The mechanical strength of inorganic porous hollow fibers is a critical constraint that limits their wide scale application. Various methods, including 3-point bending, 4-point bending, and diametrical compression are used for the quantification of the mechanical strength. Here, we show that these methods cannot be used in an interchangeable manner. For large sets of alumina hollow fibers, the parameters describing the cumulative probability of failure functions depend on the type of measurement, i.e., 3 or 4-point, the span size, and the measurement geometry. This implies that reporting data on mechanical properties of inorganic hollow fibers requires that extensive information about the experimental details is provided, and that a direct quantitative comparison between datasets is unjustifiable.The mechanical strength of the alumina hollow fibers tends to follow a normal distribution, or log-normal distribution, instead of the often used Weibull distribution. Monte Carlo simulations demonstrate that, especially at small sample set sizes, it is difficult to accurately determine the shape of the probability distribution. However, detailed knowledge of the type and the shape of this distribution function is essential when mechanical strength values are to be used in further design.

No abstract