The Pre-College ROI Project | Issue 2 of 12

By Lindsey Kundel, Editor in Chief, InGenius Prep

Few pre-college categories inspire as much aspiration—or as much exploitation—as research.

The word itself carries enormous weight in admissions culture. Selective universities talk openly about valuing intellectual initiative and independent inquiry. Admissions officers note the power of a student who has genuinely wrestled with a problem. And so families have responded logically: if research matters, their student should do it.

What they often don’t realize is how wide the gap has grown between research as an intellectual experience and research as a marketed credential. A student who spends six weeks at MIT working on an original project under a faculty mentor is doing something categorically different from one who pays $4,000 for ten Zoom sessions resulting in a “publication” in a journal that accepted every paper submitted that year. Both can list “research” on the Common App. Neither is required to explain which they’ve done.

This installment applies the four-pillar ROI framework introduced in January to the research category specifically. The goal is not to dismiss research as a pre-college experience—when well-structured, it remains among the highest-ROI activities available to students. The goal is to equip families with enough specificity to tell the difference: to evaluate role clarity, intellectual ownership, and publication credibility before committing time, money, and application real estate.

The data in this issue is clear. The questions it raises are equally clear. We hope both are useful.

Executive Summary

Pre-college research is simultaneously one of the most admissions-valuable categories available to high school students and one of the most aggressively monetized. The distance between a genuinely formative research experience and a credentialed performance of one can be difficult to see from the outside—and is sometimes invisible even to the student living it.

This gap matters because admissions readers can usually tell the difference.

A 2023 joint investigation by ProPublica and The Chronicle of Higher Education documented what many counselors had observed for years: at least 20 online research programs for high schoolers had proliferated in the United States, many connected to newly founded journals that publish student work with minimal or no peer review. One Ivy League admissions officer, speaking anonymously to ProPublica, described the growth of these operations as a “fast-growing epidemic” that had “trebled or quadrupled in the past few years.”

At the same time, a small category of research opportunities remains highly selective, institutionally credible, and genuinely transformative. RSI (Research Science Institute), hosted at MIT, admits roughly 80–100 students annually from approximately 3,000 applications—an acceptance rate of 2–3%, comparable to MIT’s own undergraduate admissions. The NIH High School Summer Internship Program accepts fewer than 7% of applicants. These programs are free. They are also extraordinarily difficult to access.

Between these poles lies a large and heterogeneous middle: paid mentorship platforms, university lab placements of variable quality, independent project-based programs, and a range of publication vehicles whose credibility varies as much as their cost.

This installment applies the four-pillar ROI framework—Growth, Narrative, Credibility, and Wellbeing—to the research category and offers families the specific tools to evaluate any research opportunity before committing to it.

 

The central finding: a single question separates high-ROI research from low-ROI research.

What did the student actually do—and how much of the thinking was theirs?

Title, affiliation, and publication status are secondary. Role clarity, intellectual independence, and the capacity to speak to one’s own process are what admissions readers look for—and what families should ask for before enrolling.

Section 1 — The Landscape: What Families Are Told and What the Market Offers

1.1 Why “Research” Became the Aspirational Category of the Decade

The rise of research as a pre-college credential follows a familiar pattern. As admissions selectivity increased—Common Application data shows total submitted applications more than doubling since the early 2010s while acceptance rates at selective institutions declined sharply—families began seeking experiences that could differentiate applicants beyond strong grades and test scores.

Admissions commentary from selective institutions provided the cue. MIT, Caltech, and similar research-intensive universities have long emphasized intellectual curiosity and independent initiative in their guidance. By the mid-2010s, that language had been translated by the pre-college industry into a more actionable message: do research. Publish something. Show you can think like a scientist or scholar.

The translation was not entirely wrong. Research can, in appropriate contexts, produce exactly the kind of evidence admissions officers are looking for: sustained intellectual engagement, demonstrable growth, and output that can be credibly described and verified. But it created demand that the market was quick to supply—not always with substance.

Caltech’s published data offers a revealing data point. For the Class of 2027, 45% of admitted students submitted materials documenting their own past research—but Caltech itself notes that research is “not an admissions requirement.” That figure reflects the composition of an extraordinarily selective admitted class, not a threshold families should treat as a target. Among the full Class of 2027 applicant pool, the figure is unknown and likely substantially lower.

The implication families often draw—that research is now expected—is not supported by evidence. The implication they should draw is that well-structured, independently conducted research, when present, can add meaningful credibility. When absent, or when replaced by low-quality substitutes, it adds nothing—and may subtract.

1.2 What the Market Actually Offers: A Taxonomy

The pre-college research market is not monolithic. It spans at least four structurally distinct categories, each with different selectivity, cost, and admissions implications. Families who treat them as interchangeable make predictably costly errors.

 

Program Type	Operator	Cost to Student	Selectivity	Intellectual Ownership	Admissions Credibility
Federally / State-Funded Elite Programs (RSI, NIH SIP, SSP, PRIMES)	Government / MIT / University	Free or stipended	Extremely high (2–7%)	High	Very high
University Lab Placements	Individual PI or dept.	Low or free	Variable	Variable	Variable
Paid Mentorship Platforms (e.g., Polygence, Lumiere, Pioneer)	Private ed. companies	$2,000–$8,000+	Low to moderate	Variable	Variable (depends on output)
Pay-to-Publish Programs	Private vendors	$1,500–$5,000+	Effectively none	Low	Low or negative

 

Figure 1.1. The pre-college research landscape spans four structurally distinct categories with different costs, selectivity profiles, and admissions implications. Cost does not reliably predict credibility.

Tier 1: Highly Selective, Institutionally Sponsored Programs

These programs are the clearest high-ROI category in the research space. They are characterized by genuine intellectual contribution, real faculty or scientist mentorship, and institutional accountability. They are also scarce by design.

• RSI (Research Science Institute), hosted at MIT, receives approximately 3,000 applications annually and admits 80–100 students—an overall acceptance rate of 2–3%. It is fully funded, including housing and meals. In 2018, the domestic acceptance rate was reported at 3%. The program is frequently described as harder to gain admission to than MIT’s undergraduate program itself.
• NIH High School Summer Internship Program (HS-SIP): accepts fewer than 7% of applicants. Geographic eligibility restrictions apply; most positions require residence within 40 miles of an NIH campus. Students receive a monthly stipend of approximately $2,300 for full-time research under a Principal Investigator.
• SSP (Summer Science Program): a residential, five-week program in astrophysics or biochemistry with highly competitive admissions. Students conduct original research in cohorts, with documented alumni outcomes that include ISEF finalists and Ivy-enrolled scholars.
• PRIMES (Program for Research in Mathematics, Engineering, and Science) at MIT: year-round, mentored, highly selective mathematics research for exceptional students.

These programs share several features: their acceptance rates reflect genuine evaluation, their outputs are independently verifiable, and they are recognized by name in admissions circles at research-intensive institutions. The student who participated in RSI does not need to explain what RSI is.

Tier 2: University Lab Placements

A meaningful but highly variable category. University affiliation does not determine ROI here—what matters is the nature of the student’s actual contribution.

Some university lab placements involve genuine bench work: data collection, experimental design, analysis under close mentorship. Others involve literature review, data organization, or observation with limited independent work. The two look nearly identical on an application if the student describes them only by affiliation.

Admissions readers know this. “I worked in a lab at Johns Hopkins” communicates location, not contribution. “I spent eight weeks processing and analyzing tissue samples as part of a glioblastoma study, which led me to revise our initial hypothesis” communicates both.

Tier 3: Paid Mentorship Platforms

This is the fastest-growing and most heterogeneous category in the market. Companies connect high school students with graduate students or early-career researchers for structured, paid project work. One such company, Polygence, reports having served more than 6,000 students across 100+ countries through a network of 2,500+ mentors since its 2019 founding.

The business model is tuition-dependent and openly scalable. Programs typically run 10 sessions at $2,000–$4,000+ for a core mentorship experience, with additional fees for “showcasing” support, publication preparation, or symposium presentation.

The ROI of a paid mentorship experience depends almost entirely on what the student did, not on which platform they used.

Tier 4: Pay-to-Publish Programs

These types of journals have grown substantially as a category. Estimates suggest that between 10,000 and 15,000 such journals were operating by 2021, compared to approximately 1,000 publishers a decade earlier. A 2022 study found that nearly a quarter of researchers surveyed across 112 countries had published in, or were uncertain whether they had published in, a pay-to-publish journal.

For high school students, the dynamic is even more opaque: they are typically first-time authors, unfamiliar with legitimate publishing norms, and operating in a market specifically designed to exploit their inexperience.

The ROI of a pay-to-publish program again depends on the credibility of both the student’s original research and the credibility of the publication in which it is published.

1.3 The Definition Problem: What “Research” Actually Means

In academic contexts, research has a specific meaning: the identification of a question not yet answered, a review of existing knowledge, the design and execution of a method for producing new data or analysis, and the drawing of conclusions from evidence.

In the pre-college marketplace, “research” is applied to: literature reviews, data visualization exercises, guided projects with predetermined outcomes, shadowing in a lab setting, and AI-assisted summaries of existing scholarship. The term is used fluidly, without a shared standard, and often chosen precisely because its perceived prestige outpaces its actual definition.

The question is not whether a student ‘did research.’ It is whether they engaged in genuine inquiry—and whether that process developed their thinking.

This distinction matters because it is exactly what admissions readers are trying to assess. Role clarity—what the student actually did, with how much independence, and with what result—is the interpretive test a research listing must pass to add credibility to an application.

Section 2 — Why Families Misread Value in This Category

2.1 The Publication Heuristic

Publication, in academic culture, signals peer validation. A paper published in a credible journal has passed external review by independent experts—a process designed to separate work that meets a scholarly standard from work that does not.

Families understandably apply this logic to pre-college publications. If my student’s work was published, it must have been evaluated and found worthy.

That logic fails when the publication vehicle lacks the mechanism that makes publication meaningful: genuine peer review, transparent standards, verifiable rejection rates, and independence from the organization that produced the work.

A paper accepted within days of submission, by a journal founded the same year as the mentorship program submitting to it, whose editorial board cannot be independently verified, does not carry the same evidentiary weight as a paper published in a long-standing, COPE-registered journal with a documented review process. Admissions officers at research-intensive institutions have read enough applications to know the difference.

The 60% of articles published in predatory journals that receive zero citations in the five years following publication—a figure drawn from published research on predatory publishing patterns—reflects something applicants may not expect: the academic community largely ignores these publications because they do not contribute to scholarly knowledge. Admissions readers are part of the academic community.

2.2 The Affiliation Heuristic

University names produce cognitive relief. “My student worked at a Johns Hopkins lab this summer” carries an immediate impression of rigor and selectivity that may not reflect the nature of the student’s actual contribution.

This is not a new observation. Admissions readers have noted for years that a university’s name on a research affiliation says almost nothing about what a student actually did there. The lab name opened a door. What happened inside the door is what gets evaluated.

The question “Where did you do it?” is considerably less important than “What problem were you working on?”, “What was your specific contribution?”, and “What did you find that you didn’t expect?” If a student cannot answer the last question, the experience may have been observation rather than inquiry.

2.3 The Mentor Credential Heuristic

Paid mentorship platforms market mentor credentials prominently. A PhD from [Name Institution], publications in [Name Journal], expertise in [Specific Subfield]. These are not fabrications. The mentor’s qualifications are often real.

What families less often ask is: what did the mentor do with my student? Was the project designed for this student, or is it a template adapted for dozens of students simultaneously? How much of the thinking was the student’s? How much of the writing? Is the mentor available to speak specifically and knowledgeably about this student’s contribution in a letter of recommendation?

A mentor’s credential is not the same as a student’s growth. The former is fixed; the latter is what the experience is for.

2.4 The Cost-Seriousness Heuristic

The price-quality heuristic—the behavioral economics tendency to associate higher cost with higher value—applies directly to this category. Paid mentorship platforms charge $2,000–$8,000+ for experiences that are not, by structure, more rigorous than lower-cost alternatives. Many of the most genuinely high-ROI research experiences in this landscape are entirely free.

RSI is fully funded. NIH SIP provides a $2,300 monthly stipend. SSP covers residential costs. State-sponsored STEM programs are publicly funded. The student who gains access to any of these has received an experience whose cost does not approach what some families spend on paid mentorship—and whose credibility in admissions review substantially exceeds it.

Cost signals investment, not quality. In this category, investment and quality are often inversely correlated at the extremes.

Section 3 — Applying the Four-Pillar ROI Framework to Research

The four-pillar framework introduced in January—Growth ROI, Narrative ROI, Credibility ROI, and Wellbeing ROI—applies to research with some category-specific nuances. This section works through each pillar as it specifically applies to research experiences.

3.1 Growth ROI

Growth ROI in research measures whether the student actually developed new intellectual, analytical, or methodological capacity. The key word is developed. Exposure to an interesting lab environment, or observation of scientists at work, does not produce Growth ROI unless the student is an active participant in the inquiry process.

 

High Growth ROI looks like:  A student who moved through the full arc of inquiry—from question to method to data to interpretation—with decreasing scaffolding over time. They can explain what they tried, what didn’t work, what they changed, and what they concluded. They developed skills they can deploy in subsequent work.

 

Low Growth ROI looks like:  A student who followed a mentor’s blueprint closely and produced a predetermined output. The project’s conclusion was knowable before the research began. The student cannot explain the methodology independently or describe what surprised them.

 

A diagnostic question worth applying:

If the student’s mentor had not been available, could the student have continued working independently? And if so, for how long before stalling?

The answer to that question predicts Growth ROI with reasonable accuracy. Real intellectual development produces real independence. Guided credential-production produces dependence that collapses without the guide.

3.2 Narrative ROI

Narrative ROI reflects how well a research experience integrates into the coherent story an application tells about the student’s curiosity, development, and direction. The relevant unit of analysis is not the research project itself, but its relationship to everything else in the application.

Research that grows organically from prior coursework, extends into subsequent study, and connects to essay themes produces high Narrative ROI. Research selected because “machine learning is popular” or “genomics sounds impressive”—with no prior connection to the student’s demonstrated interests—tends to produce the opposite.

• What strengthens Narrative ROI: the research topic reflects the student’s genuine curiosity; it connects to AP/IB coursework in the same area; the student references the research in essays with specificity and authentic voice; the experience raised questions the student continued pursuing.
• What weakens it: the topic was chosen by the mentor or platform; the student cannot speak to why this specific problem interested them; the research field is unrelated to any other part of the application; the experience appears as an isolated “research block” rather than a thread in a larger story.

A practical signal: if a student can only describe their research in the third person (“I was part of a team that studied X”) rather than the first person with intellectual ownership (“I was trying to understand why X…”), the Narrative ROI of the experience is likely low.

3.3 Credibility ROI

This is the pillar where the prestige-vs-pay-to-play distinction is sharpest—and where families are most vulnerable to misreading value.

Credibility ROI evaluates whether an admissions reader can confidently believe what a research listing represents. It is not about how impressive an experience sounds. It is about whether the experience holds up under interpretive scrutiny.

 

Program Type	Credibility Signal	Credibility Risk
Tier 1 Elite (RSI, NIH SIP, SSP)	Recognized; acceptance rate confirms selectivity; institutional accountability	None at the program level—only if student cannot speak to their work
University Lab (genuine contribution)	High if student has specific role, output, and recommender who observed the work	Lower if role was peripheral or recommendation is generic
Paid Mentorship (strong output)	Variable; depends on intellectual ownership and output specificity	Weaker as volume of similar applications increases; template-recognition is growing
Pay-to-Publish	Often recognized; may raise questions about other parts of application	Negative in admissions contexts where readers are familiar with the journal

 

Figure 3.1. Credibility ROI in research depends on program structure, student role clarity, and the quality of supporting documentation—not on title or cost.

 

The negative credibility trap is worth naming explicitly. An application that lists a publication in a journal that an admissions reader has seen listed on dozens of similar applications—particularly one whose acceptance process appears non-evaluative—can raise concerns about judgment and guidance that extend beyond the research listing itself. The question in the reader’s mind becomes not “what did this student accomplish?” but “what else in this application might be similarly manufactured?”

3.4 Wellbeing ROI

Research, when aligned with genuine curiosity, is one of the most motivating and wellbeing-supportive pre-college experiences available. The student who is genuinely wrestling with a problem they chose, with a mentor who takes their thinking seriously, is in a fundamentally different psychological state than the student who is producing a paper to check a credential box.

Wellbeing ROI in research is most at risk when: the topic was not the student’s own; the cost was high enough to create parental pressure for a “good outcome”; the student is performing productivity (attending sessions, turning in drafts) without genuine intellectual engagement; or the experience is added on top of an already overextended schedule.

The best indicator of high Wellbeing ROI: is the student still thinking about the research problem after the program ends? Do they bring it up in conversation unprompted? Have they followed any threads it opened?

These questions produce information that admissions essays cannot fake.

Section 4 — When Research Programs Actually Work (and for Whom)

4.1 The Student Profile That Produces High ROI

Research ROI is not a function of program quality alone. It depends substantially on whether the student is ready to engage with genuine inquiry—and whether the experience is structured to support that engagement rather than to simulate it.

High-ROI research profiles tend to share several characteristics:

• Sustained prior interest in a specific area, evidenced by coursework, reading, or self-directed projects—not generic interest in “science” or “research” in the abstract
• Prior academic depth: AP, IB, or dual enrollment coursework in the relevant field. Students who take their first chemistry course and simultaneously pursue a chemistry research project are operating beyond their preparation and will typically produce shallow work regardless of program quality
• Comfort with ambiguity and iteration. Research involves wrong turns, failed experiments, and revised questions. Students who require constant confirmation and predetermined success paths often experience paid research as frustrating rather than formative
• Metacognitive capacity: the ability to reflect on and articulate their own learning process. This is the quality that makes research transferable to essays, interviews, and recommendations

Lower-ROI research profiles are not less intelligent or less capable. They are frequently students who are excellent at executing clear instructions, who have not yet developed the tolerance for productive confusion that original inquiry requires—and who would benefit more from exceptional coursework or project-based independent work before pursuing formal research.

4.2 Grade and Timing Considerations

Research ROI varies significantly by grade level, and families often don’t account for this when evaluating the “correct” time to pursue an experience.

• 9th and 10th grade: research readiness varies enormously. Exposure experiences and science fair projects at this stage can be valuable for building foundational skills. But claiming “research” on an application for work that involved primarily data collection or observation under heavy scaffolding is unlikely to withstand interpretive scrutiny. The better use of 9th-10th grade is building the coursework depth that makes 11th grade research genuinely possible.
• 11th grade: the highest-value year for substantive research. Students typically have sufficient coursework preparation in at least one domain, enough time remaining in their application timeline to continue pursuing threads, and the cognitive maturity to engage with real problems. A student who begins research in 11th grade and can reference it genuinely in 12th grade applications—with a specific recommender and a real output—is in a very strong position.
• 12th grade: meaningful if part of a continuing trajectory; weak if begun specifically for admissions purposes in senior fall. Applications completed before significant research has produced output are speculative, not evidential.

4.3 Access and Equity Considerations

The distribution of high-ROI research opportunities is not equitable, and families should understand both the implications of this and the pathways that exist to address it.

Tier 1 programs (RSI, NIH SIP, SSP) are largely free—which means cost is not the barrier to access. The barrier is academic preparation: years of advanced coursework, competition experience, and demonstrated excellence in a STEM field. This preparation is itself inequitably distributed. Students from well-resourced schools with strong AP and IB programs, science olympiad teams, and math leagues have structural advantages in building the preparation that Tier 1 programs require.

Paid mentorship platforms are, in one sense, the most democratized form of research access: any student with $2,000–$8,000 can enroll. But access here is not the same as ROI. A student from a rural school with limited STEM preparation who purchases a paid mentorship experience without the foundational coursework to engage with it independently will not benefit in the same way as a student with strong preparation.

For students with genuinely limited access—geographic, financial, or institutional—admissions readers are trained to evaluate opportunity contextually. A student from a rural school who designed and executed an independent science project with their biology teacher as mentor, and who can speak to their inquiry with specificity and intellectual ownership, presents a more credible research profile than a student from a well-resourced school who paid for a publication in a journal the admissions reader recognizes.

Section 5 — Red Flags and Risk Patterns

5.1 The Pay-to-Publish Pipeline

This is the most documented risk pattern in the pre-college research category and deserves direct treatment.

The pipeline works as follows: a family pays for a research mentorship program, typically $1,500–$5,000+. The mentor (typically a graduate student or postdoc) works with the student to produce a paper on a topic selected from a menu or assigned by the program. The paper is submitted to a journal that is affiliated with, owned by, or a “partner” of the mentorship program. The journal accepts the paper within weeks. The student lists the publication on their Common App.

The volume of this practice has grown to the point where admissions officers at selective institutions have developed familiarity with specific journals associated with it. A publication in a journal that an admissions reader has seen on 15 other applications in the same cycle—all from students who paid the same program—does not add credibility to an application. In some cases, it invites scrutiny of the entire file.

How to evaluate any publication venue:

• Is the journal listed in COPE (Committee on Publication Ethics) or the Directory of Open Access Journals (DOAJ)?
• Does the journal disclose an acceptance rate or rejection rate?
• Is the editorial board independently verifiable (real academics at real institutions who can be contacted)?
• Has the journal been indexed in academic databases (PubMed, Scopus, Web of Science)?
• Was the paper submitted before the research began—or did the research process produce the paper organically?

If most of these questions cannot be answered—or if the answers are negative—the publication vehicle should not be treated as an admissions asset.

5.2 The Parallel Project Problem

A structural risk specific to large paid mentorship platforms: when a mentor works with multiple students simultaneously, often on closely related topics within the same research area, the outputs can become difficult to distinguish.

Admissions offices reading large volumes of applications sometimes encounter multiple students in the same cycle describing “original” research on near-identical topics, using similar methodologies, and citing overlapping sources—all from the same program. The pattern is recognizable. The credibility of each individual application is weakened by the pattern.

This is not an argument against paid mentorship universally. It is an argument for asking a specific question before enrolling: How many other students is my mentor working with on similar projects? What differentiates my student’s inquiry from theirs?

5.3 Role Inflation

‘Research Associate.’ ‘Junior Fellow.’ ‘Lead Researcher.’ These titles appear with increasing frequency on high school applications, attached to experiences that range from genuine inquiry to supervised data entry.

Titles do not create credibility. Specificity does. An admissions reader who sees “Research Associate, Genomics Laboratory” and then reads a description that amounts to “I assisted with experiments and learned about the field” is no better informed than before reading the description.

The practical consequence: applications that lead with impressive titles and follow with vague descriptions invite skepticism. Applications that lead with honest descriptions of specific contribution—even modest contribution—and are supported by a recommender who can corroborate them, produce meaningful credibility.

5.4 The Recommendation Gap

Research is one of the few extracurricular categories where a supporting recommendation from an observing adult can independently verify the student’s claims. When that recommendation is absent—or when it is generic rather than specific—the credibility of the research listing is substantially weakened.

If a student lists a research experience prominently, particularly as one of their primary distinguishing activities, and no recommender can speak specifically to their intellectual contribution, admissions readers notice. The absence is interpretable.

Before committing to a research experience, families should be able to identify, in advance, the adult who will be able to write a specific, substantive recommendation based on direct observation of the student’s work. If no such adult exists in the program’s structure, the Credibility ROI of the experience is limited from the start.

Section 6 — How Admissions Officers Actually Read Research

6.1 The Role Clarity Test

Experienced admissions readers often apply an informal role clarity test when evaluating a research listing. It is not a rubric; it is a set of questions that good descriptions answer and poor descriptions avoid.

• Can I understand specifically what this student contributed?
• Is the description student-driven (active language, specific outcomes, honest scope) or program-driven (passive language, prestige-forward, inflated claims)?
• Could this student, in a 10-minute alumni interview, explain the research problem, their methodology, and what they found?
• Is there an adult in this application who observed the work and can speak to it specifically?

When all four questions have affirmative answers, the research listing adds credibility. When none do, the listing typically blends into the background noise of a large applicant pool—regardless of the program’s name or cost.

6.2 What Gets Attention

• Clear articulation of the research question, in the student’s own words, with genuine specificity about the problem being investigated
• Evidence of intellectual ownership: the student describes methodology, findings, and interpretation with first-person engagement, not institutional passive voice
• A concrete output that can be described specifically: a poster presentation, a dataset contributed to a larger study, a prototype, a draft paper (regardless of publication status)
• A letter of recommendation that addresses the research directly, with specificity about the student’s contribution and growth
• Contextual connection to other parts of the application—coursework, essays, and interests that make the research feel like a natural extension rather than a strategic addition

6.3 What Disappears

• Research descriptions that list the institution and nothing else
• Publications in journals that readers recognize from previous application cycles as low-standard or program-affiliated
• Research topics that are unrelated to every other aspect of the student’s academic profile
• Experiences described entirely in third person or passive voice (“We examined…” “The study found…”) without first-person intellectual engagement
• Applications where the research is the obvious capstone of a strategic credential-building plan rather than an organic expression of sustained curiosity

6.4 How the Tiers Are Actually Read

The following reflects patterns in how admissions offices at research-intensive and highly selective institutions have described their approach in published guidance and public commentary. Individual institutional practices vary.

• Tier 1 programs (RSI, NIH SIP, SSP, PRIMES): recognized by name. Acceptance rates confirm the competitive context. High Credibility ROI almost regardless of how the student describes the work—though specific, intelligent description still matters, particularly in essays and interviews.
• University lab placements: evaluated based on description specificity, recommendation quality, and role clarity. The institution’s name does not substitute for evidence of contribution. A genuine, modest contribution described specifically is more credible than an inflated description of peripheral involvement.
• Paid mentorship platforms: growing skepticism at highly selective institutions, particularly as application volume from these programs has increased and template outputs have become recognizable. Not automatically disqualifying. Individual ROI depends on the quality of the student’s intellectual engagement, the specificity of their output, and whether a recommender can support their claims.
• Pay-to-publish credentials: increasingly recognized. May flag the application for closer review. Readers at some institutions have described patterns in which the presence of these publications correlates with other credential-inflation across the file.

6.5 Two Composite Case Comparisons

Case A: Paid Platform, Low Role Clarity

A student lists “Research Associate” at a paid mentorship program in AI. They describe their project as “analyzing natural language processing models to evaluate sentiment in social media data.” The paper was submitted to a journal that accepted it within two weeks. No recommender mentions the research. In the student’s interview, when asked what surprised them about their findings, they describe a generic challenge with data cleaning.

Admissions interpretation: The description is plausible but undifferentiated. The publication does not add credibility; the reader has seen the journal before. The absence of a supporting recommendation for what is listed as a primary activity raises questions. The interview response suggests limited intellectual ownership. The listing adds no meaningful information about the student’s readiness.

 

Case B: Independent Project, High Role Clarity

A student spent the summer developing a model to analyze air quality patterns in their county, building on an AP Environmental Science unit that raised a question they couldn’t answer in class. They used publicly available EPA data, taught themselves the analytical method through documentation and YouTube, hit a dead end in their initial approach, revised their hypothesis, and produced a clear visualization that they presented at a regional science fair. Their APES teacher, who was available for questions throughout, writes a specific letter describing the student’s persistence through the failed first approach and the intellectual risk they took in revising their methodology.

Admissions interpretation: No institutional affiliation. No publication. The entire experience is legible, specific, independently initiated, and verifiably supported. The recommendation corroborates claims that could not have been manufactured. This application presents credible evidence of intellectual initiative and Growth ROI.

 

The difference between Case A and Case B is not program quality. It is intellectual ownership. Case B is also free.

Section 7 — Family Decision Tools

7.1 Pre-Enrollment Research Quality Checklist

Before committing to any research experience—paid or otherwise—families should be able to answer the following:

1. What specific question or problem will the student be working on? (If the answer is “we’ll determine that later,” the intellectual ownership has already been partially transferred away from the student.)
2. Was the problem identified by the student, the mentor, or the program? (Student-identified problems produce higher Narrative and Growth ROI.)
3. What decisions will the student make independently during the research process? What decisions will the mentor make?
4. What output will the student produce—and how specifically can the student describe it now, before the experience begins?
5. Who is the adult mentor, and can they write a specific, substantive recommendation based on direct observation of this student’s work?
6. If the experience includes a publication component, can the publication venue be verified as peer-reviewed and independent using COPE or DOAJ?
7. How many other students is the mentor working with on closely related projects?
8. What does the student find genuinely interesting about this research area—and can they articulate it unprompted?

If more than two of these questions produce unsatisfying answers, the experience is unlikely to deliver the Credibility ROI that justifies its cost.

7.2 The Role Clarity Test (Post-Experience)

After any research experience concludes, the student should be able to answer the following questions in their own words, without preparation or notes:

• What was the research question you were working on?
• How did you decide on the methodology?
• What did you find?
• What surprised you?
• What would you do differently?
• What questions are you still thinking about?

If a student cannot answer these questions fluently and specifically, the Growth ROI of the experience is lower than it may appear on paper—regardless of what the application describes. And if they struggle with these questions in conversation with a family member, they will struggle with them in an alumni interview.

7.3 Red Flag Scoring

Apply before committing. If three or more of the following apply, pause and evaluate alternatives.

 

Risk Signal	Weight
Publication guaranteed before research begins	High
No verifiable peer review or rejection rate for the publication venue	High
No identifiable mentor who can write a specific personal recommendation	High
Research topic assigned entirely by the program	Medium
Mentor works with 10+ students on similar projects simultaneously	Medium
Selectivity language with no disclosed acceptance rate	Medium
No mechanism for iteration or revision on the student’s work	Medium
Student cannot articulate genuine interest in the research area	Medium
Cost exceeds $3,000 without proportionate evidence of program rigor	Low

 

Figure 7.1. Red flag scoring tool for pre-college research programs. Three or more signals warrant serious reconsideration.

7.4 Lower-Cost and Alternative Pathways

Some of the most credible research experiences available to high school students cost nothing. Families should evaluate these seriously before committing to paid alternatives.

• University professor outreach: A well-prepared 11th grade student with a specific question and demonstrated coursework depth in the relevant field who sends a thoughtful, specific email to a faculty member whose work they have actually read is more likely to receive a response than families expect. This pathway produces the highest possible Credibility ROI when it succeeds—a real faculty member, real lab, real contribution, real recommendation.
• Science fair and ISEF pathway: Independently structured, externally evaluated, no cost. ISEF finalists and state-level qualifiers are recognized by name at research-intensive institutions. The pathway rewards genuine intellectual ownership over time.
• Independent project with a high school teacher mentor: The most consistently undervalued option in this landscape. A student who designs and executes a real investigation with a teacher who knows them well—and who can write a specific, knowledgeable letter about the student’s intellectual process—presents one of the most credible research profiles possible at the high school level.
• Community college enrollment + research component: Dual enrollment in a course with an independent research component provides both coursework credentials and research experience at a fraction of the cost of paid mentorship platforms.
• Government science programs: NASA, EPA, USDA, and state science agencies operate programs for high school students that are free, competitive, and carry institutional credibility equivalent to or exceeding paid alternatives.

Section 8 — Methodology Note

This installment applies the same four-pillar ROI framework introduced in January—Growth, Narrative, Credibility, and Wellbeing—without modification. Evidence is drawn from the same four categories: public institutional data and reporting, admissions office guidance, pattern-based composite case analysis, and developmental and learning science.

One category-specific addition applies to this issue: publication credibility claims are evaluated against academic publishing norms as defined by the Committee on Publication Ethics (COPE), the Directory of Open Access Journals (DOAJ), and established academic indexing databases. These are the same standards the research community applies. Admissions readers at research-intensive institutions generally apply them as well.

Named programs are cited when they are unambiguous positive reference points with publicly documented selectivity and structure: RSI, NIH SIP, SSP, and PRIMES. Named programs are not cited as cautionary examples; composite cases and structural descriptions are used for that purpose, consistent with the series’ approach.

Statistical claims in this issue are sourced as follows: RSI acceptance rate of 2–3% from program documentation and reporting (MIT Math Department, CEE); NIH HS-SIP acceptance rate of approximately 7% from published program documentation; the 20+ online research programs figure from ProPublica/Chronicle of Higher Education (2023); predatory journal volume estimates from academic research on predatory publishing (Beall, subsequent scholars); Caltech Class of 2027 data from Caltech’s own admissions announcement; and the 60% no-citation figure for predatory journal publications from published research on predatory publishing patterns.

Section 9 — Looking Ahead: March and the Competitions Question

The March installment examines competitions across academic, creative, and STEM domains—and the analysis will feel familiar to anyone who has worked through this issue.

The evaluative questions are closely related: What is the selection mechanism? Is selectivity real or manufactured? What does a strong performance in this competition reveal about the student’s preparation, depth, and intellectual ownership? And what does it not reveal?

There is also a specific intersection to explore. The paid research-to-competition pipeline—in which students use paid mentorship programs specifically to generate ISEF or similar competition projects—has grown substantially in recent years. The question of whether a competition entry reflects genuine independent inquiry or a well-funded simulation of it is a live one, and one the March installment will address directly.

The same single diagnostic question applies: what did the student actually do—and how much of the thinking was theirs?

Section 10 — Conclusion

In 2023, a joint investigation by two of the most respected journalism institutions in American higher education documented what counselors had been observing for years: a fast-growing industry had emerged to help families buy their way into a credential that admissions culture had decided to value. The industry’s product was not research. It was the appearance of research.

That appearance is increasingly legible to the people it is meant to impress.

This is not a reason for cynicism about research as a pre-college category. It is a reason for precision. Genuinely formative research experiences—experiences that develop intellectual independence, produce verifiable output, and are supported by adults who observed the work—remain among the highest-ROI activities available to high school students. The programs that deliver them are, in many cases, free. They are also difficult to access, and that difficulty is real and worth acknowledging.

But the gap between what families are sold and what admissions readers value has never been wider in this category than it is right now. A single question—what did the student actually do, and how much of the thinking was theirs—cuts across every tier of this market. It should be the first question families ask before enrolling, and the last question they ask when evaluating whether an experience delivered what it promised.

Credibility is not purchased. It is earned by doing work that holds up under scrutiny.

If you find this framework useful for evaluating an upcoming decision—or if you’d like a second perspective on a specific program, experience, or plan—our team is available as a resource. Not because there’s one correct path, but because clarity is easier to find when you’re not navigating alone.

Issue 3—Competitions: Distinction, Depth, and Signal Saturation—publishes in March.

References

Section 1 — Landscape

Center for Excellence in Education. (n.d.). Research Science Institute (RSI). https://www.cee.org/programs/rsi

Massachusetts Institute of Technology, Department of Mathematics. (n.d.). RSI at MIT Math. https://math.mit.edu/research/highschool/rsi/

National Institutes of Health, Office of Intramural Training and Education. (n.d.). Summer Internship Program (SIP). https://www.training.nih.gov/research-training/pb/sip/

ProPublica & The Chronicle of Higher Education. (2023, May 18). The newest college admissions ploy: Paying to make your teen a “peer-reviewed” author. https://www.propublica.org/article/college-high-school-research-peer-review-publications

Polygence. (n.d.). About Polygence. https://www.polygence.org

California Institute of Technology. (2023). Caltech extends admissions offers to Class of 2027. https://www.caltech.edu/about/news/caltech-extends-admissions-offers-to-class-of-2027

Caltech Undergraduate Admissions. (n.d.). Research papers and publications. https://www.admissions.caltech.edu/apply/first-year-applicants/supplemental-materials/research-papers-and-publications

Common Application. (2023). Application trends and volume reports. https://www.commonapp.org

Sections 2–3 — Misreading Value / ROI Framework

Kahneman, D., & Tversky, A. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131. https://doi.org/10.1126/science.185.4157.1124

Kahneman, D. (2011). Thinking, fast and slow. Farrar, Straus and Giroux.

Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth, and happiness. Yale University Press.

National Association for College Admission Counseling. (2023). State of College Admission. https://www.nacacnet.org

Harvard College Admissions. (2022). How we read applications. https://college.harvard.edu/admissions/apply/what-we-look-for

Damour, L. (2019). Under pressure: Confronting the epidemic of stress and anxiety in girls. Ballantine Books.

Sections 4–6 — Red Flags / Admissions Reading

Committee on Publication Ethics (COPE). (2022). Core practices. https://publicationethics.org/core-practices

Beall, J. (2012). Predatory publishers are corrupting open access. Nature, 489(7415), 179. https://doi.org/10.1038/489179a

Wikipedia contributors. (2024). Predatory publishing. Wikipedia. https://en.wikipedia.org/wiki/Predatory_publishing

U.S. Department of Education. (2022). Auxiliary and enrichment education programs: Oversight and transparency considerations. https://www.ed.gov

Bray, M. (2017). Shadow education: Private supplementary tutoring and its implications for policy makers. UNESCO International Institute for Educational Planning. https://unesdoc.unesco.org

OECD. (2021). Education at a Glance. https://www.oecd.org/education/