Of all textile fibers, cotton is the market leader with about 20 million tons produced and processed annually. The sale price of the fiber is established on the basis of its quality which is evaluated either traditionally or by instrumental techniques as High Volume Instruments that are used increasingly frequently. These machines characterize about ten technological criteria that reflect the intrinsic properties of the cotton fiber. In an effort to establish price/quality rules, industry organizations have standardized the instrumental methods used to measure the main fiber parameters. Absolute measurement methods based on quantifiable, physical parameters are now being sought to establish standard values for each characteristic in order to calibrate the HVI lines. The aim of this thesis was to analyze the different types of bias in HVI measurements of fiber strength (specific resistance to rupture force / linear mass) related to other fiber properties, and assess their statistical distribution and any interactions or combinations. The results of this novel study indicate that the methods used to measure the characteristics tested lead notably to an incorrect assessment of the ruptured fiber mass during the dynamometer test. The importance of this bias should be considered in the light of the magnitude of other errors and the overall precision of the measurements. Unless substantial improvements are made to the equipment, HVI lines will prove to be unsuitable as the reference method for the measurement of strength on the basis of measurable physical parameters. From a scientific standpoint, a measurement of the major intrinsic properties of each fiber would be required to gain a full understanding of the phenomena observed. In commercial use, the confidence limits of the results produced by HVI lines nevertheless remain acceptable.