2027 Best Online Machine Learning Doctorate Programs for Government and Public Sector Careers
Co-Founder and Chief Data Scientist
Choosing an online machine learning doctorate for a government career means balancing academic rigor with public-service realities: budgets, security rules, policy impact, and full-time work. The U.S. Bureau of Labor Statistics projects 26% growth for computer and information research scientists from 2023 to 2033, signaling strong demand for advanced AI and machine learning expertise.
This guide is for public sector professionals, contractors, analysts, engineers, and administrators comparing accredited online doctoral options. You'll learn how cost, format, curriculum, dissertations, and employer recognition affect the right choice.
Key Things You Should Know
- The strongest online machine learning doctorate choices are usually accredited PhD, DSc, or applied doctoral programs in data science, artificial intelligence, computer science, information systems, or engineering with machine learning research-not necessarily programs labeled only "machine learning."
- For career value, compare programs by public sector relevance: data governance, cybersecurity, responsible AI, geospatial analytics, health informatics, defense analytics, privacy, procurement, and policy translation matter as much as algorithms.
- BLS reported a $140,910 median annual wage for computer and information research scientists in 2024, while 2025 federal GS-15 base pay runs from $123,041 to $159,950 before locality pay; doctorate ROI depends on role, agency, location, clearance, and leadership path.
What Are the Best Online Machine Learning Doctorate Programs for Government and Public Sector Careers?
The best online machine learning doctorate for a government or public sector career is the one that combines recognized accreditation, advanced quantitative training, credible faculty supervision, and a research pathway tied to public problems.
Because fully online doctorates titled specifically "machine learning" are still uncommon, students should broaden the search to machine learning-heavy doctorates in data science, artificial intelligence, computer science, computational engineering, information systems, or analytics.
If you want a research-intensive path with broad employer recognition, compare ML-focused programs with an online PhD in data science, since many data science doctorates include machine learning, statistical modeling, cloud computing, and applied research methods that fit public sector analytics roles.
The table below summarizes the strongest program models for government and public sector career alignment. Use it to identify the type of doctorate that best matches your intended role rather than relying only on a program title:
| Program model | Common degree labels | Best public sector fit | Main trade-off |
| Research doctorate with machine learning concentration | PhD in Data Science, Computer Science, Artificial Intelligence, or Computational Science | Federal labs, research agencies, policy research centers, university-affiliated research institutes, senior analytics roles | Usually requires a dissertation, strong math preparation, and several years of sustained research |
| Applied professional doctorate | Doctor of Science, Doctor of Computer Science, Doctor of Information Technology, or Doctor of Engineering | Agency technology leadership, digital transformation, public sector consulting, enterprise AI governance | May emphasize applied projects over publication-focused research, which can matter for research-heavy roles |
| Public-interest technology doctorate | Information Systems, Informatics, Public Policy Analytics, or Technology Management doctorate | State and local government modernization, civic technology, public health analytics, education data systems | May offer less depth in advanced algorithms unless electives and dissertation supervision are carefully chosen |
| Security-focused AI doctorate | Cybersecurity, Cyber Operations, or Assurance doctorate with AI/ML coursework | Defense, homeland security, intelligence support, fraud detection, cyber threat analytics | Best for students whose machine learning goals are tied to security rather than general AI research |
For most government professionals, the "best" program is not simply the fastest or cheapest. It is the program that lets you complete a defensible dissertation or applied capstone using realistic public sector data constraints, such as privacy rules, sensitive information, procurement limits, and interagency governance.
Consider a doctorate if you need advanced research authority, want to lead AI strategy, plan to evaluate or build machine learning systems for high-stakes public decisions, or hope to move into senior technical leadership.
Consider a master's degree, graduate certificate, or employer-funded training instead if your goal is primarily to learn tools, qualify for entry-level analytics roles, or move quickly into implementation work without a long dissertation timeline.
How Should You Compare Online Machine Learning Doctorate Programs for Your Public Sector Career Goals?
Comparing online machine learning doctorate programs should begin with your target public sector function, not with rankings alone. A program that is excellent for academic AI research may not be the best fit for someone leading fraud analytics in a state agency, modernizing a city data platform, or managing AI risk in a federal office.
Use the following criteria to compare programs in a practical way. Each point affects whether the doctorate will support your public sector career rather than simply add another credential:
- Define your target role first: research scientist, chief data officer, AI governance lead, cyber analytics specialist, public health informatics leader, defense contractor researcher, or policy analyst.
- Verify institutional accreditation before reviewing tuition, curriculum, or faculty; an unaccredited doctorate can create problems in federal hiring, tuition reimbursement, and academic credibility.
- Check whether the curriculum includes machine learning theory, statistical modeling, data engineering, responsible AI, privacy, cybersecurity, cloud architecture, and research methods.
- Review faculty expertise for government-relevant work, including grants, public sector consulting, federal lab collaboration, health systems, national security, education policy, or civic technology.
- Ask whether dissertation or capstone topics can use restricted, de-identified, simulated, or publicly available government datasets when agency data cannot be released.
- Compare total cost, not just tuition per credit, including residencies, technology fees, dissertation continuation fees, travel, books, and time away from work.
- Assess flexibility by looking at asynchronous coursework, evening seminars, part-time sequencing, leaves of absence, and dissertation support for working professionals.
- Request career outcome details that are specific to doctoral students, public agencies, contractors, research labs, or nonprofit policy organizations.
Students who are still deciding between doctoral study and a shorter AI credential can review affordable AI degrees online to understand whether a master's-level route could meet the same career goal at lower cost and with less time risk.
The table below highlights common decision trade-offs. It can help you avoid choosing a program based on a single factor that does not match your long-term public sector plans:
| Decision | Choose this if... | Be careful if... |
| PhD | You want research leadership, publications, academic mobility, or federal lab credibility | You need a faster practice-focused credential with less emphasis on theory |
| Professional doctorate | You want to solve applied agency problems and move into executive technology leadership | Your target role requires a traditional research doctorate or peer-reviewed publication record |
| Fully online | You need maximum flexibility due to full-time work, military service, caregiving, or location | You benefit from in-person lab access, close research cohorts, or local federal research networks |
| Hybrid with residencies | You want stronger networking, research workshops, and faculty access | Travel costs or security-sensitive employment make required campus visits difficult |
| Low-cost program | You are self-funding or using limited agency tuition assistance | Lower tuition comes with weak advising, limited ML electives, or poor dissertation support |
| Accelerated program | You already have a strong research agenda, quantitative background, and protected study time | The pace leaves too little time for a rigorous dissertation or agency-approved data access |
Common mistakes include choosing a school before confirming accreditation, assuming any AI-labeled doctorate will impress government employers, ignoring dissertation requirements, and comparing tuition without calculating total program cost. A better approach is to build a shortlist around role fit, accreditation, faculty supervision, research feasibility, and funding.
Are Online Machine Learning Doctorate Programs Accredited and Respected by Government Employers?
Online machine learning doctorate programs can be respected by government employers when they are offered by properly accredited institutions and demonstrate academic rigor. Public sector employers generally care more about institutional legitimacy, relevant expertise, research quality, and role fit than whether every course was completed on campus.
For U.S. students, institutional accreditation from a recognized accreditor is the baseline quality check. Programmatic accreditation is less common for doctoral machine learning programs, although ABET accreditation may matter more for certain engineering or undergraduate computing pathways than for research doctorates.
When verifying quality, focus on evidence that an agency, university, or public contractor can reasonably evaluate. These indicators are especially important if you plan to use tuition reimbursement, apply for federal positions, or pursue leadership roles that require credential review:
- Institutional accreditation appears in the U.S. Department of Education or Council for Higher Education Accreditation databases.
- The doctorate is clearly listed in the university catalog with credit requirements, dissertation or capstone rules, faculty qualifications, and academic policies.
- The program discloses online delivery expectations, residency requirements, time limits, transfer rules, and continuation fees.
- Faculty have doctoral credentials and current work in machine learning, data science, AI governance, cybersecurity, public policy analytics, or relevant applied domains.
- The school can explain how online doctoral students receive research advising, library access, statistical software, computing resources, and dissertation committee support.
- The program has credible alumni outcomes, employer partnerships, publications, public sector projects, or applied research examples.
Respect also depends on how you use the degree. A dissertation on explainable AI for benefits eligibility, anomaly detection for public finance, wildfire risk modeling, or privacy-preserving health analytics may be more persuasive to a government employer than a generic machine learning project with no public sector context.
Be cautious with programs that make unrealistic salary promises, advertise extremely short doctoral timelines, hide faculty information, lack dissertation transparency, or appear to rely heavily on sales-style admissions. A legitimate online doctorate should withstand the same scrutiny you would apply to an on-campus program.
Admission requirements vary by university and degree type, but online machine learning doctorate programs usually expect evidence that you can handle advanced quantitative coursework and independent research. Government and public sector professionals can be strong applicants when they bring applied data problems, leadership experience, and domain knowledge.
Most programs evaluate a combination of academic preparation, technical readiness, research fit, and professional experience. The following requirements are common, although exact standards differ by school:
- A master's degree in computer science, data science, statistics, engineering, information systems, analytics, mathematics, public policy analytics, or a closely related field.
- Minimum graduate GPA requirements, often with stronger expectations for quantitative and technical courses.
- Prerequisite knowledge in programming, statistics, calculus, linear algebra, databases, algorithms, or machine learning fundamentals.
- Professional resume showing technical, analytical, research, leadership, military, government, nonprofit, contractor, or public service experience.
- Statement of purpose explaining why doctoral study is necessary and how the research interest fits faculty expertise.
- Writing sample, research proposal, technical portfolio, publications, project examples, or capstone work from a prior graduate program.
- Letters of recommendation from faculty, supervisors, research mentors, or senior technical leaders who can speak to doctoral readiness.
- Interview with faculty or program leadership, especially for dissertation-based programs with limited advising capacity.
Some programs admit students with a bachelor's degree directly into a longer doctoral pathway, but working professionals usually benefit from completing a relevant master's first. A master's program can build the math, coding, and research foundation needed to complete doctoral work without losing momentum in the dissertation stage.
If your background is in public administration, policy, criminal justice, health administration, or military operations rather than computing, look for bridge options. Useful preparation may include graduate courses in Python, statistics, machine learning, data management, research design, and cloud computing before applying.
Applicants with government experience should translate that experience into research questions. For example, instead of saying you worked with agency data, explain that you want to study model bias in eligibility screening, predictive maintenance for infrastructure, emergency response optimization, or secure machine learning workflows for regulated environments.
The best curriculum for government careers combines technical depth with public accountability. Machine learning systems used in public agencies often affect benefits, safety, infrastructure, policing, health, finance, education, transportation, and national security, so doctoral students need more than model-building skills.
Students who are earlier in their planning can use an artificial intelligence major career overview to see how AI skills connect to roles before committing to a doctorate. At the doctoral level, the curriculum should go deeper into research design, algorithmic limitations, governance, and domain-specific application.
The table below connects common machine learning specializations to public sector career uses. Use it to check whether a program's electives and dissertation supervision match the problems you want to solve:
| Specialization | Public sector applications | Useful course topics |
| Responsible AI and explainable machine learning | Benefits eligibility, auditing, procurement, public accountability, model risk management | Explainability, fairness, bias testing, AI governance, ethics, regulatory compliance |
| Cybersecurity and adversarial machine learning | Threat detection, fraud analytics, zero-trust monitoring, critical infrastructure protection | Security analytics, adversarial AI, anomaly detection, network science, cryptography basics |
| Geospatial and environmental machine learning | Disaster response, climate resilience, transportation planning, land management, wildfire risk | Remote sensing, GIS, spatial statistics, computer vision, simulation modeling |
| Health and social services analytics | Public health surveillance, Medicaid analytics, opioid response, resource allocation | Biostatistics, privacy-preserving analytics, causal inference, health informatics, NLP |
| Natural language processing | Public comments analysis, records management, legal and regulatory review, constituent services | Text mining, large language models, information retrieval, human-in-the-loop evaluation |
| Operations research and decision intelligence | Emergency logistics, staffing, maintenance, budgeting, public transportation optimization | Optimization, simulation, reinforcement learning, decision analysis, stochastic modeling |
A strong doctoral curriculum should also teach students how to evaluate whether machine learning is appropriate at all. In public service, a simpler statistical model, rule-based system, or human-centered workflow may sometimes be more transparent, defensible, and cost-effective than a complex AI system.
Look for programs that require rigorous research methods. Quantitative methods, experimental design, causal inference, reproducibility, data governance, and evaluation methods are especially valuable because government leaders often need evidence that a model is accurate, fair, secure, and legally defensible.
Online machine learning doctorate programs usually combine remote coursework with structured research milestones. The exact format matters because doctoral study is not just taking classes online; it also requires advising, independent research, committee approval, and sustained progress through a dissertation or applied doctoral project.
Most online doctoral formats fall into several patterns. Understanding these models helps you choose a structure that fits your work schedule, travel limits, and need for faculty interaction:
- Asynchronous online courses allow students to complete lectures and assignments on a weekly schedule, which is useful for government employees with variable hours or field responsibilities.
- Synchronous seminars require live online attendance and may offer stronger discussion, cohort connection, and faculty access.
- Hybrid programs include short residencies for research workshops, exams, proposal defense, networking, or dissertation planning.
- Dissertation-based doctorates require an original research contribution, committee review, proposal approval, data collection or experimentation, defense, and final manuscript submission.
- Capstone or applied project doctorates focus on a practical organizational problem, often producing an implementation plan, evaluation, prototype, or applied research report.
- Comprehensive exams or qualifying exams may test machine learning theory, research methods, statistics, computing, and specialization knowledge before dissertation work begins.
For government professionals, the most important question is whether the dissertation or capstone can be completed without violating data rules. Agencies may restrict data export, cloud storage, publication, human subjects research, or model deployment.
Ask schools how they support students using de-identified data, synthetic data, open government datasets, secure enclaves, or employer-approved research protocols.
Dissertation programs are usually better for students aiming at research scientist, policy research, federal lab, or academic-adjacent roles. Capstone-oriented doctorates may be better for executives, technology managers, and applied leaders who want to improve agency systems rather than produce theory-driven research.
Residencies are not automatically a drawback. Short, well-designed residencies can strengthen advising and networking, but they become a problem if travel costs, clearance obligations, deployment schedules, or family responsibilities make attendance unrealistic.
Working professionals commonly need several years to complete an online machine learning doctorate. The timeline depends on transfer credits, course load, dissertation readiness, faculty availability, residency requirements, and whether the student can protect regular research time while employed full time.
The table below gives a realistic planning view rather than a promise. Individual programs may be shorter or longer, especially when a dissertation requires data access, institutional review, security approval, or extensive experimentation:
| Student situation | Typical pace | What affects completion |
| Full-time employee taking one course at a time | Longest but most sustainable | Workload, family responsibilities, dissertation momentum, employer flexibility |
| Full-time employee taking two courses in some terms | Moderate pace | Quantitative background, course difficulty, agency busy seasons, residency travel |
| Student with a relevant master's and clear research topic | Potentially faster | Transfer credit rules, advisor match, proposal approval, data availability |
| Student changing fields from policy or administration | Often longer | Prerequisite coursework, coding preparation, statistics foundation, research methods training |
| Student using employer data | Variable | Legal review, data use agreements, privacy review, publication restrictions, supervisor approval |
The most common delay is not coursework; it is the dissertation or doctoral project. Machine learning research can slow down when datasets are incomplete, models do not perform as expected, committee feedback changes the scope, or the employer cannot approve data use.
To stay on track while working full time, choose a topic early, schedule weekly research blocks, confirm data access before the proposal stage, and select a program with structured dissertation milestones. A flexible program is helpful, but too much flexibility can become a risk if there are no clear deadlines or advisor check-ins.
Students should also be honest about opportunity cost. If the doctorate slows current job performance, delays promotion applications, or adds stress without a clear career purpose, a shorter graduate certificate, second master's, or employer-sponsored AI training may be a better near-term step.
Online machine learning doctorate costs vary widely by institution type, credit requirements, transfer policies, fees, residencies, and dissertation continuation charges.
NCES data for the 2023-24 academic year reported average graduate tuition and required fees of $12,596 at public institutions and $30,377 at private nonprofit institutions, which is a useful benchmark but not a substitute for program-level cost calculations.
Doctoral students should compare the full cost of attendance, not only advertised tuition. A lower per-credit price can become less attractive if the program requires more credits, expensive residencies, or multiple paid dissertation continuation terms:
- Tuition per credit or per term.
- Number of required credits after transfer evaluation.
- Technology, library, student services, and graduation fees.
- Residency travel, lodging, meals, and time away from work.
- Software, cloud computing, textbooks, data storage, and hardware costs.
- Dissertation continuation, extension, or candidacy fees.
- Lost overtime, reduced consulting work, or unpaid leave during intensive milestones.
Funding can come from several sources, but availability varies by employer and school. Public sector professionals should ask about federal financial aid eligibility, employer tuition assistance, veterans benefits, military education benefits, graduate assistantships, scholarships, fellowships, agency workforce development funds, and contractor professional development support.
If your goal is to reduce education cost before doctoral study, comparing a cheapest online computer science degree pathway can help you decide whether you need additional technical preparation at the bachelor's or master's level before applying to a doctorate.
The table below shows how to think about doctorate ROI for public sector careers. It does not predict individual earnings, but it identifies the factors most likely to influence whether the degree is worth the investment:
| ROI factor | Why it matters for government careers | What to ask before enrolling |
| Current role and pay band | A doctorate may have different value for a GS-13 analyst, a state IT manager, a military officer, or a contractor researcher | Will the degree qualify me for roles I cannot reasonably reach with my current credentials? |
| Employer tuition support | Agency or contractor funding can substantially change the cost-benefit calculation | What annual cap, grade requirement, service obligation, or approved-school rule applies? |
| Promotion pathway | Some technical leadership roles value doctoral expertise, while others prioritize management experience or certifications | Do job postings for my target roles list a doctorate as required, preferred, or simply acceptable? |
| Research relevance | A public sector-focused dissertation can become evidence of expertise | Can I build a portfolio around a problem my agency or target employer recognizes? |
| Time to completion | Longer timelines increase cost and opportunity cost | What percentage of online doctoral students reach candidacy and finish, and what support exists if I fall behind? |
Be especially careful with debt. Federal student loans can make doctoral study accessible, but repayment should be evaluated against realistic public sector pay ranges, not private-sector AI salary headlines. Public Service Loan Forgiveness may be relevant for qualifying employment and loans, but eligibility rules are specific and should be verified before borrowing.
An online machine learning doctorate can support advanced public sector roles where research, technical authority, and strategic judgment matter. It is most valuable when paired with experience in government operations, regulated data environments, public accountability, procurement, security, or policy implementation.
BLS reported a 2024 median annual wage of $140,910 for computer and information research scientists, but public sector compensation varies by agency, grade, locality, union rules, contractor status, and leadership level.
In the federal system, 2025 GS-15 base pay ranges from $123,041 to $159,950 before locality pay, showing why candidates should compare both technical labor-market data and government pay structures.
The table below connects potential career paths to the type of doctoral preparation that can be most useful. Use it to map the degree to real responsibilities rather than assuming the doctorate alone creates a senior role:
| Career path | Typical responsibilities | Helpful doctoral focus |
| Government AI or machine learning research scientist | Develop, test, and evaluate models for defense, energy, health, transportation, environment, or scientific agencies | Machine learning theory, experimental design, high-performance computing, reproducibility |
| Public sector chief data officer or analytics director | Lead data strategy, governance, analytics teams, data sharing, and AI adoption | Data governance, leadership, responsible AI, enterprise architecture, applied analytics |
| AI governance or model risk lead | Create review processes for fairness, explainability, privacy, procurement, and accountability | Responsible AI, policy analytics, auditing, risk management, legal and ethical frameworks |
| Cyber analytics or threat intelligence specialist | Use machine learning to detect anomalies, fraud, intrusion patterns, and adversarial behavior | Cybersecurity, adversarial machine learning, anomaly detection, secure systems |
| Public health or social services analytics leader | Analyze population health, program outcomes, service delivery, and resource allocation | Biostatistics, causal inference, privacy-preserving analytics, health informatics |
| Policy research scientist or quantitative evaluator | Measure program outcomes, evaluate interventions, and translate technical findings for decision-makers | Causal inference, econometrics, statistics, machine learning evaluation, policy communication |
| Public sector AI consultant or contractor principal | Advise agencies on AI strategy, implementation, procurement, compliance, and technical evaluation | Applied machine learning, systems design, governance, project leadership, domain specialization |
A doctorate can also help in nonprofit research organizations, federally funded research and development centers, university-affiliated research centers, public utilities, transportation authorities, hospital systems, and education agencies. These employers often need people who can connect advanced analytics with public mission outcomes.
However, a doctorate is not the only path into public sector machine learning. For many implementation roles, a master's degree, strong portfolio, cloud certification, cybersecurity credential, or direct agency experience may be sufficient. The doctorate makes the most sense when your target work requires original research, technical leadership, or high-level credibility in complex AI decisions.
Which Online Machine Learning Doctorate Programs Offer the Best Public Sector Career Support and Professional Networks?
The best public sector career support comes from programs that understand government hiring, regulated data work, and applied research networks. Generic career services may be less useful for doctoral students than faculty connections, research collaborations, agency-facing projects, and alumni in relevant technical leadership roles.
When evaluating career support, look for evidence that the program can help you translate doctoral work into public sector opportunity. The following features are especially valuable for working professionals who already have experience but need stronger networks or research credibility:
- Faculty with experience on government grants, agency advisory boards, federal lab collaborations, defense projects, public health analytics, civic technology, or nonprofit research.
- Dissertation topics connected to public sector problems, such as AI accountability, emergency management, infrastructure, cybersecurity, benefits systems, or environmental modeling.
- Access to research seminars, doctoral colloquia, publication support, conference preparation, and peer review communities.
- Alumni working in federal, state, local, military, contractor, research institute, public university, or nonprofit policy environments.
- Career staff who understand federal resumes, USAJOBS language, government contractor roles, clearance-sensitive hiring, and public sector leadership competencies.
- Partnerships or project opportunities with agencies, public hospitals, transportation authorities, utilities, school systems, or federally funded research centers.
- Support for translating technical work into executive communication, policy memos, procurement evaluation, and stakeholder presentations.
Before enrolling, ask admissions teams direct questions about doctoral outcomes. Good questions include whether graduates publish, move into senior technical roles, receive agency promotions, join research labs, or lead AI initiatives. If the school cannot provide specific examples, ask to speak with faculty or alumni who work in government-related machine learning roles.
Professional networks outside the university also matter. Students should consider joining public-interest technology groups, data science associations, cybersecurity organizations, statistics societies, AI governance communities, and domain-specific groups in health, transportation, environment, or defense. The strongest career outcomes often come from combining doctoral research with visible professional contribution.
A final red flag is a program that treats public sector employment as an afterthought. If your goal is government impact, prioritize schools that can discuss data ethics, public accountability, procurement, security, accessibility, and policy translation with the same seriousness as algorithms.
Other Things You Should Know About Machine Learning
No. Most doctoral programs do not require a clearance. However, some government, defense, intelligence, or contractor roles may require one, and clearance rules can affect whether you can use certain work-related data in a dissertation.
Sometimes, but only with approval from your employer, university, and any required research review process. Many students use de-identified, public, synthetic, or simulated datasets when sensitive agency data cannot be shared or published.
They can, especially for research labs, policy research organizations, and academic-adjacent roles. For management or applied leadership roles, a strong dissertation, project portfolio, and record of implementation may matter more than journal publications alone.
Possibly. Certifications in cloud platforms, cybersecurity, project management, privacy, or data governance can complement a doctorate by proving current implementation skills, especially for public sector roles that combine research expertise with operational responsibility.
- Ph.D. in Artificial Intelligence https://www.ucumberlands.edu/academics/graduate/phd-artificial-intelligence
- Thoughts on Academia and Industry in Machine Learning Research • David Stutz https://davidstutz.de/thoughts-on-academia-and-industry-in-machine-learning-research/
- Partner with other government organizations and reach more people https://granicus.com/blog/partner-with-other-government-organizations-to-reach-more-people/
- Public-Private Partnerships https://www.acus.gov/document/public-private-partnerships
- Coordinated Research Programs — Atlas of Innovation https://atlasofinnovation.org/library/coordinated-research-programs
- Collaboration Across Borders: State, Federal, and Industry Partnerships for Mission Success | Leadership Connect https://leadershipconnect.io/state-local-government/collaboration-across-borders-state-federal-and-industry-partnerships-for-mission-success/
- Call to Serve https://ourpublicservice.org/build-better-government/talent/government-hiring/call-to-serve
- PhD in Technology - Artificial Intelligence and Machine Learning Specialization https://walshcollege.edu/programs/phd-technology-artificial-intelligence-and-machine-learning/
- Machine Learning PhD Applications — Everything You Need to Know — Tim Dettmers https://timdettmers.com/2018/11/26/phd-applications/
