Author: Taylor Warfield, Former Bain Manager and interviewer

Last Updated: October 7, 2026

 

The Mathematica Policy Research interview is a multi-round process built around a recruiter phone screen, a technical interview on research and evaluation design, a timed writing exercise, and a half-day panel of back-to-back 30-minute conversations. This guide walks through every round, the questions candidates actually get asked, and the base salary ranges Mathematica publishes in its own job postings.

 

Before reading on:

 

Most candidates struggle to land interviews and even fewer turn them into offers. Watch my free training to learn how to triple your chances of landing interviews and increase your chances of receiving an offer by 8x.

 

👉 Watch for free

 

Key Takeaways

 

Mathematica hires for research judgment and writing quality rather than case interview performance, so your preparation should center on explaining how you would evaluate a program and on producing defensible analysis under time pressure.

 

• Expect three to five rounds spread over roughly seven to eight weeks, based on Glassdoor's reported average of 52.58 days across 301 interviews

 

• The technical round is a research design conversation, not a market sizing or profitability case

 

• Almost every analyst and researcher role includes a writing exercise or a writing sample submitted with the application

 

• Posted base salaries run about $70,000 to $95,000 for research analysts and $100,000 to $135,000 for researchers

 

• Mathematica has paused both its summer internship and summer fellowship programs, so full-time postings are the main way in

 

What Is Mathematica Policy Research?

 

Mathematica Policy Research, now officially named Mathematica, is a 100 percent employee-owned research firm that evaluates government and philanthropic programs. Founded in 1968, it employed roughly 1,900 researchers, analysts, and technologists across nine United States offices at the end of 2024. Its work spans health, education, nutrition, employment, disability, and international development.

 

The name confusion trips up a surprising number of candidates. Mathematica the research firm has nothing to do with Wolfram's Mathematica software, and generic interview guides frequently blend the two into one useless article. If a prep resource mentions symbolic computation or notebooks, you are reading about the wrong company.

 

The firm also has a genuine claim to historical importance. Mathematica ran the New Jersey Income Maintenance Experiment starting in 1968, the first large-scale randomized social policy experiment in the United States, which tested a negative income tax. Interviewers reference this lineage, and knowing it signals you did more than skim the homepage.

 

Attribute	Detail
Founded	1968, spun off from a Princeton-affiliated consultancy
Ownership	100 percent employee owned since 1986, with an ESOP added in 2005
Headcount	About 1,900 staff at the end of 2024
Offices	Nine US locations, including Princeton, Washington DC, Cambridge, Chicago, Ann Arbor, Oakland, Seattle, Tucson, and Woodlawn
Clients	Federal agencies, state and local governments, foundations, universities, and international organizations
Renamed	Dropped "Policy Research" from its legal name in 2019

 

Sources: Mathematica about page and public company records, as of 2026.

 

What Is the Mathematica Policy Research Interview Process?

 

The Mathematica interview process runs three to five rounds and takes about seven to eight weeks on average. It opens with a 30-minute phone screen, moves into a technical interview on research and evaluation design, adds a writing exercise for most analyst and researcher roles, and closes with a half-day panel. Glassdoor puts the average at 52.58 days across 301 reported interviews.

 

1. Application: submit a resume and a cover letter, plus any extras the posting names such as transcripts, references, or a writing sample
   
   
2. Recruiter or hiring manager screen: a 30-minute call covering your background, your interest in the specific policy area, and the realities of the role
   
   
3. Technical interview: usually two researchers who present a program scenario and ask how you would evaluate it
   
   
4. Writing exercise: a timed or short take-home task where you summarize findings for a non-technical audience
   
   
5. Panel round: four to ten staff members in back-to-back 30-minute blocks, often two interviewers at a time, each pair probing a different area

 

The panel is the round candidates underestimate. One reported research analyst loop ran four hours with roughly ten staff members, ranging from senior researchers down to analysts at your own level. Each pair is typically assigned a theme such as technical skills, collaboration, or problem solving, which means the same story can land three times without ever feeling repetitive to them.

 

Offers are contingent on a background check. Certain federal projects also require you to obtain a security clearance as a condition of working on that contract, and Mathematica states in its postings that a denied clearance can end your employment. Raise any concerns about clearance eligibility early rather than at the offer stage.

 

Role	Typical rounds	Technical component	Writing component
Research Assistant	2 to 3	Walk-through of a submitted code sample	Occasional
Research Analyst	4	Evaluation design scenarios with two researchers	Timed exercise
Researcher (doctorate)	3 to 5	Methods discussion and dissertation defense of your own work	Writing sample at application
Data Analytics or Engineering	3 to 4	SQL, Python, and data quality problems	Rare
Survey or Field Interviewer	1 to 2	None, with logistics and equipment screening instead	None

 

Does Mathematica Ask Case Interview Questions?

 

No. Mathematica does not run McKinsey-style case interviews with market sizing, profitability trees, or an exhibit you have 60 seconds to interpret. The technical round instead asks how you would design or evaluate a study, which rewards the same structured thinking but a completely different body of knowledge.

 

This distinction matters because so many candidates arrive over-prepared for the wrong test. I have coached hundreds of people one-on-one, and the ones who struggle at policy research firms are usually the ones who try to force a profitability framework onto a question about whether a job training program worked. Standard case interview frameworks do not map onto causal inference.

 

What does transfer is the habit of structuring an answer before speaking, stating assumptions out loud, and driving toward a recommendation. If you have prepared for a public sector case interview, you already have the communication muscle. You just need to swap the content underneath it.

 

Dimension	Consulting case interview	Mathematica technical interview
Core question	Should the client do this?	Did this program actually cause the outcome?
Math tested	Mental arithmetic, breakevens, growth rates	Regression logic, sampling, statistical power
Time pressure	High, with a 30-minute case	Low, closer to a conversation among colleagues
What impresses	Speed, structure, a crisp recommendation	Naming threats to validity before you are asked
Wrong answer risk	Arithmetic errors	Overclaiming causality from correlational data

 

Plenty of candidates apply to Mathematica alongside strategy firms, and the two processes reward different preparation. If you are running both tracks at once, my case interview course covers the consulting side in as little as 7 days so you can spend your remaining time on research methods.

 

Mission-driven candidates often shortlist Mathematica next to foundations and evaluation shops, which is why practicing a social impact case interview is still worthwhile. The framing overlaps even when the technical content does not.

 

What Technical Questions Does Mathematica Ask?

 

Mathematica technical interviews present a real program and ask you to design an evaluation for it. Two researchers usually run the round together, and they care far more about how you reason through study design than whether you land on their preferred method. Expect scenario questions rather than textbook definitions.

 

Common question types include the following:

 

• A state launches a new program and wants to know if it worked. How would you evaluate it

 

• You cannot randomize because the program is already running statewide. What would you do instead

 

• Walk us through a quantitative and a qualitative project you have run, and explain why you chose each method

 

• Your outcome data has 30 percent missing values. How does that change your analysis

 

• Explain your dissertation findings to someone with no statistics background

 

Here is an example of how a strong answer sounds:

 

Interviewer: A state wants to know whether its new job training program increased employment. How would you evaluate it?

 

You: I'd start by pinning down the question: is it average effect on employment, on earnings, or on job retention? Those need different outcome measures and different follow-up windows.

 

You: If the program is oversubscribed, a lottery gives me a randomized design, which is the cleanest option. If it launched statewide at once, I'd look for a comparison group using neighboring states or a difference-in-differences setup on administrative wage records.

 

You: The biggest threat here is selection: the people who sign up are already more motivated to find work. I'd want baseline earnings histories to test whether my comparison group tracks the treatment group before the program started.

 

Notice what that answer does. It clarifies the objective first, offers more than one design and explains the tradeoff, then names the threat to validity without being prompted. That last move is what separates offers from rejections, and it is the same instinct rewarded in consulting technical interview questions.

 

On the programming side, Mathematica postings ask for familiarity with at least one tool from a list of R, Python, Stata, SAS, and SQL, with R or Python listed as preferred. Depth beats breadth every time. Analytics-track candidates should also expect hands-on data problems closer to a data analyst case interview, including SQL joins and data quality checks.

 

What Is the Mathematica Writing Exercise?

 

The writing exercise asks you to turn research findings into something a policymaker can act on, usually as a short memo or issue brief. Some roles run it as a timed exercise during the loop. Others, including researcher postings, require a writing sample uploaded with your original application alongside your resume, cover letter, and salary expectations.

 

Mathematica sells clarity to government clients, so this round is not a formality. Job postings explicitly ask for the ability to summarize complex information and translate analytic findings for non-technical audiences. A brilliant analyst who writes like a journal referee will lose to a competent analyst who writes like a briefing.

 

Use a simple structure that works under time pressure:

 

1. Lead with the finding: state the answer in the first two sentences, not the third paragraph
   
   
2. Quantify it: give the effect size in units a non-specialist understands, such as percentage points or dollars per participant
   
   
3. Name the caveat: one honest limitation, stated plainly, buys enormous credibility
   
   
4. Close with the implication: tell the reader what decision this finding should inform

 

If you are submitting a sample rather than writing on the spot, pick something short and policy-facing over your longest academic paper. A five-page memo you wrote for a real audience beats a 40-page thesis chapter. The same audience-first instinct drives a strong case interview presentation.

 

What Behavioral Questions Does Mathematica Ask?

 

Mathematica behavioral questions center on collaboration, communication with non-experts, and genuine interest in the policy area you applied to. Candidates consistently describe the tone as conversational rather than adversarial, with 70.7 percent rating their interview experience positively on Glassdoor. That friendliness is real, and it is also a trap if you mistake it for a low bar.

 

Questions reported by candidates include the following:

 

• Tell us about the quantitative and qualitative projects you have worked on

 

• Describe a time you failed. What did you learn from it

 

• Why Mathematica, and why this policy area specifically

 

• Tell me about a time you explained technical findings to someone without a research background

 

• How do you handle competing deadlines across multiple projects

 

• Walk me through your greatest research accomplishment

 

Structure each answer with the STAR method and keep the result quantified. Because the panel format repeats similar ground with different interviewers, prepare six distinct stories rather than three. Running out of material by hour three is a common and avoidable failure.

 

The "why Mathematica" answer deserves real work. Naming a specific study the firm ran, explaining what you found interesting about its design, and connecting it to your own background beats any generic statement about mission. My fit interview course covers 98 percent of these question types in a few hours.

 

How Hard Is It to Get Hired at Mathematica?

 

Mathematica interviews rate 2.85 out of 5 for difficulty on Glassdoor, which is moderate by consulting standards. The questions are fair and the interviewers are helpful. The hard part right now is the funnel, not the loop.

 

Federal contract cancellations reshaped this entire sector in 2025. The Department of Education cut close to $900 million in Institute of Education Sciences contracts that February, and Inside Higher Ed reported that the cuts fueled layoffs at several major research firms, Mathematica among them. Employee reviews describe multiple layoff rounds across that year.

 

Two entry points that used to exist are currently closed. Mathematica's own careers site states that both the summer internship program and the summer fellowship program are paused and not accepting applications. That removes the standard pipeline for undergraduates and doctoral students alike.

 

Here is the honest read. Fewer roles are open, more experienced applicants are competing for them, and hiring often depends on whether a specific contract gets awarded. Postings sometimes say outright that an offer is contingent on winning the work.

 

None of that means you should skip applying. It does mean you should widen your search to include economic consulting firms and state-facing analytics groups, and it means a referral matters more than it did two years ago. Having helped tens of thousands of people land offers, I would rather see you run four parallel tracks than wait three months on one.

 

How Much Does Mathematica Pay?

 

Mathematica publishes anticipated base salary ranges directly in its job postings, which makes compensation unusually easy to verify. Research analyst roles posted in 2026 listed bases from $70,000 to $95,000, while a researcher role listed $100,000 to $135,000 with eligibility for a discretionary bonus. Employee stock ownership plan holdings sit on top of base pay.

 

Role	Posted base salary	Source and year
Quantitative Research Analyst, Education and Employment	$70,000 to $90,000	Mathematica job posting, 2026
Research Analyst, Disability Programs	$70,000 to $95,000	Mathematica job posting, 2026
Researcher, Global Division	$100,000 to $135,000	Mathematica job posting, 2025
Company-wide range	About $73,000 to $163,000	Glassdoor estimates from 1,704 submissions, October 2025

 

Compare that to an entry-level management consultant salary at a strategy firm and Mathematica pays less in cash. It pays more in schedule predictability, methodological depth, and ownership stake, since the firm is 100 percent employee owned and distributes profit-sharing to staff.

 

Note that researcher postings ask you to submit salary expectations with your application. Give a range anchored to the posted band rather than a single number, and do it before you have any bargaining power. The usual advice on how to negotiate a consulting offer applies with one adjustment: published bands at a nonprofit-facing firm flex less than they do at a private strategy firm.

 

What Are the Best Tips to Pass the Mathematica Policy Research Interview?

 

The candidates who get offers share one trait: they show up sounding like they already work there. These seven tips get you there faster.

 

Tip #1: Read three of their actual studies before the first call

 

Mathematica publishes its evidence library, case studies, and blog posts publicly. Pick three projects in the division you applied to and learn the research question, the design, and the headline finding.

 

One candidate report put it bluntly: knowing their studies in detail impresses interviewers. This is the cheapest differentiation available to you.

 

Tip #2: Prepare a two-minute answer for every line on your resume

 

The first round is almost entirely a walk through your background. Interviewers pick items off your resume and ask you to explain what you did, why you chose that approach, and what you would change.

 

If a project is on the page, it is fair game. Strip anything you cannot defend, and make sure the consulting resume you submit leads with methods and quantified results.

 

Tip #3: Practice explaining a regression to a non-specialist

 

Every job posting names this skill and every panel tests it. Take your most technical project and explain it out loud in 90 seconds without using the words coefficient, significance, or model.

 

Record yourself. The gap between what you think you said and what a policymaker would hear is usually large.

 

Tip #4: Lead with the evaluation question, not a framework

 

In the technical round, spend your first 30 seconds clarifying what outcome the client actually cares about. Interviewers reward that instinct heavily.

 

Jumping straight to a method signals you memorized techniques instead of learning to think. This is the opposite of how brainstorming in case interviews works, where breadth earns points quickly.

 

Tip #5: Treat the writing exercise like a client deliverable

 

Budget the last five minutes for editing rather than writing. Cut every sentence that does not help a busy reader make a decision.

 

Short paragraphs, a clear headline finding, and one honest limitation will outperform dense prose almost every time.

 

Tip #6: Bring different questions for each panel interviewer

 

You may meet ten people in one day. Asking each of them the same question about culture wastes your best signal and theirs.

 

Map your questions to ask at the end of an interview to each person's seniority, so senior researchers get design questions and analysts get day-to-day ones.

 

Tip #7: Name your policy area and commit to it

 

Mathematica hires into specific divisions such as health, education, disability, nutrition, and international development. Candidates who say they are open to anything read as unfocused.

 

Pick the area you applied to, explain what pulled you toward it, and let your consulting cover letter tell the same story your interview answers do.

 

The Mathematica Policy Research interview rewards preparation that looks nothing like standard consulting prep, so build your study plan around evaluation design, plain-language writing, and three of their published studies rather than case drills. Start with the writing, because it is the round most candidates ignore and the one that separates finalists. Block two hours this week to draft a one-page policy memo from a study you already know well.

 

Frequently Asked Questions

 

How long does the Mathematica Policy Research interview process take?

 

Glassdoor reports an average of 52.58 days from application to decision across 301 submitted Mathematica interviews, which works out to roughly seven and a half weeks. Candidates frequently report longer timelines of three to four months, especially when a role depends on a contract that has not been awarded yet. Build your job search around the assumption that Mathematica will be slow, and keep other applications moving in parallel.

 

Do you need a PhD to work at Mathematica Policy Research?

 

No. Research analyst postings list a bachelor's degree in public policy, economics, sociology, psychology, statistics, data science, or a related field as the baseline requirement. A doctorate is expected for researcher and senior researcher titles, where you own study design and publish under your own name. Master's degree holders typically enter as analysts or associates and move up from there.

 

Why is Mathematica Policy Research now called Mathematica?

 

The firm dropped Policy Research from its legal name in 2019 and now operates as Mathematica, Inc. The change reflected its expansion beyond traditional policy evaluation into data science, software engineering, and analytics products. Job postings, the careers portal, and the interviewers themselves all use the shorter name, so match their language once you are in the room.

 

How much does a Mathematica research analyst make?

 

Mathematica publishes salary ranges directly in its job postings. A 2026 posting for a Quantitative Research Analyst on the Education and Employment team listed an anticipated base of $70,000 to $90,000, and a Research Analyst posting on the Disability Programs team listed $70,000 to $95,000. Employee stock ownership plan holdings sit on top of base pay, since Mathematica is 100 percent employee owned.

 

Is it hard to get a job at Mathematica Policy Research?

 

Glassdoor candidates rate Mathematica interview difficulty at 2.85 out of 5, which is moderate compared with strategy consulting firms. The harder part is the funnel rather than the questions, because federal contract cancellations in 2025 triggered multiple rounds of layoffs across policy research firms and Mathematica has paused both its summer internship and summer fellowship programs. Fewer entry points means more competition for each full-time opening.

 

What programming languages should you know for a Mathematica interview?

 

Mathematica job postings ask for familiarity with at least one statistical programming language from a list that includes R, Python, Stata, SAS, and SQL, with R or Python listed as preferred. Depth in one language beats shallow exposure to four. Bring a specific project you can walk through line by line, since technical interviewers often ask you to explain code you actually wrote.

 

Everything You Need to Land a Consulting Offer

 

Need help passing your interviews?

• Case Interview Course: Become a top 10% case interview candidate in 7 days while saving yourself 100+ hours
  
  
• Fit Interview Course: Master 98% of consulting fit interview questions in a few hours
  
  
• Interview Coaching: Accelerate your prep with 1-on-1 coaching with Taylor Warfield, former Bain interviewer and best-selling author

 

Need help landing interviews?

• Resume Review & Editing: Craft the perfect resume with unlimited revisions and 24-hour turnaround

 

Need help with everything?

• Consulting Offer Program: Go from zero to offer-ready with a complete system

 

Not sure where to start?

• Free 40-Minute Training: Triple your chances of landing consulting interviews and 8x your chances of passing them