Abstract
Increased levels of walking have been associated with a range of individual and societal benefits, including reductions in energy use and greenhouse gas emissions, and improved public health. Recent scholarship has sought to provide evidence that good street connectivity encourages walking, and though some correlation is evident, the statistical significance of these studies is generally weak. This paper asserts that the most commonly used methods to measure street connectivity, intersection density and block length, are weakened by random variation, and more significantly, utterly fail to detect street patterns. An alternative measure, modified route directness, is introduced, and shown to be capable of measuring the impact of street pattern on permeability. A more effective measure of connectivity could help demonstrate the correlation between street connectivity and walking, enabling stronger arguments for changes regulatory activity and public investment that supports improved connectivity.
Notes
1. As measurements from an intersection on a grid form right angles, Pythagorean theory can explain these results. Any point along a line proceeding at an angle deviating 45° from the grid will have a route directness measure of 1.41 (dotted lines in Figure ), while any point along a line following the grid (0- or 90° angle) will have a measure of 1.00. The parcels in the second row fall closer to the 45° angle than those in the 10th row (average deviation of 17.8° vs. 26.2°), hence, their route directness scores are worse. If blocks have the same length width ratio, these same angles will result, regardless of block size.
2. This same phenomenon also resulted in the score for an 800′ × 800′ grid scoring slightly better than an 800′ × 600′ grid. These imperfections are a modest price to pay for the ability to detect street pattern as will be shown in the next section.
3. The area shown extends approximately 1/4 mile in each direction beyond the study area, because connectivity measures for street segments near the edge of the study area extended to these streets. The measurements taken for block size and intersection density included all of the same blocks that were part of the route directness measure.