The Ivy Institute | College Admissions Research | November 2026
Executive Summary
When a teenager gives a different answer to “What do you want to study?”, ask what stayed interesting. In all seven student histories we could compare, the student’s later plans kept a connection to an earlier field, question, method or goal. The name of the major could change while the reason for exploring it still made sense.
We reviewed 15 students’ histories. Seven had two dated, student-endorsed choices we could compare. The other eight lacked a clear second choice or contained possibilities the student had not adopted. The middle gap between observations was 119 days, with a range of 16 to 424 days. These numbers measure the time between two conversations or decisions. They do not tell us how long a student spent deciding.
Some students refined a broad interest. Others combined fields or described related routes at different colleges. We could not verify a major on a submitted application for any of the seven. The evidence tells us how these students’ intentions developed, and leaves their eventual college majors open.
For families, a useful next conversation starts with a recent class, project or question the student enjoyed. What about it held their attention? Which course or experience would let them test that interest next? The Ivy Institute conducted and funded this study and publishes it as a provider of paid admissions advising. Its clients may differ from other applicants. The Ivy Institute, internal materials, data, and information.
Contents
Research Questions
Did the student name a new subject, narrow an earlier interest or find another route to the same question? We also asked what the student actually chose, what an adult only suggested, and what could be verified at the application stage.
First, among the student histories that support a reliable before-and-after comparison, how did documented academic directions change over time? The analysis distinguishes greater specificity, interdisciplinary combination, parallel institution-specific planning, a change in central direction with continuity, and complete change of direction.
Second, did later directions preserve a traceable bridge to earlier interests? A bridge could be the same broad field, a recurring problem, a method of inquiry, or a stable career purpose. This question matters because labels alone can make a developing interest look more discontinuous than it is.
Third, how strong was the evidence at each endpoint? A student-stated interest is different from a counselor suggestion. A planned application choice is different from a field printed on a completed application. A completed application is different from a submitted application. The study therefore treats evidence level as part of the result.
Fourth, when the records offered a reason for development, what could be said without inferring motivation? The analysis accepted a reason only when the student’s preference or explanation was documented. It did not assign motives from timing, activities, family circumstances, or counselor recommendations.
Fifth, how should families interpret an evolving intended major in light of external evidence and different college policies? This final question places the internal findings in context while preserving the distinction between comparison and validation.
The study did not ask whether one major is strategically easier for admission, whether students should choose a field to improve admission chances, or whether advising produced the observed pattern. It did not compare admission outcomes by intended major. Those questions require different data and designs.
What does “intended major” mean?
The same word can mean an interest, a possible career or a formal choice on an application. Ask which one the student means before deciding that their plans have changed.
A pre-professional aspiration is not necessarily a major. Pre-law and pre-med commonly describe goals or advising tracks rather than standalone undergraduate fields. A student interested in medicine might consider biology, chemistry, neuroscience, public health, engineering, history of science, or another field. A student interested in law might study English, economics, political science, philosophy, engineering, or many other disciplines. Coding “medicine” as the same kind of variable as “biochemistry” would blur career purpose and degree field.
An academic division is also not always a major. A student may be choosing between a college of engineering and a college of arts and sciences while still exploring the precise field. Another student may name a field that appears in different schools at different universities. The same intellectual interest can therefore produce different application labels because institutions organize knowledge differently.
Timing creates another problem. A statement from early high school may express broad curiosity. A statement in application season may reflect a field selected for a particular form. A later statement can be more specific because the student learned more, because a deadline forced a choice, or because the college’s available programs created a new vocabulary. The records alone do not always separate these mechanisms.
Speaker attribution matters as well. Advising conversations often contain possibilities. A counselor may propose a field to investigate, a parent may ask about career prospects, and a student may test an idea aloud. If every mentioned field is treated as the student’s intention, the analysis will manufacture changes that did not occur. This study therefore required tied to the student evidence and rejected unadopted suggestions as endpoints.
Finally, an intended major is not always binding. Some universities admit to the institution as a whole and let students declare later. Others consider the first-choice major in admission or use direct-entry pathways for constrained programs. The same declaration can be exploratory at one college and consequential at another. Any practical interpretation has to start with official policy for the relevant cycle.
These distinctions explain the conservative design. The study prefers a smaller denominator with defensible comparisons over a larger count built from ambiguous mentions. It also reports missing finality as a finding. A record that supports “current direction” but does not support “submitted major” should not be promoted to a stronger claim.
How we did the study
We needed two dated observations for the same student and a clear reason to treat each as that student’s academic direction. An adult could suggest a major, but the suggestion counted only if the student adopted it. Each eligible student counted once, even if several colleges organized the field differently.
Read the detailed method and checks
Setting and company role
The analysis uses authorized records from The Ivy Institute’s work with students and families. The company conducted the study, funded the work, and intends to publish the report on its own website. The Ivy Institute is a commercial admissions-advising organization and has an interest in how its services are understood. That relationship creates a risk of favorable interpretation. The study responds by using explicit eligibility rules, reporting exclusions, separating observation from interpretation, and avoiding claims about admission effectiveness.
The records were created during ordinary advising and application work. They were not collected from a research protocol designed in advance. This makes the study retrospective and observational. The available evidence can describe what was documented. It cannot establish what a student thought between meetings or what would have happened without advising.
Bounded screening frame
The screening frame contained fifteen student histories previously identified during related internal research. This was a purposive starting frame. It was neither a census of all company students nor a probability sample. The frame included students at different stages and with different amounts of documented follow-up.
Each person could contribute at most one primary comparison. Multiple colleges, essays, or meetings did not create multiple students. When a student considered different fields at different colleges, those choices remained nested within one history. This rule prevents students with more extensive records from dominating the totals.
The screened material covered dates from 2024 through 2026. The seven eligible comparisons began between August 2025 and October 2026 and ended between December 2025 and November 2026. Because the frame was selected before the final coding and included incomplete histories, it also provided a direct view of missingness.
Inclusion criteria
A history entered the paired analysis when all of the following were present.
- The same student could be followed at two different dates.
- Each date was tied to an academic direction rather than only an activity, course, essay prompt, or career title.
- The direction was linked to the student’s own preference, plan, or endorsed interest.
- The order of the two observations was reliable.
- The later point supplied enough information to compare with the earlier point.
A history was excluded from the paired denominator when it had only one eligible date, when the later possibility was an unadopted counselor suggestion, when the evidence described general exploration without another field, or when no academic-direction pair could be established. Exclusion does not mean the student had no interests. It means the evidence did not support the comparison required by this study.
Unit of analysis
The unit was the individual student. The study did not count colleges, applications, meetings, essays, fields, or mentions as independent observations. This is crucial because one student can discuss several fields repeatedly. A mention-based analysis could make an extensively documented student appear to represent many cases.
The study also separated three levels of analysis. The student level asks how a person’s direction developed. The observation level identifies the dated evidence used for the early and later endpoint. The field level describes the academic content at those endpoints. Counts in the public report refer to students unless explicitly labeled otherwise.
Evidence hierarchy
The coding hierarchy moved from weaker to stronger evidence.
- A student-stated interest or a clear summary of the student’s preference.
- A student-endorsed current academic direction.
- A planned application direction tied to a college.
- A direction selected in an application in progress.
- A verified direction on a submitted application.
The eligible records reached the levels of exploration or current application planning. None provided enough evidence for the final submitted-choice category under the study’s strict rule. This missing endpoint constrains every conclusion.
Development categories
Broad-to-specific development means a general domain later appeared as a narrower field. Interdisciplinary development means an earlier interest remained while the later direction combined it with another discipline, method, or application. Parallel school-specific framing means the student considered related but different pathways at different institutions during the same period. Changed direction with continuity means the intellectual center moved while a prior field, problem, or method remained visible. Wholesale replacement requires a later direction with no documented bridge to the earlier one.
These categories describe the record. They are not judgments about whether the change was good, authentic, strategic, or mature. A more specific label is not automatically better. An interdisciplinary label is not automatically coherent. A stable label is not necessarily deeply considered. Those evaluations would require additional evidence.
Coding and review
The review proceeded chronologically within each student. Early and later endpoints were first identified. Each endpoint was then rated for evidence strength and speaker attribution. The analyst recorded the academic direction, any documented bridge, and any explicit reason. Alternative explanations were considered. For example, a new label could reflect a college’s program menu rather than a change in the student’s underlying interest.
Potential transitions were rejected when the later field came only from a recommendation or when a date reflected record creation rather than the date of the student’s choice. Omitted fields were not assumed to have been abandoned. Continued silence about an earlier interest was not coded as rejection. Missing follow-up was not coded as stability.
Because the eligible sample is small, subtype counts and grade-specific cells are not published. Combining several small cells into a single continuity result protects privacy and avoids a table whose complementary counts could reveal suppressed values. The private source ledger retains the detailed coding and calculations.
Calculations
Fifteen histories were screened. Seven met all inclusion criteria and eight did not. The elapsed intervals for the eligible pairs were 16, 17, 64, 119, 132, 415, and 424 days. The median was the fourth sorted value, 119 days. These figures describe calendar distance between recorded endpoints. They do not measure advising intensity, active reflection, or hours worked.
Every eligible history contained a documented bridge between early and later directions. The continuity numerator was therefore seven and the denominator seven. No eligible history met the strict wholesale-replacement definition. No later endpoint was verified as a submitted application choice. Percentages are shown only where the denominator is explicit and the measure is not likely to imply population representativeness.
Privacy protections
The public report removes names, institutions tied to individual students, private writing, and case narratives. It does not publish the subtype distribution, grade-specific counts, exact combinations of schools and interests, or personal reasons attached to a single case. Synthetic examples later in the report illustrate coding logic without reproducing any student’s language or exact combination of facts.
All company-based evidence is cited publicly only as “The Ivy Institute, internal materials, data, and information.” Exact source locations and calculations remain in a separate restricted ledger and are not included in the public publication kit.
Which students could we compare?
We started with 15 students. We could make a two-date comparison for seven. The remaining eight could still be exploring important questions. The evidence available to us did not establish a second eligible choice.
The first result is the size of the usable evidence base. Fifteen histories were screened, and seven supported a paired comparison. Eight did not. The 47 percent eligibility rate is not a quality score for the underlying advising. It reflects the strict research question. A useful advising record can still be insufficient for a longitudinal comparison if it contains only one dated statement of direction.
| Evidence-funnel measure | Students | Denominator |
|---|---|---|
| Histories screened | 15 | 15 |
| Eligible paired histories | 7 | 15 |
| Excluded from paired analysis | 8 | 15 |
The exclusions are analytically important. Some histories had a clear current intended field but no earlier eligible field. Others contained an early interest and a later counselor-proposed option that the student had not clearly adopted. One documented continued exploration, which is a meaningful student state but does not create a second field for transition coding. Another contained a cross-year possibility whose speaker endorsement was unclear.
These cases demonstrate why a title search or keyword count is not enough. A word such as “major” can appear in a prompt, a planning checklist, a counselor’s suggestion, or a student’s decision. Longitudinal research requires dates, identity matching, speaker attribution, and comparable concepts at both ends.
The funnel also limits the headline. The study can say what happened in seven comparable histories. It cannot silently treat the excluded eight as unchanged, changed, or unresolved. Their status is missing for the paired question.
Main Findings
All seven later choices kept something from the earlier interest. Sometimes the common thread was a subject. Other times it was a problem the student wanted to solve or a way they liked to work. No later choice could be verified as the major on a submitted application.
1. Every eligible comparison retained continuity
All seven later directions retained at least one bridge to the earlier direction. A bridge did not require an identical label. It could be a shared domain, a persistent problem, a method the student continued to value, or a long-term purpose connecting multiple disciplines.
This finding changes the interpretation of apparent switching. If a student moves from a broad scientific interest to a specific computing or engineering direction, the label has changed while the underlying technical orientation remains. If a student connects humanities interests with film, law, or intellectual property, the later direction may represent integration. If a student connects business with agriculture, environmental questions, mathematics, or technology, the direction may become more interdisciplinary without abandoning business.
The study therefore found developmental continuity rather than simple label stability. The later records generally contained more structure. They showed fields in relation to one another, fields connected to a recurring problem, or fields translated into an institution’s academic vocabulary.
| Continuity measure | Students | Denominator |
|---|---|---|
| Retained an earlier field, problem, method, or purpose | 7 | 7 |
| Met the strict definition of complete change of direction | 0 | 7 |
This does not mean no one changed. It means the observed changes were connected. A zero in the replacement row also does not establish that wholesale changes are rare in the company’s full population. The screened frame was small and purposive.
2. Greater specificity was one visible form of development
Several histories moved from a broad cluster to a more specific direction. General interests in mathematics and science later appeared as computing, systems, or engineering-oriented planning. Broad combinations of business, agriculture, and technology later incorporated more specific quantitative or operational fields. General curiosity about how people think later developed into decision analysis connected to computing or economics.
Specificity can emerge for several reasons. A student may acquire new experience, learn the names of fields, investigate college programs, or respond to an essay that demands a direct answer. The records can show that specificity appeared. They cannot always identify which mechanism caused it.
The analysis also avoids assuming that specificity equals finality. A more precise field at one date may still be provisional. The later endpoint in this study remains an intended direction unless submission is verified.
3. Interdisciplinary combination was common in the qualitative pattern
The eligible histories repeatedly connected disciplines. Scientific interests were connected with engineering or medicine. Humanities fields were connected with film and law. Business was connected with culture, history, agriculture, public questions, mathematics, or technology. Psychology was connected with finance, decision-making, or neuroscience.
The report does not publish a subtype count because the cells are small. The qualitative recurrence is still important. It shows that forcing each student into a single isolated label can misrepresent what the student is trying to study. An intended major may be the administrative home for a broader intellectual agenda.
Interdisciplinary language can also become vague. “I want to combine everything” is not the same as a coherent direction. The stronger internal examples had a bridge that explained why the fields belonged together, such as a recurring question, a real project, a method, or a long-term purpose. The weaker possibilities remained lists of attractive subjects.
4. Institution-specific pathways complicated the idea of one intended major
At least one eligible history clearly involved different but related academic framing across institutions. The student was not necessarily changing over time. The college structure changed the available route. A field located in a business school at one university might be approached through a social-science or humanities department elsewhere. Another university might offer a specialized interdisciplinary program.
This is a measurement issue and a planning issue. If researchers compare labels without considering institutional organization, they may code a false switch. If families assume every college offers the same route, they may choose a label that does not match the actual program.
The appropriate unit is therefore the student’s underlying direction plus the institution-specific pathway. Applications may require different labels while remaining coherent when each label is supported by the same chain of experiences and questions.
5. A clearer change in center of gravity still preserved a bridge
One eligible comparison showed a more visible shift from earlier possibilities to a later focus on decision processes. Even there, the later direction remained connected to earlier computing, finance, and systems-oriented interests. The result was coded as changed direction with continuity rather than complete change of direction.
This distinction protects against two opposite errors. One error is to call every new label a complete change. The other is to call every connected development stable. The student did change what sat at the center of the academic plan. The bridge explains how the change developed.
6. Reasons were useful but unevenly attributable
The records sometimes linked development to a project, research experience, course, personal question, family experience, or emerging career purpose. These reasons helped explain why a later field made sense. They were not counted into a public frequency table because attribution and completeness varied, and small combinations could identify students.
The study treated reasons conservatively. A counselor’s explanation of why a field could fit did not become the student’s motive unless the student adopted it. An activity related to a field did not prove that the activity caused the interest. A sequence in time did not prove causation.
For practice, this means counselors should ask the student directly what changed and why. A dated one-sentence reflection can be more analytically useful than a long list of possibilities.
7. The observed endpoint was intention, not submission
The most consequential limitation appears in the finality check. All seven included histories had a later documented direction. None had a later endpoint that this review could verify as the major on a submitted application.
| Finality measure | Students | Denominator |
|---|---|---|
| Documented later intended direction | 7 | 7 |
| Verified major on a submitted application | 0 | 7 |
This distinction prevents a planning discussion from being reported as a completed choice. It also means the report cannot say whether a direction survived the final application process, whether the student later changed it, or whether the student enrolled in that field.
8. Observation intervals varied widely
The shortest eligible comparison was 16 days. The median was 119 days. The longest was 424 days. A short interval can capture rapid clarification during application season. A long interval can capture development across grades or cycles. These are different processes placed in the same descriptive analysis.
| Interval statistic | Days | Eligible pairs |
|---|---|---|
| Minimum | 16 | 7 |
| Median | 119 | 7 |
| Maximum | 424 | 7 |
The interval does not tell us when the student actually changed. The transition could have occurred immediately after the early observation, just before the later observation, or gradually between them. More frequent standardized measurement would be needed to estimate timing.
What stayed the same as interests changed?
A family can look at a change in three parts. What interested the student at first? What experience gave them new information? What part of that interest still matters now? A new major name does not have to erase the earlier story.
The first layer is the durable question. This is the recurring problem or curiosity that continues even when the label changes. Examples might include how people make decisions, how technology can solve practical problems, how stories shape communities, or how scientific knowledge can improve health. These are synthetic illustrations, not student quotations.
The second layer is the domain. Domains are broad areas such as science, humanities, social science, business, technology, or the arts. A student may remain within a domain while becoming more specific.
The third layer is the field. Fields are more precise academic homes such as neuroscience, economics, film studies, computer science, or anthropology. The field can change while the durable question remains.
The fourth layer is the institutional pathway. A college may place a field in a particular undergraduate school, combine it with another discipline, or restrict direct entry. The pathway can change the application label without changing the student’s underlying question.
This layered model helps interpret several kinds of development.
Narrowing occurs when the durable question and domain remain while the field becomes more precise. Integration occurs when the student finds that two fields illuminate the same question. Translation occurs when the same direction is expressed through different institutional pathways. Reorientation occurs when the central question or field changes but retains a meaningful bridge. Replacement would occur when no bridge remains.
The model also suggests better evidence. A coherent application should be able to answer four questions. What does the student want to understand or do? Which broad domain contains that work? Which field currently provides the best tools? How does the specific college organize that path? The answers can evolve without becoming contradictory.
What the Study Did Not Establish
These histories do not tell us how often applicants nationwide switch majors, which routes colleges prefer or how advising affected a student. They also do not tell us whether a student later enrolled in the field they discussed.
The study did not show that the later direction was better. “Better” could mean more authentic, more feasible, more employable, more coherent, or more advantageous in admission. Each definition would require its own measures.
The study did not test admissions outcomes. No finding should be read as evidence that changing, retaining, or combining interests affected acceptance. An application can be coherent with a stable direction or an evolving one. Admission decisions depend on many factors and institutional priorities.
The study did not verify submitted majors. It therefore cannot calculate the share of students who changed between first advising and submission. It also cannot compare intended fields with enrolled majors.
The study did not estimate prevalence beyond the seven eligible histories. The result that seven of seven retained continuity is exact for the coded sample. It is not a population rate.
The study did not evaluate the truth of every institution-specific statement contained in the internal material. Public policy claims in this report were checked against official sources. Private planning comments were used only to analyze student direction.
Comparison With External Research
National statistics about college students changing a declared major answer a different question. Our seven students were considering what to study before college. The national figures below provide perspective, while the original populations and time periods stay visible.
Postsecondary major change is common, but it is a different outcome
The National Center for Education Statistics analyzed a nationally representative cohort of about 25,000 first-time college-level students. Among students who had declared a major in an associate’s or bachelor’s program, about 30 percent changed majors within three years. The rate was 33 percent among students in bachelor’s programs and 28 percent among those in associate’s programs. About one in ten changed majors more than once. The report also found variation by original field. NCES, 2017.
Those figures should not be compared numerically with this study. NCES measured formal changes after college enrollment. The Ivy Institute analysis measured documented high-school intentions before a verified submission endpoint. The national study had a probability-based design and a large sample. This study has seven deliberately selected eligible pairs. The NCES result provides context for the general proposition that academic direction can change. It does not validate the internal seven-of-seven continuity result.
The distinction also prevents a familiar statistical error. One cannot subtract a high-school intention rate from a college major-change rate and interpret the difference as advising impact. The populations, measures, time periods, and denominators are different.
Adolescent interests can be stable and still become more differentiated
Tracey, Robbins, and Hofsess followed adolescents from eighth through twelfth grade and found that patterns of vocational interests were fairly stable while interests also became more crystallized over time. That combination closely matches the conceptual distinction in this report between continuity and development. Stable broad interests can acquire greater resolution. Tracey, Robbins, and Hofsess, 2005.
Hirschi’s longitudinal work on vocational-interest differentiation found that increases in differentiation predicted increases in career decidedness. The study involved Swiss adolescents and should not be treated as a direct comparison to U.S. college applicants. It supports the general idea that a student may become more decided as interests become better differentiated. Hirschi, 2009.
Germeijs and Verschueren followed 535 students across grade twelve and found average increases in orientation, exploration, decisional status, and commitment, with substantial variation among students. Their model emphasizes broad exploration earlier and in-depth exploration later. The Ivy histories contain a similar descriptive sequence in some cases, but the internal study did not administer standardized exploration or commitment scales. Germeijs and Verschueren, 2006.
Rogers and Creed’s longitudinal study of 631 students in grades ten through twelve found that self-efficacy and goals predicted career planning and exploration, including change over time. The result highlights why confidence and agency may matter in academic planning. This internal study did not measure self-efficacy, so it cannot test that relationship. Rogers and Creed, 2011.
Research on educational and vocational identity also suggests reciprocal development. Negru-Subtirica and Pop followed 1,030 adolescents across three waves and found that educational commitments supported vocational commitments, while exploration in one domain related to exploration in the other. The finding supports treating academic and career planning as related processes rather than a one-time label selection. Negru-Subtirica and Pop, 2018.
Longitudinal STEM research offers a more specific example. Jiang, Simpkins, and Eccles used nationally representative high-school data and found that early math and science motivational beliefs were associated with later STEM choices. A later study by Jiang and Simpkins examined changes in those beliefs from ninth through eleventh grade and their relation to college major choice. These studies demonstrate that both early orientation and development across high school can matter. They do not imply that a student must maintain a single field label. Jiang, Simpkins, and Eccles, 2020 and Jiang and Simpkins, 2024.
Across these sources, the most defensible comparison is conceptual. Interests can show continuity while becoming more differentiated. Exploration and commitment can develop together. Experiences and beliefs can relate to later choices. None of the external studies independently checks The Ivy Institute’s records.
Institutional Policy Matters
An intended major can carry different weight depending on where a student applies. Use the official policies linked below to ask whether students enter a major immediately, enter an undergraduate school, or declare later. Check the current rules for the actual program.
MIT states that applicants apply for general admission, identify only a course of interest, and choose a major at the end of the first year. MIT also says the field of interest does not affect admission and that approximately half of its students ultimately major in something different from the application field. MIT Admissions.
Yale explains that students enter undeclared and generally have until the end of sophomore year to select a major. Its admissions discussion emphasizes that Yale admits students to one undergraduate college rather than to individual majors. Yale Undergraduate Admissions and Yale Admissions podcast transcript.
Illinois describes a different structure. Its official review policy says applicants are reviewed within the academic community or program to which they apply, and interest in the intended major may be considered. The first-year application asks for a first-choice major and allows an optional second choice. University of Illinois Undergraduate Admissions and first-year application guidance.
The University of Washington likewise says requested majors may be factors in admission and identifies programs with direct first-year pathways. For computer science, the official page says the primary pathway is direct first-year admission and instructs applicants to list it as the first-choice major. University of Washington Office of Admissions and Computer Science major page.
These examples are not a ranking and do not cover every exception. They show why a universal rule about intended majors is unsafe. Families should verify the current cycle, the undergraduate school, direct-entry requirements, internal-transfer rules, and whether an alternate path is genuinely acceptable to the student.
The internal finding of parallel institution-specific framing makes sense in this policy environment. The same student may need to translate one intellectual direction into different program structures. Authenticity requires that each pathway be one the student would actually pursue. A student should not select a less desired field on the assumption that changing later will be easy.
Why some students could not be counted
When we lacked a clear second academic choice, we left the student out of the paired calculation. We did not count that student as having stayed the same, switched fields or given up on an interest.
Several histories had only one eligible dated direction. Some of those single points were detailed and credible. A student could have a well-developed current interest, a preferred field, and a clear explanation, yet still lack a comparable earlier point. It would be incorrect to label that student stable simply because no earlier difference was recorded. Stability requires observation at more than one time.
Another history contained a later statement of continued exploration. That statement is meaningful because uncertainty can be a genuine developmental position. It did not qualify as a second field under the prespecified rule. Counting it as a “change to undecided” would treat uncertainty as though it were an academic discipline. Counting it as stable would ignore the shift in confidence. The analysis therefore kept it outside the paired field-transition denominator.
One history contained an intriguing later combination, but the evidence did not clearly establish that the student had adopted it. Advising records naturally include hypotheses. A counselor may suggest combining a creative field with a technical field, or may identify a program that appears relevant. Research coding must preserve the difference between exposure to an option and endorsement of an option.
Other excluded histories contained fields at one meeting but no second dated endpoint, or contained academic language embedded in an essay-planning context without a clear student decision. These are examples of “information present, comparison absent.” They show why the amount of text cannot substitute for the quality of the endpoint.
The excluded cases also illustrate potential selection bias. Students with long advising relationships, frequent meetings, and detailed planning may be more likely to produce two comparable points. Students who begin late, communicate less often, or arrive with an already developed direction may be less likely to qualify. The seven eligible histories may therefore overrepresent visible development and extensive documentation.
Future research can reduce this problem by asking the same short questions at fixed times. A student could rate current interest in up to three fields, confidence in each, the main question connecting them, and the evidence that informed the rating. The same form could be repeated at intake, the end of each grade, application planning, and submission. That prospective design would convert many of today’s exclusions into interpretable observations.
Examples of how we read a change
These examples are invented. Use them to see how two descriptions can be connected without assuming that the writer has already chosen a college or career.
Example A: broad-to-specific
In March, a student says she likes “science and solving environmental problems.” In October, she says she plans to study chemical engineering because she wants to develop water-treatment materials. The later field is more specific. The environmental purpose and scientific method supply the bridge. The case would be coded broad-to-specific, provided both statements were tied to the student and reliably dated.
Example B: interdisciplinary development
In tenth grade, a student identifies history as his strongest subject. In twelfth grade, he plans to study digital humanities to analyze how public archives shape collective memory. History remains central, while computing becomes a method. The case would be coded interdisciplinary development.
Example C: parallel institutional pathways
A student is interested in how cities finance public infrastructure. At one college, she considers economics in a liberal-arts division. At another, she considers an undergraduate public-policy program. The labels differ because the institutions organize the subject differently. If both are current at the same time, this is parallel school-specific framing rather than a chronological switch.
Example D: changed direction with continuity
A student first plans to study computer science because he enjoys building applications. After a research experience, his central interest becomes cognitive science and human-computer interaction. Computing remains a method, but the center moves toward human behavior. The case would be coded changed direction with continuity.
Example E: complete change of direction
A student first identifies music composition and later identifies geology, with a clear statement that the earlier direction no longer fits and no shared problem, method, or purpose. If both endpoints are student-endorsed, this would meet the wholesale-replacement definition. No eligible internal case met that rule.
Example F: ineligible suggestion
A counselor tells a student that statistics might be a useful alternative. The student does not respond or later choose it. Statistics is a proposed option, not the student’s later direction. It cannot serve as an endpoint.
Example G: ineligible submission inference
A student drafts an essay about neuroscience. The record later says the application is complete, but it does not show the selected major. The study may identify neuroscience as an intended direction if other evidence supports it. It cannot claim that neuroscience was the submitted major.
These examples demonstrate that the coding system focuses on meaning and evidence level. It does not reward change or stability. It asks what can be supported.
How much depends on the definition?
The result depends on what qualifies as a connection. A narrower definition could classify the same two interests differently. We kept the original rule visible and did not assign motives when the student’s explanation was missing.
Read the detailed method and checks
The first alternative is a label-only rule. Under that rule, any difference between early and later labels would count as change. This would produce more apparent switching because “science” and “computer science” would be different, as would “history” and “film studies.” The rule would ignore the bridges that motivated this study and would overstate discontinuity.
The second alternative is a domain-only rule. Under that rule, all fields within STEM, humanities, social science, or business might be treated as stable. This would understate development. Moving from a broad scientific interest to a particular engineering field can be a substantial decision even if both sit within STEM.
The third alternative is to count every counselor suggestion as a student possibility. This would increase the number of endpoints and transitions, but the resulting measure would describe the advising menu rather than student intention.
The fourth alternative is to require a verified submitted choice. That rule is appropriate for a submission study. In the present records it would reduce the eligible denominator to zero. The finding itself is useful. The available evidence supports a study of evolving intentions, not final application selections.
The fifth alternative is to classify continued exploration as an endpoint. Doing so could add a category called “moved to undecided,” but only if the research question were confidence or decisional status rather than field transition. A future study should measure field and certainty separately so that a student can remain interested in the same fields while becoming more or less decided.
The sixth alternative is to limit observation to a narrow application-season window. That would make intervals more comparable but would exclude the longest developmental histories. It might emphasize rapid application framing and miss earlier exploration. Conversely, using only multi-year histories would reduce the sample and might miss meaningful short-term clarification.
Across these alternatives, the most robust conclusion is narrow. The seven included histories all contain continuity, but the precise subtype depends on how one balances labels, domains, and underlying questions. The report therefore foregrounds the continuity bridge and keeps small subtype counts private.
A Family Workbook for Academic Direction
Families can use the following sequence as a structured conversation. It is designed to support exploration and documentation rather than force a final answer.
Step 1: list current fields
Ask the student to name up to three fields that feel genuinely interesting today. Add a confidence rating from one to five. “One” means newly curious. “Five” means the student has tested the interest and would be comfortable choosing it now. Do not require all three slots.
Step 2: identify the evidence
For each field, list the experiences that informed it. Separate direct experience from observation. Taking a course, completing a project, conducting research, working, or creating something provides different evidence than hearing that a career is prestigious or lucrative.
Step 3: name the recurring question
Write one question that each field helps the student pursue. If several fields connect to the same question, the student may have an interdisciplinary direction. If each field points to an unrelated question, more exploration may be useful.
Step 4: distinguish attraction from commitment
For each option, ask what is attractive and what remains uncertain. Students often like an outcome without knowing whether they like the daily academic work. A student may like the idea of medicine but dislike laboratory science, or like technology companies without enjoying technical problem-solving. This is information, not failure.
Step 5: choose a next test
Select one feasible activity that would provide new evidence. It might be reading a syllabus, completing a small project, interviewing a professional, observing a class, or comparing introductory course assignments. The test should answer a specific uncertainty.
Step 6: check institutional structure
For each college, identify where the field lives, how first-year admission works, when majors are declared, and what happens if the student changes direction. Use official sources and record the date checked.
Step 7: document the present decision
Write a short statement with three parts. “Right now I am leaning toward ____. The strongest reasons are ____. I am still exploring ____.” This format allows clarity and openness at the same time.
Step 8: revisit at a defined date
Schedule the next review after a meaningful event, not every week. Revisit after a course, project, summer program, research experience, or college visit. Compare the new statement with the earlier one and record what changed.
Step 9: verify before submission
At the final application stage, confirm the selected academic unit and field in the actual form. Check that essays and short answers use consistent language. Save the submission confirmation and the selected field only in authorized, secure records.
Step 10: keep permission to evolve
The final application choice is a present plan. Its consequences depend on the college. Where the choice is nonbinding, the student may continue exploring. Where direct admission matters, the student should understand the constraints. In both situations, intellectual growth can continue.
Guidance for Students and Families
Start with recurring questions
Ask what the student repeatedly wants to understand, build, change, explain, or protect. A recurring question is often more stable than a major label. Write down two or three questions in the student’s own words. Revisit them after courses, projects, or summer experiences.
Separate field, career, and professional track
Create three columns. In the first, list subjects the student wants to study. In the second, list possible careers. In the third, list professional pathways such as medicine or law. Connect the columns, but do not collapse them. This exercise prevents “doctor” from being treated as an undergraduate major and helps students see several academic routes to a goal.
Build an evidence chain
For each possible field, identify the evidence already present. Evidence can include coursework, independent reading, research, employment, caregiving, creative work, a project, a competition, or sustained curiosity. The goal is not to manufacture a perfect résumé. It is to see which interests have been tested and which remain hypothetical.
Use four prompts.
- What experience first made this field interesting?
- What have you done since then?
- What question remains unresolved?
- What would you want to study next?
A field can be genuine even when the evidence is new. The answers simply reveal the current level of exploration.
Record changes without treating them as failures
Maintain a dated academic-interest log. After a meaningful experience, write three sentences about what became more interesting, less interesting, or newly connected. This produces a clearer history than reconstructing motivation during application season.
When an interest changes, ask whether the shift is narrowing, integration, translation, reorientation, or replacement. The category is less important than the explanation. Students should be allowed to say, “I learned enough to revise my plan.”
Research colleges at the pathway level
For every college, verify five points on official pages.
- Is admission to the university, an undergraduate school, or a specific major?
- Is the intended field binding, nonbinding, or part of direct admission?
- Are there capacity-constrained programs?
- What is the process for changing or entering the field later?
- Would the student attend if admitted outside the first choice?
Record the URL and access date. Policies can change by cycle.
Choose application fields honestly
The application field should reflect a direction the student can genuinely explain and would be willing to pursue. Strategic considerations may be part of planning, especially where admission is program-specific. They should not turn a field the student does not want into a false declaration.
If a student has several connected interests, the application can name a current academic home and explain the broader questions around it. Specificity can coexist with openness.
Use essays to show development, not pressure to sound certain
An academic-interest essay can show how a question developed, what evidence deepened it, and why the college’s resources are relevant. It need not promise a lifelong career. Avoid claims that sound irreversible unless they are true.
A useful structure is past question, tested experience, current direction, next inquiry, and specific college fit. This structure shows intellectual movement while keeping the focus on what the student has actually done and wants to do next.
Recheck before submission
Before each application is submitted, confirm the selected field, undergraduate school, alternate choice, and related essays. Make sure the labels and narrative align. Document the submitted choice separately from planning notes. This final verification would also make future research stronger.
Guidance for Counselors and Researchers
Keep the student’s own words and the dates of the decisions clear in private notes. Distinguish what the student wants to try from what a counselor recommends. Verify the selected major separately if the question concerns submission.
Read the detailed method and checks
Speaker attribution should be explicit. “Counselor suggested economics” and “student chose economics” are different events. A dropdown or short tag can preserve that distinction without turning advising into a research interview.
Final application verification deserves its own field. The absence of a verified submitted choice was the largest limitation in this study. A submission checklist could capture the selected undergraduate school, first-choice field, second choice, submission date, and any later change. This should be stored securely and used for research only with appropriate authorization.
Researchers should define the outcome before counting. Possible outcomes include first expressed interest, current intended field, application-selected field, submitted field, enrolled field, and declared college major. Combining these outcomes produces misleading transition rates.
Future internal studies could use a prospective design. Students could complete a brief academic-interest inventory at entry, at defined grade milestones, before application submission, and after enrollment. The instrument should allow multiple interests and confidence ratings. A separate question should capture the durable problem or purpose connecting the fields.
To study advising effects, a future project would need more than before-and-after records. Options include a carefully matched comparison group, a staggered introduction of a structured exploration module, or a within-student design with repeated standardized measures. Any design involving minors should receive appropriate consent and privacy review.
Qualitative analysis would remain valuable. Numbers can show that a label changed. They cannot fully show whether the change reflected discovery, pressure, a new opportunity, or a more accurate vocabulary. Structured interviews could examine those mechanisms while avoiding the assumption that stability is always success.
Limitations
Only seven of the 15 histories supplied two usable choices. The observation gaps differed, and none supplied a verified submitted major. The findings describe a small group of advising clients, with no independent group for comparison.
The second limitation is sample size. Seven paired histories allow transparent descriptive analysis, but they do not support stable subgroup estimates. The public report suppresses subtype and grade-specific counts. No inferential tests were appropriate.
The third limitation is retrospective measurement. The records were created for advising. Wording, frequency, and speaker attribution varied. The study did not ask students standardized questions at fixed intervals.
The fourth limitation is endpoint finality. No later endpoint was verified as a submitted-major choice under the study’s rule. The study therefore concerns intended directions and application planning.
The fifth limitation is unequal observation time. Intervals ranged from 16 to 424 days. Short and long intervals may capture different developmental processes. The median does not remove that variation.
The sixth limitation is incomplete observation. A missing mention is not evidence that an interest disappeared. A student may have changed privately, discussed a field elsewhere, or held several interests simultaneously. The analysis uses only what was documented and authorized.
The seventh limitation is classification judgment. Terms such as interdisciplinary development and continuity bridge require interpretation. The private ledger preserves the rationale, but a second independent coder was not used. Future work should report inter-rater agreement.
The eighth limitation is adviser influence. The records contain advice as part of the process. Even when a student endorsed a direction, the study cannot separate independent development from the influence of advising, family, teachers, peers, or institutional constraints.
The ninth limitation is policy variability. Official university policies differ and can change. The examples in this report were verified for publication context, but families must confirm the current cycle.
The tenth limitation is commercial conflict. The Ivy Institute studied its own clients and will publish the report. The company benefits when its work appears thoughtful or useful. Transparent denominators and limitations reduce that risk but do not remove it. Independent replication would be stronger.
Finally, the external literature is contextual. National college-level data and peer-reviewed adolescent-development studies use different populations and measures. Similarity in concepts does not verify the internal sample.
Conclusion
In every one of the seven eligible comparisons, an earlier interest remained visible in the student’s later plans. For a parent, asking what has stayed interesting can open a better conversation than asking the student to defend a change of label.
The result is modest and useful. It suggests that families should look beneath the label. A student can change the name of a field while preserving the question that matters. A student can keep the same broad interest while substantially deepening it. Different college structures can produce different application labels for the same underlying direction.
The study also shows the importance of evidentiary restraint. Fifteen histories were screened, eight could not support the paired question, observation intervals varied, and no submitted-major endpoint was verified. The report therefore does not claim a general rate, a causal advising effect, or an admissions benefit.
The practical standard is honest coherence. Students should explore, document what they learn, distinguish fields from careers, verify program policies, and choose application pathways they would genuinely pursue. Colleges ask for a present direction. They do not erase the developmental reality that thoughtful students can learn enough to change their minds. (The Ivy Institute, internal materials, data, and information.)
References
Germeijs, V., & Verschueren, K. (2006). High school students’ career decision-making process: A longitudinal study of one choice. Journal of Vocational Behavior, 68(2), 189–204. https://doi.org/10.1016/j.jvb.2005.08.004
Hirschi, A. (2009). Development and criterion validity of differentiated and elevated vocational interests in adolescence. Journal of Career Assessment, 17(4), 384–401. https://doi.org/10.1177/1069072709334237
Jiang, S., & Simpkins, S. D. (2024). Examining changes in adolescents’ high school math and science motivational beliefs and their relations to parental STEM support and STEM major choice. Developmental Psychology, 60(4), 693–710. https://doi.org/10.1037/dev0001683
Jiang, S., Simpkins, S. D., & Eccles, J. S. (2020). Individuals’ math and science motivation and their subsequent STEM choices and achievement in high school and college. Developmental Psychology, 56(11), 2137–2151. https://doi.org/10.1037/dev0001110
Massachusetts Institute of Technology Admissions. (n.d.). Biographical and personal information: Field of study. https://mitadmissions.org/apply/firstyear/biographical-information/
National Center for Education Statistics. (2017). Beginning college students who change their majors within 3 years of enrollment (NCES 2018-434). U.S. Department of Education. https://nces.ed.gov/pubs2018/2018434/
Negru-Subtirica, O., & Pop, E. I. (2018). Reciprocal associations between educational identity and vocational identity in adolescence: A three-wave longitudinal investigation. Journal of Youth and Adolescence, 47(4), 703–716. https://doi.org/10.1007/s10964-017-0789-y
Rogers, M. E., & Creed, P. A. (2011). A longitudinal examination of adolescent career planning and exploration using a social cognitive career theory framework. Journal of Adolescence, 34(1), 163–172. https://doi.org/10.1016/j.adolescence.2009.12.010
The Ivy Institute, internal materials, data, and information.
Tracey, T. J. G., Robbins, S. B., & Hofsess, C. D. (2005). Stability and change in interests: A longitudinal study of adolescents from grades 8 through 12. Journal of Vocational Behavior, 66(1), 1–25. https://doi.org/10.1016/j.jvb.2003.11.002
University of Illinois Urbana-Champaign Undergraduate Admissions. (n.d.). Admission review policy. https://www.admissions.illinois.edu/admission-review-policy/
University of Illinois Urbana-Champaign Undergraduate Admissions. (n.d.). How to apply: First-year. https://www.admissions.illinois.edu/first-year-apply/
University of Washington Office of Admissions. (n.d.). Admission to majors. https://admit.washington.edu/apply/admission-to-majors/
University of Washington Office of Admissions. (n.d.). Computer science. https://admit.washington.edu/majors/computer-science/
Yale Undergraduate Admissions. (n.d.). Majors and academic programs. https://admissions.yale.edu/majors
