Measuring small populations within the United States can be difficult. Here is an example: even though two separate studies agree the US Jewish population is roughly 6.5 million, they used different methodologies to arrive at this number:

Many federations around the country commission scientific studies to better understand their local Jewish populations. These reports typically rely on random digit dialing, in which researchers come up with a percentage of Jews in the community based on the results of telephone surveys. In other instances, researchers will estimate the number of Jews based on the number of people with Jewish last names.

These reports provided the backbone for Sheskin and Dashefsky’s own annual estimate. But since not every federation studies its own population, the two conducted original research in some localities. In this, they were often aided by knowledgeable community members or by local estimates they found online. Lastly, they used data collected by the U.S. Census of three solidly Hasidic Jewish towns in New York state: Kiryas Joel, Kaser Village and New Square. (Aside from these exceptions, the U.S. Census does not count Jews.)

Adding these figures together, Sheskin and Dashefsky came up with a national estimate — albeit a patchwork one — that far exceeded previous figures. And in some ways exceeded their own expectations. Their national total of 6,588,000 is an overestimate, they contend, because some Jews — such as college students who live in one place and go to school elsewhere, or retirees who live part-time in one city and part-time in another — were likely counted twice…

Saxe came to his national estimate of 6.4 million through very different means.

Daunted by the steep expense and lengthy time required by random digit dialing, Saxe and his team ferreted out data that already existed to reach his conclusion. This included information from more than 150 government surveys on topics completely unrelated to Judaism, such as health care or education. Each study had a sample size of at least 1,000 people, and each study asked the question: What is your religion?

“From this, we are now absolutely confident — and it has been vetted by all sorts of groups and people — that about 1.8% of the adult American population says that their religion is Judaism,” he said.

Saxe adjusted his sample to account for children and came to a total of 6.4 million Jews in America.

In order to count and know more about relatively smaller populations in the United States, say Muslims in the United States when asking questions about religion, survey researchers often try to oversample these groups so that they can draw conclusions from a larger N. But as this article notes, finding people of smaller groups through random-digit dialing can take a long time.

Both of these researchers worked with existing data in order make generalizations: one worked with local figures and the other used a sample of large-scale surveys. In both cases, this is a clever use of existing data because doing a large-scale survey would have likely been a lot more costly in terms of time and money.

I would guess both sets of researchers are happy that their figures are close to those of the other study as this enhances the validity of their numbers.