Systemic inaccuracies, proportional to the concentrations of serum proteins and the thyroxine (T4) they carry, have been reported in direct free T4 immunoassays. However, analytical recoveries of free T4 have not been carefully examined in most current methods, and they have not previously been examined across the pathophysiological range of serum T4 binding. In the present study we investigated ranges of serum T4 binding using free and total T4 measurements from 1359 individuals. Carefully characterized, gravimetrically calibrated, serum-based free T4 test solutions were then prepared with a constant normal free T4 concentration (12 ng/L) and varied serum T4 binding (approximately 300:1 to 24,000:1, ng protein bound T4: ng free T4). These standardized test solutions were analyzed using five T4 analog based free T4 methods. Analytical recoveries were calculated as ratios of actual free T4 measurements to the target value, and expressed as a percent of the target. Analytical recoveries were directly proportional to the extent of serum T4 binding and ranged 2% to 155%, 25% to 131%, 53% to 106%, 37% to 93%, and 37% to 73%, lowest to highest, in different methods. These systemic inaccuracies will confound interpretations of free T4 test results in clinical conditions with altered T4 binding. Future investigations into free T4 status must examine the analytical recovery of the free T4 method(s) used, as they relate to the extent of serum T4 binding in the clinical condition(s) studied.