Laserfiche WebLink
Nor~lization Technique <br /> <br />Two survey results indicated that data normalization ms desirable for this <br />study: 1) Smae data groups have 8ignifican~ly hisher percentases of return <br /> than others, with the hishest percentases coming from groups which <br /> represent smaller portions of the 24-hour traffic volume (AADT), <br /> i.e,, AH Peak - 66% return, 13% of AADT, Hid-day - 49% return, $8% <br /> of AADT. This situation would tend to over or under emphasize the <br /> respective data groups. <br /> <br />Plottin8 the response to question 5 with respect to time and <br />comparin$ the graph to the 24-hour distribution of traffic also <br />indicated that the mid-day was under-represented. <br /> <br />The normalization technique employed adjusts the effect of each sample group so <br />as to not unduly influence the aggregate result. <br /> <br />First, the 24-hour distribution of traffic ~as plotted. See the graph on the <br />follo~n8 paEeo Next, the peak periods were identified and the percentage of <br />the 24-hour traffic included by each peak period ~as determined. The ~eehiay <br />raw data vas combined into 3 time of day groups, i.e. A1 + C1 - AH peak, A2 & <br />C2 - off peak, A3 + C3 - PH peak. Next, the percent return by group vas <br />determined as a percentage of returns. A factor was determined to reconcile <br />the percent return by group to equal the percent of 24 hour traffic the group <br />was to represent. This factor was applied to the ra~ data and the normalized <br />veekday aggregate percentages ~ere calculated. <br /> <br />Note: 1984 AI~DT - 31,500 <br /> 866 returns - 2.7% <br /> <br /> <br />