Articulation Matrix: The 8-Week Camp and the ASSIP 2026 Calendar

How the two fit together

The Aspiring Scientists Summer Internship Program (ASSIP) at George Mason University’s College of Science is a full-time, eight-week REU-style apprenticeship that runs Monday through Friday, 9:00 am to 5:00 pm, from Monday, June 22, 2026 through the ASSIP Research Symposium on Wednesday, August 12, 2026 (orientation Thursday, June 18; holidays observed Friday, June 19 for Juneteenth and Friday, July 3 for Independence Day). It is directed by Dr. Amanda Haymond Still (cosassip@gmu.edu), accepts high-school students aged 15 and older (16+ for in-person wet-lab placements) as well as undergraduates not yet graduated, sets no GPA minimum, allows a maximum of two summer sessions, costs $25 to apply plus $1,299 tuition with full waivers for Pell-eligible and reduced-lunch students, and awards three GMU College of Science credits on completion. Prof. Gao’s 2026 placement is officially remote-only, which makes the operational discipline of this camp — version-controlled code, written pre-analysis plans, asynchronous code review — load-bearing rather than optional. The mentor model assigns one faculty mentor per group; with a cohort sitting at the high end of typical ASSIP group sizes (the program ranges from 2 to 15 students per group), the camp is staffed with layered mentorship — Prof. Gao as faculty mentor, PhD students and research assistants as sub-leaders — so that every intern gets contact time, feedback on writing, and a careful read of their identifying assumption.

The two pieces of the experience are deliberately complementary. ASSIP frames: it supplies the real research project, the institutional credibility of a Mason College of Science program, the layered mentorship, the symposium deliverable, the JSSR publication track, the scientific-writing forum, the career programming, and the transcript line. The camp deepens: it supplies the econometric rigor — probability and inference, OLS and its standard-error zoo, the causal-inference toolkit, the identification reading discipline, and the AI co-pilot module — that turns a summer placement into a defensible paper rather than a slide deck. A student doing ASSIP alone gets an apprenticeship; a student doing the camp inside ASSIP gets an apprenticeship plus the methods to answer their own research question honestly.

The ASSIP 2026 calendar

Week	ASSIP Dates	Camp Focus	Mentor Activity	Cohort Deliverable
1	Jun 22 – Jun 26 (orientation Thu Jun 18; Jun 19 Juneteenth observed off during orientation week)	Foundations — probability, sampling distributions, the Central Limit Theorem, hypothesis testing and what a p-value actually is; loading real public data (yfinance, FRED, SEC EDGAR) reproducibly	Orientation, lab onboarding, GitHub + environment setup, one-on-one introductions, research-interest intake form, layered-mentorship pairings (faculty mentor + PhD/RA sub-leader)	Project-idea brainstorm: each intern submits a one-page brainstorm of 2–3 candidate research questions tied to a public dataset
2	Jun 29 – Jul 3 (Fri Jul 3 Independence Day observance — 4 working days)	The OLS engine — regression in matrix form, Gauss–Markov, Frisch–Waugh–Lovell, the standard-error zoo (classical, HC1–HC3, HAC/Newey–West, clustered), omitted-variable bias diagnostics	Guided literature scan against each intern’s candidate question; mentor curates a 5–8 paper reading list per group; intake of data-access constraints (CRSP/Compustat tickets if needed)	Data-source choice: each intern locks in a primary dataset and writes a 1-page data memo (source, vintage, unit of observation, sample window, known caveats)
3	Jul 6 – Jul 10	Causal Inference I — potential outcomes and the Rubin causal model, selection bias, propensity-score matching, instrumental variables, the exclusion restriction	Pre-analysis plan (PAP) kickoff: mentor walks each intern through the PAP template — research question, hypotheses, sample, treatment, controls, fixed effects, clustering, primary spec, robustness plan	PAP draft (v0): committed to the intern’s repo under `/pap/`
4	Jul 13 – Jul 17	Causal Inference II — difference-in-differences (2×2 and staggered), regression discontinuity (sharp and fuzzy), synthetic control, event studies; the fair-lending case after Sun & Gao (2019, PNAS)	Data acquisition and cleaning: mentor pair-codes the build pipeline (raw → clean → analysis), reviews variable-construction choices, signs off on the sample-construction filter cascade	Analysis dataset: a frozen, documented, version-controlled analysis-ready dataset with a data dictionary
5	Jul 20 – Jul 24	Reading the Frontier I — dissecting published empirical papers for data, design, identifying assumption, and threats; guided reads of Fama–French (1992, 1993), Jegadeesh–Titman (1993), Petersen (2009), Bertrand–Duflo–Mullainathan (2004)	Identification memo + threats table: mentor red-teams each intern’s PAP, lists every threat to identification, and the intern writes a memo addressing each in writing	Filed PAP (tagged git commit, e.g. `pap-v1.0`) + first-look frozen regressions: the primary specification run on the locked dataset, results table committed
6	Jul 27 – Jul 31	Reading the Frontier II + AI co-pilot — text-as-data, ML-assisted empirics, responsible use of LLMs in research, replication and robustness discipline; mentor anchor on Gao, Han, Kim & Pan (2024, JCF, 84:102520)	Project-specific methods deep-dive: if the project uses 10-K / 8-K text, mentor walks the intern through a text classifier; if it uses panel ML, mentor walks through cross-validated prediction with honest test-set discipline	Robustness draft (alternative specifications, subsamples, placebo tests) + working text classifier or ML pipeline (if the project calls for it)
7	Aug 3 – Aug 7	Research Project I — writing the paper: structure (intro, data, design, results, robustness, conclusion), honest empirical writing, the replication package, LaTeX in Overleaf	Paper drafting + peer review: mentor reads each draft in full, line-edits the intro and identification section, runs a structured peer-review round between interns	Full paper draft (12–20 pages, LaTeX, with figures and tables compiled from code) + first poster draft (36×48 in. Mason template)
8	Aug 10 – Aug 12 (3-day symposium week)	Research Project II — defense + symposium — polishing the poster, rehearsing the 8-minute talk, finalizing the JSSR abstract, replication-package sanity check	Poster polish (typography, figure legibility, accessibility), 8-minute talk dry-run with hostile-question simulation, abstract line-edit for JSSR submission	Wed Aug 12 ASSIP Research Symposium poster (in-person poster session with virtual breakouts for remote interns; audience: Mason faculty, mentors, families, College of Science leadership, Mason News press) + JSSR abstract submission to the Journal of Student-Scientists’ Research (GMU Mason Publishing)

What ASSIP brings that the camp content does not

The camp is a textbook with notebooks and problem sets. ASSIP is an institution. The five things ASSIP supplies that no book can are:

Layered mentorship. One faculty mentor per group, with PhD students and research assistants as sub-leaders, gives every intern multiple humans reading their work — a faculty perspective on the research question, a graduate-student perspective on the methods, an RA perspective on the day-to-day code. The camp’s content is consumed alone or in pairs; ASSIP’s mentorship is consumed in conversation.
The JSSR publication track. The Journal of Student-Scientists’ Research, hosted by GMU Mason Publishing, is a real venue with a real archival record. Every ASSIP intern is expected to submit at minimum an abstract; the optional full paper goes through a student-scientist peer-review process. Nothing in a textbook gets archived in Mason Publishing.
The 3-credit transcript line. Three GMU College of Science credits show up on an official transcript — they articulate to other universities, they count toward dual-enrollment GPAs, and they signal that an outside institution has vouched for the work.
The symposium. Wednesday, August 12, 2026, an in-person poster session at GMU (with virtual breakouts for remote interns including Prof. Gao’s 2026 group), where Mason faculty, mentors, families, College of Science leadership, and Mason News press read your work and ask hard questions. The textbook ends in a problem set; ASSIP ends in a defense.
Institutional credibility. “I worked with a faculty mentor at George Mason University’s College of Science in the Aspiring Scientists Summer Internship Program” is a sentence a 16-year-old can put on a college application, and it lands differently than “I read a book.” That credibility is what the $25 application + $1,299 tuition (or, for Pell-eligible / reduced-lunch students, the full waiver) actually buys.

What the camp content brings that an ASSIP placement alone might not

ASSIP is full-time and intense, but the methodological content varies by mentor and lab. A finance-and-econometrics placement without a structured curriculum can end up teaching one project’s methods rather than a field’s methods. The five things the camp supplies that a generic placement might miss are:

A structured 8-week curriculum of econometric methods, sequenced from probability and OLS through causal inference and frontier reading, deliberately aligned to the ASSIP calendar above.
Approximately 40 reproducible Jupyter notebooks — one per major concept, every dataset publicly accessible, every result reproducible from a fresh clone of the repository.
Approximately 40 problem sets with answer keys, sequenced so that the Friday problem set previews the next Monday’s lecture rather than recapping the last one.
Five worked capstone exemplars — five end-to-end research projects with data, code, paper, and poster, so that an intern doesn’t enter ASSIP guessing at what “done” looks like.
The AI co-pilot module — a deliberate, ethics-first treatment of how to use Claude, ChatGPT, and Copilot as research assistants: where they help (boilerplate, syntax, regex, prose polishing), where they hurt (made-up citations, plausible-but-wrong identification arguments), and how to keep a paper trail of what the model did versus what you did.

The handshake

The book is the textbook. The wiki (see /wiki/) is the operational portal — the place that holds the ASSIP-specific logistics, the rolling weekly agenda, the office-hours schedule, the Slack norms, the data-access tickets, the symposium templates, and the JSSR submission walkthrough. The GitHub repository holds both — the book sources and the wiki content — and it also holds each intern’s pre-analysis plan as a tagged commit (pap-v1.0 filed at the end of Week 5), so that “what you said you would do” is provably timestamped against “what you ended up doing.” When the August 12 symposium poster goes up, the reviewer who wants to see whether the result was pre-specified or fished can pull up the Week-5 commit and verify it in thirty seconds. That handshake — textbook in the book, logistics in the wiki, evidence in the repo — is what lets a high-school cohort produce work that an econometrician will read seriously.

# Articulation Matrix: The 8-Week Camp and the ASSIP 2026 Calendar ## How the two fit together The **Aspiring Scientists Summer Internship Program (ASSIP)** at George Mason University's College of Science is a full-time, eight-week REU-style apprenticeship that runs Monday through Friday, 9:00 am to 5:00 pm, from Monday, June 22, 2026 through the **ASSIP Research Symposium on Wednesday, August 12, 2026** (orientation Thursday, June 18; holidays observed Friday, June 19 for Juneteenth and Friday, July 3 for Independence Day). It is directed by Dr. Amanda Haymond Still (cosassip@gmu.edu), accepts high-school students aged 15 and older (16+ for in-person wet-lab placements) as well as undergraduates not yet graduated, sets no GPA minimum, allows a maximum of two summer sessions, costs \$25 to apply plus \$1,299 tuition with full waivers for Pell-eligible and reduced-lunch students, and awards three GMU College of Science credits on completion. Prof. Gao's 2026 placement is officially **remote-only**, which makes the operational discipline of this camp — version-controlled code, written pre-analysis plans, asynchronous code review — load-bearing rather than optional. The mentor model assigns one faculty mentor per group; with a cohort sitting at the high end of typical ASSIP group sizes (the program ranges from 2 to 15 students per group), the camp is staffed with layered mentorship — Prof. Gao as faculty mentor, PhD students and research assistants as sub-leaders — so that every intern gets contact time, feedback on writing, and a careful read of their identifying assumption. The two pieces of the experience are deliberately complementary. **ASSIP frames:** it supplies the real research project, the institutional credibility of a Mason College of Science program, the layered mentorship, the symposium deliverable, the JSSR publication track, the scientific-writing forum, the career programming, and the transcript line. **The camp deepens:** it supplies the econometric rigor — probability and inference, OLS and its standard-error zoo, the causal-inference toolkit, the identification reading discipline, and the AI co-pilot module — that turns a summer placement into a defensible paper rather than a slide deck. A student doing ASSIP alone gets an apprenticeship; a student doing the camp inside ASSIP gets an apprenticeship plus the methods to answer their own research question honestly. ## The ASSIP 2026 calendar | Week | ASSIP Dates | Camp Focus | Mentor Activity | Cohort Deliverable | |------|-------------|------------|-----------------|--------------------| | **1** | Jun 22 – Jun 26 (orientation Thu Jun 18; Jun 19 Juneteenth observed off during orientation week) | **Foundations** — probability, sampling distributions, the Central Limit Theorem, hypothesis testing and what a *p*-value actually is; loading real public data (yfinance, FRED, SEC EDGAR) reproducibly | Orientation, lab onboarding, GitHub + environment setup, one-on-one introductions, research-interest intake form, layered-mentorship pairings (faculty mentor + PhD/RA sub-leader) | **Project-idea brainstorm**: each intern submits a one-page brainstorm of 2–3 candidate research questions tied to a public dataset | | **2** | Jun 29 – Jul 3 (Fri Jul 3 Independence Day observance — 4 working days) | **The OLS engine** — regression in matrix form, Gauss–Markov, Frisch–Waugh–Lovell, the standard-error zoo (classical, HC1–HC3, HAC/Newey–West, clustered), omitted-variable bias diagnostics | Guided literature scan against each intern's candidate question; mentor curates a 5–8 paper reading list per group; intake of data-access constraints (CRSP/Compustat tickets if needed) | **Data-source choice**: each intern locks in a primary dataset and writes a 1-page data memo (source, vintage, unit of observation, sample window, known caveats) | | **3** | Jul 6 – Jul 10 | **Causal Inference I** — potential outcomes and the Rubin causal model, selection bias, propensity-score matching, instrumental variables, the exclusion restriction | Pre-analysis plan (PAP) kickoff: mentor walks each intern through the PAP template — research question, hypotheses, sample, treatment, controls, fixed effects, clustering, primary spec, robustness plan | **PAP draft** (v0): committed to the intern's repo under `/pap/` | | **4** | Jul 13 – Jul 17 | **Causal Inference II** — difference-in-differences (2×2 and staggered), regression discontinuity (sharp and fuzzy), synthetic control, event studies; the fair-lending case after Sun & Gao (2019, *PNAS*) | Data acquisition and cleaning: mentor pair-codes the build pipeline (raw → clean → analysis), reviews variable-construction choices, signs off on the sample-construction filter cascade | **Analysis dataset**: a frozen, documented, version-controlled analysis-ready dataset with a data dictionary | | **5** | Jul 20 – Jul 24 | **Reading the Frontier I** — dissecting published empirical papers for data, design, identifying assumption, and threats; guided reads of Fama–French (1992, 1993), Jegadeesh–Titman (1993), Petersen (2009), Bertrand–Duflo–Mullainathan (2004) | Identification memo + threats table: mentor red-teams each intern's PAP, lists every threat to identification, and the intern writes a memo addressing each in writing | **Filed PAP** (tagged git commit, e.g. `pap-v1.0`) **+ first-look frozen regressions**: the primary specification run on the locked dataset, results table committed | | **6** | Jul 27 – Jul 31 | **Reading the Frontier II + AI co-pilot** — text-as-data, ML-assisted empirics, responsible use of LLMs in research, replication and robustness discipline; mentor anchor on Gao, Han, Kim & Pan (2024, *JCF*, 84:102520) | Project-specific methods deep-dive: if the project uses 10-K / 8-K text, mentor walks the intern through a text classifier; if it uses panel ML, mentor walks through cross-validated prediction with honest test-set discipline | **Robustness draft** (alternative specifications, subsamples, placebo tests) **+ working text classifier or ML pipeline** (if the project calls for it) | | **7** | Aug 3 – Aug 7 | **Research Project I** — writing the paper: structure (intro, data, design, results, robustness, conclusion), honest empirical writing, the replication package, LaTeX in Overleaf | Paper drafting + peer review: mentor reads each draft in full, line-edits the intro and identification section, runs a structured peer-review round between interns | **Full paper draft** (12–20 pages, LaTeX, with figures and tables compiled from code) **+ first poster draft** (36×48 in. Mason template) | | **8** | Aug 10 – Aug 12 (3-day symposium week) | **Research Project II — defense + symposium** — polishing the poster, rehearsing the 8-minute talk, finalizing the JSSR abstract, replication-package sanity check | Poster polish (typography, figure legibility, accessibility), 8-minute talk dry-run with hostile-question simulation, abstract line-edit for JSSR submission | **Wed Aug 12 ASSIP Research Symposium poster** (in-person poster session with virtual breakouts for remote interns; audience: Mason faculty, mentors, families, College of Science leadership, Mason News press) **+ JSSR abstract submission** to the *Journal of Student-Scientists' Research* (GMU Mason Publishing) | ## What ASSIP brings that the camp content does not The camp is a textbook with notebooks and problem sets. ASSIP is an institution. The five things ASSIP supplies that no book can are: 1. **Layered mentorship.** One faculty mentor per group, with PhD students and research assistants as sub-leaders, gives every intern multiple humans reading their work — a faculty perspective on the research question, a graduate-student perspective on the methods, an RA perspective on the day-to-day code. The camp's content is consumed alone or in pairs; ASSIP's mentorship is consumed in conversation. 2. **The JSSR publication track.** The *Journal of Student-Scientists' Research*, hosted by GMU Mason Publishing, is a real venue with a real archival record. Every ASSIP intern is expected to submit at minimum an abstract; the optional full paper goes through a student-scientist peer-review process. Nothing in a textbook gets archived in Mason Publishing. 3. **The 3-credit transcript line.** Three GMU College of Science credits show up on an official transcript — they articulate to other universities, they count toward dual-enrollment GPAs, and they signal that an outside institution has vouched for the work. 4. **The symposium.** Wednesday, August 12, 2026, an in-person poster session at GMU (with virtual breakouts for remote interns including Prof. Gao's 2026 group), where Mason faculty, mentors, families, College of Science leadership, and Mason News press read your work and ask hard questions. The textbook ends in a problem set; ASSIP ends in a defense. 5. **Institutional credibility.** "I worked with a faculty mentor at George Mason University's College of Science in the Aspiring Scientists Summer Internship Program" is a sentence a 16-year-old can put on a college application, and it lands differently than "I read a book." That credibility is what the \$25 application + \$1,299 tuition (or, for Pell-eligible / reduced-lunch students, the full waiver) actually buys. ## What the camp content brings that an ASSIP placement alone might not ASSIP is full-time and intense, but the methodological content varies by mentor and lab. A finance-and-econometrics placement without a structured curriculum can end up teaching one project's methods rather than a field's methods. The five things the camp supplies that a generic placement might miss are: 1. **A structured 8-week curriculum** of econometric methods, sequenced from probability and OLS through causal inference and frontier reading, deliberately aligned to the ASSIP calendar above. 2. **Approximately 40 reproducible Jupyter notebooks** — one per major concept, every dataset publicly accessible, every result reproducible from a fresh clone of the repository. 3. **Approximately 40 problem sets** with answer keys, sequenced so that the Friday problem set previews the next Monday's lecture rather than recapping the last one. 4. **Five worked capstone exemplars** — five end-to-end research projects with data, code, paper, and poster, so that an intern doesn't enter ASSIP guessing at what "done" looks like. 5. **The AI co-pilot module** — a deliberate, ethics-first treatment of how to use Claude, ChatGPT, and Copilot as research assistants: where they help (boilerplate, syntax, regex, prose polishing), where they hurt (made-up citations, plausible-but-wrong identification arguments), and how to keep a paper trail of what the model did versus what you did. ## The handshake The book is the textbook. The wiki (see `/wiki/`) is the operational portal — the place that holds the ASSIP-specific logistics, the rolling weekly agenda, the office-hours schedule, the Slack norms, the data-access tickets, the symposium templates, and the JSSR submission walkthrough. The GitHub repository holds both — the book sources and the wiki content — and it also holds each intern's pre-analysis plan as a tagged commit (`pap-v1.0` filed at the end of Week 5), so that "what you said you would do" is provably timestamped against "what you ended up doing." When the August 12 symposium poster goes up, the reviewer who wants to see whether the result was pre-specified or fished can pull up the Week-5 commit and verify it in thirty seconds. That handshake — textbook in the book, logistics in the wiki, evidence in the repo — is what lets a high-school cohort produce work that an econometrician will read seriously.