While physiotherapists commonly use manual tests to assess the stiffness of the spine, these tests have been shown to have poor to fair reliability. This paper reviews the literature on stiffness perception and suggests two possible approaches to enhance the objectivity and accuracy of stiffness assessment. The first approach is to abandon manual assessment of stiffness and instead rely upon instruments to measure stiffness. The second approach involves continuing with manual assessment but adopting strategies that have been shown in other disciplines to enhance manual stiffness assessment. These strategies include clearly defining what physiotherapists mean by stiffness, identifying and controlling factors in the clinical environment that affect stiffness perception, and instituting training programmes that provide quantitative feedback on performance. Changing the method of rating stiffness magnitude, so that stiffness is rated relative to reference stiffness stimuli rather than the physiotherapist's expectation of normal, may also improve the value of manual assessment of stiffness.