Midblock conflict study was undertaken to explore the relationship between fatal crash sites and conflict rates. Peak hour traffic at fourteen selected locations was video taped. Trained observers recorded traffic compositions at midblock, average space mean speeds by modes and conflicts by type and mode.

The studies showed a weak crash-conflict association. Comparison of conflict data for various sites, and for different combinations of conflict types revealed that the presence of only a few non-motorised modes can cause conflicts between motorised vehicles and non-motorised vehicles. The study did not provide a conclusive relationship between midblock conflicts and fatal crash sites, but the conflict studies provided useful insights into the interaction between different traffic entities present in the traffic streams.

An important conclusion of this study is that a traffic-plannning emphasis on conflict-removal, without accompanying epidemiological studies, does not result in fatality reduction, even if conflicts are removed.