Observations of traffic inspections at a U.S. land border port of entry in El Paso, TX indicate that the process is highly variable. In a series of 24 half-hour observation periods of ordinary non-commercial traffic, the average inspection duration ranged from 16.6 s to 56.6 s. The proportion of inspections which involved some physical search of the vehicle, as indicated by the inspector leaving the inspection booth, varied from 5% to 56% in different observation periods. In 4 out of 10 cases, the log-mean of inspection duration in simultaneous observations of parallel lanes of traffic differed significantly (p<0.05). This suggests that differences in inspector behavior are responsible for much of the variability of the inspection process. Similar results are found for the SENTRI program. A survey of public perception reveals that a majority of English-language respondents perceive the inspectors to be fair while a majority of Spanish-language respondents (both U.S. and Mexican citizens) perceive the process to be more arbitrary, as they state that fairness "depends on the inspector." Spanish-language respondents are also more likely to report having to submit to additional searches than English-language respondents. A common theme that emerges from the analysis of these two datasets is that efforts to standardize some aspects of inspections, while preserving inspector autonomy, may improve the performance of the process by eliminating variability which organized criminal groups may be able to exploit.