Assessment of spinal stiffness is frequently used by manual therapy practitioners as a part of clinical diagnosis and treatment decisions. While studies have commonly found poor reliability of such procedures, conflicting evidence indicates that assessment of spinal stiffness may help predict response to specific treatments. This study evaluated the criterion validity of manual assessments of spinal stiffness by comparing them to indentation measurements in patients with low back pain (LBP). As part of a standard examination, an experienced clinician assessed passive accessory spinal stiffness of the L3 vertebrae using posterior to anterior (PA) force on the spinous process of L3 in 50 subjects (54% female, mean (SD) age=33.0 (12.8) years, BMI=27.0 (6.0) kg/m2) with LBP. A criterion measure of spinal stiffness was performed using mechanized indentation by a blinded second examiner. Results suggested that manual assessments were uncorrelated to criterion measures of stiffness (spearman rho=0.06, p=0.67). Similarly, sensitivity and specificity estimates of judgments of hypomobility were low (0.20-0.45) and likelihood ratios were generally not statistically significant. Sensitivity and specificity of judgments of hypermobility were not calculated due to limited prevalence. Additional analysis found that BMI explained 32% of the variance in the criterion measure of stiffness, yet failed to improve the relationship between assessments. Further studies should investigate whether manual assessment of stiffness relates to other clinical and biomechanical constructs, such as symptom reproduction, angular rotation, quality of motion, or end feel.