The stress experienced by dogs admitted to animal shelters not only impacts welfare, but may lead to behavior that threatens successful adoptions. This stress is reflected in an elevation in the plasma cortisol levels of newly admitted dogs. We previously found 30 min of human interaction reduced the plasma cortisol response dogs showed to shelter housing. The goal of the present study was to assess the practicality and effectiveness of the human interaction procedure as a means that volunteers and staff might use to reduce the stress of shelter dogs. We, therefore, investigated several parameters of the effect of human interaction on plasma cortisol concentrations and behavior that were chosen for their relevance to this goal. The current study found that a second day of 30 min of petting reduced cortisol levels as effectively as the first (P < 0.05) and that 15 min of this interaction was as effective as 30 min. However, the reduction in cortisol concentrations did not persist when dogs were returned to the home kennel. During petting, signs of excitation (vocalizations) and anxiety (panting) as well as escape attempts were reduced, and social solicitation (tail-wagging) increased (P < 0.05 for all behavioral effects). Cortisol reductions were pronounced in dogs admitted as strays (P < 0.01), but human interaction did not reduce cortisol levels in a subpopulation relinquished by their owners. We also measured hair cortisol levels to assess stress prior to shelter admittance. Strays and dogs released by their owners showed comparable hair cortisol concentrations that were intermediate to those of pet dogs living in a home and those of dogs diagnosed with Cushing’s disease. These findings show that as little as 15 min of human interaction can moderate cortisol levels of shelter dogs. The cortisol reduction is repeatable and associated with positive behavioral change. However the effect is of short duration and appears to vary with the source of the dog. Finally the accumulation of cortisol in hair may be a useful method to estimate the condition of the dog prior to shelter admittance.