Browsed by
Category: Analysis

COVID – Media, and Anchoring Bias

COVID – Media, and Anchoring Bias

Comments 0 Comment

In January 2020, when COVID first emerged as an unknown virus in Wuhan, I distinctly remember voicing my concerns to a friend who was planning to travel to Southeast Asia a few weeks later. At the time, there were only a handful of cases reported outside Wuhan. However, I had seen videos of police welding doors shut, citizens falling to the ground, livestock being culled, makeshift hospitals constructed to support tens of thousands of patients, and reported case numbers growing exponentially – daily, on the John’s Hopkins worldwide COVID-19 dashboard. I had developed an idea and image of what the virus was…and understood it was capable of becoming the monstrous pandemic we had only seen in disaster movies.

Despite the rapid growth, severity, and seemingly unstoppable spread beyond the initial ground zero in Wuhan, the World Health Organization (WHO) seemed to downplay the scope and severity of the outbreak – declining to call COVID-19 a pandemic until months later.

“The WHO chief has walked a tightrope, trying to lead the global health-care response, keep communication channels open with China and avoid unnecessary panic.”

https://www.bloomberg.com/opinion/articles/2020-02-27/coronoavirus-why-the-who-won-t-call-covid-19-a-pandemic

As is often the case, the lack of clarity and misinformation from the source can cause us to develop a worst case scenario understanding of the situation. China was not forthcoming with information or data, at best. As a result, each data point, each article from the press, each speech from our leadership, and every first hand account gained heightened attention as we constructed the image of the “invisible enemy” in our minds.

So, with the dearth of information and data clarity, we (as a society) developed an anchoring bias with COVID-19, with our primary lens through the media. The areas impacted early became a focal point for our foundational understanding. Over a year later and through two massive waves of COVID cases, New York City stands apart as one of the hardest hit areas. With the horrific fatalities rates of nursing home residents, the city as a whole suffered worse than any jurisdiction in the US. Over 35% of all deaths in New York City involved COVID-19 in 2020. That is more than three times the average of other states and jurisdictions around the country. For comparison, California and Florida saw around 10% of deaths involve COVID-19 as one of the causes of death in 2020. Given 12 percent of all US newsroom employees live in New York City, much of the media had a front row seat to the worst of the worst.

With such a huge impact to New York City at such a critical time for the country, it isn’t surprising the media developed an understanding that was shaped by personal experience and direct exposure as much as the objective science and situations in other jurisdictions. It seems apparent climate, population density, and other environmental factors play a critical role in the transmission and severity of COVID-19. Many suburban or rural communities have seen a much smaller impact and government mandates, restrictions, and media reporting should reflect the reality for localities whenever possible. Top-down fearmongering has been, and will continue to be, an unproductive way to overcome this invisible adversary.

A Tale of Two States: North Dakota and South Dakota

A Tale of Two States: North Dakota and South Dakota

Comments 0 Comment

When attempting to determine the effectiveness of responses to issues or the results of experiments, using a simple A-B test often yields straightforward answers to more complex problems. If we have two similar samples, a binary “Do X for sample A, and do Y for sample B” can be an easy way to see which intervention is more effective.

With COVID-19, government policy mandates have been unprecedented. As with most other countries around the world, the United States seen dramatic examples of top-down restrictions designed to protect citizens from the invisible virus – and each state has had different approaches. But have these interventions been successful? Have they saved lives? How many? Or destroyed livelihoods? Has the cure been worse than the disease?

Although the answers to these questions remain largely undetermined and are difficult to discern with a high degree of confidence, we can look at relatively similar sample populations which had substantially different policy interventions and examine the results. For this example, let’s compare COVID’s recent impact in North Dakota vs. South Dakota:

Both states share similar climate profiles, population size and densities, and overall demographics. While different in some respects, the core similarities of each state provide us with a unique opportunity to isolate policy interventions as the most significant underlying variable with the most recent wave of COVID infections.

North Dakota’s Government Response: While initially advocating for a limited government response, on November 13th, 2020 Governor Burgum signed an executive order for:

1) Mask mandates (all indoor and public settings – 30 days)
2) Bars/Restaurants 50% capacity, 150 persons. Curfew 10pm to 4am.
3) Banquets, Halls, events 25% of capacity.

These orders were eased somewhat and extended until mid-January.

South Dakota’s Government Response: Governor Noem has advocated or public advisories, protecting vulnerable populations, and championing individual liberties and awareness.

So, what were the results?

The results have been nearly identical from a practical standpoint. In isolation, it would appear North Dakota’s government interventions led to a direct reduction in COVID cases. However, the timing of the downturn in cases also happened almost exactly at the same time in South Dakota (without the restrictions).

Below are the COVID case and fatality trends for each state:

source: worldometers.info

source: data.cdc.gov

Note: The CDC cause of death data is updated on a slightly delayed basis based on the medical record reporting. As a result, the weekly death totals will likely rise to a normalized trendline for recent weeks. However, the directional trends related to cause (COVID vs. non-COVID) should remain the same.

Digging deeper into the numbers it appears the relatively small variances in cases per population, and overall case fatality rates could be largely driven by demographic differences as seen in the census numbers below:

https://www.census.gov/quickfacts/ND

https://www.census.gov/quickfacts/SD

So…Do government mask mandates and restrictions work? Given the available data, a case could be made either way. However, it is clear the results (restrictions vs. no restrictions) are similar enough to warrant a much closer look. It appears mother nature is firmly in control – despite our best efforts to mitigate the impacts through mask mandates and school/business closures and restrictions.

Data, Problem Solving, COVID-19, and Solutions

Data, Problem Solving, COVID-19, and Solutions

Comments 0 Comment

Data. For the last 25 years in finance and technology I have used data to explore, uncover facts, and understand narratives about what is happening in different business environments and the world. Data has answers. It can tell stories and is an immensely powerful tool. Everything from simple facts to incredible sagas can all be discovered and learned from a myriad of data sources.

Just as with business problems, data can be used to understand COVID-19, identify issues, and develop solutions.. We can find answers using the same frameworks of knowledge and skill sets developed from the business world. At the fundamental level, the analysis is basically the same.

I was propelled into data analytics working at an overburdened call center during the dot-com heyday. One reoccurring question management constantly asked was: “Why are our call volumes up?”…or rather “What’s hitting us?!?!!!!”. They wanted answers to fix things that were broken…and improve service, now! They would hear about issues from the floor of associates and supervisors shouting heads-ups, or exchanging comments and notes. But it was chaotic…it was a mess. Management was shooting in all directions. There had to be a better way.

When I began as a broker in 1997, the company was experiencing hyper-growth. We were breaking volume records every other day. Business was booming and the call center was constantly playing catch up. We couldn’t hire and train people fast enough. Hold times were atrocious. Within the midst of ticker tape light boards and screens showing hour-long phone queues, I and other front line reps were tackling new customer issues every five or ten minutes throughout an entire shift…plus overtime. We did all this while dealing with a bubbling cauldron lottery of frustrated and abusive customers – any one call could single-handedly ruin your day.

At one point in the chaos, a director pulled me aside and tasked me with getting answers and developing solutions to improve the customer experience. “Figure it out” was the marching order. The relief of being off the phones was awesome…no more aggravated customers directly venting frustrations all day. I had a project – to find answers and fix things.

Fortunately, management was also driven to find solutions. They wanted to solve problems…but were often overwhelmed and needed ammunition. What’s hitting us? .What is happening? What can we do to solve this issue? …and these service delays? If they could inform Systems and Operations of issues, fix problems, and slow inquiry volumes, the staffing and service issues could be mitigated. At the time, initiatives were developed and championed on the fly, based on gut instinct. They didn’t have time for multivariate observational studies, surveys, or interviews…or detailed research projects. Answers were needed yesterday.

I wanted to help get answers…. I also wanted my project to succeed to stay off the phones. So I had to prove results – quickly. My first goal was to understand the scope and nature of the inquiries coming in. If we had 5,000 calls in a day, how many of those were operations related? …systems related? …deposit related? …password issues? …complaints?, etc. No one had any idea. We had hands-on experience talking with customers and understanding those issues at the most granular level. However, we couldn’t see the big picture.

I knew data could help us scope and understand the real customer impacts and opportunities to improve the situation. Armed with the facts, I could develop reports, get answers, and develop logical narratives about what was happening. This would enable management to allocate resources and promote initiatives with the highest impact.

I was fired up for the challenge.

I dove into whatever data I could find…beginning with the thousands of emails we were receiving each day. I developed queries to scope issues and segment emails….and discovered more and more insights. The first few days of the project turned into weeks, months, and years. I learned about call center reporting, databases, email systems, call routing, and workflows. Query tools and analytics became second nature. I worked with incredible engineers to develop pattern matching text queries to flag complaints, categorize, rank, and spot variances in email inquiry trends. We built call tracking to segment calls and provide real-time transparency and visibility to the front line managers. We created new workflows and service level dashboards to monitor and improve operations. Ultimately, I was able to build powerful forecasting tools to accurately predict inquiry volumes and staffing needs weeks and months ahead.

Over the years, I have worked with some incredible people and gained a broad understanding of call centers, customer service, marketing, operations, finance, technology, and business in general. I have become a business and data/analytics expert…and have been doing it for the last 20+ years. About half of that was with the original company, and then several startups in Silicon Valley and overseas.

The COVID-19 pandemic emerged as I was taking a break while recovering from a torn bicep and spine surgery. In late January, I recognized the potential severity of the COVID-19 outbreak as I followed the horrific stories out of Wuhan. I then discovered the Johns Hopkins dashboards and some other limited data sources. Given my experience working with analytics and recognizing the challenge ahead, I was compelled to dig deep into COVID-19 data, research, and focus my efforts to help.

Tragically, there have been over 22 million cases and almost 800,000 worldwide fatalities as I write this. One side benefit is there is now a mountain of data about cases, testing, treatments, hospitalizations, and deaths. However, it is messy and the data accuracy is constantly a consideration as I research. Countries and jurisdictions can distort numbers and even falsify reports to satisfy or communicate narratives. Sadly, the World Health Organization only uses reported numbers and does not have authority to validate any data. We missed an opportunity to stop it early – in China. We can and should do better – globally.

Over the past few months, I have spent countless hours digging through data, reading reports and listening to the medical and scientific experts. Stories have been uncovered… new narratives have emerged, and facts continue to surface.

On the news each night, broadcasts show the big numbers and we watch and listen hopelessly as things never seem to improve – and we must continue lockdowns, wear masks, and close businesses and schools. This should not be the case.

Societies and the world will be better off dealing with COVID-19 by seeing the big picture and calibrating initiatives and responses accordingly. Filtering through fearmongering headlines, the information overload can be overwhelming. However, there are answers in the data. We will get through this crisis by learning – understanding and developing solutions to mitigate the impact of COVID-19 and any future viruses to our health, well-being, and societies. I hope to help by uncovering and sharing facts and narratives, and providing answers and solutions. We will be better prepared next time. The story will end well.