One of the many pieces of excrement flung against the wall by RG4N in the desperate hope something would stick was an ITE Journal article in which the author asserted a disproportionate (to square footage) traffic impact for “free-standing discount superstores” over 200,000 square feet. The conclusion, in other words, was that 199,999 square feet stores should have a trip generation figure of X per square foot; while 200,000 square foot stores should have a trip generation figure of Y, where Y is much larger than X.
This is counter-intuitive to say the least. One could argue that the increased size results in more trips overall – which would be the result of continuing to apply X trips per square feet (X times 200,000 is obviously more than X times 100,000). One could even argue that the increased size results in fewer trips than the same number of square feet in _two_ stores (“one-stop shopping”). But the theory that a bigger store results in, and I emphasize units here, more trips per square foot has always seemed ludicrous to me.
Anyways, as it turns out, Wal-Mart went with a slightly smaller store – which the army of anonymous RG4N trolls have used for quite a while as conspiracy fodder – claiming that they snuck it in under the threshold to avoid these supposedly more valid rules (which, again, as far as I can tell, the ITE still hasn’t seriously considered adopting).
As it turns out, I wasn’t alone in my skepticism. In addition to several disagreements about methodology, the respondent (another traffic engineer) points out that the study was too small to be statistically rigorous; the stores were too different to draw any firm conclusions; and that the author’s supposed intuitive conclusion isn’t. Some excerpts follow, since I’m not sure how long this article stays up for free. I’ll leave out the most esoteric stuff.
As a transportation consultant who is involved in both the performance and the review of traffic studies, my colleagues and I at McMahon Associates, Inc. are extremely concerned that the August 2006ITE Journal article entitled “Trip Generation Characteristics of FreeStanding Discount Superstores” lacks the rigorous scientific analysis and thoroughness that we have come to expect in ITE Journal articles.
As such, although ITE Journal states: “Opinions expressed herein are those of the authors and do not reflect official ITE or magazine policy unless so stated,” the article may be utilized by transportation professionals and others as “gospel” even though its analysis is flawed, in our opinion, in many respects.
2. Additionally, the square footage of a gas station is not a good choice for independent variable, as compared to the number of fueling positions, when determining its estimated trip generation; i.e., a 225-square-foot building could serve four fueling positions or 14 fueling positions.
5. We also question whether the author confirmed, in her comparison to the ITE Land Use Code 813 rates, that the latter (ITE) square footage baselines are the same as she assumed, especially with regard to the garden center, which typically has significant (15,000 to 20,000 square feet) square footage. While we agree that the rates should be applied to “total” square footage, inclusive of a garden center, it is our understanding that the ITE samples were largely (or totally) based on building foundation square footage, not inclusive of outside garden centers. Our observations about baselines and “with and without gas pumps” are intended to reinforce our opinion that the author’s analysis appears to be an “apples to oranges” comparison rather than “apples to apples.”
7. There is also a fairly large discrepancy between the number of vehicle trips collected between different days at some of the supercenter locations. Site 3 shows an increase of almost 17 percent in site traffic between the day 1 and day 2 counts. The increases in site traffic between the day 1 and day 2 counts at site 1 and site 5 are both about 10 percent. The fluctuation in these counts suggests that there could be flaws in the data or that other factors may have been involved in the traffic generation of the site on one or both days of the counts. These discrepancies may reflect seasonal variations, as the article indicates that the first weekday count was taken in July while the second count was taken in October.
and here’s the one that I think is the most important to laypeople:
9. We also take issue with the author’s statement that “free-standing discount superstores intuitively should have a higher trip generation rate than free-standing discount stores, which by definition do not contain a full-service grocery store but have most of the other amenities of the superstore.” Are not shopping centers evidence that larger stores, with presumably more services or products in one location, result in documented lower trip rates, because customers shop longer and their shopping needs can be accommodated in fewer trips due to greater availability of goods and services? In fact, the author’s argument is shown not to be the case in Table 1 of the article, where the author’s own comparisons show that, as retail store sizes become larger and more services/products are offered, trip generation rates decrease. We also note that the number of samples for ITE free-standing discount store (47) and ITE shopping center (407) is large enough so as to make these land uses’ rates statistically more reliable than ITE’s rates for free-standing discount superstore (10 samples) or the author’s study (five samples).
In conclusion, while the author’s study and article adds to the body of knowledge on trip generation characteristics of superstores in excess of 200,000 square feet, its data and analysis of that data, we submit, are not rigorous or conclusive enough to support the article’s recommendation that the rates derived from the author’s analysis should be used as the future norm for 200,000 square-feet-plus superstores. Until such time that more samples are collected (we would recommend at least 20); preferably from various locations in the country, as she also recommends, to test geographic differences, if any; and are computed on common baselines first (separately, without, or with gas pumps) before combined (i.e., if not statistically different), we suggest that the jury is still out on the validity of this article’s rates, conclusions and future use.