Crime rates in large American cities continued to fall in the first half of 2026, with homicides down 18% from the same period in 2025 and 51% since the most recent peak in 2022.. This study examined patterns for 13 crime types in 36 cities that have consistently and rapidly published monthly data over the past nine years. Researchers found that reported levels of 9 of 13 offenses were lower in the first half of 2026 than in the first half of 2025, with reported carjacking incidents seeing the largest drop (-47%) and drug offenses seeing the largest rise (+12%). Reported incidents of domestic violence rose 8%. Only shoplifting (+5%) remained elevated compared to levels in the first half of 2019, before the onset of the COVID-19 pandemic and nationwide social justice protests of 2020. The reported rates of four offenses have dropped by 50% or more from their peaks over the past nine years: homicide, carjacking, residential burglary, and motor vehicle theft . Lastly, lethality—the share of serious violent crime that is fatal—fell 23% from the first half of 2025 and 20% from the first half of 2019 among 17 cities with sufficient data.

Women’s incarceration rates have grown substantially since 2000, but there are still few programs for women reentering the community after incarceration. Los Angeles County developed the Providing Opportunities for Women in Reentry (POWR) program, a trauma-informed cognitive behavioral intervention with comprehensive case management and housing support, to address this need. This study evaluates outcomes of the first two POWR programs, covering March 2023 through June 2024, and examines service receipt, program operations, staff and participant perspectives, and the legal outcomes of participants one year after enrollment. Key findings include that program staff conducted comprehensive needs assessments with participants and created individually tailored, goal focused care plans with them. Program staff also coordinated referrals to outside services, most commonly for basic needs such as food and for mental health services, and followed up to ensure participants kept appointments. Participants also often requested referrals for resources for their children, such as summer programs or places to receive Christmas gifts or school supplies. Nearly 80% of participants were not rearrested in Los Angeles County in the year after enrollment, and even more avoided a felony arrest. About the same proportion avoided any reincarceration in county jail.

Contemporary approaches to crime prevention and community development have traditionally relied on reactive law enforcement, social interventions, urban planning, or public health initiatives operating within separate disciplinary domains. While these approaches have achieved varying degrees of success, persistent challenges such as crime, substance abuse, social fragmentation, and the underutilization of public spaces highlight the need for more integrated and preventive strategies. This paper introduces third space engineering, an interdisciplinary framework that advances spatial advocacy as a proactive, place-based approach to influencing behavior through the intentional design, activation, and stewardship of positive social environments. Researchers found that third space engineering is not the engineering of human behavior. It is the intentional engineering of physical and social spaces to positively influence human behavior and create environments in which individuals and communities are more likely to thrive. Its relevance therefore extends far beyond architecture, policing, hospitality, or urban planning. Every profession that creates, manages, regulates, finances, researches, protects, or activates human environments has a role to play in advancing the framework.

More than 37 million adults under 65 have some college but no credential because they stopped out during the course of their degree. While a stop-out—an interruption in a learner’s continuous enrollment with the expectation of eventual return—is the result of both personal and systemic challenges, the systemic barriers are within institutional control. Key findings of this brief include that college and university leaders consistently cited learners’ financial concerns as the primary stop-out reason. Over the past 30 years, the average cost of college at four-year institutions has increased around 70%, significantly impacting degree affordability. Participants also noted a recurring theme among students who struggled academically before stopping out: a lack of interest in their classes, chosen major, or the career they were planning to enter after graduation. Waning career interest is particularly indicative of a stop-out, as what once led a student to pursue education is no longer a significant motivator. Career advancement was one of the most consistently cited motivating factors when considering why students return to higher education. A return to formal collegiate education is often seen as the most efficient way to further develop a chosen skillset, motivating stop-outs to return to an institution.

This report identifies high-demand and high-opportunity green job fields in Pablo, Montana’s Flathead region and across the state. This report also discusses the skills students need for good, sustainable jobs, the trainings needed to prepare them for those jobs, and opportunities to expand or adapt academic curricula. Researchers identified six relevant green jobs clusters: smart manufacturing, clean energy, green building, responsible food systems, water resources management, and land and forest management. Across these clusters, researchers found that some jobs with applications across clusters have higher quality or pay than expected relative to the required education and skill r. These include electricians, electric power line installers and repairers, and welders, cutters, and other metalworkers. . In addition, any skills and credentials are transferable across green job clusters. Researchers also identified multiple strategies educational institutions can use to prepare students for green jobs, such as adapting existing coursework and training to include more exercises in real-time problem-solving and workflow communication and forming strong partnerships with other educators to help educational institutions access, leverage, and design new coursework. Lastly, educational institutions could explore the logistics of how to expose more students to experiential learning, training opportunities, and job opportunities, especially in rural areas.

Blocked cluster randomized trials are widely used to evaluate educational interventions. For example, several schools can be demarcated as clusters across a number school districts or blocks that are randomly assigned to implement an intervention or serve as controls, with student outcomes tracked over time. In such trials, researchers face critical choices about how to define and estimate the average treatment effect. This working paper provides an applied guide to estimands (e.g., person- vs. cluster-weighted) and estimators (e.g., regression models, multilevel models) for blocked cluster randomized trials and empirically examines their performance using 26 large-scale blocked cluster randomized trials. Findings show that estimator choice can substantially affect point estimates and precision: ranges in the average treatment effect estimates exceed 0.10 standard deviation in 18% of cases, and standard error estimates differ by more than 50% in 16% of cases. Some estimators fail or produce unstable results in common real-world designs.

Federal agencies manage public lands and waters, including national forests and parks, that are tribal ancestral territories. These places have special significance for Tribes. Various federal land and water management agencies and Tribes have entered into shared decision-making agreements. Under these agreements, Tribes provide substantive input on federal natural and cultural resource management decisions over many years. The relevant federal agencies are the U.S. Departments of Agriculture, U.S. Department of Commerce, U.S. Department of the Interior, and their subcomponents. The report identifies five lessons that federal agencies and Tribes learned from shared decision-making: developing detailed written agreements and supporting documents; developing mutual understanding and trust; reconciling differences in policies and procedures for agreement implementation; involving federal staff with appropriate knowledge, skills, and decision-making ability; and securing sufficient resources. These lessons can be important for various reasons. For example, with respect to developing mutual understanding and trust, tribal officials party to an agreement said agency staff met with them in person, had open minds, and treated them with respect. This encouraged candid conversations and helped participants better understand one another’s perspectives.

There are current concerns about the availability of affordable housing both nationally and in Maine. Although there are varying standards by which affordable housing is defined, there seems to be general agreement that there is not enough of it nationwide. In Maine, there is a high level of unmet need for affordable housing across the state. MaineHousing’s need assessment data shows between 80% and 90% of affordable housing need was unmet in most counties as of October 2025. The study examines affordable housing tax credits in Maine and other select states. Key findings include that Maine’s low-income housing tax credit supported the completion of 429 affordable housing units between 2023 and 2025. These 429 units had total development costs of approximately $179.2 million, an average of about $375,000 per unit. In addition, 19 out of 33 state-level affordable housing tax credits, including Florida, have sunsets or appropriation processes providing mechanisms for states to periodically re-examine program performance and state funding priorities.

Although many publications describe what artificial intelligence (AI) could do for emergency management, almost none describe what AI-enabled products are available or what it would take to adopt and diffuse those products. AI-enabled products matter in emergency management EM because they have the potential to improve the decisions and actions taken by emergency managers and others who are responsible for preparing for, responding to, and recovering from disasters. Key findings include that the AI-enabled product landscape is substantial but unevenly distributed among emergency management functional areas. Approximately half of the identified products are either purpose-built for emergency management or have documented use in emergency management contexts; more than 40% are general-purpose solutions with no emergency management-specific design. Most products address narrow slices of emergency management work and would require stacking with other products to meet the full scope of a task area. In addition, one in five products depends on another product to function. Most products require technical integration that presupposes information technology capacity, ongoing data inputs, and continuous internet connectivity. Lastly, privacy risk is near-universal across the landscape; legal and cyber concerns affect a meaningful minority of products, and verifiable AI governance assurance is rare.

This report includes a description of federal programs to address underage drinking; the extent of progress in preventing and reducing underage drinking; surveillance data on underage drinking patterns, and underage access to alcohol. Comprehensive underage drinking prevention and reduction efforts include a balance of evidence-based prevention, intervention, treatment, and enforcement policies and programs implemented at multiple levels. The U.S. Substance Abuse and Mental Health Services Administration created a program in response to the federal STOP Act mandate to create a national media campaign to prevent underage drinking. The program aimed to provide parents with the resources they need to address the risks of alcohol use with their children. Historical data on past-month alcohol use among individuals ages 12–20 show a 35.5% decline from 2004 to 2019. Past-month alcohol use among individuals ages 12–20 was 15.1% in 2022 and 14.6% in 2023. Since no underage youth should be consuming alcohol, these rates remain unacceptably high. Alcohol is still the most widely consumed substance among America’s youth—used more often than cannabis and tobacco. Underage youth who drink tend to binge drink and to consume more on a single drinking occasion than adults do, increasing the risk of harm.

This fact sheet provides a snapshot of the use of Supplemental Security Income (SSI), a means tested, cash assistance program for persons with limited income and resources, including those under the age of 65 who are blind or disabled. Key findings include that 8.8 million people received SSI benefits. Working-age adults ages 18 to 64 accounted for the largest share of SSI recipients (61%), followed by adults age 65 and over (28%) and children under the age of 18 (12%). The average monthly SSI payment received was $766. Adults age 65 and over had lower average payments than working-age adults: $712 compared with $785. Over one-third, 35%, of SSI recipient households had income below their poverty threshold. Another 26% of SSI recipient households had income above poverty but below 200% of their poverty threshold. The majority, 82%, of SSI recipient households were also covered by Medicaid or the Children’s Health Insurance Program. Other benefits commonly received by SSI recipient households included the Supplemental Nutrition Assistance Program (63%), rental subsidies (24%), and energy assistance (16%).

The study aims to determine if high-dose vitamin D3 added to standard chemotherapy improves outcomes in patients with previously untreated metastatic colorectal cancer. Results show that among 455 randomized patients with a median follow-up of 20 months, the median progression-free survival for high-dose vitamin D3 (n = 228) was 11.8 months versus 10.3 months for standard-dose vitamin D3 (n = 227). There were no significant differences in objective response rate between high-dose and standard-dose vitamin D3, or in overall survival. There were no clinically meaningful differences in the most common grade 3 or greater adverse events between the high- and standard-dose groups, including neutropenia and hypertension, or in the incidence of vitamin D–associated toxicities. Overall, among patients with previously untreated metastatic colorectal cancer, addition of high-dose vitamin D3, versus standard-dose vitamin D3, to standard chemotherapy plus bevacizumab did not improve progression-free survival.