Researchers acknowledge that they didnt quantify how much fluids people drank.
They only measured serum sodium, which is a reflection of the salt and water balance in the body, but not really a reflection of overall water stores or hydration status. Relatively low serum sodium actually can be pathologic, and is classified based on total body water status, i.e. hypervolemic, hypovolemic, or euvolemic hyponatremia. The actual serum sodium level will not tell you the patients fluid volume status.
The most common clinical scenario in which we will see high serum sodiums is in nursing home patients who do not have access to water and are free water deplete. However, the osmotic balance in anyones body varies and is generally keeping the serum sodium between 135 and 145. The study does not necessarily take into account all the entities that contribute to a relatively lower serum sodium, such as a reset osmostat as we age, SIADH (a sodium lowering entity caused by many medications), diuretic use (for hypertension or chronic edema) which also can lower serum sodium.
Using serum sodium is at best and indirect measure of hydration status and not really a good reflection of how much water a person drinks.