Connect with us

The Conversation

What’s that microplastic? Advances in machine learning are making identifying plastics in the environment more reliable

Published

on

theconversation.com – Ambuj Tewari, Professor of Statistics, University of Michigan – 2025-03-06 07:35:00

Microplastics are tiny bits of plastic that show up in the environment.
Svetlozar Hristov/iStock via Getty Images Plus

Ambuj Tewari, University of Michigan

Microplastics – the tiny particles of plastic shed when litter breaks down – are everywhere, from the deep sea to Mount Everest, and many researchers worry that they could harm human health.

I am a machine learning researcher. With a team of scientists, I have developed a tool to make identification of microplastics using their unique chemical fingerprint more reliable. We hope that this work will help us learn about the types of microplastics floating through the air in our study area, Michigan.

Microplastics – a global problem

The term plastic refers to a wide variety of artificially created polymers. Polyethylene, or PET, is used for making bottles; polypropylene, or PP, is used in food containers; and polyvinyl chloride, or PVC, is used in pipes and tubes.

Microplastics are small plastic particles that range in size from 1 micrometer to 5 millimeters. The width of a human hair, for comparison, ranges from 20 to 200 micrometers.

Most scientific studies focus on microplastics in water. However, microplastics are also found in the air. Scientists know much less about microplastics in the atmosphere.

When scientists collect samples from the environment to study microplastics, they usually want to know more about the chemical identities of the microplastic particles found in the samples.

A pile of empty plastic bottles and containers of varying colors.
Plastic bottles are often made of polyethylene, while food containers usually containe polypropylene.
Anton Petrus/Moment via Getty Images

Fingerprinting microplastics

Just as fingerprinting uniquely identifies a person, scientists use spectroscopy to determine the chemical identity of microplastics. In spectroscopy, a substance either absorbs or scatters light, depending on how its molecules vibrate. The absorbed or scattered light creates a unique pattern called the spectrum, which is effectively the substance’s fingerprint.

A diagram showing how electromagnetic radiation interacting with a sample chemical generates a spectrum.
Spectroscopy can match a substance with its unique fingerprint.
VectorMine/iStock via Getty Images Plus

Just like a forensic analyst can match an unknown fingerprint against a fingerprint database to identify the person, researchers can match the spectrum of an unknown microplastic particle against a database of known spectra.

However, forensic analysts can get false matches in fingerprint matching. Similarly, spectral matching against a database isn’t foolproof. Many plastic polymers have similar structures, so two different polymers can have similar spectra. This overlap can lead to ambiguity in the identification process.

So, an identification method for polymers should provide a measure of uncertainty in its output. That way, the user can know how much to trust the polymer fingerprint match. Unfortunately, current methods don’t usually provide an uncertainty measure.

Data from microplastic analyses can inform health recommendations and policy decisions, so it’s important for the people making those calls to know how reliable the analysis is.

Conformal prediction

Machine learning is one tool researchers have started using for microplastic identification.

First, researchers collect a large dataset of spectra whose identities are known. Then, they use this dataset to train a machine learning algorithm that learns to predict a substance’s chemical identity from its spectrum.

Sophisticated algorithms whose inner workings can be opaque make these predictions, so the lack of an uncertainty measure becomes an even greater problem when machine learning is involved.

Our recent work addresses this issue by creating a tool with an uncertainty quantification for microplastic identification. We use a machine learning technique called conformal prediction.

Conformal prediction is like a wrapper around an existing, already trained machine learning algorithm that adds an uncertainty quantification. It does not require the user of the machine learning algorithm to have any detailed knowledge of the algorithm or its training data. The user just needs to be able to run the prediction algorithm on a new set of spectra.

To set up conformal prediction, researchers collect a calibration set containing spectra and their true identities. The calibration set is often much smaller than the training data required for training machine learning algorithms. Usually just a few hundred spectra are enough for calibration.

Then, conformal prediction analyzes the discrepancies between the predictions and correct answers in the calibration set. Using this analysis, it adds other plausible identities to the algorithm’s single output on a particular particle’s spectrum. Instead of outputting one, possibly incorrect, prediction like “this particle is polyethylene,” it now outputs a set of predictions – for example, “this particle could be polyethylene or polypropylene.”

The prediction sets contain the true identity with a level of confidence that users can set themselves – say, 90%. Users can then rerun the conformal prediction with a higher confidence – say, 95%. But the higher the confidence level, the more polymer predictions given by the model in the output.

It might seem that a method that outputs a set rather than a single identity isn’t as useful. But the size of the set serves as a way to assess uncertainty – a small set indicates less uncertainty.

On the other hand, if the algorithm predicts that the sample could be many different polymers, there’s substantial uncertainty. In this case, you could bring in a human expert to examine the polymer closely.

Testing the tool

To run our conformal prediction, my team used libraries of microplastic spectra from the Rochman Lab at the University of Toronto as the calibration set.

Once calibrated, we collected samples from a parking lot in Brighton, Michigan, obtained their spectra, and ran them through the algorithm. We also asked an expert to manually label the spectra with the correct polymer identities. We found that conformal prediction did produce sets that included the label the human expert gave it.

Two very similar looking line graphs, each with a large peaks and a few smaller peaks.
Some spectra, such as polyethylene on the left and polypropylene on the right, look very similar and can easily be confused. That’s why having an uncertainty measure can be helpful.
Ambuj Tewari

Microplastics are an emerging concern worldwide. Some places such as California have begun to gather evidence for future legislation to help curb microplastic pollution.

Evidence-based science can help researchers and policymakers fully understand the extent of microplastic pollution and the threats it poses to human welfare. Building and openly sharing machine learning-based tools is one way to help make that happen.The Conversation

Ambuj Tewari, Professor of Statistics, University of Michigan

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Read More

The post What’s that microplastic? Advances in machine learning are making identifying plastics in the environment more reliable appeared first on theconversation.com

The Conversation

Social media before bedtime wreaks havoc on our sleep − a sleep researcher explains why screens alone aren’t the main culprit

Published

on

theconversation.com – Brian N. Chin, Assistant Professor of Psychology, Trinity College – 2025-04-08 07:49:00

Social media use before bedtime can be stimulating in ways that screen time alone is not.
Adam Hester/Tetra Images via Getty Images

Brian N. Chin, Trinity College

“Avoid screens before bed” is one of the most common pieces of sleep advice. But what if the real problem isn’t screen time − it’s the way we use social media at night?

Sleep deprivation is one of the most widespread yet overlooked public health issues, especially among young adults and adolescents.

Despite needing eight to 10 hours of sleep, most adolescents fall short, while nearly two-thirds of young adults regularly get less than the recommended seven to nine hours.

Poor sleep isn’t just about feeling tired − it’s linked to worsened mental health, emotion regulation, memory, academic performance and even increased risk for chronic illness and early mortality.

At the same time, social media is nearly universal among young adults, with 84% using at least one platform daily. While research has long focused on screen time as the culprit for poor sleep, growing evidence suggests that how often people check social media − and how emotionally engaged they are − matters even more than how long they spend online.

As a social psychologist and sleep researcher, I study how social behaviors, including social media habits, affect sleep and well-being. Sleep isn’t just an individual behavior; it’s shaped by our social environments and relationships.

And one of the most common yet underestimated factors shaping modern sleep? How we engage with social media before bed.

Emotional investment in social media

Beyond simply measuring time spent on social media, researchers have started looking at how emotionally connected people feel to their social media use.

Some studies suggest that the way people emotionally engage with social media may have a greater impact on sleep quality than the total time they spend online.

In a 2024 study of 830 young adults, my colleagues and I examined how different types of social media engagement predicted sleep problems. We found that frequent social media visits and emotional investment were stronger predictors of poor sleep than total screen time. Additionally, presleep cognitive arousal and social comparison played a key role in linking social media engagement to sleep disruption, suggesting that social media’s effects on sleep extend beyond simple screen exposure.

I believe these findings suggest that cutting screen time alone may not be enough − reducing how often people check social media and how emotionally connected they feel to it may be more effective in promoting healthier sleep habits.

How social media disrupts sleep

If you’ve ever struggled to fall asleep after scrolling through social media, it’s not just the screen keeping you awake. While blue light can delay melatonin production, my team’s research and that of others suggests that the way people interact with social media may play an even bigger role in sleep disruption.

Here are some of the biggest ways social media interferes with your sleep:

  • Presleep arousal: Doomscrolling and emotionally charged content on social media keeps your brain in a state of heightened alertness, making it harder to relax and fall asleep. Whether it’s political debates, distressing news or even exciting personal updates, emotionally stimulating content can trigger increased cognitive and physiological arousal that delays sleep onset.

  • Social comparison: Viewing idealized social media posts before bed can lead to upward social comparison, increasing stress and making it harder to sleep. People tend to compare themselves to highly curated versions of others’ lives − vacations, fitness progress, career milestones − which can lead to feelings of inadequacy and anxiety that disrupt sleep.

  • Habitual checking: Social media use after lights out is a strong predictor of poor sleep, as checking notifications and scrolling before bed can quickly become an automatic habit. Studies have shown that nighttime-specific social media use, especially after lights are out, is linked to shorter sleep duration, later bedtimes and lower sleep quality. This pattern reflects bedtime procrastination, where people delay sleep despite knowing it would be better for their health and well-being.

  • Fear of missing out, or FOMO: The urge to stay connected also keeps many people scrolling long past their intended bedtime, making sleep feel secondary to staying updated. Research shows that higher FOMO levels are linked to more frequent nighttime social media use and poorer sleep quality. The anticipation of new messages, posts or updates can create a sense of social pressure to stay online and reinforce the habit of delaying sleep.

Taken together, these factors make social media more than just a passive distraction − it becomes an active barrier to restful sleep. In other words, that late-night scroll isn’t harmless − it’s quietly rewiring your sleep and well-being.

How to use social media without sleep disruption

You don’t need to quit social media, but restructuring how you engage with it at night could help. Research suggests that small behavioral changes to your bedtime routine can make a significant difference in sleep quality. I suggest trying these practical, evidence-backed strategies for improving your sleep:

  • Give your brain time to wind down: Avoid emotionally charged content 30 to 60 minutes before bed to help your mind relax and prepare for sleep.

  • Create separation between social media and sleep: Set your phone to “Do Not Disturb” or leave it outside the bedroom to avoid the temptation of late-night checking.

  • Reduce mindless scrolling: If you catch yourself endlessly refreshing, take a small, mindful pause and ask yourself: “Do I actually want to be on this app right now?”

A brief moment of awareness can help break the habit loop.The Conversation

Brian N. Chin, Assistant Professor of Psychology, Trinity College

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Read More

The post Social media before bedtime wreaks havoc on our sleep − a sleep researcher explains why screens alone aren’t the main culprit appeared first on theconversation.com

Continue Reading

The Conversation

Providing farmworkers with health insurance is worth it for their employers − new research

Published

on

theconversation.com – John Lowrey, Assistant Professor of Supply Chain and Health Sciences, Northeastern University – 2025-04-08 07:48:00

Farmworkers at Del Bosque Farms pick and pack melons on a mobile platform in Firebaugh, Calif., in July 2021.
AP Photo/Terry Chea

John Lowrey, Northeastern University; Timothy Richards, Arizona State University, and Zachariah Rutledge, Michigan State University

Agricultural employers who provide farmworkers with health insurance earn higher profits, even after accounting for the cost of that coverage. In addition, farmworkers who get health insurance through their employers are more productive and earn more money than those who do not.

These are the key findings from our study published in the March 2025 issue of the American Journal of Agricultural Economics.

To conduct this research, we crunched over three decades of data from the Labor Department’s National Agricultural Workers Survey. We focused on California, the nation’s largest producer of fruits, nuts and other labor-intensive agricultural products in the U.S., from 1989 to 2022.

We determined that if 20% more farmworkers got health insurance coverage, they would have earned $23,063 a year in 2022, up from $22,482 if they did not. Their employers, meanwhile, would earn $7,303 in net profits per worker annually in this same scenario, versus $6,598.

Why it matters

Roughly half of California’s agricultural employers are facing labor shortages at a time when the average age of U.S. farmworkers is also rising.

Some of them, including grape producers, are responding by investing more heavily in labor-saving equipment, which helps reduce the need for seasonal manual labor. However, automated harvesting isn’t yet a viable or affordable option for labor-intensive specialty crops such as melons and strawberries.

Despite labor shortages, agricultural employers may be reluctant to increase total compensation for farmworkers. They may also be wary of providing additional benefits such as health insurance for two main reasons.

First, seasonal workers are, by definition, transient, meaning that the employer who provides coverage may not necessarily be the same one who benefits from a healthier worker. Second, it costs an employer money but doesn’t necessarily benefit them in the future if the worker moves on.

Most U.S. farmworkers are immigrants from Mexico or Central America. Roughly 42% are immigrants who are in the U.S. without legal authorization, down from 55% in the early 2000s.

As the share of farmworkers who are unauthorized immigrants has declined, the share who are U.S. citizens – including those born here – has grown and now stands at about 39%.

The low wages farmworkers earn offer little incentive for more U.S. citizens and permanent residents to take these jobs. These jobs might become more attractive if employers offered health care coverage to protect the health of the worker and their household.

Farmworkers who lack legal authorization to be in the U.S. are not eligible for private health insurance policies, and many can’t enroll in Medicaid, a government-run health insurance program that’s primarily for low-income Americans and people with disabilities. Regardless, some employers do take steps to help them gain access to health care services. As of 2025, a large share of farmworkers remain uninsured, including many citizens and immigrants with legal status.

Limited access to health care is an unfortunate reality for farmworkers, whose jobs are physically demanding and dangerous. In addition, farmworkers are paid at or near the minimum wage and are constantly searching for their next employment opportunity. This uncertainty causes high levels of stress, which can contribute to chronic health issues such as hypertension.

What still isn’t known

It is hard to estimate the effect of employer-provided health insurance on workers and employers, since labor market outcomes are a result of highly complex interactions.

For example, wages, productivity and how long someone keeps their job are highly interdependent variables determined by the interaction between what workers seek and what employers offer. And wages do not always reflect a worker’s skills and abilities, as some people are more willing to accept a job with low pay if their compensation includes good benefits such as health insurance.

The Research Brief is a short take about interesting academic work.The Conversation

John Lowrey, Assistant Professor of Supply Chain and Health Sciences, Northeastern University; Timothy Richards, Professor of Agribusiness, Arizona State University, and Zachariah Rutledge, Assistant Professor of Agricultural, Food and Resource Economics, Michigan State University

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Read More

The post Providing farmworkers with health insurance is worth it for their employers − new research appeared first on theconversation.com

Continue Reading

The Conversation

Being alone has its benefits − a psychologist flips the script on the ‘loneliness epidemic’

Published

on

theconversation.com – Virginia Thomas, Assistant Professor of Psychology, Middlebury – 2025-04-04 07:18:00

Studies show that choosing ‘me time’ is not a recipe for loneliness but can boost your creativity and emotional well-being.
FotoDuets/iStock via Getty Images Plus

Virginia Thomas, Middlebury

Over the past few years, experts have been sounding the alarm over how much time Americans spend alone.

Statistics show that we’re choosing to be solitary for more of our waking hours than ever before, tucked away at home rather than mingling in public. Increasing numbers of us are dining alone and traveling solo, and rates of living alone have nearly doubled in the past 50 years.

These trends coincided with the surgeon general’s 2023 declaration of a loneliness epidemic, leading to recent claims that the U.S. is living in an “anti-social century.”

Loneliness and isolation are indeed social problems that warrant serious attention, especially since chronic states of loneliness are linked with poor outcomes such as depression and a shortened lifespan.

But there is another side to this story, one that deserves a closer look. For some people, the shift toward aloneness represents a desire for what researchers call “positive solitude,” a state that is associated with well-being, not loneliness.

As a psychologist, I’ve spent the past decade researching why people like to be alone – and spending a fair amount of time there myself – so I’m deeply familiar with the joys of solitude. My findings join a host of others that have documented a long list of benefits gained when we choose to spend time by ourselves, ranging from opportunities to recharge our batteries and experience personal growth to making time to connect with our emotions and our creativity.

YouTube video
Being alone can help remind people who they are.

So it makes sense to me why people live alone as soon as their financial circumstances allow, and when asked why they prefer to dine solo, people say simply, “I want more me time.”

It’s also why I’m not surprised that a 2024 national survey found that 56% of Americans considered alone time essential for their mental health. Or that Costco is now selling “solitude sheds” where for around US$2,000 you can buy yourself some peace and quiet.

It’s clear there is a desire, and a market, for solitude right now in American culture. But why does this side of the story often get lost amid the warnings about social isolation?

I suspect it has to do with a collective anxiety about being alone.

The stigma of solitude

This anxiety stems in large part from our culture’s deficit view of solitude. In this type of thinking, the desire to be alone is seen as unnatural and unhealthy, something to be pitied or feared rather than valued or encouraged.

This isn’t just my own observation. A study published in February 2025 found that U.S. news headlines are 10 times more likely to frame being alone negatively than positively. This type of bias shapes people’s beliefs, with studies showing that adults and children alike have clear judgments about when it is – and importantly when it is not – acceptable for their peers to be alone.

This makes sense given that American culture holds up extraversion as the ideal – indeed as the basis for what’s normal. The hallmarks of extraversion include being sociable and assertive, as well as expressing more positive emotions and seeking more stimulation than the opposite personality – the more reserved and risk-averse introverts. Even though not all Americans are extraverts, most of us have been conditioned to cultivate that trait, and those who do reap social and professional rewards. In this cultural milieu, preferring to be alone carries stigma.

But the desire for solitude is not pathological, and it’s not just for introverts. Nor does it automatically spell social isolation and a lonely life. In fact, the data doesn’t fully support current fears of a loneliness epidemic, something scholars and journalists have recently acknowledged.

In other words, although Americans are indeed spending more time alone than previous generations did, it’s not clear that we are actually getting lonelier. And despite our fears for the eldest members of our society, research shows that older adults are happier in solitude than the loneliness narrative would lead us to believe.

YouTube video
It’s all a balancing act – along with solitude, you need to socialize.

Social media disrupts our solitude

However, solitude’s benefits don’t automatically appear whenever we take a break from the social world. They arrive when we are truly alone – when we intentionally carve out the time and space to connect with ourselves – not when we are alone on our devices.

My research has found that solitude’s positive effects on well-being are far less likely to materialize if the majority of our alone time is spent staring at our screens, especially when we’re passively scrolling social media.

This is where I believe the collective anxiety is well placed, especially the focus on young adults who are increasingly forgoing face-to-face social interaction in favor of a virtual life – and who may face significant distress as a result.

Social media is by definition social. It’s in the name. We cannot be truly alone when we’re on it. What’s more, it’s not the type of nourishing “me time” I suspect many people are longing for.

True solitude turns attention inward. It’s a time to slow down and reflect. A time to do as we please, not to please anyone else. A time to be emotionally available to ourselves, rather than to others. When we spend our solitude in these ways, the benefits accrue: We feel rested and rejuvenated, we gain clarity and emotional balance, we feel freer and more connected to ourselves.

But if we’re addicted to being busy, it can be hard to slow down. If we’re used to looking at a screen, it can be scary to look inside. And if we don’t have the skills to validate being alone as a normal and healthy human need, then we waste our alone time feeling guilty, weird or selfish.

The importance of reframing solitude

Americans choosing to spend more time alone is indeed a challenge to the cultural script, and the stigmatization of solitude can be difficult to change. Nevertheless, a small but growing body of research indicates that it is possible, and effective, to reframe the way we think about solitude.

For example, viewing solitude as a beneficial experience rather than a lonely one has been shown to help alleviate negative feelings about being alone, even for the participants who were severely lonely. People who perceive their time alone as “full” rather than “empty” are more likely to experience their alone time as meaningful, using it for growth-oriented purposes such as self-reflection or spiritual connection.

Even something as simple as a linguistic shift – replacing “isolation” with “me time” – causes people to view their alone time more positively and likely affects how their friends and family view it as well.

It is true that if we don’t have a community of close relationships to return to after being alone, solitude can lead to social isolation. But it’s also true that too much social interaction is taxing, and such overload negatively affects the quality of our relationships. The country’s recent gravitational pull toward more alone time may partially reflect a desire for more balance in a life that is too busy, too scheduled and, yes, too social.

Just as connection with others is essential for our well-being, so is connection with ourselves.The Conversation

Virginia Thomas, Assistant Professor of Psychology, Middlebury

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Read More

The post Being alone has its benefits − a psychologist flips the script on the ‘loneliness epidemic’ appeared first on theconversation.com

Continue Reading

Trending