New article outlines considerations for clinical trial design to enhance OS collection and analysis. Key findings include the importance of prespecified statistical analyses of OS as a safety endpoint and robust assessments to support approval.
A comprehensive report by the American Statistical Association highlights concerns about the federal statistical system's future, citing weaknesses in core supports and outdated methodology. The report recommends 15 specific measures to modernize the system and ensure timely, accurate data.
The 2020 Joint Statistical Meetings feature cutting-edge research on COVID-19 modeling, including nowcasting and forecasting. Researchers also presented advancements in precision medicine for stem cell transplants, improving patient outcomes by extending lives.
Researchers suggest starting with Final Four picks and building backward and forward from there for optimal strategies. The study's models produce many brackets that would have placed in the top 100 of the ESPN bracket challenge, especially those that start by picking Elite Eight or Final Four teams.
The US Census Bureau's economic census, conducted every five years, will now use the North American Product Classification System (NAPCS) to standardize vital economic statistics. This change aims to improve data precision and provide businesses with a sense of confidence in measurement accuracy.
The Analytics and Data Science Institute created nine research laboratories to tackle business and public service data problems. Students from various backgrounds collaborate to understand statistical and computer science perspectives, enabling them to communicate effectively with non-experts.
Researchers have developed a new method called matching-adjusted indirect comparison (MAIC) to compare the effectiveness of drugs with limited data, enabling more accurate reimbursement decisions. By analyzing patient characteristics, MAIC provides unbiased estimates of treatment effects, helping to optimize resource allocation.
Researchers used data analysis techniques, including decomposing songs into musical features, to determine Beatles songwriting styles. The approach identified distinct patterns in Lennon and McCartney's compositions, allowing for the prediction of disputed authorship.
The study found that provinces receiving high PEPFAR funding experienced significant reductions in infant mortality, saving the lives of many Kenyan children. However, the effectiveness of the program varies across countries and regions with different HIV prevalence rates.
A team of Stanford statisticians, led by Brad Efron and Mike Baiocchi, developed an evidence-based program to prevent rape in Nairobi's poorest schools. The program, which involves training girls to be aware of common scenarios and empower them to register their will, has been shown to cut the rate of sexual assault in half.
A recent study found that many researchers overlook the importance of weighting and sample design in survey data, leading to incorrect estimates. The American Statistical Association has identified this issue as a systemic problem, with only half of secondary analyses accounting for weighting.
A recent study analyzed NFL injuries from 2007-2016 and found that the increase in injuries occurred faster after practice restrictions took effect, not as a result of them. The study also showed no significant differences in injury rates for different muscle groups or types of injuries.
Researchers used cohort change ratios to estimate working life expectancies in MLB pitchers, revealing that most careers are much shorter than the few stars who stick around. This method could be applied to other sports, like American football and basketball, where contact is more prevalent.
The 'customer-based corporate valuation' (CBCV) model, developed by Daniel McCarthy and Pete Fader, analyzes the stability of sales to assess a company's long-run success. This approach has been shown to accurately predict company value using publicly available data, with implications for investors and private equity firms.
Home run numbers increased significantly between 2016-2017 due to a drop in vertical break of pitches, suggesting pitchers aimed to reduce home runs. However, this strategy failed to work as other factors contributed more to the surge.
A key statistic in baseball, Wins Above Replacement Player (WARP), has been found to not accurately measure player quality due to the game's unique nature. The analysis revealed that team performance is more indicative of a team's quality than individual player value.
Researchers propose new statistical methodology to understand climate system and attribute extreme climatic events. This approach views extreme events as a given and assesses the degree of changes in thermodynamic state variables influenced by climate change.
A new statistical method, ADGN, matches multiple records with precision and identifies the scope of discrimination when applied to voter ID laws. The research found that white registered voters are significantly more likely to possess a voter ID than African-American or Hispanic voters.
Groundbreaking research employs innovative statistical methods to accurately quantify modern slavery's prevalence, revealing an estimated 40.3 million people enslaved worldwide in 2016. The work has been widely accepted by the UN and advocates, offering a solution to estimate the extent of slavery in well-developed countries.
Researchers developed a new statistical method to calculate sea level rise, showing it's not only increasing but also accelerating. The study found the rate of sea level change on the East Coast is the highest in 2,000 years, with areas like New York City facing rapid increases.
Advanced statistical modeling tools can analyze large electronic health records to inform medical decisions and develop individualized treatments for high-risk patients. The methods leverage machine learning and flexible models to draw valid inference from representative populations.
A recent statistical analysis suggests that the changed strike zone could negatively affect batter productivity. The study found pitch location-specific differences in run expectancies, which may lead to internal struggles among players and impact salaries.
A study by University of Virginia Assistant Professor Matthew Gerber found that Twitter can help predict crime by analyzing common topics of discussion, such as sports and entertainment. The result was more precise than older forecasting models, with accurate predictions for 19 out of 25 crime types.
A new statistical method for network anomaly detection uses machine learning to identify unusual behavior, potentially providing a more robust approach to cybersecurity. The model has been developed by researchers at the University of Oxford and is being presented at the 2017 Joint Statistical Meetings.
Researchers have deciphered the mechanism of general anesthesia using statistical analysis, revealing oscillations that disrupt brain communication. The discovery enables real-time brain state monitoring and may lead to new recovery methods for elderly patients experiencing cognitive dysfunction after anesthesia.
Researchers propose new methods for addressing the reproducibility crisis in scientific research, emphasizing the need for greater understanding of data analysis as a human behavior. Experts argue that current p-value approaches may be flawed and advocate for Bayesian inference as a solution.
Cox was recognized for his 1972 paper on the proportional hazards model, widely used in survival data analysis. The model has been applied to various fields, leading to life-changing breakthroughs and societal impact.
A new statistical model has been developed to help guide forest management decisions under climate change. The model uses future climate scenarios to produce more accurate predictions than previous methods, allowing for the design of strategies that mitigate climate change impacts on loblolly pine plantations.
A new analysis of Medicare's Hospital Compare portal found that the statistical methodology used to rate and compare hospitals underestimates mortality rates of acute myocardial infarction (AMI) at small hospitals. The research proposes an alternative model that incorporates additional factors such as hospital volume, nursing staff, an...
The American Statistical Association publishes guidelines for researchers to treat statistics as a science, not a recipe. The 10 simple rules emphasize the need for principled statistical analysis to avoid misperceptions and inaccurate claims.
The American Statistical Association has released six principles for improving the use and interpretation of p-values. The statement addresses misconceptions and misuse of p-values, emphasizing the importance of proper study design, data analysis, and reporting.
The number of statistics graduates has increased by 300% since the 1990s, with bachelor's degrees growing 17% from 2013 to 2014. Demand for statisticians is expected to outpace growth, with a potential shortage of 140,000 workers with deep analytical skills by 2018.
Advances in statistics are helping to understand the complex roles microbes play in human biology. The study found no significant association between BMI and bacterial species abundance in the gut microbiome.
A representative survey of Mogadishu's population used remote sensing and smartphone technology to gather data on household dynamics, access to electricity, and child education. The findings revealed significant challenges in collecting accurate data due to security concerns and fragmented authority structures.
Researchers tracked foodborne illness outbreaks using social media, finding that online reviews of restaurants and businesses can aid traditional surveillance systems. Meanwhile, statisticians analyzed social media data after Typhoon Haiyan to develop best practices for emergency response.
A statistical model predicts NFL teams' play-calling tendency with high accuracy, allowing casual fans and coaches to anticipate opponent's next play. The model achieved an average prediction accuracy of 75% in testing, with some games showing as high as 91.6% correct predictions.
Experts discuss the use of value-added models (VAMs) to evaluate teacher performance, with some arguing they are not effective and others defending their usefulness. The panel discussion highlighted the importance of accurate data systems and long-term approaches to estimating educational effects.
A new statistical model unveiled at the Joint Statistical Meetings in Seattle helps police analysts quickly identify crime series and apprehend suspects. The model uses machine learning algorithms to analyze crime data and predict the likely location of next events in a series.
The field of risk analysis is entering a new phase with real-time probabilistic risk analysis, enabling better management of complex systems. New methodologies and technologies are being used to predict risks and improve global resilience.
A new statistical model developed by Michael Braun and collaborators can help businesses estimate the value of customer relationships and optimize marketing efforts. The model considers factors such as service quality and price paid during transactions to assess the impact of marketing activities on customer retention.
An analysis of pediatric primary health care accessibility and availability in multiple states found systematic disparities between and within states. The study reveals that rural and urban communities, as well as public versus private insurance groups, face varying levels of accessibility challenges.
A new technique helps doctors optimize cancer treatment doses in phase I/II clinical trials. The Optimal Two-Cycle Dose-Finding Design uses a patient's cycle 1 data to set a dose level for the second cycle, improving patient benefit and controlling toxicity risk.
Statisticians combine academic and commercial techniques to analyze large data sets, reducing processing time and increasing insight. This technology has the potential to aid researchers in various fields, including astronomy, by providing quicker insights from Big Data.
The world population is expected to surpass 11 billion people by 2100, driven primarily by Africa's rapid population growth. The UN projects that Nigeria will have a 90% chance of exceeding 752 million people in 2100, with the US population projected to reach 450 million by century's end.
A new statistical method, Bayesian feature allocation models, can analyze next-generation sequencing data to determine the genetic composition of a patient's cancerous tumor. This information can help physicians select targeted drugs tailored to individual tumors.
A study linked household survey data with New York state administrative records to better measure government transfers and their impact on reducing poverty. The analysis identified widespread underreporting of social welfare program payments, resulting in overstated poverty rates and ineffective policy decisions.
A recent analysis found that women's professional tennis players exhibit less consistency in their performance compared to men, primarily due to differences in match format. The study revealed that women's matches, which are best-of-three sets, resulted in lower frequencies of upsets and letdowns at Grand Slam events.
A new article by Roger D. Peng proposes a two-pronged approach to improve scientific data analysis, combining massive-scale education efforts with the identification of evidence-based data-analytic strategies to increase replicability and reproducibility.
A statistical analysis of poll performance in battleground states over the last three presidential elections shows polling firms produced estimates that were fairly accurate in 2004 and 2008, but underestimated support for President Obama in 2012. The authors attribute this to outdated sampling methodology that excluded cellphone-only ...
A study by Xuan Chen and Carlos A. Flores found that Job Corps participants experienced average wage effects of 5.7-13.9% four years after graduation, with non-Hispanics seeing greater gains.