The Ivy Institute | College Admissions Research | November 2026
Executive Summary
Counting schools can hide an important change in a college list. We compared two dated working lists for each of nine students. The earlier lists contained 174 student–college choices in total. The later lists also contained 174. Yet some colleges appeared on a later list for the first time, and others disappeared. The additions and removals offset one another across the group.
A student–college choice means that one student listed one college. If three students listed the same college, that counted as three choices. A campus listed twice by the same student counted once. These figures tell us what was under consideration at two points in time. They do not mean that 174 applications were submitted, that the students had 174 distinct colleges between them, or that every college ever considered was recorded.
The nine pairs came from the 2025 and 2026 fall application seasons. The median gap between the earlier and later observations was 19 days. We could compare the names of colleges and calculate changes in list size. We could also see academic interests, geographic options and cost questions in the broader materials. We could not reliably assign a personal reason to each addition or removal. In particular, a college’s location does not explain why a student considered it, and a discussion of price does not establish a family’s finances.
For a family revisiting a list, a useful question is “Which choices changed, and what do we still need to learn about them?” A second question is “Which of these schools would you actually apply to and attend under conditions that work for you?” The first can be answered by comparing two lists. The second belongs to the student and family.
The Ivy Institute conducted and funded this study and publishes it as a provider of paid admissions advising. The nine students had working lists that we could compare. They are not representative of all applicants. Source for internal findings: The Ivy Institute, internal materials, data, and information.
Table of contents
- What families mean by a college list
- How the comparison worked
- What changed across the pairs?
- Geography, academic fit and cost
- Which colleges actually received applications?
- What other research tells us
- A practical way to revisit a list
- Four decisions that a list count misses
- Limitations and conclusion
- References
What families mean by a college list
Picture two students who both say they have twelve schools on their list. One has twelve colleges they are actively preparing to apply to. The other has six they intend to pursue, four on which they still need information, and two a relative suggested. Their number is the same, but their next week may look quite different.
Even one student’s list can have several purposes. An early list may collect possibilities without committing the student to anything. A shorter set may guide campus visits. A working application list may organize essays and deadlines. A list of schools with submission confirmations answers a different question again. A sound comparison begins by asking whether the two lists represent the same kind of choice.
That issue shaped the research. We wanted to know how lists change during the application season. We could not simply subtract the number of colleges discussed in a September meeting from the number discussed in an October meeting. People rarely repeat every school under consideration every time they speak. A school absent from the October conversation could still be central to the plan.
Instead, we looked for two dated views of the same student’s identifiable working list. We asked which colleges were present at both points, which appeared only later, and which appeared only earlier. We separately examined whether the available information explained a change or confirmed that a student had applied. Where it did not, we kept those questions open.
How the comparison worked
A previously assembled inventory contained 47 entries for student applications. At least two entries concerned the same applicant in different rounds of work. We set aside the known repeat before looking for pairs. We did not use 47 as a denominator for distinct students or for a rate of list change.
We selected 26 candidates for a closer check, in two waves. Selection was deliberate. We looked for an identifiable working list, including both straightforward lists and more complicated arrangements with earlier options, priorities or recommendations. We did not choose cases by calculating their eventual change in size. This process favored students whose application planning was recorded clearly enough to inspect. A family that discussed its entire plan in conversation alone would be less likely to enter the paired sample.
We looked for observations from July 1 through October 31 within the same 2025 or 2026 season, at least seven days apart. For each candidate we checked the earliest and latest available dated views in a qualifying season. The seven-day gap was an analytic choice to keep a quick formatting edit from being treated as a second period of planning. It does not imply that real decisions take at least seven days.
Twelve candidates had a date pair we could inspect under the time rule. For nine, we could identify a comparable named set at both ends. Other candidates lacked a suitable date pair, had an empty or noncomparable earlier list, or contained material that referred to a different application cycle. A blank earlier list was treated as missing, never as zero colleges. We included one paired comparison per student. The Ivy Institute, internal materials, data, and information.
What went into a named set
We counted the main set the student was using at that point. If a clear heading distinguished a current list from a previous list, we used the current set. If a student kept a named shortlist and explicitly described it as a list of top schools, we could compare the same shortlist across dates. It was not relabeled as an exhaustive list of intended applications. General reference lists, optional possibilities kept separately, counselor suggestions and lists of interesting programs were excluded from the primary comparison.
We kept colleges separate when they were different undergraduate institutions. A parent university and an explicitly named undergraduate program generally counted as one institution, provided they clearly referred to the same campus. Two separate campuses in the same public system counted as two when the student named both. We did not expand “all campuses” or a broad system label into an invented set of schools.
We matched obvious spelling, punctuation and common-name differences within a student’s own materials. Changing a college’s annotation from “early” to “regular” did not create a new institution. Moving it from a reach section to a target section did not remove and re-add it. We treated a consistent but unresolved campus abbreviation conservatively and excluded it from both dates. We report below how including that label would affect the totals.
These choices defined list membership for this study. They did not certify the accuracy of every program name, deadline, admission classification or academic claim that appeared alongside the college. A student’s label of “likely,” for example, is not evidence of the probability of admission.
How to read an addition or removal
A college counted as an addition when it appeared in the defined later set and was absent from the comparable earlier set. A removal meant the reverse. Those words describe the visible difference. They do not prove that the student consciously rejected a college, changed their mind on a particular day, or never returned to it. A list may lag behind a decision.
The same caution applies to a match. If the names were unchanged at both ends, the student could still have changed an application round, learned more about the major, revised priorities or narrowed interests between those dates. Two snapshots show the endpoints. They cannot capture every conversation or every intermediate version.
For families, one advantage of comparing named sets is that a simple mistake becomes easier to spot. If two aliases for the same school appear as an addition and a removal, review the labels before interpreting the change. If an entire public system appears where specific campuses once did, the scope of the list may have changed. It may be possible to resolve that with the student. If it cannot be resolved, “unknown” is a more accurate result than a guessed change.
What changed across the pairs?
Earlier named choices (chart scale) — 174 of 174
Later named choices (chart scale) — 174 of 174
| Measure | Earlier observation | Later observation |
|---|---|---|
| Students with comparable lists | 9 | 9 |
| Named student–college choices | 174 | 174 |
| Mean named choices per observed student | 19.3 | 19.3 |
The average uses 174 divided by nine. It is an average for this selected sample only. It is not a recommended list length. The research includes some long working lists and a named shortlist that was still open to revision. Rounding the average to one decimal makes it easier to read but does not change the underlying counts. The Ivy Institute, internal materials, data, and information.
Comparing each college name showed movement in both directions. Across the group, the number of additions equaled the number of removals, giving a net change of zero. The students contributing the changes formed too small a group for us to publish exact change counts or a campus-by-campus distribution. We withhold the complementary count of retained choices as well, because it would allow readers to recover the suppressed counts by subtraction.
This is a concrete example of why it helps to count both names and totals. If we had examined only average list size, the report might wrongly suggest that nothing changed. If we had recorded only the fact that some names changed, the report might make movement look more extensive than the available evidence supports. The two observations belong together. Total list size stayed steady across the group, while the composition of the lists changed.
What a net change of zero looks like
Consider a fictional student whose earlier list contains North College, Harbor University, Valley College and Prairie Institute. A later list contains North College, Harbor University, Valley College and City University. The size stayed at four. Three schools were retained, one was removed and one was added. There were two membership changes even though the net size change was zero.
Now imagine two students. One adds a school and keeps every previous choice. The other removes a school and adds nothing. Combined list size is again unchanged, even though each student’s own total changed. Both examples are possible ways an aggregate can remain steady. They are invented demonstrations of the arithmetic, not descriptions of individual Ivy students.
The reproducible calculation is straightforward. Within each student, count the colleges present at both dates, the colleges present only at the earlier date and the colleges present only at the later date. The earlier size must equal retained plus removed. The later size must equal retained plus added. The difference in size must equal additions minus removals. We checked these identities before adding the students’ results together.
How much did the timing matter?
The shortest observation gap in the nine pairs was 10 days, the longest was 109 days and the median was 19 days. These are intervals between dated views of a list. They are not periods spent actively deciding about colleges. A student with an unchanged list over a long interval may have done extensive research. A student with different names over a short interval may have made the decision before the earlier date and updated the list later.
A common uncertainty is whether a list counted as “initial” really is the student’s first list. In many cases we could only establish the earliest dated comparable view that was available within the study window. That could already reflect months of exploration. Similarly, the later view fell in the application season before November 1. For the full set, we could not establish that it preceded each student’s first application submission. This is a study of working lists in the run-up to deadlines. It is not a reconstruction of every student’s first idea and final submitted set.
That boundary is especially important for a reader comparing the result with their own family. A student who is still exploring in spring may add and remove more names than a student whose working list is largely settled by October. Our sampling procedure gave us a clearer view of the latter part of planning. It cannot establish which months produce the most changes.
What the findings depend on
We repeated the calculation treating one consistently written but unresolved campus label as a single matching item at both dates. The total would become 175 at each endpoint, with the same net change of zero. We kept the more conservative 174 as the principal result because we could not confidently verify the campus identity. This check illustrates the difference between a measurement choice that shifts the level and one that changes the conclusion.
We also separately checked the clearly identified top-school shortlist. Omitting a named shortlist that was not presented as a final plan leaves equal combined totals at both dates. We avoid publishing the reduced totals because subtracting them from the main table would reveal the exact size of one student’s shortlist. The result on net size survives both checks. These are robustness checks on this sample, not evidence of a population trend.
The nine students form a selected group, not a probability sample. A national confidence interval would give the wrong impression of precision. Our validity checks instead include stable student identity, independent dates for each endpoint, consistent definitions of the active list, deduplication of aliases, explicit handling of uncertain labels, exact arithmetic checks and the two sensitivity tests above. One reviewer made the membership decisions. Independent agreement is unknown.
Geography, academic fit and cost
We reviewed the geography of the schools that appeared or disappeared and checked their locations against the institutions themselves. A geographic change is a property of the named set. A reason for that change belongs to the student. Even if a newly listed university is in another state or country, its appearance does not establish that the family wanted a longer journey, a different climate or a cheaper tuition classification. These alternatives may matter, but a map cannot tell us which one mattered here.
For the same reason, we do not publish the small set of individual origin and destination combinations. A school’s country and an unusual program, taken together, could make an applicant identifiable to people who know the family. We can report that geographic options were considered without presenting a particular student’s decisions as a public itinerary.
Academic information appeared alongside many institutional names. A student might specify an undergraduate school, a business program, a health-related field or an honors option. Such a detail can help explain what the student intended to investigate. It does not automatically prove why the college entered or left the list. To count an academic reason, we would need a clear statement linking a particular change to that academic feature. The available reasons were too sparse for a publishable ranking.
Cost information required similar care. We encountered explicit attention to application-fee timing, scholarship consideration and questions about the possible cost of attending. These are three distinct issues. Saving an application fee can affect when a student applies without changing the schools they like. Seeking scholarship consideration can add a deadline without making an award likely. A family wondering which college might cost more has raised a question, not necessarily made a decision. We did not infer income, resources, aid eligibility or affordability from these discussions.
One practical way to avoid overclaiming is to use the student’s own distinction between a concern and a reason. “We need to check whether this major is available” is an open question. “We removed the school after finding the major was unavailable” is a reported decision with a reason. “The school was gone from the later list and the major was unavailable” contains two observations but still lacks the student’s explanation. We did not turn the last pattern into a motive.
A school can remain while its role changes
A college considered for two different majors can stay on both lists while the application changes considerably. A student can move it from early to regular, choose a different undergraduate school, add an honors application, or treat it as a lower priority. A membership comparison will label the college retained. It will not capture those other decisions unless they are measured separately.
Families may find it useful to keep a short note beside a retained school: “What would make me happy to attend here?” That answer may change even when the college name does not. An earlier answer could focus on one lab, while a later answer emphasizes access to an introductory course, a realistic commute or a more flexible major. Revisiting the answer is a way to learn whether the college still serves the student’s goals.
An adult looking at two unchanged lists might wonder whether the student has done any work at all. The study gives no basis for that judgment. Researching a difficult question and deciding to keep the school is a real decision. A family can recognize it by recording what was learned, rather than requiring a visible addition or deletion as proof of progress.
Which colleges actually received applications?
A planned list is an expression of intent. A submitted application is an action. A college’s confirmation or a clearly dated submission record is evidence of that action. The existence of a school-specific essay, a named place on a list, a completed draft or an internal “ready” label is insufficient on its own.
Some broader materials referred to individual submissions or described a particular application as sent. Those reports are not a complete series of receipts for all schools at both dates. We could not establish the full set of submitted applications for each of the nine students. We therefore report no conversion rate from list membership to submission and no rate of late or missed applications. We also do not assume that a college removed from a working list never received an application through another route.
For a family organizing its own records, the distinction can be made simply. Use separate labels for “considering,” “plan to apply,” “in progress,” “submitted” and “received or complete under the college’s instructions.” If the student changes their mind, close or update the planned item. If the student submits, retain the college’s confirmation. The family then has a usable history without having to remember what a check mark meant three months ago.
This kind of history would make future research stronger. It would allow researchers to compare initial and later preferences, actual submissions and the reasons students themselves gave at the time. It would also allow a family to see which extra requirements came with an application it truly intended to submit. Our study cannot fill those gaps retrospectively by interpreting a student’s silence.
What other research tells us
Amanda Pallais studied a change in the number of free ACT score reports students could send to colleges. Her research found that a small change in the cost and structure of sending scores influenced college-application behavior. It provides evidence that seemingly modest details of the application process can matter. It examined a different population and a particular policy change. Our study did not vary application fees, assign students to different costs or establish that a fee caused any observed list change. Pallais, Journal of Labor Economics, 2015
Caroline Hoxby and Christopher Avery studied application patterns among high-achieving students from low-income families. Their research documented substantial differences in the sets of colleges students applied to and examined geographic and informational context. That work is relevant because a list reflects opportunities students know about as well as their academic interests. It does not justify assigning a socioeconomic category to any Ivy student or assuming that families in a paid advising sample face the same circumstances. Hoxby and Avery, Brookings Papers on Economic Activity, 2013
Federal Student Aid’s guidance on evaluating aid offers helps families compare costs using the relevant components of the offer. A family’s estimate during list building and an actual aid offer after admission are different kinds of information. The estimate can help identify questions. Once offers exist, comparing grants and scholarships, expected costs and borrowing may change which colleges are viable. Our analysis ends too early to treat those later decisions as observed list changes. Federal Student Aid
Together these sources make it reasonable to examine cost, information and the structure of applications. They do not independently validate the 174 choices counted in these nine pairs. The evidence for those counts is The Ivy Institute, internal materials, data, and information.
A practical way to revisit a list
Ask the student to choose one older version and one current version. Write down the date on each, even if the date simply means “we reviewed it today.” Decide what the list contains. Is it every college the student has heard about, a shortlist for visits, or the set they are preparing to apply to? Put options still being researched in their own space if that makes the active list easier to understand.
| Family question | What to record |
|---|---|
| What did we add? | Name the college or campus precisely. Record the student’s reason if they have one. |
| What did we remove? | Note the decision in the student’s words. If the change was only a correction or duplicate, label it that way. |
| What did we keep? | Identify what still makes the school a realistic option and any new information that could change that view. |
| What remains undecided? | Give the open question a next step and an owner. |
| What will applying actually require? | Check the specific round, major, writing, materials and any separate program or scholarship deadlines. |
| What has actually been submitted? | Track the submission or receipt separately from the planning list. |
First ask whether the student would attend
A student can keep a college for months because it entered the list early, a friend mentioned it or someone described it as a sensible choice. Ask the student to imagine receiving an offer under realistic conditions. Would they want to investigate seriously? If so, what makes it worth that attention? If the answer is no, ask whether new information could change that response.
This conversation does not have to produce a deletion. A student who says “I don’t know enough yet” has identified a research task. They may need to understand the academic program, the cost estimate, the trip home or the everyday student experience. The next action can be smaller than deciding whether to apply.
For an adult, the harder part may be listening when the student’s answer differs from the adult’s. A school that looks appealing on paper may not offer the combination the student wants. Conversely, a student might value a feature an adult considers minor. The family can investigate the underlying facts together while allowing the student’s priorities to remain their own.
Compare an academic route, not just the name of a major
Two colleges can list a similar subject while giving students different ways to study it. One route might admit directly to a particular undergraduate school. Another might ask students to enter a broader college and declare later. A student who wants to combine two fields needs to know whether the curriculum and timetable make the combination realistic.
Begin with a concrete learning question. What course, method or experience would the student hope to try early in college? Then look at the admission route, introductory sequence and prerequisites. If the student changes interests, what alternatives would remain available? The answers can help the family judge fit even when both institutions use the same major name.
If the information is unclear, label it as a question. A college need not leave the list the first time a requirement proves confusing. It may move into a “confirm before applying” category. The point of the exercise is to prevent a provisional assumption from quietly becoming the foundation of an application.
Talk about geography in terms the family can use
“Close to home” can mean a short drive, an affordable trip, a reliable train route, a familiar climate or the ability to return quickly during a family emergency. Two campuses in the same region can feel very different under those criteria. A student can prefer an urban environment while caring more about travel time than state boundaries.
Try asking which trips the family actually expects to make. Would the student want to come home for a short break? Would a parent or other family member need an accessible route to visit? Are internships, performances, fieldwork or community connections tied to a particular location? These questions make the geography decision specific to the family. They do not turn geography into an automatic measure of academic fit or affordability.
If an institution moves on or off the list after this discussion, write down the actual reason only if the student wants to record it. A simple “travel time does not work for me” is more informative than an outsider’s inference that a student rejected a whole region.
Keep costs in more than one column
The cost of applying and the cost of attending are different. A fee waiver may remove an application cost without making attendance affordable. A strong need-based or merit offer might make attendance feasible, but the amount may be unknown when the student is building the list. Comparing sticker prices alone can be misleading. Federal Student Aid offers a framework for examining the components of an aid offer once a family has one. Federal Student Aid
A family can write down what it knows, what it estimates and what remains unknown. For example, the school may publish a cost of attendance, the family may have an estimate from an official calculator, and a scholarship decision may remain uncertain. These should not be collapsed into a single promised price. If financing would depend on an award, the student can decide whether applying for the award is realistic and worthwhile.
This conversation can be sensitive. Families do not owe a public account of their finances. A student can record that a cost question remains open without publishing numbers or assigning blame. In research, the same restraint is essential. A public tuition figure is not evidence of what a particular household can afford.
Give possible additions a fair test
When someone recommends a college late in the process, ask what it would add to the student’s choices. Perhaps it offers a different academic pathway, a more manageable travel option or an opportunity the student has not found elsewhere. Perhaps it simply repeats features already represented on the list. The answer can be worth investigating without immediately starting another application.
The student can identify one fact that would make the recommendation genuinely actionable. If the suggested program is restricted to a different undergraduate unit, that matters. If the scholarship deadline has passed, that may change the reason the family was considering the school. If the student would apply only because someone called the college prestigious, they may want to reconsider before adding the work.
The goal is not to keep the list at a fixed number. Sometimes adding a college makes the set more useful to the family. Sometimes removing one gives the student time to understand another choice more fully. The study provides no universal target list length or optimal mix.
Four decisions that a list count misses
Keeping a school after learning more. Imagine an applicant interested in both environmental science and public writing. The family wonders whether students in the science program can contribute to a campus publication. The student investigates, finds a workable route and keeps the college. Neither the earlier count nor the later count changes. Yet the family has replaced a guess with a specific reason for applying. An unchanged list can reflect work.
Moving an option out of the active group. Another student may still like a college but recognize that the extra audition or portfolio would require more preparation than is realistic this fall. The student moves it to a separate group to revisit. If the active list is what the family is counting, the number changes. If the family counts every school under consideration, it does not. Writing the purpose at the top of each list makes both answers easier to interpret.
Changing the undergraduate route. A student may still be committed to a university but realize that a different undergraduate school suits their interests better. The college name remains. The requirements, essays or academic opportunities may change. A family should revisit the relevant official instructions even if no name is added or removed. A study focused only on membership, including this one, cannot quantify how much that change affected the application.
Deciding after an offer arrives. A student’s preferred set may change after they learn more about actual costs or admission to a specific program. That is an important college-choice decision, but it occurs at a different stage from preparing a working application list. This study did not examine choices among offers. Families can still plan for that later decision by preserving the questions that remain unanswered in the fall.
A ten-minute meeting that does not need to produce a deletion
Choose one school that the student feels particularly sure about and one that still raises a question. Ask the student to describe the most attractive feature of each in a single sentence. Then ask which fact, if it were different, would most change their view. A strong answer might involve curriculum access, the route into a major, the cost estimate, travel time or the kind of support available.
This exercise is not a test of whether the student has researched enough. It helps the family choose a useful next step. A student who says “I like the engineering program, but I can’t tell how students enter my intended major” has identified a specific question. A student who says “I would only attend if a particular award came through” has identified a financial condition. Neither statement has to become a final decision that afternoon.
At the end, record a date for the next look only if a future decision requires one. A parent’s well-intentioned weekly reminder can become distracting after the important question has been resolved. A family may prefer to revisit the list after a school visit, after reading the current requirements or when a counselor can answer a remaining question. The plan can reflect the student’s actual decision process.
How to handle a list drawn up by several people
Students may receive advice from parents, teachers, counselors, friends and relatives. All can contribute something worthwhile. The difficulty comes when a combined list no longer distinguishes a recommendation from a choice the student has made. A table of possible names can grow rapidly without showing what the student intends.
Ask each person suggesting a school to give one reason that connects to the student. Then give the student room to respond. They might immediately see the connection, want to investigate a fact, or decide the school adds little. The recommendation can remain in a research section without becoming an active application. That small separation gives the student control over the workload and a clearer voice in the decision.
In our research, this distinction prevented an earlier suggestion from appearing falsely as a later removal. We used the active named set where the distinction was clear. Where the boundary between suggestions and decisions was unclear at the earlier date, we left the pair out of the quantitative comparison. A family organizing its own papers can use the same idea without keeping anything more elaborate than a dated note.
If an application is already in progress
The fact that a student has invested time in an application can make revisiting the choice uncomfortable. The family can ask two separate questions. Is the school still a realistic option the student would consider attending? How much work remains, and what would that work displace? The first concerns the value of the choice. The second concerns the present cost of keeping it active.
A student may decide to finish because the remaining work is small and an unresolved academic question can be answered later. Another may decide to stop because a clear program requirement or personal constraint changes the decision. The study does not rank those responses. It also cannot infer from a removal whether a student had already invested an hour or several weeks in the application.
If the family decides to stop, preserve any work the student wants to reuse and close out reminders that no longer apply. If the family decides to continue, record the next requirement and any unresolved condition. Both decisions become easier to understand later when the family can see why it made them.
What a stronger future study would collect
To learn more about why lists change, a future study could invite students to record their initial set before advising begins and ask for updates at agreed intervals. The student could say whether an addition was their own choice, a recommendation under investigation or a replacement for an earlier option. They could decline to explain private financial or personal considerations. The study would need appropriate consent and clear limits on how those explanations could be published.
At the end, a researcher could compare the final plan with verified submissions and ask which differences were intentional. Two reviewers could independently match college and campus names, then resolve disagreements without seeing admissions outcomes. With enough participants and privacy protection, the study might distinguish changes in distance, type of institution, academic route or financial expectations. Those measurements require additional evidence. They cannot be recovered by treating the admissions outcome or the college’s published price as a proxy for the family’s earlier reasoning.
Limitations and conclusion
The paired analysis contains only nine students. Selection favored clear, dated, comparable working lists from one paid advising organization. The earliest date in a pair may already follow substantial exploration. The later date is a dated planning observation, not necessarily the last list before the first or final submission. We did not know the complete application outcome for each named school.
Different students also used their lists differently. We made explicit decisions about which set was active, how to treat separately labeled possibilities, when to combine aliases and when an unresolved campus label should be withheld. These decisions are documented for review. The sensitivity checks preserve the result on net size but do not establish that all possible definitions would produce identical results.
Geography, academic fit and cost were considered only when supported by the available evidence. We did not infer a student’s finances or motives from a college’s reputation, location, program or price. Very small groups and information that could reconstruct them were withheld. The result is less detailed at the individual level and safer for the families whose experiences informed it.
Across the nine comparable pairs, total planned choices stayed at 174 while schools moved onto and off lists. If a family wants to understand how its plan is changing, the useful comparison is between two dated sets of names, followed by a conversation about why each choice still belongs. Submission becomes a separate question, one that can be answered with a confirmation when the student is ready. The Ivy Institute, internal materials, data, and information.
