三校不超过 500 词),伯克利(约 1.5 至 2 页)、普渡与加州大学圣塔芭芭拉称个人经历陈述(Personal History Statement),加州大学戴维斯称个人经历与多元化陈述(4,000 字符),杜克称人生经历陈述(Life Experiences Statement)。布朗、波士顿大学和南加州大学的个人陈述就是主文书,承担目的陈述的作用,所以请逐校阅读表格第三栏。

In engineering and computer science the personal statement is often short or doubles as the main essay. UCLA Computer Science asks for up to 500 words beside a 500-word SOP, UC Berkeley Mechanical Engineering a Personal History of 1-2 pages that must not overlap the SOP, UC Santa Barbara at least 250 words with no duplication, and Purdue and UC Davis Computer Science a personal history statement; Cornell Engineering and Michigan CSE add a personal statement to the SOP. At Penn Engineering (up to 2 pages answering six prompts), Columbia SEAS (250-1,000 words recommended), Brown, Boston University and USC Viterbi the personal statement is the main essay, and MIT Chemistry asks for a 500-1,000-word personal statement as well as its statement of objectives.

在工程与计算机领域,个人陈述往往很短,或者就是主文书。UCLA 计算机要求不超过 500 词的个人陈述配合 500 词 SOP,伯克利机械工程要求 1-2 页且不得与 SOP 重复的个人经历陈述,加州大学圣塔芭芭拉要求至少 250 词且不得重复,普渡和加州大学戴维斯计算机要求个人经历陈述;康奈尔工程与密歇根 CSE 在 SOP 之外另加个人陈述。宾大工程学院(不超过 2 页,回答六个问题)、哥伦比亚 SEAS(建议 250-1,000 词)、布朗、波士顿大学和南加州大学 Viterbi 的个人陈述是主文书;麻省理工化学系在目标陈述之外还要求 500-1,000 词的个人陈述。

Science & Engineering: what each programme asks for理工科:各项目要求

UNIVERSITY 大学WHAT THE PROGRAMME ASKS FOR 项目要求PERSONAL STATEMENT RULE (UNIVERSITY-WIDE) 全校个人陈述规定OFFICIAL PAGE 官方页面
MIT
麻省理工学院
QS 2027: 1
Statement of objectives; Chemistry also a 500-1,000-word personal statement + 1,000-1,500-word objectives naming 3 faculty
Chemistry: GRE not accepted; MechE letters due 1 Nov
References: 3
Some departments (Chemistry 500-1,000 words)
oge.mit.edu
oge.mit.edu
Stanford
斯坦福大学
QS 2027: =2
Electrical Engineering: SOP ≤2 pages single-spaced, naming research interests and faculty
References: 3
not stated on official pageee-www.stanford.edu
Harvard
哈佛大学
QS 2027: 5
SEAS Computer Science: SOP ≤1,000 words naming labs
PhD deadline 15 Dec
References: 3
Often required (HKS, Health Policy, GSD)
hks.harvard.edu
seas.harvard.edu
Caltech
加州理工学院
QS 2027: 7
SOP ~2 pages + CV ~2 pages (Geological & Planetary Sciences); application not reviewed until all 3 letters arrive
Deadlines 1-15 Dec by option
References: 3 (max 3)
Not stated centrallygps.caltech.edu
Penn
宾夕法尼亚大学
QS 2027: 15
Penn Engineering PhD (incl. CIS): personal statement max 2 pages answering six prompts + resume; no writing sample
GRE optional; IELTS 7.5 / TOEFL 100 recommended, no minimum; window 15 Sep to 15 Dec 2026
References: 2 (CIS 3; 1+ from faculty)
Wharton: personal statement essay; English: personal statement max 1,200 words
doctoral.wharton.upenn.edu
gradadm.engineering.upenn.edu
Cornell
康奈尔大学
QS 2027: =16
Engineering: academic SOP ≤2 pages naming faculty + personal statement
MAE 1 Dec; CEE 7 Jan
References: 3
Required, separate from the SOP
duffield.cornell.edu
engineering.cornell.edu
Yale
耶鲁大学
QS 2027: 16
Chemistry: personal statement of purpose
GRE + TOEFL; due 1 Dec
References: 3+
Not stated centrallychem.yale.edu
Johns Hopkins
约翰斯·霍普金斯大学
QS 2027: 20
Computer Science: statement, letters and grades are the main components
All materials incl. letters by 1 Dec
Optional in Arts & Sciences
arthist.jhu.edu
cs.jhu.edu
UC Berkeley
加州大学伯克利分校
QS 2027: =20
Mechanical Engineering: SOP 1-2 pages double-spaced + Personal History 1-2 pages (no overlap)
References: 3
Personal History Statement, required, ~1.5-2 pages
psychology.berkeley.edu
me.berkeley.edu
Chicago
芝加哥大学
QS 2027: 24
Mathematics: statement; do not submit GRE
IELTS 7.0 / TOEFL 90
References: 3+
Not stated centrallymathematics.uchicago.edu
Princeton
普林斯顿大学
QS 2027: 27
Computer Science: no minimum English score for PhD admission
Below IELTS 8 speaking → oral test in year 1
Some departments (SPIA max 250 words)
spia.princeton.edu
cs.princeton.edu
Columbia
哥伦比亚大学
QS 2027: =43
SEAS doctoral (incl. Computer Science): personal statement 250-1,000 words recommended + CV
GRE not required (2026 cycle); doctoral deadline 15 Dec
References: 3
Psychology asks for a personal statement as well as the SOP
gsas.columbia.edu
engineering.columbia.edu
Northwestern
西北大学
QS 2027: =45
Computer Science: academic statement explaining fit with a named adviser; resume; may name up to 2 faculty
GRE required; due 1 Dec 2026
References: 2 minimum, 3 recommended
Personal statement ≤500 words (TGS)
sesp.northwestern.edu
mccormick.northwestern.edu
UCLA
加州大学洛杉矶分校
QS 2027: 49
Computer Science: statement of purpose ≤500 words + personal statement ≤500 words; no CV required
GRE optional; due 15 Dec; IELTS 7.0
References: 3
Personal statement ≤500 words (Psychology, CS)
cs.ucla.edu
cs.ucla.edu
Michigan
密歇根大学
QS 2027: 51
CSE: statement of purpose (research interests, goals) + personal statement; choose faculty in application
GRE not considered
References: 3 suggested
Personal statement ≤500 words (English allows 2 pages)
prod.lsa.umich.edu
cse.engin.umich.edu
Carnegie Mellon
卡内基梅隆大学
QS 2027: 55
Computer Science Dept: statement of purpose 1-2 pages + CV; select faculty in application
GRE optional; final deadline 9 Dec 2026; no adviser needed before applying
References: 3
Not stated centrallycsd.cmu.edu
NYU
纽约大学
QS 2027: 58
Courant Computer Science: statement of purpose + CV
GRE not required; due 12 Dec
References: 3
Not stated centrallycs.nyu.edu
Brown
布朗大学
QS 2027: 66
Computer Science: personal statement; research motivation valued; no need to name faculty
GRE optional; due 15 Dec 2026
References: 3
Personal statement is the main essay
graduateschool.brown.edu
cs.brown.edu
Duke
杜克大学
QS 2027: 70
Computer Science: statement of purpose + resume; no writing sample
GRE optional; due 15 Dec (fall only); TOEFL, IELTS or Duolingo for non-native speakers
References: 3
Life Experiences Statement
gradschool.duke.edu
gradschool.duke.edu
UT Austin
得克萨斯大学奥斯汀分校
QS 2027: 72
Computer Science: statement of purpose (why UT CS)
GRE optional; due 15 Dec, letters 20 Dec
References: 3 or more
Not stated centrallycs.utexas.edu
UIUC
伊利诺伊大学厄巴纳-香槟分校
QS 2027: 74
Siebel Computer Science: academic statement of purpose 1-2 pages + CV
GRE optional; late applications not accepted
References: 3
Some programmes add a personal statement (Curriculum & Instruction)
catalog.illinois.edu
siebelschool.illinois.edu
UC San Diego
加州大学圣地亚哥分校
QS 2027: 81
CSE: statement of purpose (no word limit) + CV
No GRE for Fall 2027; due 16 Dec 2026
References: 3
Not stated centrallycse.ucsd.edu
Penn State
宾夕法尼亚州立大学
QS 2027: =92
Computer Science & Engineering: statement of purpose + CV/resume
GRE not required for PhD; TOEFL/IELTS per Graduate School (GCAC-305); deadline on CSE admissions site
References: 3
Not stated centrallybulletins.psu.edu
U Washington
华盛顿大学
QS 2027: =92
Allen School CSE: no prior faculty contact or chosen adviser needed
Deadline not stated (fee requests by ~6 Dec)
Psychology: personal experience statement ≤2 pages
psych.uw.edu
cs.washington.edu
Boston University
波士顿大学
QS 2027: 94
Computer Science: personal statement + CV/resume
GRE not mandatory; due 1 Dec
References: 3
Personal statement
bu.edu
bu.edu
Purdue
普渡大学
QS 2027: =100
Computer Science: academic statement of purpose + personal history statement + resume
GRE not required; due 1 Dec 2026; IELTS 7.5 recommended
Personal history statement
education.purdue.edu
cs.purdue.edu
University of Wisconsin-Madison
威斯康星大学麦迪逊分校
QS 2027: =131
Computer Sciences: statement of purpose + resume/CV + supplemental sections
GRE not required; due 15 Dec
References: 3
Personal statement in Psychology, Ed Policy, Epidemiology
guide.wisc.edu
guide.wisc.edu
University of California, Davis
加州大学戴维斯分校
QS 2027: 137
Computer Science: SOP relating plans to UC Davis research + personal history statement + CV
Due 15 Dec, no spring entry; min GPA 3.0; TOEFL/IELTS unless English-medium degree
References: 3
Personal history and diversity statement max 4,000 characters
grad.ucdavis.edu
cs.ucdavis.edu
Georgia Institute of Technology
佐治亚理工学院
QS 2027: =142
PhD CS: online application; SOP length not stated on page; desirable GPA 3.5
GRE optional; IELTS 7.5 (R/L/S 6.5, W 5.5) or TOEFL 100; due 15 Dec
References: 3 (academic preferred)
Not stated centrallycc.gatech.edu
University of Southern California
南加州大学
QS 2027: 153
Viterbi PhD (incl. CS): personal statement (no length) + resume/CV; research ability expected
GRE not required for CS; due 15 Dec 2026
References: 3 (faculty)
Combined with the statement; no separate diversity statement in sampled programmes
dornsife.usc.edu
viterbigradadmission.usc.edu
University of North Carolina at Chapel Hill
北卡罗来纳大学教堂山分校
QS 2027: =158
Computer Science: statement of purpose + resume/CV (guidance on cs.unc.edu)
GRE not required; listing shows 9 Mar 2027 deadline (fellowships 15 Dec)
References: 3
Not stated centrallygradschool.sites.unc.edu
Washington University in St. Louis
圣路易斯华盛顿大学
QS 2027: =162
McKelvey Engineering PhD (incl. CS): SOP 1 to 2 pages + CV
GRE optional (no effect); IELTS 6.5 / TOEFL 90; deadline on separate page
References: 3
Education asks for a personal statement; others use SOP only
education.washu.edu
engineering.washu.edu
Texas A&M University
得克萨斯农工大学
QS 2027: 169
CSE: one-page SOP on research interests and closely related faculty + CV (min 1 page, research and publications)
GRE optional; TOEFL below 100 makes admission unlikely; due 15 Dec
References: Not stated (research advisers preferred)
Not stated centrallyengineering.tamu.edu
Arizona State University
亚利桑那州立大学
QS 2027: 172
SCAI Computer Science: SOP about 1 to 2 pages naming 2 to 3 ASU faculty; CV encouraged; min GPA 3.5
IELTS 7 / TOEFL 90; due 31 Dec; GRE policy unclear on page
References: 3
Not stated centrallyscai.engineering.asu.edu
University of California, Santa Barbara
加州大学圣塔芭芭拉分校
QS 2027: 173
Computer Science: SOP + personal history statement (min 250 words, no duplication) + CV
IELTS 7 / TOEFL 100; due 15 Dec (page shows 2025 date); GRE not mentioned
References: 3
Personal history statement required, min 250 words
graddiv.ucsb.edu
graddiv.ucsb.edu
Michigan State University
密歇根州立大学
QS 2027: 182
Chemistry PhD: academic statement + personal statement/SOP (no lengths) + CV + Chemistry Applicant Datasheet
GRE General and Subject recommended, not required; international deadline 15 Dec (page shows Fall 2026 cycle)
References: 3 minimum
Personal statement (English 750 to 1,000 words; Epidemiology 800 to 1,000 words)
english.msu.edu
chemistry.msu.edu
Emory University
埃默里大学
QS 2027: 183
Computer Science and Informatics: SOP (research questions, focus, plans) + CV
GRE optional; IELTS 7 / TOEFL 85; due 15 Dec
References: 3
Statement on Building Collaborative and Respectful Graduate Communities, max 500 words, if programme requires
graduateschool.emory.edu
computerscience.emory.edu

Requirements are quoted from the official pages linked in each row, in English as published. The third column is the university-wide rule for the personal or second statement, with its own official link; “Not stated centrally” means the programme sets it. 表中要求按各行所附官方页面的英文原文摘录;第三栏为全校个人陈述(第二篇文书)规定,并附其官方链接;“Not stated centrally”表示由项目自行规定。

Golden rule 核心原则

The programme page always beats this guide. Use this guide to decide what to say; use the programme page to decide what the second essay is called, how long it is, and whether it is separate from the statement of purpose or replaces it.

项目官网的要求永远优先于本指南。用本指南决定“写什么”,用项目页面决定第二篇文书叫什么、写多长,以及它是独立于目的陈述还是取代目的陈述。

Two essays, no overlap 两篇文书,互不重复

UC Berkeley Mechanical Engineering asks for a Personal History that does not overlap the SOP, and UC Santa Barbara Computer Science for at least 250 words with no duplication. Write the statement of purpose first, then build the personal statement from what it leaves out. UC Davis caps each of the two at 4,000 characters, about 650 English words.

伯克利机械工程要求个人经历陈述不得与 SOP 重复,加州大学圣塔芭芭拉计算机要求至少 250 词且不得重复。先写目的陈述,再用它没有涵盖的内容组织个人陈述。加州大学戴维斯规定两篇各不超过 4,000 字符,约 650 个英文单词。

Your statement of purpose 你的目的陈述

The research plan and faculty fit belong in the statement of purpose, covered by AO_US_ScienceEngineering_StatementOfPurpose_Guide. Use this guide for the personal statement beside it, or for programmes where the personal statement is the only essay.

研究计划与导师匹配属于目的陈述,请参阅 AO_US_ScienceEngineering_StatementOfPurpose_Guide。本指南用于与之并列的个人陈述,以及个人陈述是唯一文书的项目。

Table built from 37 official programme pages and the university-wide personal-statement rule of each university (data checked 2026-11-18). Rules change each cycle; the programme page always wins. 表格依据 37 个官方项目页面及各校的全校个人陈述规定整理(数据核对日期 2026-11-18)。规则每年可能调整,以项目页面为准。

Start here 开始之前

The statement in a science and engineering application理工科申请中的个人陈述

In science and engineering the statement often carries more weight than in other fields. For an advertised project or studentship it may be the only written document, so it must show that you can do the advertised work. For a self-proposed project it sits beside a proposal agreed with a supervisor. In both cases panels look for technical skills shown through results: what you built, measured or trained, how well it worked and what you did when it did not.

在理工科,陈述往往比其他学科更重要。申请已发布的项目或奖学金职位时,它可能是唯一的书面材料,必须证明你能胜任所招聘的工作;自主选题时,它与和导师商定的研究计划一起提交。两种情况下,评审都看你如何用结果展示技术能力:你搭建、测量或训练了什么,效果如何,失败时又做了什么。

How the statement changes across the field本领域内不同方向的差异

FIELD 方向WHAT THE PREPARATION PARAGRAPH SHOWS 准备部分要体现什么WHAT FIT MEANS 匹配的含义
Electrical · Mechanical · Civil
电气 · 机械 · 土木
Simulation and test tools named, rigs or field data, results with numbers, and any paper or preprint.
写明仿真与测试工具、试验台或现场数据、带数字的结果,以及论文或预印本。
A group whose facility (laboratory, rig, simulator) your project needs, and industry partners.
项目所需设施(实验室、试验台、仿真器)所在的课题组,以及行业合作方。
Computer science
计算机科学
Models or systems you built, code you can link to, datasets and benchmarks, and contributions to open source.
你构建的模型或系统、可以给出链接的代码、数据集与基准测试,以及开源贡献。
A supervisor whose recent papers your method extends, and the compute you need.
近期论文与你的方法相承接的导师,以及你所需的计算资源。
Biology · Life sciences
生物 · 生命科学
Bench techniques you can run alone, model organisms or cell lines, and data analysis.
能独立操作的实验技术、模式生物或细胞系,以及数据分析。
A lab whose system and techniques match, often an advertised project.
研究体系与技术相符的实验室,常为已发布项目。
Chemistry · Materials
化学 · 材料
Synthesis or characterisation methods, instruments used, and safety training.
合成或表征方法、用过的仪器,以及安全培训。
A group with the instruments and the chemistry your project requires.
拥有项目所需仪器与研究方向的课题组。

What a statement is not陈述不是什么

Not a second proposal

The proposal explains the design. The statement explains why you can carry it out.

研究计划讲研究设计,陈述要说明的是你为什么有能力完成它。

Not a CV in sentences

The panel has your CV. Pick three or four experiences and say what each taught you.

评审手里有你的简历。挑三四段经历,说明每段经历让你学到了什么。

Not a life story

Childhood and family appear only if they explain a research choice, and then in one sentence.

童年与家庭只在能解释某个研究选择时出现,而且只用一句话。

What the document is called 这份文件叫什么

In the US the second essay is called a personal statement (Cornell, Northwestern, UCLA, Michigan), a Personal History Statement (UC Berkeley, Purdue, UC Santa Barbara), a personal history and diversity statement (UC Davis) or a Life Experiences Statement (Duke). At Brown, Boston University and USC the personal statement is the main essay. Some departments call it a statement of suitability or a research statement; for advertised projects, answer the advert’s person specification point by point.

在美国,第二篇文书可能叫个人陈述(康奈尔、西北大学、UCLA、密歇根)、个人经历陈述(Personal History Statement,伯克利、普渡、加州大学圣塔芭芭拉)、个人经历与多元化陈述(加州大学戴维斯)或人生经历陈述(杜克)。布朗、波士顿大学和南加州大学的个人陈述就是主文书。有的院系称之为适合度陈述或研究陈述;申请已发布项目时,请逐条回应招聘启事中的人员要求。

Before you write 动笔之前

Two statements: purpose and personal两篇陈述:目的陈述与个人陈述

The statement of purpose answers an academic question: can this person do this research, here, with these faculty? The personal statement answers a second one: who is this person, how did they get here, and what will they add to the department? Write the statement of purpose first, using AO_US_ScienceEngineering_StatementOfPurpose_Guide, then write the personal statement from what it leaves out. Committees read both, so a story told twice wastes space in each.

目的陈述回答一个学术问题:这个人能否在这里、与这些老师一起完成这项研究?个人陈述回答另一个问题:这个人是谁、怎样走到今天、能为院系带来什么?先参照 AO_US_ScienceEngineering_StatementOfPurpose_Guide 写好目的陈述,再用它没有涵盖的内容写个人陈述。招生委员会两篇都会读,同一个故事讲两遍,两篇都会浪费篇幅。

STATEMENT OF PURPOSE 目的陈述PERSONAL STATEMENT 个人陈述
Main question
核心问题
Can you do this research here?
你能否在这里完成这项研究?
Who are you, and what will you bring?
你是谁,能带来什么?
Content
内容
Research question, methods and preparation, named faculty, goals.
研究问题、方法与准备、具体老师、目标。
Your path into research, what shaped how you work, obstacles you handled, teaching, mentoring or community work, what you would add.
你走向研究的经历、塑造你工作方式的因素、你克服过的困难、教学、指导或社区工作,以及你能带来的贡献。
Typical length
常见篇幅
About 1-2 pages or up to 1,000 words (Northwestern up to 1,000; Michigan about 2 pages).
约 1-2 页或不超过 1,000 词(西北大学不超过 1,000 词;密歇根约 2 页)。
Up to 500 words (Northwestern, UCLA, Michigan); 1.5 to 2 pages (UC Berkeley); 4,000 characters (UC Davis).
不超过 500 词(西北大学、UCLA、密歇根);1.5 至 2 页(伯克利);4,000 字符(加州大学戴维斯)。
Faculty named
是否写导师
Yes, in the statement or the form.
写,在陈述或申请表中。
Only where it is the main essay, or where the programme asks for fit in it.
只有在它是主文书、或项目要求写匹配时才写。
Where you meet both
两篇都要的项目
UCLA Computer Science, Cornell Engineering, Michigan CSE, MIT Chemistry; Personal History at UC Berkeley, Purdue, UC Davis, UC Santa Barbara
UCLA 计算机、康奈尔工程、密歇根 CSE、麻省理工化学;伯克利、普渡、加州大学戴维斯、圣塔芭芭拉要求个人经历陈述
No overlap: UC Berkeley Mechanical Engineering; no duplication: UC Santa Barbara.
不得重叠:伯克利机械工程;不得重复:加州大学圣塔芭芭拉。

How to use the sections of this guide如何使用本指南的各部分

1It is the main essay

Use all five sections at full strength. In your field: Penn Engineering, Columbia SEAS, Brown, Boston University, USC Viterbi.

五个部分全部按完整要求写。本领域的例子:宾大工程、哥伦比亚 SEAS、布朗、波士顿大学、南加州大学 Viterbi。

2Both are asked for

Research and faculty fit stay in the SOP. Here, use sections 1, 2 and 4 in personal form: where the interest began, how you got here, what you will contribute.

研究与导师匹配留在 SOP 中。这里按个人视角使用第 1、2、4 部分:兴趣从何而来、你如何走到今天、你将贡献什么。

3Diversity or community

UC Davis, Duke (Life Experiences) and Emory (optional, up to 500 words) ask how your experience prepares you to support others in a research community. Give one or two concrete examples.

加州大学戴维斯、杜克(人生经历)和埃默里(可选,不超过 500 词)想了解你的经历如何让你在研究群体中支持他人。给出一两个具体例子。

Start here 开始之前

How long, and how the words are shared篇幅与字数分配

Write one master version of about 1,000 words, then cut it to each limit. A strong statement has five parts. Panels spend most of their time on preparation and fit, so those two parts get more than half the words. Shorter limits keep the same order and cut each part in proportion. Each section page then gives a blueprint, a worked example, a weak versus stronger comparison and lines from other fields. In the US the personal statement is usually shorter than the statement of purpose (often 500 words); where it is the main essay it runs longer.

先写一份约 1,000 词的母版,再按各校限制裁剪。一份好的陈述由五部分组成。评审把大部分时间花在“准备”和“匹配”上,所以这两部分占一半以上的篇幅。字数限制更短时,保持同样顺序,各部分按比例压缩。之后每个部分都包含结构说明、范例、“较弱”与“更好”写法的对比,以及其他方向的示例。在美国,个人陈述通常比目的陈述短(常为 500 词);作为主文书时会更长。

100350250200100
1 Opening2 Motivation & preparation3 Fit4 Readiness & goals5 Closing

Cutting to size按篇幅压缩

LIMIT 限制WHAT TO KEEP 保留什么
About 1,000 words or 2 pages
约 1,000 词或 2 页
All five parts, with two or three examples of preparation.
五个部分全部保留,准备部分写两三个例子。
About 500-750 words or 1 page
约 500-750 词或 1 页
One sentence of opening, one strong example of preparation, a fit paragraph naming one supervisor, one sentence of goals.
开头一句、准备部分一个有力的例子、点名一位导师的匹配段、目标一句。
1,500 to 5,000 characters
1,500 至 5,000 字符
About 250 to 800 words: the question, your best piece of evidence and the named fit. No greeting and no closing paragraph.
约 250 至 800 词:研究问题、最有力的一条证据和具体的匹配。不写称呼,也不写结尾段。
Limits in your field 本专业的篇幅要求

US science and engineering limits for the personal statement: UCLA Computer Science up to 500 words; UC Santa Barbara at least 250 words; UC Berkeley Mechanical Engineering 1-2 pages; Columbia SEAS 250-1,000 words recommended; Penn Engineering up to 2 pages; MIT Chemistry 500-1,000 words.

美国理工科的个人陈述篇幅:UCLA 计算机不超过 500 词;加州大学圣塔芭芭拉至少 250 词;伯克利机械工程 1-2 页;哥伦比亚 SEAS 建议 250-1,000 词;宾大工程学院不超过 2 页;麻省理工化学 500-1,000 词。

The five questions every science and engineering panel asks理工科评审必问的五个问题

#PANEL QUESTION 评审问题WHERE YOU ANSWER IT 在哪里回答
1Why research, and why this question? 为什么做研究,为什么是这个问题?Opening · Motivation 开头 · 研究动机
2What has this person already done that shows they can do research? 申请人已经做过什么,能证明其研究能力?Motivation & preparation 研究动机与准备
3Why this supervisor, group and university? 为什么是这位导师、这个课题组和这所大学?Fit 匹配度
4Will they cope with a PhD and finish it, and where are they heading? 能否胜任并完成博士?今后方向是什么?Readiness & goals 准备度与目标
5Does the statement add to the proposal and CV, or repeat them? 陈述是在补充研究计划与简历,还是在重复?Whole document 全文
Before you write 动笔之前

The golden thread逻辑主线

A statement works when each part explains the next: a question you met in your own work led you to train in certain methods, those methods and that question point to one supervisor, and that supervisor’s group is the step towards a clear goal. If you cannot write these four sentences, the statement will read as a list.

好的陈述每一部分都能引出下一部分:你在自己的工作中遇到一个问题,于是接受了相应的方法训练;这些方法和这个问题指向某一位导师;而这位导师的课题组正是通往清晰目标的一步。如果写不出这四句话,陈述读起来就会像一份清单。

WHY研究动机

My graph model located 94% of simulated faults within one bus, but its accuracy fell sharply once I added rooftop solar.

PREPARED已有准备

I trained the model on 1,200 simulated faults and processed 18 months of phasor-measurement data as a research assistant.

HERE为何在此

Professor A’s work on protection in inverter-dominated networks and Dr B’s graph-based state estimation match my problem and my method.

NEXT今后方向

A fault locator that works as solar grows, an open benchmark, then research on grid protection.

The worked example used in this guide本指南使用的范例

Example applicant 范例申请人

Background 背景
BEng Electrical Engineering and Automation (Mainland China); MSc Electrical Power Engineering (UK), with a dissertation on graph neural networks for fault location on the IEEE 33-bus test system; one year as a research assistant on a phasor-measurement project; second author of one conference paper.

Applying to 申请
PhD in Electrical Engineering at US universities: a personal statement beside the statement of purpose, from about 250 words to 2 pages and kept to each school’s limit.

Working title 暂定题目
Physics-Informed Graph Learning for Fault Detection and Location in Distribution Networks with High Inverter-Based Generation

How to use the example 如何使用范例

The same applicant appears in every section so you can watch one statement move from opening to closing; the whole text is printed after the sections. Don’t copy the wording. Study the moves, read the numbered notes beside each extract, then make the same moves with your own experience. Citations shown as (Authors, Year) are placeholders, and “Dr A” and “Professor A” stand for the people you will name. The examples are in English only, because your application will be written in English.

同一位申请人贯穿每个部分,方便你看到一份陈述如何从开头写到结尾;各部分之后附有完整全文。不要照搬措辞,要学习写法:先看每段旁边的编号解析,再用你自己的经历做同样的事。(Authors, Year) 为占位符,“Dr A”“Professor A”代表你将写出的具体导师。范例仅用英文,因为你的申请材料也用英文写。

Section 1 第 1 部分
1

Opening开头

Typical length 典型篇幅 60-120 words · about 10% 约占 10%

The first paragraph tells the reader what you want to research and why that question matters to you. Start from your own work: an observation in your data, a problem in a project, a result that did not fit. State the programme you are applying for by the end of the paragraph.

第一段要告诉读者你想研究什么,以及这个问题为什么对你重要。从你自己的研究出发:数据中的一个发现、项目中的一个难题、一个不符合预期的结果。在段落结尾写明你申请的项目。

Blueprint结构

1Move 1 · The source

A specific moment in your own work that raised the question.

你自己工作中引出问题的具体时刻。

2Move 2 · The question

The question in one plain sentence.

用一句平实的话写出问题。

3Move 3 · The application

The degree, the department and, where allowed, the supervisor.

学位、院系,允许的话写上导师。

Worked example · Opening paragraph 范例 · 开头段

1My MSc dissertation model, a graph neural network trained on simulated faults, located 94% of faults within one bus on the IEEE 33-bus feeder, but its accuracy fell sharply once I added rooftop solar to the network. 2The question I want to study is how to detect and locate faults on distribution networks where inverter-based generation makes fault currents small and unusual, using the few measurements a utility can afford. 3I am applying for the PhD in Electrical Engineering to develop the method set out in my proposal, a physics-informed graph model validated with hardware-in-the-loop tests.

Why it works 写法解析
1A result and its limit 结果与局限

Opens with a number and the condition under which it broke down.

以一个数字和它失效的条件开头。

2The engineering question 工程问题

States the problem, the constraint and the practical setting.

写出问题、约束条件和实际场景。

3The application 申请对象

Ends with the degree and the method in one line.

以学位和方法收尾,一行写完。

Advertised projects 已发布项目

For an advertised studentship, open with the project title and reference instead of your own question, then say which part of your work prepares you for it.

申请已发布的奖学金项目时,开头写项目名称和编号,而不是你自己的问题,再说明你的哪部分工作为此做好了准备。

Section 1 continued 第 1 部分(续)

First lines that a panel keeps reading让评审读下去的第一句

✕ Weak 较弱“Since I was young I have loved taking machines apart, and I dream of becoming a scientist.”

A childhood interest that every engineering applicant can claim.

每位工科申请人都能写的童年兴趣。

✓ Stronger 更好“My graph model located 94% of simulated faults within one bus, but accuracy fell sharply once I added rooftop solar.”

A measured result and a limit, which is where research begins.

一个可测量的结果和它的局限,研究正从这里开始。

✕ Weak 较弱“With the development of new energy, the power system faces great challenges.”

True of the whole field and says nothing about the applicant.

对整个领域都成立,与申请人无关。

✓ Stronger 更好“Inverter-based generation makes fault currents small and unusual, so overcurrent protection can miss them.”

Names the technical mechanism behind the problem.

写出问题背后的技术机制。

The same move in other fields其他方向的同一写法

Biology 生物 “My CRISPR knockout removed the expected band, yet the mutant zebrafish developed normally, which raised the question of what compensates.”
Chemistry 化学 “The catalyst I made for my MSc reached 80% yield in batch, but only 35% once I moved the reaction into flow.”
Computer sci. 计算机 “My code-summarisation model scored well on the benchmark but failed on the private repositories of the company where I interned.”

Common opening mistakes开头常见错误

A list of interests 罗列兴趣

“I am interested in AI, smart grids, renewable energy and robotics.” Name one problem.

“我对人工智能、智能电网、新能源和机器人都感兴趣。”请只写一个问题。

Childhood story 童年故事

A paragraph about a parent, a book or a dream. Keep at most one sentence, and only if it explains a choice.

用一整段写父母、一本书或一个梦想。最多一句,而且只在能解释某个选择时才写。

Restating the title 重复题目

Copying the proposal’s title and abstract. The panel has already read them.

照抄研究计划的题目和摘要。评审已经读过了。

Section 2 第 2 部分
2

Research Motivation & Preparation研究动机与准备

Typical length 典型篇幅 300-400 words · about 35% 约占 35%

This is the longest part. Panels use it to judge whether you can do research. Choose two or three experiences in the order they built your skills. For each one, say what you did yourself, which methods and tools you used, what result you reached and what it taught you about the next question.

这是篇幅最长的部分,评审用它判断你是否具备研究能力。按技能逐步积累的顺序,挑选两三段经历。每一段都要写明:哪些工作是你亲手做的、用了什么方法和工具、得到了什么结果,以及它如何引出下一个问题。

Blueprint结构

1Foundation

Undergraduate training that matters for the project, in one or two sentences.

与项目相关的本科训练,一两句话。

2First research

Dissertation or project: your role, method, tools, result.

毕业论文或项目:你的角色、方法、工具、结果。

3Further research

RA work, publication, conference, or research at work.

研究助理、发表、会议或工作中的研究。

4What it taught you

How these experiences lead to the proposed project.

这些经历如何引向拟议的项目。

Worked example · Motivation and preparation 范例 · 研究动机与准备

1My BEng in Electrical Engineering and Automation in Mainland China covered power-system analysis and relay protection, and my final-year project on transformer fault diagnosis was my first experience of training a classifier on measured data.

2In the MSc in Electrical Power Engineering in the UK I moved to distribution networks. For my dissertation I generated 1,200 fault scenarios on the IEEE 33-bus feeder in OpenDSS, built a graph neural network in PyTorch that treats buses as nodes and lines as edges, and compared it with an impedance-based method. The model located 94% of faults within one bus, but when I added rooftop solar at 30% of buses its accuracy fell below 70%. Finding out why, by inspecting which faults it misplaced, taught me more than the headline result.

3For the past year I have worked as a research assistant on a phasor-measurement project. I wrote a pipeline in Python that cleans 18 months of measurement data with timestamp gaps and outliers, and I am second author of a conference paper on event detection from those data. The code is on a public repository. 4These two experiences are the basis of my proposal: the dissertation showed the problem that solar creates, and the measurement work showed me how sparse and noisy real data are, which is why the method combines graph learning with the physics of the network.

Why it works 写法解析
1Foundation 基础

Relevant courses and a first project with data, briefly.

简要写出相关课程和第一个使用数据的项目。

2Tools, numbers, failure 工具、数字与失败

Software, dataset size, accuracy and the condition that broke the model.

软件、数据规模、准确率,以及导致模型失效的条件。

3Real data and a paper 真实数据与论文

A pipeline, a public repository and an authorship role a panel can check.

数据处理流程、公开代码库和可核实的作者身份。

4Back to the method 回到方法

Explains why the proposal is physics-informed.

解释研究计划为何采用物理信息约束。

Section 2 continued 第 2 部分(续)

Showing preparation, not claiming it用事实展示准备,而不是自我评价

✕ Weak 较弱“I have strong programming ability and master many deep-learning frameworks.”

A claim with no artefact or result.

只有论断,没有成果或结果。

✓ Stronger 更好“I built a graph neural network in PyTorch that located 94% of 1,200 simulated faults within one bus; the code is public.”

A model, a number and something the panel can open.

一个模型、一个数字,以及评审可以打开查看的东西。

✕ Weak 较弱“I participated in a national project on smart grids.”

“Participated” hides the applicant’s role and the project’s size hides it further.

“参与”掩盖了申请人的角色,项目的规模更让人看不清。

✓ Stronger 更好“I wrote the pipeline that cleans 18 months of phasor data and am second author of the paper that used it.”

One task, one output, one role.

一项任务、一项产出、一个角色。

The same move in other fields其他方向的同一写法

Biology 生物 “I ran the knockout screen, imaged 300 embryos on a confocal microscope and analysed the images in Fiji.”
Chemistry 化学 “I synthesised and characterised nine catalysts by XRD and NMR and built the flow reactor used in the yield tests.”
Computer sci. 计算机 “I fine-tuned three code models, built a test set of 2,000 private functions with the company, and wrote the evaluation scripts.”

Common preparation mistakes准备部分常见错误

Competition prizes as research 把竞赛奖项当研究

Several paragraphs on modelling contests. Keep one line unless the work became a research result.

用几段写建模竞赛。除非成果发展成了研究结果,否则只写一行。

Everything included 样样都写

Every course, internship and competition, at equal length. Pick the two or three that lead to the project.

把每门课、每段实习和每次竞赛都写进去,而且篇幅相同。只选通向项目的两三段。

Team results as your own 团队成果算作个人

Writing “we found” without saying what you did. Panels compare the statement with your referees’ letters.

只写“我们发现”却不说你做了什么。评审会把陈述与推荐信相互印证。

Section 3 第 3 部分
3

Fit匹配度

Typical length 典型篇幅 200-300 words · about 25% 约占 25%

Supervisors read the fit paragraph most closely. It tells them whether you have read their work and whether you would contribute to their group. Name the supervisor, link one of their papers to your question and another to your method, then name a resource at the university that makes the project possible.

导师读得最仔细的是匹配段。从这一段能看出你是否读过他们的研究,能否为课题组出力。写出导师姓名,把对方的一篇论文与你的问题联系起来,另一篇与你的方法联系起来,再写出这所大学中让项目可行的具体资源。

Blueprint结构

1The supervisor

Name, and a specific paper linked to your question.

姓名,以及一篇与你的问题相关的具体论文。

2The second link

A co-supervisor, method or source linked to your design.

与你的设计相关的第二导师、方法或资料。

3The place

Group, centre, facility, archive, data or training.

课题组、研究中心、设施、档案、数据或培训。

4What you bring

What you add to the group in return.

你能反过来为课题组带来什么。

Worked example · Fit 范例 · 匹配度

1Professor A’s work on protection in inverter-dominated networks (Authors, Year) defines the problem my project addresses, and her finding that fault currents from inverters are limited and controller-dependent explains why my dissertation model failed.

2Dr B’s research on graph-based state estimation (Authors, Year) provides the modelling approach I would extend, by adding network physics to the learning model. 3The school’s real-time digital simulation laboratory would make hardware-in-the-loop validation possible, and its links with network operators offer a route to field data.

4In return, I bring a working codebase for fault simulation and graph learning, experience of cleaning measurement data, and a benchmark dataset that other students in the group could use.

Why it works 写法解析
1Problem link 问题联系

Professor A’s finding explains the applicant’s own failed result.

A 教授的发现解释了申请人自己的失败结果。

2Method link 方法联系

Dr B’s approach is named and what will be added to it.

写出 B 老师的方法以及将在其基础上补充什么。

3Facility 设施

The simulator and partners are what make validation possible.

仿真实验室和合作方让验证成为可能。

4What you bring 你能带来什么

Code and data the group can reuse.

课题组可以复用的代码与数据。

Section 3 continued 第 3 部分(续)

Writing fit that a supervisor recognises写出导师认得出的匹配

✕ Weak 较弱“Your lab has advanced equipment and a strong research team.”

Says nothing about which equipment or why it matters for you.

没说是什么设备,也没说为什么对你重要。

✓ Stronger 更好“The real-time digital simulation laboratory would make hardware-in-the-loop validation of the method possible.”

Names the facility and the stage of the project that needs it.

写出设施以及项目中需要它的环节。

✕ Weak 较弱“I have read all of Professor A’s papers and admire her achievements.”

Unlikely and uncheckable; shows no understanding.

不太可信也无法核实,看不出理解。

✓ Stronger 更好“Professor A’s finding that inverter fault currents are controller-dependent explains why my model failed.”

One finding, understood and applied.

一个被理解并用上的发现。

Before you name a supervisor 点名导师之前

Check three things: that the person is taking students this year, that you have read and can discuss the work you cite, and that the programme lets you name supervisors. You usually choose faculty in the application rather than email them: Carnegie Mellon needs no adviser before applying, the U Washington Allen School needs no prior contact, Michigan CSE has you choose faculty in the application, and Brown Computer Science says there is no need to name faculty.

先核对三件事:对方今年是否招生;你引用的研究是否读过、能否展开讨论;项目是否允许写出导师姓名。通常在申请表中选择老师,而不是发邮件:卡内基梅隆申请前无须确定导师,华盛顿大学 Allen 学院无须事先联系,密歇根 CSE 在申请表中选择老师,布朗计算机说明无须写出老师姓名。

The same move in other fields其他方向的同一写法

Biology 生物 “Dr A’s lab studies genetic compensation in zebrafish (Authors, Year), and the institute’s imaging facility supports the live imaging I plan.”
Chemistry 化学 “Professor A’s work on flow photochemistry (Authors, Year) matches my problem, and the department’s flow-chemistry suite has the reactors I need.”
Computer sci. 计算机 “Dr A’s papers on evaluating code models out of distribution (Authors, Year) match my question, and the group’s GPU cluster makes the experiments feasible.”

Common fit mistakes匹配部分常见错误

A project that does not exist 项目与对方无关

Proposing a project unrelated to the group’s current funding or equipment. Check the group page and recent papers first.

提出与课题组现有经费或设备无关的项目。先查看课题组页面和近期论文。

One name, no reason 只有名字没有理由

Naming a professor without linking any of their work to yours.

点了教授的名字,却没有把对方的任何研究与你的研究联系起来。

Copied fit paragraph 复制匹配段

The same paragraph sent to six universities with the name changed. Supervisors notice.

同一段话改个名字就发给六所大学,导师一眼就能看出来。

Section 4 第 4 部分
4

Readiness & Goals准备度与目标

Typical length 典型篇幅 150-250 words · about 20% 约占 20%

Panels want to know that you understand what three or four years of research involve and that you will finish. Show this with evidence of independent work and of coping with problems, then say where the PhD leads. A goal is more convincing when it is specific and follows from the project.

评审想知道你是否了解三四年的研究生活意味着什么,以及你能否坚持到底。用独立工作和解决问题的经历来证明这一点,再说明博士之后的去向。目标越具体、越是从项目中自然延伸出来,就越有说服力。

Blueprint结构

1Why now

Why this is the right point to start a PhD.

为什么现在是读博的合适时机。

2Realistic view

A setback you handled, and what it showed you about research.

一次你处理过的挫折,以及它让你对研究有何认识。

3Outputs

What you aim to produce during the PhD.

你计划在博士期间产出什么。

4After the PhD

A specific next step and a longer aim.

具体的下一步与更长远的目标。

Worked example · Readiness and goals 范例 · 准备度与目标

1A year of research assistance has shown me that most of the work in a project lies in data and tools rather than in the model, and I am applying now because I have a defined problem, a working codebase and a supervisor-agreed plan. 2When 11% of the phasor data turned out to have timestamp gaps, I wrote a gap-detection step, tested three ways of filling short gaps against held-out data, and documented the choice; the paper’s results were later shown to be unchanged with and without the filled data.

3During the PhD I aim to publish the method and the hardware-in-the-loop validation as journal papers and to release the benchmark dataset openly. 4Afterwards I plan to work on grid protection in a university or a network operator’s research team, helping utilities place the few sensors they can afford where they detect the most faults.

Why it works 写法解析
1Why now 为何是现在

Problem, code and plan are in place.

问题、代码和计划都已到位。

2A technical setback 技术上的挫折

A data problem, a tested fix and a check on the result.

一个数据问题、经过检验的修复方法,以及对结果的核查。

3Outputs 产出

Papers plus an open dataset, which engineering panels value.

论文加开放数据集,这是工科评审看重的。

4Two possible routes 两条可能的路径

Academia or industry research, both linked to the project.

学术界或企业研究,都与项目相关。

Section 4 continued 第 4 部分(续)

Goals a panel believes让评审信服的目标

✕ Weak 较弱“I can work 12 hours a day in the lab.”

Hours are not evidence of good research.

工作时长不能证明研究做得好。

✓ Stronger 更好“I tested three ways of filling gaps against held-out data and showed the results were unchanged.”

Shows judgement and checking, which matter more than hours.

展示判断与核查,比时长更重要。

✕ Weak 较弱“In the future I want to make great contributions to the energy industry.”

A wish with no route.

只有愿望,没有路径。

✓ Stronger 更好“I plan to work on grid protection in a network operator’s research team.”

A specific place for the skills the PhD builds.

为博士阶段培养的能力指出具体去处。

The same move in other fields其他方向的同一写法

Biology 生物 “I aim to publish the compensation mechanism and to continue in developmental genetics as a postdoctoral researcher.”
Chemistry 化学 “I plan to publish the flow method and to work in process chemistry, where scale-up is the main problem.”
Computer sci. 计算机 “I aim to release the evaluation suite and to work on reliable code models in a research lab.”

Common readiness mistakes准备度常见错误

Industry goal that skips research 忽视研究的职业目标

“I will join a big tech company as an engineer.” Link the goal to research skills the PhD builds.

“我将进入大型科技公司当工程师。”请把目标与博士阶段培养的研究能力联系起来。

Only ambition 只谈抱负

Goals with no evidence that you know what research involves day to day.

只写目标,没有证据表明你了解日常研究工作。

Hiding the gap 回避空档

A year out of study left unexplained. Say briefly what you did and what it added.

对一年的学业空档不作解释。简要说明你做了什么、有何收获。

Section 5 第 5 部分
5

Closing结尾

Typical length 典型篇幅 40-100 words · about 10% 约占 10%

The closing paragraph brings the three threads together in two or three sentences: the question, your preparation and the fit. Do not introduce new material, and do not thank the panel at length. If the course allows, say you would welcome the chance to discuss the project.

结尾段用两三句话把三条线索收拢:研究问题、你的准备、你与这里的匹配。不要引入新内容,也不要长篇致谢。如果项目允许,可以写一句欢迎进一步讨论项目。

Blueprint结构

1Return to the question

One clause that names it again.

用一个分句再次点出问题。

2Why you, why here

Preparation and fit in a single sentence.

用一句话概括准备与匹配。

3Next step

Interview or discussion, briefly.

简短提及面试或进一步讨论。

Worked example · Closing paragraph 范例 · 结尾段

1My proposal sets out a fault detection and location method that keeps working as solar generation grows. 2I am applying because the problem grew out of my own dissertation results, because my MSc and measurement work have given me the tools to build and test the method, and because Professor A and Dr B’s group has the expertise and the simulation facility the project needs. 3I would welcome the opportunity to discuss the project at interview.

Why it works 写法解析
1The problem 问题

Restated as a result the project will deliver.

以项目将交付的成果重述问题。

2Three reasons 三个理由

Origin, tools and fit in one sentence.

来由、工具与匹配,一句写完。

3A modest close 得体收尾

One sentence; the next step.

一句话,写下一步。

✕ Weak 较弱“I am confident that I will become an excellent PhD student in your lab.”

Confidence without content.

只有自信,没有内容。

✓ Stronger 更好“I would welcome the opportunity to discuss the project at interview.”

Short and specific.

简短而具体。

Common closing mistakes结尾常见错误

A new idea at the end 结尾出现新内容

A second research interest or a new experience in the final paragraph. Move it up or cut it.

在最后一段提出第二个研究兴趣或新的经历。要么前移,要么删掉。

Repeating the opening 重复开头

Copying the first paragraph. Summarise the case instead.

照抄第一段。请改为概括全文。

Letter format in a form 表格里写成书信

“Dear Sir or Madam” and a signature in a form field. A statement is not a letter unless the course asks for one.

在表格栏目里写“Dear Sir or Madam”和署名。除非项目要求写信,陈述不是书信。

Section 6 第 6 部分
6

Length & Formatting篇幅与格式

Applies to every version 适用于所有版本

A statement over the limit can be cut off by the form or set aside without being read. Within the limit, use most of the space, and make the document easy to read on a screen. UC Davis caps each statement at 4,000 characters and UC Santa Barbara sets a minimum of 250 words for the personal history statement, so check both ends of the limit.

超出限制的陈述可能被表格截断,或者不经阅读就被搁置。在限制之内尽量用足篇幅,并让文件在屏幕上便于阅读。加州大学戴维斯规定每篇陈述不超过 4,000 字符,加州大学圣塔芭芭拉规定个人经历陈述至少 250 词,所以上下限都要核对。

ITEM 项目UPLOADED DOCUMENT 上传文件TEXT BOX IN THE FORM 表格栏目
Length
篇幅
Use 90 to 100% of the limit. If no limit is given, 1 to 2 pages or about 1,000 words.
用足限制的 90% 至 100%。没有限制时,写 1 至 2 页或约 1,000 词。
Check whether spaces count. Paste early and look at where the text is cut.
确认空格是否计入。尽早粘贴,看看文字在哪里被截断。
Heading
标题
Your name, the programme and the document’s name as the course page gives it.
写上姓名、项目名称,以及课程页面使用的文件名称。
None. Start with the first sentence.
不需要,直接从第一句写起。
Font and layout
字体与版式
A plain font at 11 to 12 point, normal margins, single or 1.15 spacing unless told otherwise; paragraphs, not bullet points.
普通字体 11 至 12 号,常规页边距,除非另有要求用单倍或 1.15 倍行距;用段落而不是项目符号。
Plain text. Bold, italics and bullet symbols may be lost; leave a blank line between paragraphs.
纯文本。粗体、斜体和项目符号可能会丢失;段落之间空一行。
File and spelling
文件与拼写
PDF named clearly, for example Surname_Firstname_Statement.pdf; one spelling convention throughout.
PDF,命名清楚,例如 Surname_Firstname_Statement.pdf;全文使用同一种拼写习惯。
Keep a copy of the text you submitted; you will need it before the interview.
保留一份提交的文本,面试前需要用到。

Language habits to check需要检查的语言习惯

✕ Weak 较弱“As we all know, artificial intelligence is changing every field of science.”

A general opener that adds nothing.

笼统的开头,没有增加任何信息。

✓ Stronger 更好“On the IEEE 33-bus feeder with 30% rooftop solar, my model’s accuracy fell below 70%.”

A specific, checkable statement in fewer words.

用更少的字写出具体、可核查的内容。

The worked example in full 完整范例

The full statement, start to finish完整陈述范例

Here is the worked example as one document of about 655 words: the master version you cut from. Read it once without the notes to hear how the parts join, then check that each paragraph does one job. Do not reuse its sentences; use it to check proportions. For a 500-word or one-page version, keep the opening, the dissertation paragraph, the first fit paragraph and one sentence of goals; for an advertised studentship, replace the opening with the project title and answer its requirements in order.

下面把范例合成一份完整文件,约 655 词,即你裁剪各版本时使用的母版。先不看解析通读一遍,感受各部分如何衔接,再检查每一段是否只承担一项任务。不要套用其中的句子,用它来检查各部分的比例。如果只能写 500 词或一页,保留开头、毕业论文段、匹配部分第一段和一句目标;申请已发布的奖学金项目时,把开头换成项目名称,并按顺序回应其要求。

9622711114576
1 Opening2 Research Motivation & Preparation3 Fit4 Readiness & Goals5 Closing

Worked example · Personal statement, PhD in Electrical Engineering范例 · 完整陈述

My MSc dissertation model, a graph neural network trained on simulated faults, located 94% of faults within one bus on the IEEE 33-bus feeder, but its accuracy fell sharply once I added rooftop solar to the network. The question I want to study is how to detect and locate faults on distribution networks where inverter-based generation makes fault currents small and unusual, using the few measurements a utility can afford. I am applying for the PhD in Electrical Engineering to develop the method set out in my proposal, a physics-informed graph model validated with hardware-in-the-loop tests.

My BEng in Electrical Engineering and Automation in Mainland China covered power-system analysis and relay protection, and my final-year project on transformer fault diagnosis was my first experience of training a classifier on measured data.

In the MSc in Electrical Power Engineering in the UK I moved to distribution networks. For my dissertation I generated 1,200 fault scenarios on the IEEE 33-bus feeder in OpenDSS, built a graph neural network in PyTorch that treats buses as nodes and lines as edges, and compared it with an impedance-based method. The model located 94% of faults within one bus, but when I added rooftop solar at 30% of buses its accuracy fell below 70%. Finding out why, by inspecting which faults it misplaced, taught me more than the headline result.

For the past year I have worked as a research assistant on a phasor-measurement project. I wrote a pipeline in Python that cleans 18 months of measurement data with timestamp gaps and outliers, and I am second author of a conference paper on event detection from those data. The code is on a public repository. These two experiences are the basis of my proposal: the dissertation showed the problem that solar creates, and the measurement work showed me how sparse and noisy real data are, which is why the method combines graph learning with the physics of the network.

Professor A’s work on protection in inverter-dominated networks (Authors, Year) defines the problem my project addresses, and her finding that fault currents from inverters are limited and controller-dependent explains why my dissertation model failed.

Dr B’s research on graph-based state estimation (Authors, Year) provides the modelling approach I would extend, by adding network physics to the learning model. The school’s real-time digital simulation laboratory would make hardware-in-the-loop validation possible, and its links with network operators offer a route to field data.

In return, I bring a working codebase for fault simulation and graph learning, experience of cleaning measurement data, and a benchmark dataset that other students in the group could use.

A year of research assistance has shown me that most of the work in a project lies in data and tools rather than in the model, and I am applying now because I have a defined problem, a working codebase and a supervisor-agreed plan. When 11% of the phasor data turned out to have timestamp gaps, I wrote a gap-detection step, tested three ways of filling short gaps against held-out data, and documented the choice; the paper’s results were later shown to be unchanged with and without the filled data.

During the PhD I aim to publish the method and the hardware-in-the-loop validation as journal papers and to release the benchmark dataset openly. Afterwards I plan to work on grid protection in a university or a network operator’s research team, helping utilities place the few sensors they can afford where they detect the most faults.

My proposal sets out a fault detection and location method that keeps working as solar generation grows. I am applying because the problem grew out of my own dissertation results, because my MSc and measurement work have given me the tools to build and test the method, and because Professor A and Dr B’s group has the expertise and the simulation facility the project needs. I would welcome the opportunity to discuss the project at interview.

Putting it to work 实际应用

An email to faculty, where it helps在有帮助时给老师写邮件

Most US programmes do not expect an email before you apply: faculty are named in the statement of purpose or chosen in the form. Carnegie Mellon Computer Science needs no adviser before applying, the U Washington Allen School needs no prior contact, Duke Computer Science says contact is not needed, and Cornell English asks you not to contact faculty before an offer. Some psychology programmes welcome it: UC Davis strongly recommends emailing potential mentors, Carnegie Mellon and Georgia Tech encourage contact, and Wisconsin-Madison asks you to check whether faculty admit that year. Where you write, there is no word limit: give the professor enough detail to judge the fit. Say who you are and which programme you are applying to, state your question and your strongest piece of preparation, link one specific piece of their work to your plan, ask one clear question (for example whether they are taking students this cycle), and attach your CV. Write the first draft yourself; your academic will check the fit paragraph.

多数美国项目并不期待申请前的邮件:意向老师写在目的陈述里或在申请表中选择。卡内基梅隆计算机申请前无须确定导师,华盛顿大学 Allen 学院无须事先联系,杜克计算机说明无须联系,康奈尔英文系要求录取前不要联系老师。一些心理学项目则欢迎联系:加州大学戴维斯强烈建议给潜在导师发邮件,卡内基梅隆和佐治亚理工鼓励联系,威斯康星麦迪逊要求先确认老师当年是否招生。如果写信,邮件没有字数上限,要写得足够详细,让老师能判断是否匹配:说明你是谁、申请哪个项目,写出研究问题和最有力的一项准备,把对方的一项具体研究与你的计划联系起来,提一个明确的问题(例如今年是否招生),并附上简历。第一稿由你自己撰写,导师负责核对匹配段落。

Worked example · First email to a supervisor 范例 · 第一封联系导师的邮件

Subject: Prospective PhD student (2027 entry): fault location in distribution networks with high solar

Dear Professor [Surname],

1I am completing an MSc in Electrical Power Engineering at a UK university and plan to apply for a PhD in your group for 2027 entry. I am writing to ask whether you would consider supervising a self-proposed project, or whether a related studentship is planned.

2Your paper on protection in inverter-dominated networks (Authors, Year) shows that inverter fault currents are limited and depend on the controller. In my dissertation I trained a graph neural network on 1,200 simulated faults on the IEEE 33-bus feeder; it located 94% of them within one bus, but its accuracy fell below 70% once I added rooftop solar, which your finding helps explain.

3I would like to add network physics to the graph model and validate it with hardware-in-the-loop tests. As a research assistant I processed 18 months of phasor-measurement data and I am second author of a conference paper based on it; my code is at [link].

4Would a project of this kind fit your group’s plans? I attach my CV and a two-page outline.

Kind regards,
[Full name]

Why it works 写法解析
1Two routes in one line 一句写明两条路径

Asks about both a self-proposed project and studentships.

同时询问自主选题和奖学金项目。

2Their finding explains your result 对方发现解释你的结果

The strongest kind of link in a science email.

理工科邮件中最有力的一种联系。

3Evidence they can open 可以打开的证据

A code link and a paper.

代码链接和论文。

4One question 一个问题

Easy to answer with yes, no or a suggestion.

对方用“是”“否”或一个建议就能回答。

What to attach

Your CV and a one- or two-page summary of the project, as PDFs named with your surname.

简历和一至两页的项目摘要,PDF 格式,文件名带上你的姓。

When to send it

Well before the first deadline, so that there is time for a reply, a call and a revised draft.

远早于第一个截止日期发出,留出回复、通话和修改草稿的时间。

After a reply

Name the supervisor in the fit paragraph and say what you discussed; do not quote the email.

在匹配段写出导师姓名并说明讨论内容;不要引用邮件原文。

Putting it to work 实际应用

Adapting to each university按校调整

Keep the 1,000-word master version and cut from it. Never write a new statement for each university; rewrite the fit paragraph and trim to the limit, in the way the row for each target describes.

保留 1,000 词母版,在此基础上裁剪。不要为每所大学从头重写,只需改写匹配段并按下表各行的做法压缩到限制以内。

TARGET 目标WHAT TO DO 做法
A form box or character limit
表格栏目或字符限制
UC Santa Barbara Computer Science (personal history statement of at least 250 words, no duplication) 加州大学圣塔芭芭拉
Keep the question, your best piece of evidence and the named fit, in about 250 to 800 words. Drop the greeting and the closing paragraph, use plain text, and paste the draft into the form early to see where it is cut.
保留研究问题、最有力的一条证据和具体的匹配,约 250 至 800 词。不写称呼和结尾段,用纯文本,并尽早把草稿粘贴进表格,看看在哪里被截断。
About 500-750 words or one page
约 500-750 词或一页
UCLA Computer Science (personal statement up to 500 words beside a 500-word SOP) UCLA 计算机
One sentence of opening, one strong example of preparation, a fit paragraph naming one supervisor, one sentence of goals.
开头一句、准备部分一个有力的例子、点名一位导师的匹配段、目标一句。
About 1,000 words or up to two pages
约 1,000 词或不超过两页
UC Berkeley Mechanical Engineering (Personal History 1-2 pages, no overlap); MIT Chemistry (500-1,000 words) 伯克利机械工程、麻省理工化学
Use the master version. Rewrite the fit paragraph for each university and check whether the limit is in words or pages.
使用母版。针对每所大学改写匹配段,并确认限制按词数还是页数计算。
A statement in place of a proposal
陈述代替研究计划
Penn Engineering (2 pages answering six prompts); Columbia SEAS (250-1,000 words); Brown, Boston University and USC Viterbi (personal statement as the main essay) 宾大工程、哥伦比亚、布朗、波士顿大学、南加州大学
Add a research-interests paragraph of 150 to 250 words after the opening: the question, the data or sources, and the method. Name the faculty whose work fits, because no proposal will do it for you.
在开头之后加一段 150 至 250 词的研究兴趣:研究问题、数据或材料、研究方法。写出研究方向匹配的老师,因为没有研究计划替你完成这一步。
Before you submit 提交之前

Read it like a panel用评审的眼光重读

Reread your statement as a supervisor with five minutes and thirty others to read. Score each question 1 (weak), 2 (adequate) or 3 (strong), and fix the lowest score first.

像一位只有五分钟、还有三十份要读的导师那样重读。每项打 1(弱)、2(一般)、3(强)分,先修改得分最低的一项。

#CHECK 检查SELF-SCORE 自评
1Does the first paragraph name a research question and where it came from? 第一段是否写出了研究问题及其来由?1 2 3
2Does each experience give a tool, a number and what I built or measured myself? 每段经历是否写明了工具、数字以及我亲手搭建或测量的内容?1 2 3
3Does every claim about my ability have an example behind it? 每个关于能力的说法背后都有例子吗?1 2 3
4Does the fit paragraph link named work by the supervisor to my question or method? 匹配段是否把导师的具体研究与我的问题或方法联系起来?1 2 3
5Are my goals specific, and do they follow from the project? 我的目标是否具体,并从项目中延伸而来?1 2 3
6Does it add to the proposal and CV rather than repeat them? 它是在补充研究计划和简历,而不是重复吗?1 2 3
7Is it the right document type and within the limit, in words, pages or characters? 文件类型是否正确,词数、页数或字符数是否在限制以内?1 2 3

Three tests三个测试

The name test

Cover the university’s name. Could the statement be sent anywhere? If so, the fit paragraph needs rewriting.

遮住学校名称,这份陈述是否可以发给任何学校?如果可以,匹配段需要重写。

The evidence test

Underline every adjective about yourself. Replace each one with the event that shows it, or delete it.

把所有描述自己的形容词划出来,逐一替换为能体现它的事件,或者删掉。

The interview test

Could you talk for two minutes about every paper, method and experience you mention? If not, cut it or prepare.

你能就文中提到的每篇论文、每种方法和每段经历讲两分钟吗?不能的话,要么删掉,要么做好准备。