A step-by-step methodology for computing delays and user costs in highway work zones is presented in this paper. The methodology is based on the relationship between speed and capacity in work zones. Speed is adjusted to account for the adverse effects of roadway geometrics, work intensity, speed enforcement, and other work zone factors. Speed reductions for work intensity and narrow lane widths were determined previously by use of data from work zones. Values from the "Highway Capacity Manual" were used where data from work zones were not available. The fundamental difference between this methodology and previous approaches is that it accounts for the various work zone factors. Two applications of the methodology are presented, one with queuing and one without queuing, to real-world work zones. Results are compared with field data. For the nonqueuing site, the computed speed and capacity were 44.4 mph and 1,778 vehicles per hour per lane (vphpl), respectively, while the field speed and capacity were 44 mph and 1,708 vphpl, respectively. For the queuing site, computed capacity was 1,012 vphpl while the field capacity was 1,220 vphpl. Considering the assumptions made in the methodology for some of the parameters, these results build confidence in the methodology. Extensive validation of this methodology is strongly recommended.