# Summer Melt Mitigation ## expand\_more ## Key Issue Overview Summer melt refers to the phenomenon where high school graduates who intend to go college, and who often have already been accepted to college, do not matriculate in the fall. Nationally, about 10% to 40% of these college-intending students are affected by summer melt, with higher shares of summer melt among low-income, first-generation, and community college-intending students ([Castleman & Page, 2014](https://www.jstor.org/stable/26612158)). The summer months after high school graduation can be challenging for students, as they often no longer have access to their high school counselors and have not yet established connections with college support staff. Yet during this transition period, there are numerous tasks that students must complete related to the enrollment process, such as registering for and attending orientation; sending in immunization forms; taking academic placement tests; and completing college loan counseling. Students may also encounter financial obstacles, such as needing to secure additional funding to cover gaps between their financial aid package and their college costs. ### Research Evidence expand\_more Addressing the issue of summer melt has spurred a decade of research into effective interventions. In these interventions, supports are offered in the summer to help students complete necessary tasks and address barriers to matriculation. Proactive Counselor Outreach Text Message Outreach Artificial Intelligence (AI)-Assisted Support from Colleges ## Proactive Counselor Outreach **What the Studies Tested:** - These studies examined the impact of counselors proactively offering one-on-one advising, either in person or virtually, to college-intending high school graduates during the summer following graduation. These studies tested offering support from high school counselors, college staff, and near-peer mentors. **What the Studies Found:** - Offering students one-on-one advising increased fall college enrollment, including an increase at four-year colleges. Results showed stronger effects for specific populations of students, including low-income students and Latinx males. While one study found greater effects among students with less defined college plans and limited access to college-planning support during the school year, another study observed the largest effects among students who had more established college plans and had already taken steps towards enrollment. *Click the dropdowns below to read summaries of each study included in the overall synthesis. Unless noted otherwise, all reported effects are statistically significant at the p<.05 level. Studies are linked (see author name and publication date). When available, we link an open access version of the study*. [Explore our methodology](/rc-methodology) Open all sections Close all sections ### Counselor Outreach in Providence, Rhode Island (Castleman et al., 2012) expand\_more **Study:** **Stemming the Tide of Summer Melt: An Experimental Study of the Effects of Post-High School Summer Intervention on Low-Income Students' College Enrollment** ([Castleman et al., 2012](https://www.tandfonline.com/doi/abs/10.1080/19345747.2011.618214)) **Intervention:** This study was the initial pilot test of a summer melt mitigation counseling intervention. The counselors proactively reached out to treatment group students and “went to considerable lengths to connect with students” (p. 7), reaching out via phone, email, instant messaging, and Facebook. There were 3 broad areas of focus that the summer advising aimed to address: (1) any remaining financial aid gaps, (2) informational barriers and ensuring students were aware of required tasks, and (3) any social and/or emotional barriers to enrollment. **Context:** - The study sample was class of 2008 graduates from small urban high schools in Providence, Rhode Island that serve a large proportion of low-income students. Notably, the context of these schools—part of the Big Picture network of schools—was distinct from many public schools. These were public college-prep schools that admitted students via lottery. The schools had more dedicated college counseling staff than is typical and required all students to submit at least one college application before graduating. For the summer counseling intervention, two college transition counselors from the high schools were hired to work full time over the summer. **Research Design:** Randomized Controlled Trial **Findings**: *Receiving the outreach...* - Led to a marginally significant (p<0.1) 14 percentage point increase in on-time fall enrollment. - Led to a 19-percentage point increase in students following through on their college intentions at the time of graduation. - Increased the rate of students enrolling at a four-year institution and students enrolling full-time by 15 percentage points each. ### Counselor Outreach in Boston, MA and Fulton County, GA (Castleman et al., 2014) expand\_more **Study:** **The Forgotten Summer: Does the Offer of College Counseling after High School Mitigate Summer Melt among College-Intending, Low-Income High School Graduates?** ([Castleman et al., 2014](https://www.maearlycollege.com/wp-content/uploads/2021/08/The-Forgotten-Summer-Does-the-Offer-of-College-Counseling-After-High-School-Mitigate-Summer-Melt-Among-College-Intending-Low-Income-High-School-Graduates.pdf)) **Intervention**: High-school based counseling offered to college-intending high school graduates during the summer after graduation. Counselors proactively reached out to students, using numerous methods such as phone, email, text, and Facebook, encouraging them to schedule a 1:1 meeting. **Context:** - Study sample was class of 2011 high school graduates from Boston, Massachusetts and Fulton, County Georgia. The intervention was implemented in partnership with uAspire in Boston and Fulton County Schools in Georgia. There was variation in the intervention implemented across the two sites. - For the Boston sample, uAspire advisors were provided with a protocol for providing outreach. This protocol included steps to review students’ financial aid letter and address any unmet need as well; review key pre-enrollment tasks and their deadlines; and assess whether students were experiencing social or emotional challenges. Students in the Boston sample had had access to uAspire advisors in their school throughout their senior year. - For the Fulton County sample, Fulton County School counselors staffed the intervention. Though these counselors received supplemental training ahead of the summer counseling intervention, they did not receive a protocol and were encouraged to use existing protocols and approaches used during the academic year. - Though all students had the option to meet in-person, most meetings among Boston graduates occurred in person and most meetings with Fulton County School graduates were conducted over the phone. Students in the Boston sample received a $25 gift card as an incentive to attend an in-person meeting. **Research Design:** Randomized Controlled Trial **Findings**: - Overall, the offer of summer counseling increased the probability of fall enrollment by 3 percentage points, with larger effects among the Boston sample that primarily received in-person support. **Subgroup Findings:** - Low-income students experienced the most substantial benefits of the intervention with increases in fall enrollment ranging from 8 to 12 percentage points. ### College Matriculation Support from High Schools Versus Colleges (Castleman et al., 2015) expand\_more **Study:** **Stay Late or Start Early? Experimental Evidence on the Benefits of College Matriculation Support from High Schools Versus Colleges** ([Castleman et al., 2015](https://www.acenet.edu/Documents/Stay-Late-or-Start-Early-Summer-Melt-Article.pdf)) **Intervention**: This study tested whether students were more receptive to summer counseling outreach from high –school- versus university-based counselors. In both cases, the counselors reached out to students via phone, email, and/or text and offered students help with completing matriculation tasks such as finalizing financial aid plans and registering for orientation and placement tests. Students could schedule 1:1 meetings with the counselors either in-person or over the phone. **Context**: - The study sample was 2012 graduates from Albuquerque Public schools who had been accepted to the University of New Mexico (the state’s public flagship university). - Students were assigned to the university-based counselor, high-school-based counselor, or a control group. All counselors participating in the intervention were school-year Albuquerque Public School counselors, but they were assigned to be based at either the university- or high-school for the summer intervention. The counselors worked for 10-20 hours per week for 5-6 weeks. They received pre-intervention training on topics such as reviewing financial aid award letters and were provided checklists and other resources to support their summer advising efforts. **Research Design**: Randomized Controlled Trial **Findings**: - There was no effect of summer counseling (when looking at the two treatment arms separately or combined) on fall enrollment among the full sample of students. **Subgroup Findings**: - The intervention did increase fall matriculation among Latino male students by 9 percentage points. This is a population of students underrepresented at the university and who face higher rates of summer melt. Among these students, the intervention was more effective when counselors were university-based. ### Text Messages and Peer Mentor Outreach (Castleman & Page, 2013) expand\_more **Study:** **Summer Nudging: Can Personalized Text Messages and Peer Mentor Outreach Increase College Going Among Low-Income High School Graduates?** ([Castleman & Page, 2013](https://www.evanstonscholars.org/uploads/1/4/8/7/14874378/castleman_page_-_summer_nudging_-_april_2013.pdf)) *This study was published as* [*"Summer Nudging: Can Personalized Text Messages and Peer Mentor Outreach Increase College Going Among Low-Income High School Graduates?"*](https://www.sciencedirect.com/science/article/abs/pii/S0167268114003217) *in 2015 but is behind a paywall. This study examined multiple interventions and is also included in the “Text Message Outreach” section.* **Intervention**: In this peer mentoring intervention, current college students provided summer outreach and support to recent high school graduates. The college students conducted the initial round of outreach and support, providing information and guidance to college-intending graduates as able, and connected students to professional support when needed. The peer mentors asked students about the status of key fall matriculation tasks such as applying for the FAFSA, reviewing their financial aid letter, and registering for orientation and placement tests. The peer mentors then scheduled 1:1 meetings with students to assist them with any incomplete tasks. **Context**: - The study sample was college-intending class of 2013 high school graduates from public schools in Massachusetts (Boston, Lawrence, and Springfield) and charter schools in Philadelphia, PA. - The intervention was implemented in collaboration with uAspire in Massachusetts (an education non-profit based in Boston, MA and with Mastery Charter Schools in Philadelphia. Both educational agencies had existing partnerships in the targeted schools. Most students in the study sample were low-income students of color. **Research Design**: Randomized Controlled Trial **Findings**: *Having access to a peer mentor led to...* - A marginally significant (p<0.1) 4.5 percentage point increase in four-year college enrollment. - The strongest enrollment effects were among students with less established college plans and less access to college-planning support during the school year. ### Teacher and Counselor Outreach in the Southwest (Daugherty, 2012) expand\_more **Study:** **Summer Link: A Counseling Intervention to Address the Transition from High School to College in a Large Urban District** ([Daugherty, 2012]()) **Intervention**: This summer texting study built on prior summer melt mitigation counseling interventions by developing and testing a district-wide, lower cost summer counseling intervention with district staff, including teachers and counselors, reaching out to students. Staff offered assistance and guidance with whatever areas students needed, including helping students register for housing; providing emotional support; and helping with financial aid applications. **Context**: - The study sample included class of 2011 high school graduates from a large urban school district in the Southwest who were planning to attend college and had applied and been accepted to at least one college. - The district budgeted two hours of support per students, although many students who participated in the counseling spent much less time with an advisor (with the majority receiving under ten minutes of support). The most common area of support requested was financial aid assistance. **Research Design**: Difference-in-Differences **Findings**: - The intervention increased the likelihood of fall college matriculation by 9 to 11 percentage points. - Effects were largest among students planning to attend four-year colleges and among those with more defined college plans who had already completed some enrollment steps, such as financial aid applications. ## Text Message Outreach **What the Studies Tested:** - Given the promising evidence summarized above from interventions offering summer counseling outreach, the next set of studies tested interventions using automated text messaging to provide college-intending students with reminders and support about college enrollment tasks that needed to be completed over the summer such as completing loan counseling, attending orientation, and completing health forms. These interventions also allowed for two-way communication, whereby students could request additional help from advisors. The goal was to reduce the per-student cost of the intervention by automating the initial student outreach. These studies tend to build on the findings of prior efforts, testing similar interventions in different contexts, and iterating to reduce costs or fine-tune the intervention. **What the Studies Found:** - Sending students texts about key milestones with the option to follow-up with an advisor increased fall college matriculation. The studies found varying effectiveness of the texting interventions across populations and contexts. This strategy seems to be particularly beneficial for students with limited access to college planning resources—those who have not yet received significant advising and support for college-going—and those less far along in the enrollment process by the time they graduate from high school. *Click the dropdowns below to read summaries of each study included in the overall synthesis. Unless noted otherwise, all reported effects are statistically significant at the p<.05 level. Studies are linked (see author name and publication date). When available, we link an open access version of the study*. [Explore our methodology](/rc-methodology) Open all sections Close all sections ### Text Messages and Peer Mentor Outreach (Castleman & Page, 2013) expand\_more **Summer Nudging: Can Personalized Text Messages and Peer Mentor Outreach Increase College Going Among Low-Income High School Graduates?** ([Castleman & Page, 2013](https://www.evanstonscholars.org/uploads/1/4/8/7/14874378/castleman_page_-_summer_nudging_-_april_2013.pdf)) *This study was published as* [*"Summer Nudging: Can Personalized Text Messages and Peer Mentor Outreach Increase College Going Among Low-Income High School Graduates?"*](https://www.sciencedirect.com/science/article/abs/pii/S0167268114003217) *in 2015 but is behind a paywall. This study examined multiple interventions and is also included in the “Proactive Counselor Outreach” section.* **Intervention**: Summer texting campaign involving ten automated text messages to students and their parents about key pre-matriculation tasks (e.g., complete housing forms, register for placement tests). The messages were tailored to provide campus-specific information and included offers for follow-up assistance from counselors if students wanted additional assistance. **Context**: - The study sample was college-intending students in the high school graduating class of 2013 from urban districts in Texas (Dallas) and Massachusetts (Boston, Springfield, and Lawrence). The intervention was implemented in partnership with the Dallas Independent School District in Dallas, TX and with uAspire in Massachusetts, an education non-profit based in Boston, MA. Most students in the study sample were low-income students of color. **Research Design**: Randomized Controlled Trial **Findings**: *Compared to students who did not receive the outreach, receiving the outreach led to...* - A 3-percentage point increase in fall enrollment at two-year colleges. There were no effects on overall college enrollment or enrollment at four-year colleges. **Subgroup Findings:** - The enrollment effects of the text message intervention were larger for students who had less access to college planning resources during high school and for those students who did not have a specific college plan by the end of their senior year. - The enrollment effects were larger for students in the middle of the GPA academic distribution *(defined as a GPA between 3.0 and 3.5. Based on senior year administrative GPA in some sites and self-reported cumulative GPA in other sites)*. ### Text Messages to Students and Families (Castleman & Page, 2016) expand\_more **Parental Influences on Postsecondary Decision Making: Evidence from a Text Messaging Experiment** ([Castleman & Page, 2016](https://education.virginia.edu/documents/epw50-parental-influences-postsecondary-decision-making2016-05pdf)) *This study was published as* [*"Parental Influences on Postsecondary Decision Making: Evidence from a Text Messaging Experiment"*](https://journals.sagepub.com/doi/abs/10.3102/0162373716687393) *in 2017 but is behind a paywall.* **Intervention**: Texts sent over the summer to students and their families encouraging them to complete summer tasks for college enrollment. The messages were customized as possible based on the college that a student as planning to attend. Text recipients could request additional assistance from an advisor via text. In addition to testing the effectiveness of the overall texting campaign, this study also examined whether texting both parents and students was a more effective strategy than texting just students alone. **Context**: - The study sample was college-intending students from the high school class of 2004 and their parents. The intervention was conducted in partnership with uAspire, an education non-profit based in Boston, MA. Study participants were recent graduates from Massachusetts (Boston, Springfield, Lawrence, and Fall River) and Florida (Miami) who had participated in uAspire programming during high school. **Research Design**: Randomized Controlled Trial **Findings**: *Compared to students/families not receiving the messages, receiving the text messages led to...* - A 3 percentage point increase in on-time college enrollment (spread across four- and two-year institutions) among all text recipients. **Subgroup Findings:** - The largest enrollment effects were among low-income and first-generation college students. - There was no difference in effects when messages were sent to both parents and students as opposed to when they were sent to students only. ### Testing Different Texting Framing Strategies (Kramer, 2020) expand\_more **Study: Experimental Evidence on the Effects (or Lack Thereof) of Informational Framing During the College Transition (**[**Kramer, 2020**](https://journals.sagepub.com/doi/full/10.1177/2332858420908536)**)** **Intervention**: Tested whether strategic framing of summer melt mitigation messages using three different behavioral frames (loss aversion, reduction of implementation ambiguity, peer support) impacts the effectiveness of the messages relative to the standard text messages sent the previous year. **Context**: - The author partnered with tnAchieves, a Tennessee nonprofit organization. tnAchieves sends text messages to Tennessee high school seniors eligible for the TN Promise scholarship, the state’s tuition-free two-year college program, during students’ senior year of high school through their first year of college. The sample for this study was college-intending high school graduates from the class of 2017 who were eligible for the TN Promise scholarship. **Research Design**: Randomized Controlled Trial **Findings**: - Results showed no effects of any of the behavioral frames compared to the standard messages. ### Text Messages to GEAR UP Participants (Linkow et al., 2021) expand\_more **Study: Study of College Transition Messaging in GEAR UP: Impacts on Enrolling and Staying in College (**[**Linkow et al., 2021**](https://ies.ed.gov/ncee/pubs/2021005/)**)** **Intervention:** Students were sent text messages during the summer following high school graduation with reminders of key milestones, alongside the offer of additional assistance from an advisor if requested. The texts also included positive / growth mindset messages. Students received 37 text messages that were tailored based on their intended college. **Context:** - The study sample was college-intending class of 2016 and 2017 graduates from high-need high schools that had active GEAR UP programs. The study sample had a large proportion of would-be first-generation college students (70%) and about 50% of the sample was Hispanic. - Since this study was embedded within active GEAR UP programs, students in the study sample had better access to college information and support than many other low-income students. Many of the students had already met with college advisers, completed financial aid applications, and had a solid understanding of the steps needed for college enrollment at the time of high school graduation. - Students in the non-messaged group also had access to GEAR UP advisors and may have received some outreach over the summer (though not the systematic text messages). **Research Design:** Randomized Controlled Trial **Effects:** - The intervention had no effect on fall college enrollment. - The authors suggest that the lack of effect may be in part due to GEAR UP students receiving intensive college advising and preparation during high school, which gave them better access to college information and support than many other low-income students. ## Artificial Intelligence (AI)-Assisted Support from Colleges **What the Study Tested:** - The most recent set of studies explored AI integration to further improve efficiency and reduce costs while still providing tailored support to students during the summer after high school graduation. These studiesinvestigated the use of university-based, AI-powered chatbots that sent personalized text-message outreach to students already accepted to the college during the summer after high school graduation. These chatbots assisted students with key pre-enrollment tasks. Designed to guide students through tasks they had not yet completed and to answer common questions, the chatbot would forward complex questions to a university counselor, improving the system over time. **What the Study Found:** - Both studies found promising evidence that the chatbots reduced summer melt, with the greatest impact seen among students already committed to attending the university (as opposed to all students who had been accepted) as well as among first-generation college students. *Click the dropdowns below to read summaries of each study included in the overall synthesis. Unless noted otherwise, all reported effects are statistically significant at the p<.05 level. Studies are linked (see author name and publication date). When available, we link an open access version of the study*. [Explore our methodology](/rc-methodology) Open all sections Close all sections ### AI Chatbot at East Carolina University (Nurshatayeva et al., 2021) expand\_more **Study: Are Artificially Intelligent Conversational Chatbots Uniformly Effective in Reducing Summer Melt? Evidence from a Randomized Controlled** **Trial** ([Nurshatayeva et al., 2021](https://link.springer.com/article/10.1007/s11162-021-09633-z)) **Intervention:** A university-based, artificially intelligent (AI) chatbot that proactively sent messages to college-intending students during the summer after high school graduation. The messages reminded students to complete required pre-matriculation tasks and offered follow-up support as needed. The messages were tailored to students based on which tasks they had already completed. On average, treatment group students received 26 unique messages. **Context:** - This intervention was tested at East Carolina University (ECU), a public four-year university located in Greenville, North Carolina committed to serving the rural communities of eastern North Carolina. Messages were sent throughout the summer of 2018 to students who were intending to enroll at ECU in the fall of 2018. **Research Design:** Randomized Controlled Trial **Findings:** - For the full sample of students, the main effect of the chatbot was a 4-percentage point increase in loan acceptance. - The rate of summer melt was relatively low at ECU and most students in the control group were already completing most tasks, so there was limited opportunity for the chatbot to affect task completion and enrollment. **Subgroup Findings:** - Effects of the chatbot were strongest among first-generation college students. For this subset of ECU-intending students, the chatbot increased loan acceptance by 8 percentage points. There were marginally statistically significant (p<0.1) increases in on-time matriculation and course registration (both 3 percentage points). ### AI Chatbot at Georgia State University (Page & Gehlbach, 2017) expand\_more **Study:** **How an Artificially Intelligent Virtual Assistant Helps Students Navigate the Road to College** ([Page & Gehlbach, 2017](https://journals.sagepub.com/doi/full/10.1177/2332858417749220)) **Intervention:** A university-based, artificially intelligent (AI) chatbot that sent personalized text outreach during the summer after high school graduation to students who had been admitted to the college. The texts offered reminders and guidance for tasks that students were required to complete over the summer, with students receiving outreach only about tasks that they had not yet completed. On average, treatment group students received 43 unique messages. **Context**: - This intervention was tested at Georgia State University (GSU), a public four-year college located in Atlanta, Georgia. Messages were sent between April and August 2016 to students who had been admitted to GSU’s fall 2016 freshman class. **Research Design:** Randomized Controlled Trial **Findings:** *Compared to GSU-committed students who did not receive chatbot outreach, receiving the chatbot messages...* - Increased on-time fall enrollment among GSU-committed students by 3 percentage points. - Increased the completion of both financial (e.g., accepting a college loan, completing college loan counseling) and non-financial (e.g., submitting final transcripts, submitting immunization records) pre-enrollment tasks. - Helped students follow through with their plans to attend GSU rather than opting for a two-year institution. ### Implementation Considerations expand\_more - **Measuring Summer Melt**: Does your district collect data on seniors’ postsecondary plans, such as through a senior exit survey or counselor tracking, as well as actual enrollment? This will enable your district to measure your summer melt rate, understand the scope of the problem, and customize student outreach. - **Summer Staff Capacity**: The success of these interventions relies on having staff available to assist students who request help in response to the automated messages. Does your school have staff on hand during the summer to provide this support or the budget to hire additional staff, even on a part-time basis? Are there potential opportunities to partner with a local college that many graduates attend to split the advising work? - **Advisor Knowledge and Training:** Advisors who provide follow-up support to students need both general and college-specific knowledge to guide students through the transition process. Your district should ensure advisors are well-trained and equipped to assist students navigating requirements across multiple colleges. - **Partnerships with Local Colleges:** Some of the interventions provided student-customized outreach based on the students’ intended college. Your district may be able to ease the workload for staff in collecting this information by collaborating with local colleges to gather college-specific requirements and deadlines. These partnerships may also involve connecting graduates with college staff for certain questions, particularly those pertaining to financial aid. - **Ensuring Communication Access:** Does your district have a communication platform in place to reach students and families? Using an existing platform can help reduce both costs and logistical challenges. If phone numbers are needed for outreach, it is crucial to gather accurate contact information for as many students as possible. Your district should ensure you have easy access to this information or another reliable method for connecting with students and families. - **Vendor Support:** Many of the texting interventions used a vendor that managed the setup and delivery of messages. This typically allowed for automated message distribution, with district staff only needed to respond to follow-up requests. While this approach streamlines the intervention process, it does include costs associated with the vendor. - **Targeting Specific Students:** Providing customized messages and follow-up advising for each individual student can be resource-intensive, especially if students are attending a wide variety of colleges. Some potential ideas to lighten the advising load at the secondary level include: - Targeting the intervention to students planning to attend local community colleges and/or the most common colleges among graduates. - Creating a partnership with local colleges wherein the local college provides advising to students who have committed to that college and the district focuses on providing advising to the remaining students. - Focusing the intervention on specific populations of students such as would-be first-generation college students or low-income students. - **Implementation Information:** Many of the studies reviewed in this section provide comprehensive implementation details, such as information about vendors used, cost breakdowns, and messaging content, that can be used to build out a summer melt mitigation campaign. Studies are linked below, with open access versions linked when available. - The Center for Education Policy Research (CEPR) at Harvard University offers a[ Summer Melt Handbook](https://sdp.cepr.harvard.edu/sites/g/files/omnuum4446/files/sdp/files/sdp-summer-melt-handbook_0.pdf) that guides education leaders through measuring and responding to summer melt happening in their school(s). The Handbook includes several tools for addressing summer melt such as an [Initial Outreach Checklist ](https://sdp.cepr.harvard.edu/sites/g/files/omnuum4446/files/sdp/files/sdp-summer-melt-sample-initial-outreach-checklist.pdf)and a[ Student Intake Form](https://sdp.cepr.harvard.edu/sites/g/files/omnuum4446/files/sdp/files/sdp-summer-melt-sample-student-intake-form.pdf). - **Using Artificial Intelligence (AI)**: AI-enabled technology shows promise for delivering personalized support at lower costs. Do any of the most commonly attended colleges/ universities in your area have AI chat-bot support available to help students with enrollment tasks? Can you connect students to these resources?