The more persistent challenge in labor research nowadays is disentangling the effects of the COVID pandemic from almost any data series there is (read for instance Is the US Labor Market Entering a Real Low-Hire, Low-Fire Equilibrium?) . The disruption is such that not only does almost any data include some sort of noise tracing back to the pandemic, but it also includes caveats from researchers noting confounding effects. To illustrate the challenge, this blog post reviews data and a few research questions from one of the industries most affected by COVID. Nursing employment costs—and healthcare assistance in general—skyrocketed during the pandemic, while experiencing a sharp slowdown afterward. In regular times, analysts would aim at addressing straightforward questions such as what is behind a given decline, increase, technology adoption, or any variation in data points. However, the COVID confounding makes almost any research question twofold, thereby making estimation biases more difficult to address.

For instance, here are four questions of how the COVID confounding obscures research on the industry’s data variation:

• Does the industry compensation data show a decline due to a slowdown in demand after COVID?
• Does the industry compensation data show a decline due to an increase in labor supply after COVID?
• Does the industry compensation data show a decline due to an increase in AI adoption?
• Does the industry compensation data show a decline due to compensation adjustments after COVID?

Let us see how the rationale of labor research questions gets complicated due to the confounding:

First, is labor supply curbing compensation in nursing?

Nursing and healthcare services have become the target careers of many Americans as labor competition intensifies from global markets. High demand, an aging population, and a perceived higher job security skyrocketed compensation for healthcare professionals between the years 2010 and 2023. However, high salaries and high compensation have started to curve down according to quarterly data released by the US Bureau of Labor Statistics. Whether the easing in compensation pressures stems from an increase in labor supply is a question worth investigating, but COVID confounding obscures it, as the decline could be due to a slowdown in demand after the health crisis.

Second, is technology adoption curbing compensation in nursing?

As AI reshapes labor across America, healthcare jobs are perceived as a refuge from the much-anticipated job losses brought by the new technology. The profession experienced a decade-long period of almost uninterrupted growth starting from 2010 to 2020. Early in April 2026, The Wall Street Journal described nursing as the sure path to prosperity. Chart 1 shows how nursing deviated from the aggregated compensation growth of all industries over the last two decades. The chart tells a story of higher growth compared to all other industries. The decline in growth seen in the last three years may mean that new technology adoption is curbing compensation as AI automation takes place. Again, whether the easing in compensation pressures stems from an increase in AI automation is a question worth investigating, but COVID confounding obscures it, as the decline could be due to a slowdown in demand after the crisis.

Chart 1

Third, is the end of the health crisis curbing compensation in nursing?

Likewise, the nursing work compensation increases, which peaked in early 2020, are often thought of as an outcome of the dramatic shortage of professionals during the COVID pandemic. Chart 2 pictures the common understanding after the health shock, which in turn nuances the question of whether the decline is an effect of COVID, AI, or labor supply.

Chart 2

This final question in itself exemplifies the challenge of the COVID confounding. Disentangling COVID becomes the noisiest data factor biasing research estimations on the current US labor market. Analysts cannot ignore the COVID confounding and jump straight to estimation methods without accounting for it.

At Econometricus we deliver quality mixed-methods research that ranks high on both academic and social media impact. We build cogent data-rooted rationales that machines (AI) can not match. Sources wise, we have more than fifteen years of experience leveraging most public sources datasets.

Feel free to get in touch to discuss your research needs.