Why is exposure data often lognormal in IH?

Why is exposure data often lognormal in IH?

• A. Because the data are normally distributed after log transformation.
• B. Data are multiplicative and skewed toward higher values.
• C. Exposure data are bounded between 0 and 1.
• D. Exposure data are deterministic.

Answer: (B)

Exposure levels arise from many independent, positive factors that multiply together to determine what a person is exposed to. Each factor can vary across situations—emission rate, ventilation, air mixing, breathing rate, duration, protective measures, and so on. Because these influences compound multiplicatively, the overall exposure tends to be right-skewed, with a long tail toward higher values. When you take the logarithm of exposure, multiplication becomes addition, and the sum of many independent contributions tends to be approximately normally distributed. That combination—multiplicative, variable factors producing a right-skewed original scale and a roughly normal log scale—is what makes exposure data frequently lognormal. It’s a natural way to model the variability and range seen in real-world exposure. The alternative that data become normally distributed after log transformation describes the consequence of lognormality, not the reason behind it. The other options don’t fit because exposure isn’t inherently bounded between 0 and 1, it isn’t purely deterministic, and the raw data are not typically normally distributed.

Exposure levels arise from many independent, positive factors that multiply together to determine what a person is exposed to. Each factor can vary across situations—emission rate, ventilation, air mixing, breathing rate, duration, protective measures, and so on. Because these influences compound multiplicatively, the overall exposure tends to be right-skewed, with a long tail toward higher values.

When you take the logarithm of exposure, multiplication becomes addition, and the sum of many independent contributions tends to be approximately normally distributed. That combination—multiplicative, variable factors producing a right-skewed original scale and a roughly normal log scale—is what makes exposure data frequently lognormal. It’s a natural way to model the variability and range seen in real-world exposure.

The alternative that data become normally distributed after log transformation describes the consequence of lognormality, not the reason behind it. The other options don’t fit because exposure isn’t inherently bounded between 0 and 1, it isn’t purely deterministic, and the raw data are not typically normally distributed.