Different amounts of a low (L/h = 100) and a high aspect ratio (L/h = 22,000) synthetic clay (hectorite) have been compounded into a polyetherimide matrix by solution casting. By applying these fillers with largely different aspect ratios we were able to validate the different and partially contradicting models found in the literature to predict coefficients of thermal expansion (CTE) of nanocomposites. A comparison with experimentally observed CTEs identified Lu's model to give the best agreement. In this model, the CTE is highly non-linear in respect to both the aspect ratio and the filler content allowing to identify technically benign combinations that assure a given target-CTE. For instance, for printed circuit boards made of thermoplastic PEI, an aspect ratio of around 1000 and a filler content of about 13 vol% are predicted to perfectly match the CTE of the nanocomposite to the CTE of copper while maintaining processability.