PHOENIX -- On election night, Jodi Owings and her family watched the results reported live on television in their Oklahoma home.

That's when she noticed the wording on a CNN graphic that displayed returns by race as white, Latino, Black, “something else” and Asian.

Owings, a citizen of the Seminole Nation of Oklahoma, asked her family if “something else” referred to them. The wording stood out because there's often a lack of reliable data on Native Americans, she said.

Native Americans make up less than 2% of the U.S. population and often are listed in datasets as “other” or denoted with an asterisk. Even when surveyed, the results can be considered statistically insignificant because the sample size isn’t large enough or the margin of error is too great to accurately reflect the population.

Stephanie Fryberg, a psychology professor at the University of Michigan who has surveyed Native American populations, said survey methods need to be reimagined so they can better capture Indian Country and the diversity within it.

Random sampling doesn’t work well, particularly in reaching tribal members who live on reservations or rural areas, she said. She and others have found success in getting people to respond to surveys by forming partnerships with Native organizations, tribes, tribal colleges and Native news outlets.

Read the full article at ABC News.