Science Policy For All

Because science policy affects everyone.

Posts Tagged ‘open science

Science Policy Around the Web – May 8, 2018

leave a comment »

By: Saurav Seshadri, PhD


source: pixabay


EPA Cites ‘Replication Crisis’ in Justifying Open Science Proposal

The U.S. Environmental Protection Agency (EPA) may soon be using far less scientific evidence to inform its policy positions.  EPA administrator Scott Pruitt recently announced that, in an effort promote reproducibility and open access to information, the EPA will no longer consider studies whose underlying data or models are not publicly available.  However, such studies often represent the ‘best available’ data, which the EPA is legally obliged to consider, and form the basis of, among others, policies limiting particulate matter in the air.  Several studies that support the health and economic benefits of lower particulate limits do so by using detailed medical information whose disclosure would compromise patient confidentiality.  The so-called HONEST (Honest and Open New EPA Science Treatment) Act, put forth by House Republicans, aims to suppress such ‘secret science’; its detractors say that it’s a poorly disguised gift to industry interests, conveniently timed to take effect just before a scheduled review of pollution limits.

Opposition to the policy has been building steadily.  A letter signed by 63 House democrats, asking for an extension to the open comment period for the policy, has so far been unsuccessful. A separate letter, signed by almost a thousand scientists, and comments from several professional associations, have also been ignored – perhaps unsurprisingly, given Pruitt’s parallel effort to bar relevant scientists from EPA advisory boards.  The scientist behind the article calling attention to the ‘reproducibility crisis’ cited by Pruitt has also spoken out, writing that simply ‘ignoring science that has not yet attained’ rigorous reproducibility standards would be ‘a nightmare’.

Perhaps the most effective response has come from scientists who are outpacing the bureaucracy.  In a pair of papers published last year, a biostatistics and public health group at Harvard used air quality data, Medicare records, and other public sources to reiterate the health risks posed by air pollution.  Such studies could not be excluded by the new EPA policy and may influence regulators to keep particulate limits low.  Another potential roadblock to implementing changes could be the controversy surrounding Pruitt himself.  The administrator has been the target of several federal probes, following a series of scandals regarding his use of government funds for purposes such as a 24-hour security detail, soundproof office, and first class travel.  Bipartisan calls for his resignation have made his future at the EPA, and the quick implementation of a Republican agenda there, uncertain.

(Mitch Ambrose, American Institute of Physics)

Science funding

NIH’s neuroscience institute will limit grants to well-funded labs

With a budget of $2.1 billion, the National Institute of Neurological Disorders and Stroke (NINDS) is the fifth largest institute at NIH.  Yet each year many investigators are constrained by a lack of funds, while some large labs have accumulated so many grants that their principal investigator can only spend a few weeks per year on a given project.  To address this disparity, NINDS recently announced a plan to revamp implementation of an existing NIH policy, in which grant applications from well-funded labs must go through an additional review by a special council. While the current secondary review rarely rejects such applications, NINDS’ policy takes two steps to make the process more stringent: first, it increases the number of labs that would undergo review, to include labs that would cross the $1 million threshold with the current grant; second, it sets higher standards for review, requiring applications from such labs to score in the top 7% of all proposals to be successful.

Responses to the idea have been tentative, despite widespread support for its objective.  One potential cause for concern is its perceived similarity to the Grant Support Index (GSI), a previous NIH initiative with a similar goal (i.e., reallocating resources to sustain less-established but deserving researchers). The GSI sought to achieve this goal by placing a cap on the number of grants that a lab could receive, using a point system. However, this caused an uproar among scientists, who, among other issues, saw it as punishing or handicapping labs for being productive – it was quickly revised to create the Next Generation Researchers Initiative, a fund earmarked for early and mid-stage investigators, for which each institute is responsible for finding money.  The new policy appears to be a step towards meeting this obligation, and not, NINDS insists, a return to the GSI.

The impact of the new policy will probably be clearer after NINDS’ next round of grant reviews takes place, in January 2019.  So far, only the National Institute of General Medical Sciences (NIGMS) has a comparable policy, which has been in place since 2016.  The success of these approaches may well shape future cohorts of NIH-funded scientists – cutoffs and uncertainty are not unique to neuroscience, and other institutes are likely to be paying close attention.

(Jocelyn Kaiser, Science)

Have an interesting science policy link? Share it in the comments!


Written by sciencepolicyforall

May 8, 2018 at 6:11 pm

Science Policy Around the Web – June 06, 2017

leave a comment »

By: Kseniya Golovnina, PhD

Source: Flickr, by USDA, via Creative Commons     (CC BY 2.0)

Food Security

What if Food Crops Failed at the Same Time?

When one group of people is fighting with climate change and another considers it “mythical”, researchers specialized in the study of social-ecological systems are developing food supply risk assessment models. Food crops are one of the most important sources of human being existence, and less than one-fourth of the planet (“breadbaskets”) produces three-fourth of the staple crops that feed the world’s population. In fact, climate change could cause crop losses in most of the breadbaskets.

Two important factors included in the models are shocks to major land crop production and economy. Shocks like droughts and heat waves in Ukraine and Russia in 2007 and 2009 almost wiped out wheat crops, and caused global wheat prices to spike. And demand assessments project that food production may have to double by 2050 to feed a growing population. Together, the potential environmental and economic stresses are making the world food production system less resilient, and will affect both rich and poor nations. To measure the fragility of the system, researchers developed scenarios of small shocks (10 percent crop loss) and large shocks (50 percent crop loss). These were then applied to corn, wheat or rice output using an integrated assessment model, the Global Change Assessment Model, which was developed by the U.S. Department of Energy.

Among the critical findings are that “breadbasket” regions respond to shocks in different ways. For example, South Asia, where most of the arable land is already in use, is quite unresponsive to shocks occurring elsewhere in the world, because the total amount of land in agricultural production cannot be changed significantly. In Brazil the situation is opposite, it has a lot of potential to bring new land into production if large shocks occur. However, cleaning Brazil’s forests requires significant effort and would add significantly to global climate change. Within the research agenda of the Pardee Center, these risks and preventive actions are discussed in more detail. The warning is clear: humankind needs to be aware and prepared for potential multiple “breadbaskets” failure if we want to reduce the potential for catastrophe. (Anthony Janetos, The Conversation)

Reproducibility in Science

Research Transparency: Open Science

Increasing amounts of scientific data, complexity of experiments, and the hidden or proprietary nature of data has given rise to the “reproducibility crisis” in science. Reproducibility studies in cancer biology have revealed that only 40 % or less peer-reviewed analyses are replicable. Another large-scale project attempting to replicate 100 recent psychology studies was successful in replicating less than 50% of the original results.

These findings are driving scientists to look for ways to increase study reliability, and make research practices more efficient and available for evaluation. A philosophy of open science, where scientists share their primary materials and data, makes analytical approaches more transparent and allows common research practices and standards to emerge more quickly. For scientific journals and associations, open science methods enable the creation of different ways to store and utilize data. Some journals are specifically dedicated to publishing data sets for reuse (Scientific DataJournal of Open Psychology Data), others require or reward open science practices like publicly posting materials and data.

The widespread use of online repositories to share study materials and data helps to store large data sets and physical materials to help mitigate the problems of reproducibility. However, open science practice is still very much in development, and faces some significant disincentives. Habits and reward structures are two major forces work against. Researchers are used to being close, and hide their data from being stolen. Journal editors tend to favor publishing papers that tell a tidy story with perfectly clear results. This causes researchers to omit “failed” studies that don’t clearly support their theories.

While efforts to overcome these obstacles are difficult, development of fully transparent science should be encouraged, as openness helps improve understanding, and acknowledges the truth that real data are often messy. (Elizabeth Gilbert and Katie Corker, The Conversation)


Have an interesting science policy link?  Share it in the comments!

Written by sciencepolicyforall

June 6, 2017 at 9:00 am