Forum Topic

Tom: And on your further point, that the increase in the control group mileage might be down to the LTN ignores the fact that the households nearest to the the LTN, and therefore most likely to avoid the restrictions in their road access by driving longer routes, saw no increase their their mileage.That is not what the report states.  Table 2 which deals with their entire sample shows that average daily mileage INCREASED by 0.3 km for the near LTN households (and control area residents saw driving increase by an average of 0.6 km)I would however highlight table 3 which deals with a subset of the data - for "stable households" only - which shows that Near LTN residents reduced average daily driving by 0.3 km.  That subset was used to test a specific hypothesis namely there was resident self selection (namely that households that drove a lot sold up and left) and I think it demonstrates that reasonably well but that is as much as the authors of the report were prepared to use that subset, so I am respecting their decision.Coming back to table 2 and the INCREASE in daily driving by Near LTN residents, that could be as a result of having to take the long way round to their destination but there could be a multitude of other reasons.  Similarly the control group increase in average mileage could be as a result of the LTN forcing longer journeys or it could be that people are driving more generally, or any number of other factors.I stand by my original post.  The report demonstrates a clear correlation between LTNs and slightly less driving but it does not show that LTNs are the cause.

Justin Stephenson ● 1194d

Lets take a step back from all the various assumptions. So, firstly as Tom and Michael have made clear, LTNs are designed to make it harder for residents within the areas to drive. Reduce access or exit points and other restrictions within an area. So it would be quite astonishing if they didnt to some degree achieve that.Secondly, the data covers the pandemic but neither the author nor Tom mention it. Who would have thought that all cycle lobbyists seem to be COVID deniers, yet a global pandemic that altered daily life for all of us is not worth a mention.THirdly, the approach used relies on vehicles being owned throughout the period of study. So around half of the total vehicles in the area are excluded as there is no usable data. So lets set this aside and see what happened. So yes, measures aimed to make it harder to drive did see a modest around 3% fall in distance driven. Whilst this may be deemed statistically significant it is also very small.However, it becomes interesting when looking at other area. Tom has told us that distance travelling by car is falling in london, so why is it going up in Lambeth outside the LTNs? And as to the control group, Tom says, his option, that the increase in distance driven was greater in the control areas that are slightly furhter away was 'quite the opposite of what would be expected'. But this assumption by Tom is quite likely completely false. Google maps will often offering differing routes that diverge in some significnat manner, it is quite possible indeed likely that such mapping apps would avoid the whole area of LTNs due to possible congestion around them. In which case, that would suggest LTNs are indeed promoting drivers to travel more distance. The counter argument relies on a narrow focus on boundary areas, where there may indeed be more limited impact as some journeys are made to different locations rarther than attempting to circum navigate the LTNs.So we have a very modest fall in traffic in an LTN, but an increase in distance for the control area a relatively small distance away. There may of course be no causility here, but equally there may be.The final point, utterly ignored by Tom, is the very significant increase of 10% in vehicle registrations in the study areas. Which goes against wider London trends. So the most obvious conclusion is that LTNs encourage car ownership.

Bill George ● 1194d

Bill, it would be helpful if you stuck to what I wrote rather than make up your own version. I explicitly mentioned the pandemic and how the study accounted for it, so your claim otherwise is demonstrably false. For those just outside the boundary, geometrically half of their driving directions are cut off by LTN restrictions. That geometrical constraint inevitably reduces as you move further from the LTN. Of course the actual effect depends on the road network, but you end up having to make a very convoluted case if you want to argue that for some reason the LTN has an increased effect on driven mileage with increased distance from the LTN.Yet  you do more than that, you claim that on the basis of your unevidenced and overly complicated construction that the straightforward explanation, that LTNs reduce rather than increase driven mileage, is “probably false”. And you end up with again ignoring that I had already mentioned, that the reported increase in vehicles is only for those taking MOTs, not new vehicle registrations which fell in Lambeth as a whole. All the evidence is therefore consistent with fewer new vehicles leading to an aging distribution of cars. The fact that you are resorting to misrepresenting or ignoring what I said, and introducing what are in effect unevidenced nonlinearities in the response of mileage to distance from LTNs.Simply the study provides evidence of what one would hope an effective LTN might achieve. It should give those who claim to oppose LTNs on the basis that they only increase mileage pause for thought. In contrast, opponents who have an ideological prejudice against LTNs will instead deploy just the arguments you have made.

Tom Pike ● 1194d

Tom, its not 'overly complicated' to assume that people a short distance away from an LTN may indeed travel slightly further as either an individual's concern about congestion in an area or a mapping app that makes that decision diverts them around an areawith numerous LTNs rather than directly skirting it. I think that would probably mirror many people's real world experiences. People have posted on here about choosing to shop in Richmond for example rather than W4 due to LTNs. Those that are nearer and do have half of their directions cut by LTNs may simply change destination for some of their travel rather than choosing to go around the LTN. For those that are further away the actual impact is as you note more modest, so they may choose to maintain the same travel pattern even on a slightly longer trip as the overall change is limited. Again you are making all kinds of assumptions about individual behaviours and travel choices based on your own perspective.So including new vehicle registrations do you have data showing total number of active permits for these areas over time? THat would be helpful. As for COVID, its a huge surprise that given lockdown and other restrictions all areas did not show a reduction in milage driven in the post LTN period. It certainly had a huge impact on many aspects of life so does unfortunately raise a fundamental question of what can be concluded from data in the time period. The study shows a very modest fall in distance travelled in an area that has gone to great lengths to make it harder to drive. Did we really need all this huffing and effort to show the blinking obvious. It also seems to be in contrast to some other studies recently that have shown increases in distance travelled by residents in LTNs. But what you dismiss as 'unevidenced nonlinerearities' is a quite feasible interpretation of real world behaviour.

Bill George ● 1194d

Tom the report is a classic example of statistical slight of hand.The report compares the LTN and Near LTN to a control group.  The purpose of the control group is to test the null hypothesis namely that the changes in driving averages in the LTN and Near LTN households were a product of chance when compared to an area unconnected and unaffected by an LTN.  That means your control area needs to be unconnected and unaffected by the LTN. In this case there is no evidence that the control group is unconnected and unaffected nor any methodology to ensure that, in fact it is the opposite because of the closeness of some of the control group to LTNs (in some case 201m away but a lot would depend on normal driving routes of the control group) it is highly likely that an unknown proportion of the control group are directly affected by the LTN.That is not a proper control group - it is the equivalent of doing a medical trial to test safety of a drug where you have one group take a full drug dose, a second group taking a 50% dose and the control group rather than being given a placebo instead get a 20% dose.  That is just a bad methodology. I admit I am not a fan of using p- values, especially when there could be a number of factors influencing the data, but that is also the opinion of a large number of professional statisticians.  There are several other tests, well beyond my mathematical competence, which are more useful when there are potentially several variables which influence the data.Based on the actual data all we can say with a high degree of confidence is that the pre v post average daily mileage shows a difference in each of the 3 groups which is probably statistically significant for that group.  I would go further and state that the differences between the groups are potentially significant, especially for those households in the LTN, but what we cannot why nor can we conclude with any confidence that the differences are down to the existence of the LTN.

Justin Stephenson ● 1193d

Tom, I will defer to Justin whose knowledge of statistics is clearly stronger than mine, but I agree with his point and in particular the probelm of the proximity of the control group. You dismiss the idea that the LTNs could have any impact on the journeys of those in control groups has no basis. The LTN areas were presumably selected in part due to volumes of through traffic, so removing this would impact a number of journeys many of which would have originated outside the immediate boundary area. My point here is that you robustly denied my comment that vehicle ownership had increased significantly in the study areas, with the report showing a 10% increase in vehicles. You claimed that it was because the study only looked at older vehicles that had MOTs. However, this is in fact a basic error and failure to read the report. The report clearly states that 'Among cars/vans with an active permit as of 1st March 2020, we identified 8575 cars/vans in scope for our analysis of ‘pre-LTN’ mileage'. As made clear in table 2, the analysis starts with the number of 'active parking permits', a figure that did indeed increase by 10% between March 2020 (18.2k vehicles) and March 2023 (20k vehicles). This universe is then reduced as vehicles that did not have MOT data were removed (the majority of which were described as 'too new for sufficient MOTs').You asserted that the study only had vehicle data for older vehicles with full MOTs, and that:'the reported increase in vehicles is only for those taking MOTs, not new vehicle registrations which fell in Lambeth as a whole. All the evidence is therefore consistent with fewer new vehicles leading to an aging distribution of cars. 'But as I demonstrate above, the data started with the total number of parking permits, not the number of vehicles with MOTs. Your second point that this shows an aging distribution of cars is not supported by the data and another factual inaccuracy.Tom, the ease with which you both dismiss the reasoned views of others and then make robust statements that are not supported by, or in this case completely contradicted by the data is deeply concerning given the prominence which some people attempt to give to your views.

Bill George ● 1193d

Bill, as I said, the paper does not give numbers for new vehicle registrations so I took that from the DVLA data for Lambeth as a whole and that showed a 4% fall. You are correct that the  number of parking permits issued increased, but that will be both a function of the number of cars owned and the coverage and functioning of the schemes since 2018 to date. Whatever the cause, the number of permits issued increased the most in the control area outside of the LTN, which hardly suggests the LTN itself was the cause for new permits being issued.I based by comments on an ageing car profile on national figures, but the numbers on the paper also support this shift: while the total number of permits increased by 10% over this period, the number of vehicles that had MOT data increased by 20%, consistent with an ageing population.Taking together what we know from the DVLA about new vehicle registrations as a whole in Lambeth and what we know from this paper about the proportion of cars with parking permits that have MOTs, it supports my comment that we are looking at an ageing distribution of cars.Again you misrepresent what I have written. I have not claimed that households in the control area are unaffected by the LTN, just that they are less affected than those living within 200m of it. Indeed, Justin's criticism that a control group should be completely unaffected by the LTN, similar to those given a placebo in a medical trial,  is not realistic. The area of the control group would need to be far enough away from the LTN such that no trips were affected by it: the justified criticism would then be that the control group was no longer a good match to the group within the LTN. That would make it difficult to attribute any differences to the LTN itself.Justin's further comment that it can be stated "with a high degree of confidence" that the changes seen in all three groups are statistically significant ignores what the paper actually states, that the change seen in the boundary group was 0.3±0.5 km. You don't need a great deal of statistical knowledge to understand that this result is statically indistinguishable from zero.

Tom Pike ● 1193d

Michael, the debate was not whether CPZs have increased generally in Lambeth, but whether there had been any cnages in the study areas. The report specifically says:'The north of Lambeth is covered by controlled parking zones, with residents requiring a permit for on-street car/van parking on their local streets. Lambeth Council provided us with numberplate plus home postcode data for active ‘resident’ and ‘disabled’ parking permits as of 1st March 2018, 1st March 2020, and 1st March 2023.'I am sure you can agree this statement is pretty clear and doesnt imply any change in the geographic make up of these areas. Changes in CPZ areas  would I expect then to have been called out by the authors, as it will bring different households into the study for the later period who could not have been included in the earlier period. This would raise questions over the comparability of the datasets used for the two periods.You are an intelligent person, please do tell me if my points are reasonable? If there is no change in the coverage of the CPZs then my observation that it is surprising to see a 10% increase in permits issued over the period also remains valid.Tom and a lot of his fellow travellers seem to be quite selective in their use of data and then very narrow in the conclusions they draw, bringing their inherent bias into the conclusions.I similarly Michael, ingore you tiresome efforts to argue obsessively about cycling, talk down the relevance of public transport and fail to reach any common ground with anyone who doesnt share your narrow focus. If Tom stops posting his reports and data with dubious conclusions, then I wont challenge them. Of course, this is a discussion forum... or should we just bow down and accept whatever you and Tom say?

Bill George ● 1193d

Michael, you are right that there have been changes in the CPZ boundaries in Lambeth, but looking into this more I don't think that explains the 10% increase in residents permits issued. As I said before, the increase was actually largest in the control area outside of the LTNs, so clearly this has nothing to do with the LTNs themselves. There is in fact a very good reason why households who already had cars but didn't previously need a residents permit applied them in 2020 - they changed to working at home rather than commuting and so their cars were parked on the street during the operational hours of the CPZ. This is confirmed by data released by Lambeth which shows a 12% increase in applications for residents parking permits in 2020, much larger than any change in previous years.I'm sure some will claim that this invalidates the entire study, but that would be to ignore the analysis performed. First of all, any shared effect will be mitigated by looking at the changes with respect to the control area, reducing the contribution of common factors such as this. In addition the authors performed a confirmatory analysis using only vehicles that provided a mileage both before and after the LTNs were introduced. This analysis further reduces any potential bias introduced by a change in the mix of cars that had parking permits. That second analysis gave an even larger reduction in mileage in the LTNs compared to the control areas, a drop of 2.3 rather than 1.3 km a day.The claim, based on an increase in parking permits, that LTNs themselves encouraged people to buy cars was never credible. It was simply another side effect of the pandemic.

Tom Pike ● 1193d