Taxi Driver Online

Note that my first table in the post above uses a different order for the regions than the DfT spreadsheet extract at the bottom of the post. And, of course, London is excluded in my table.

My table follows the ordering of the TPTP table further up the thread, thus with Wales at the top etc. The DfT lists them alphabetically.

The DfT uses the same regions for the fuel analysis as per its better known spreadsheets for the basic vehicle and driver numbers, which of course relate to the individual local authorities listed per region on that spreadsheet.

For comparison purpose I've assumed TPTP has used the same regions, but who knows? But the descriptions used are more or less identical, so it's presumably a fair assumption. (Except that TPTP uses 'Cymru Wales' instead of just good old 'Wales' )

Think I'll maybe leave my other two comparison tables for next week - I'm kind of losing sight of the wood for the trees, so if anyone else is actually reading this then it might all seem a bit, er, something or other

But maybe a more straightforward way of demonstrating that TPTP's fuel analysis tables are a bit bonkers is to look at another table they've usefully provided. Which is a UK national summary, and ignores the regional/Wales/Scotland/NI breakdown, but instead provides a slightly more detailed breakdown of the fuel types. Recall that TPTP's main table breaks the fuel types into three - fossil fuel, hybrids and full electric.

So there's slightly more detail here as regards the categories, and they're in order of number:



Which aren't difficult to reconcile to the main table:

Diesel - 26.8%
Petrol - 24.5%

That'll be 'fossil fuels' in the main table, which is shown as 51% in 'all regions combined'.

Then above there's 'petrol hybrid' (26.6%), 'plug-in hybrid' (6.9%) and 'diesel hybrid' (1.5%). Those total bang on 35%, which is obviously the same as the overall 'hybrid' figure in the main table with the regional breakdown.

Then there's 'electric' above, which is 13.7%, and obviously rounded to 14% and described as 'full electric' in the main table.

So that's just to reconcile the figures on the two tables on page 27 of the TPTP report, basically.

But the table above immediately reveals why the regional figures look a bit doolally. I mean, across the UK, and across both the HC and PHV codes, there are almost as many petrol cars as diesel?

And for every two diesel HCs or PHVs across the UK, there's one EV?

Make it make sense

I mean, go back to the DfT's summary table for the two codes. Of course, you'd really need to weight these numbers to be truly accurate, and unlike the TPTP table above these don't include Scotland and Northern Ireland.

But compare the %ages in the petrol and diesel columns below to the TPTP petrol and diesel columns above - they're in completely different ball parks

HCs in England @ 31 March 2024, extract from DfT stats




PHVs in England @ 31 March 2024, extract from DfT stats



And then compare TPTP's figure of 14% for 'electric' with the DfT's figures ('battery electric') - there's just no comparison...

Note also that TPTP says that around 7% of vehicles are 'plug-in hybrid' but, apart from the DfT's figure of 14.2% in that category for London PHVs, elsewhere in the country they're effectively non-existent according to DfT. And the 14.2% figure for PHVs in London doesn't really explain TPTP's 7% figure across the UK for both codes.

Ah, but of course where do the LEVCs fit into all of this, which are presumably the 'range-extended electric' column in the DfT's figures?

Well, they'll presumably be included as 'plug-in hybrids' in TPTP's numbers, thus helping explain the 7% plug-in hybrid figure for the whole UK

And indeed elsewhere in the TPTP report they usefully state that, of 1,499 total responses to the survey, 42 are London LEVCs, and 25 are non-London LEVCs.

So LEVCs equate to 4.5% of the total responses to the TPTP survey.

Ball park figures - across England and Wales, including London there were about 323,000 vehicles over both trades according to DfT.

There must be about 8,000 LEVCs in London, using the DfT's figure of 53.7% 'range extendable electric' HC figure for the capital is to be believed.

Can't be bothered crunching the numbers precisely, but outside of London the equivalent figure in the DfT's stats looks like it's around 1.0%. Which would mean 600 or so LEVCs in England and Wales outside of London.

So you're comparing 8,000 LEVCs in London with 600 outside London.

Yet, accordion to the TPTP survey, 42 out of that 8,000 responded, while a huge 25 of the 600 outside London responded.

That doesn't look right at all - I thought that maybe the TPTP results overall might have been skewed by overrepresentation of London HCDs, but in fact London LEVC drivers represent about 2.8% of total respondents to the TPTP, while the 8,000 or so LEVCs in London represent about 2.5% of the total 323,000 HCs and PHVs in the DfT stats. So in fact London HCD representation in the TPTP survey looks about right. On the other hand, LEVC drivers are hugely underrepresented in the survey outside of London - if there genuinely are 600 or so such vehicles outside London, then that's just 0.19%, or around three such drivers should have responded to the survey, but they actually got 25

Of course, maybe the DfT figures are wrong - 600 LEVCs outside of London about one year ago?

I'd have guessed there would be more, so who knows. (And, of course, the Scotland and Northern Ireland LEVC responses will be included in TPTP, but not in DfT, but that surely wouldn't explain the discrepancy...).

Of course, the overrepresentation of LEVC drivers outside of London would be eminently predictable from the kick off - if you're driving an LEVC, you're likely to be much more invested in the trade than the average driver, thus more likely to take an interest in the industry press like TaxiPoint, and thus much more likely to respond to the survey. On the other hand, to that extent you'd expect a similar overrepresentation of London LEVC drivers, but that just doesn't seem to be the case at all.

Anyway, you thought statistics was an easy subject?

And I've barely even scratched the surface at the DfT's methodology, and how that might skew the results...

But forget all that, and just look at the TPTP table at the top of the thread. And, in particular, the claim that petrol and diesel numbers are broadly similar, and that for every two diesels there's one EV

No, me neither...

Well this is my second table of the three, and basically compares TPTP's survey figure for the percentage of 'hybrids' to an equivalent figure extracted from the DfT's spreadsheets.

And, as previously excludes London, Scotland and Northern Ireland, as explained earlier.

But the main takeaway is that there's little consistency between the two. In Wales, for most obviously, TPTP thinks the proportion of hybrids is more than double that of DfT. And in most regions its a similar exaggeration of hybrid numbers by TPTP, or understatement by DfT - take your pick which way round you want to view it.

On the other hand, in the bottom two regions it's the other way round - and, in particular, the Yorkshire & Humberside TPTP hybrid figure is only around half of DfT's.

Put it another way - in both Wales and Yorkshire & Humber, TPTP's figures for hybrids are in the 30% ball park. DfT thinks the former is 14%; the latter is 57%

But the reason for Wales's low hybrid showing in the DfT stats should be obvious from the two DfT tables below - for both HCs and PHVs, they're very low in the dominant 'hybrid electric' category. And the other side of that coin is higher than average diesels in Wales, particularly for PHVs, but also for HCs to a slightly lesser extent.

On the other hand, Yorkshire & Humber is easily the highest hybrid proportion according to DfT. And that's mainly due to the PHVs rather than HCs, and the other side of the very low diesel PHV numbers discussed earlier.

Wonder why that might be? No point spending much time looking into it much more, but I think that for example in the PHV stronghold of Bradford the council has provided a lot of subsidies to the trade there to encourage a more 'sustainable fleet'?

Anyway, that probably isn't sufficient to explain the hybrid numbers for the whole region, but, you know...



HCs in England @ 31 March 2024, extract from DfT stats



PHVs in England @ 31 March 2024, extract from DfT stats



(Of course, ignore the London figures in the two table above - should really do two new graphics with them removed, but can't be bothered. Also, as per earlier, the 'others' column isn't shown in the two DfT tables above - the figures are either zero or not particularly significant, and are about gas and the like, I think...)

And been thinking about how cross-border stuff might affect all the stats - if approximately 35,000 of the West Midland's 46,000 PHVs are actually working elsewhere, then that's bound to distort the figures. But the DfT says it fuel figures are based on the postcode of the registered keeper via the DVLA database, so to that extent its regional stats for fuel should approximate to where the cars are actually working? To that degree the figures are comparable to TPTP's - they'll be for where the driver/vehicle is actually working, right?

Well to be honest, I can't remember the actual questions in that regard, and I don't think the actual survey document is enlightening about it, so just have to assume TPTP's regions are where the car/driver is actually working, as opposed to where it's licensed

(Although anyone who's really been paying attention will notice that cross-border working would in fact skew my weighted averages for the HC/PHV proportions, because that relates to where they're actually licensed, and the weighted average comes from the DfT figures for where the cars are actually working. I think )

So this is my final wee table comparing the TPTP numbers with DfT's - this time it's what TPTP calls 'fully electric', or 'battery electric' on the DfT spreadsheets. Elsewhere DfT says: "battery electric: this includes fully electric vehicles only". So surely no ambiguity here that the figures from both sources are referring to the same thing...

But, anyway, this table is both the one with the most obvious and consistently huge gap between the two sets of stats, and also the one in which - for me at least, and I'd guess most others - the DfT stats seems instinctively in the right ball park, while the TPTP numbers instinctively seem complete bananas

Only the Yorks & Humber figure comes anywhere close, and even that's hugely different - TPTP's proportion is nearly threefold that of DfT. In terms of the simple percentage gap, obviously East of England is the clear winner at 18.5%. In terms of exaggeration in another form, the TPTP figure for West Midlands is more than 17-fold that of DfT



Anyway, no point digging much into why there's such a huge gap between the two sets of stats.

However, couldn't help notice the bit about Yorkshire & Humberside in TPTP's narrative accompanying the actual numbers, which says:


In fact, and as described a few posts ago, the DfT stats show that Yorkshire & Humber is the least reliant on fossil fuels, and is also big on hybrids

By the same token, TPTP's 'least sustainable' claim is also based on Yorkshire & Humber's rock bottom EV status, as in the table above. But, in reality - or, at least, the DfT's 'reality' - it's not a whole lot different to the other regions, and TPTP's portrayal of it as being rock bottom in the EV league is surely built on wholly false comparisons.

And no point getting into all the charging cost and infrastructure stuff, but what about this as a rationale from TPTP?:


Oh, really? Of course, there is such an environment in large parts of Yorkshire etc. But check out the local authorities comprising the Yorkshire & Humber region, as per the DfT's standard 'taxi stats' spreadsheet:

North Yorkshire
York
East Riding of Yorkshire
Kingston upon Hull, City of
North East Lincolnshire
North Lincolnshire
Bradford
Calderdale [mainly Halifax]
Kirklees [mainly Huddersfield]
Leeds
Wakefield
Barnsley
Doncaster
Rotherham
Sheffield

In fact I'd guess that the Yorkshire & Humber region is actually very urbanised/metropolitan, at least where most of the HCs and PHVs will be located

I'd forgotten about the TPTP Uber stats.

Maybe all this is sending me cuckoo, but according to the TPTP survey, 17.3% of Scottish HC badge holders are on Uber

There are c.20,000 HC badges in the whole of Scotland. So that means something around 3,500 HCDs in Scotland are signed up to Uber.

Uber is only in Glasgow and Edinburgh, effectively. And, in total, there are something like 3,600 HCs between the two

Something doesn't add up. But either the vast majority of HCs in Glasgow and Edinburgh are signed up to Uber, or the TPTP stats are just nonsense. I'd go for the latter - would in fact be very surprised if there are any HCDs in Scotland signed up to Uber. But if there are, I'd guess the numbers are minuscule

Yes, it started at page one and continued to last last full stop.

_________________
IDFIMH

Haven't quite finished with this thread yet and, in the interests of fairness and clarity, I meant to post something about a few caveats to the earlier stuff. But drafted a few hundred words the other night, couldn't be bothered copying it to the clipboard, clicked on 'Submit' and it all disappeared

Will get back to it sometime, but in the meantime another big issue from the TPTP survey I just don't get - the whole dual-licence thing, and how we're given stats for that as well as HC and PH, as if it's somehow relevant. Well, of course, it is relevant to a degree, but not in the way it's been analysed and presented in the survey, in my opinion at least

Here's a telling line in the section about vehicle preferences, for example. Didn't even notice this initially, because didn't bother actually reading it, because obviously no-one's going to dispute the popularity of the Prius. But the survey says:


It was the last bit that made me a bit, as if dual licences holders were somehow some kind of distinct demographic in the trade with significant relevance to vehicle choice, but then HC and PHV badge holders also get a mention

Of course, be more specific about what the dual licence stuff is all about by changing one word, and that maybe reveals how illogical it all is; it then becomes:


Which surely underlines the disconnect between being a dual badge holder and vehicle preference - you may be able to switch codes slightly easier in that you could in theory be driving an HC one shift then a PHV the other shift without getting a new badge. But that's obviously not the case for the vehicle you're driving - it can only be either an HC or PHV - it can't be both, and you just can't replate the vehicle from shift to shift, for example

So maybe it's just me, but I think that as regards vehicle choice, the preferences should just have been divided between HCs and PHVs. (And in that regard, I'm not going to question the survey's 'Key Observation' that the "Citroen Berlingo is the top choice for private hire drivers, indicating a preference for more spacious vehicles." Not because that rings true for me, but because I don't get out much, in taxi driving terms at least )

Also, as regards the dual badge thing, the survey missed the point of the Scottish system by assuming that there are no England-type dual badges available in Scotland. Correct, in a purely descriptive sense. But in Scotland an HC badge entitles the holder to drive either/or, thus although it's called a 'taxi' badge and is distinct from a private hire badge, it's conceptually the same thing as a dual-badge in England...

Which, at least regarding the Scottish stats from the survey, may help explain some of the, er, questionable results. For example, some councils only issue HC badges, and then the drivers can drive either an HC or PHV, and that's entirely up to them.

But, as far as the TPTP survey is concerned, you're not a dual badge holder, although you actually are. And you must be driving an HC, according to the TPTP methodology, when in fact you're actually driving a PHV.

Strange, but true

And while on the subject of vehicle preferences and the survey's 'Key Observations', I did notice this when first having a quick skim over some of the pages:


Oh, really? I'd guess that its popularity is more because of stuff like the turning circle requirement, emissions specs, wheelchair-accessibility requirements and the de facto monopoly that results.

Correct me if I'm wrong, but I doubt London drivers (most obviously) have much choice in the matter.

And, I mean, I wonder which publication wrote this a couple of years ago?


‘MONOPOLISED’: Leading taxi industry names meet to discuss plan for alternative viable EV black cab

https://www.taxi-point.co.uk/post/monop ... -black-cab

Some of the leading names in taxi innovation met to discuss plans on how to bring new electric taxi manufacturers into the ‘monopolised’ taxi market.

Since the departure of Dynamo Motors in 2022 there is only one black cab manufacturer and model, the LEVC TX, that passes strict London regulations.



Can't claim to be at all knowledgeable about all that, but did anything really change between that piece and the survey about a year-and-a-half later justifying the claim that the LEVC was simply "popular among taxi drivers, suggesting its suitability as a traditional taxi vehicle."

I mean, where stuff like WAV requirements and the turning circle aren't specified, how popular is the LEVC with 'taxi drivers'? I'd guess numbers would equate to approximately zero.

And even where there are WAV and turning circle requirements, I'd guess the LEVC still wouldn't be particularly popular...except if there was also an environmental spec or age rule or whatever effectively requiring them. Otherwise, I'd guess the older TXs would be the default choice.

(I'm not sure if TPTP's description 'LEVC TX' includes the older TXs rather than just the newer LEVC model once known as the TX5. Which would certainly explain the effective non-mention of older TXs in the survey. But, in any case, that would only change my analysis above slightly, and the basic point still holds - surely a bit more nuance is required than simply saying it's 'popular' because of its 'suitability' as a 'traditional taxi' )

As regards the fuel types analysis and the DfT stats compared to the TPTP numbers earlier in this thread, it's perhaps worth underlining this statement prefacing the DfT's spreadsheets:


Also:


To that extent it's surely fair to say that the DfT's stats aren't the finished article, and should be taken with a pinch of salt (to use everday language rather than statistical jargon).

Of course, some of the limitations and uncertainties attaching to the DfT's stats should be obvious from some of the analysis earlier in this thread, even from an amateur statistician like me (in fact even 'amateur statistician' is probably exaggerating things slightly, but I have studied a bit of statistics in an earlier life )

Anyway, it's perhaps in turn a good idea to look at some possible limitations to the TPTP stuff from the more purely statistical angle. First is maybe the sample size. Although 1,500 or so responses to their survey is quite a good return, and a good sample size from a statistical perspective (I think a baseline for this is maybe 1,000 contributions, thus the survey certainly exceeded that), on the other hand it needs to be borne in mind that to a large extent the survey relates to very different and discrete parts (London, NI, Scotland, and England & Wales, effectively). To that extent the survey numbers have to be looked at more in terms of several different samples - you can't really lump in Scotland with London, say, because the regulatory regimes and stuff like emissions specifications are totally different.

Which is why, of course, Scotland and London are treated as different regions in TPTP's analysis. Which is all very well, but that's when the sample becomes split in terms of size, and to the extent that the numbers become a lot smaller, then the results from the survey will be less reliable.

So, for example, Wales is treated as a distinct statistical entity in the survey, for obvious reasons. But if there are only 30 responses (say) to the survey from Wales, then it automatically becomes less reliable, and more likely to produce unrepresentative and unreliable results. Which is just basic common sense, really - if you ask 5 people something in a poll, say, then you're more likely to get an unreliable result than if you ask 500 (you're more likely to get outliers with the small sample, which will distort the result).

So if you ask 50 drivers in Wales, you're likely to get a less reliable result than if you ask 500, and you'll get an even better result if you ask 5,000. Of course, if you ask everyone you get the perfect result, and zero uncertainty! (It's what statisticians call 'confidence', I think - the more people you ask in a poll, for example, the more confidence you can have that the poll will show a realistic result.)

Anyway, that's just one possible reason the TPTP results may be unreliable - in some regions in particular the results may be based on quite a small number of responses to the survey, and to that extent the results may be unreliable.

In fact I don't think the TPTP survey actually states the number of responses to their survey from each region (correct me if I'm wrong). And maybe the above is the reason why - if the numbers in certain regions are low, then it wouldn't look like a good sample size, and to that extent would make the results look unreliable. In fact, if you look at the dual licence section of the table below, for example, you can tell that the sample size is very probably pretty small - a lot of the percentages look too neat and tidy to be derived from a large number.

11.11% is 1 in 9, for example - so that could mean 9 out of of 81, say, but more likely much smaller number.

Or in London it's 60.00%. So very probably 3 out of 5, or 6 out of 10, but the 'population' was probably not much more.

And, of course, if it's dual badge holders in London, then they were very lucky to find any at all, so they did well to find at least five

(Which made me think - maybe dual-licence holders means having both an HC and PHV badge, for whatever reason that would be required.

But, again, that begs the question why TPTP decided to use dual badge as a separate category - the numbers across the country of drivers having a badge for each code must be tiny, and thus not really significant in terms of a statistical analysis of the national trade )

Think I've maybe made this point previously (although maybe the longer version was the bit that disappeared due to, er, technical reasons ), but another difference between the DfT fuel type stats and the TPTP survey that means they're not directly comparable is that the former were as at 31 March 2024 (thus almost a year old now), whereas the latter was conducted over several weeks towards the autumn period of that year. So to that extent they're very probably going to be different, particularly when considering a relatively fast moving phenomenon like the move from fossil fuel/ICE vehicles towards more, er, sustainable stuff.

On the other hand, a few months seems highly unlikely to explain the significant differences between the two sets of stats, as examined at some length above.

Which indeed points to another deficiency in the TPTP stuff - I've been a bit vague above about precisely when the survey was undertaken because, and correct me if I'm wrong, the survey doesn't actually say when precisely the data was gathered. Think the precise dates the survey was undertaken is available, but because the survey report itself doesn't actually state the dates, then maybe no point in looking it up from other sources.

Instead, the document's main title includes the words 'Driver Survey 2024/25', and that's included at the bottom of each page. But towards the end of the document it's twice referred to as the '2024 National Taxi and Private Hire Driver Survey'. So, as regards dates, it's all a bit vague.

Which in some respects (such as the driver attitudes stuff) may not matter much. But as regards the likes of vehicle stats, it might have been useful to provide at least some approximation regarding when the survey was undertaken.

To take an extreme example like an opinion poll, it's often critical to know precisely when the poll was undertaken. For example, political analysts will need to know whether polling took place before or after an important political event, like Reform's implosion last night

Less critical in terms of timing might be something like a public attitudes survey towards matters like immigration or environmental issues. One I was looking at the other day simply said 'Fieldwork November 2024'. Thus the precise date wasn't really relevant to that kind of stuff, but still important to know vaguely the timeframe when the questions were asked and results gathered.

Of course, the TPTP stuff didn't have to be tied down to a particular date because of the nature of the survey. But for comparison purposes, maybe a bit more precision than 2024/25 would have been appropriate, particularly when it seems that in actual fact the 'fieldwork' was all completed well before the end of 2024.

Anyway, even assuming a gap of a few months between the DfT and TPTP stats, it surely does little to help reconcile the differences between the two sets of data.

(Drafted this a few days ago, but have sort of lost the impetus with this stuff. But will also post some other stuff soon I actually drafted on the weekend after the study was released. So some of it will be a bit back to front, but no point deleting it all now.)


Of course, opinion polls and surveys and the like depend fundamentally on the precise questions asked. In the Scottish independence referendum the SNP were desperate that the choices should be Yes/No, because they knew that people were more likely to select the word 'Yes' because it has positive connotations. Or something like that.

By the same token, I think the Brexit options of Remain and Leave were selected on the basis of neutrality, and to avoid the Yes/No thing above. Or something like that

Anyway, perhaps some of the questionable stats in the TPTP survey could be at least partially explained in terms of the questions asked. But which can't be interrogated because we don't have access to the actual questions - not sure there's any trace online of the wording of the questions asked, nor is there any evidence in that regards sent to participants - all I've got is an email acknowledgment after I completed the survey.

But since most of my analysis above relates to the Fuel Type stats, it's perhaps instructive to consider my own car, particularly since I can't even remember my answer to the relevant question

So my Octavia's basic classification is 'mild hybrid'. On the V5C (which, it should be recalled, the DfT uses for its fuel type numbers) the model is 'TSI E-TEC MHEV' - the latter stands for mild hybrid electric vehicle

The 'Type of Fuel' according to the V5C is HYBRID ELEC

So I'd guess that DfT classifies it as 'hybrid electric' rather than 'petrol' in its classifications in the spreadsheets posted earlier.

But which isn't really how I view it - more of a petrol car with some slightly gimmicky 'alternative fuel' stuff, or whatever. (The V5C also says the taxation class is 'ALTERNATIVE FUEL CAR' )

So in terms of TPTP's three basic classifications for Regional Fuel Types (Fossil Fuel/Hybrid/Full Electric), I certainly wouldn't class it as Full Electric in terms of responding to the survey. Obviously.

And I can't make up my mind between Fossil Fuel and Hybrid - it's certainly not a hybrid as per those Toyota Aurises, or whatever but, as the V5C says, it's not really a full blown fossil fuel car either.

But that's the thing - having spent quite a long time attempting to analyse the whole thing in various posts above, I'm not really sure how I'd classify my own car in terms of the TPTP classifications.

And I can't recall how I responded to the relevant question in the survey, but obviously in view of all of the above, perhaps I might respond differently to the question now, as opposed to when replying to the survey originally.

But, of course, to a large extent that may come down to the point raised at the start of this post - how I responded might depend on the precise question asked. But which I can't remember now anyway, so who knows?

Anyway, that's all just to simply point out how the survey questions might have elicited dubious responses, in turn feeding through to questionable stats in the final survey report

There's a Suffolk County Council report on transitioning to EVs and stuff like that (incentives and barriers, blah, blah) which has made a bit of news, and which analyses the current fleet in terms of fuel type, and says:


So that refers to the DfT stats I used for much of the analysis in this thread, and is largely consistent with what I said - fully electric vehicles are less than 1% of the total in Suffolk

(The 15 EVs are across four licensing councils, the largest being Ipswich, and total vehicles are 1,999. So it's 0.7% EVs, to be precise. Or 0.750%, as the TPTP survey might have put it )

But the TPTP figure for 'full electric' was 20% in the East of England, which includes the councils above

Of course, the East of England is a bit bigger than the four councils encompassed by Suffolk County Council. But just saying...

Full report here if anyone's interested. Doubt I'll be reading it, but as usual a quick skim reveals some dodgy-looking stats. I mean, according to the pie chart on page 15, 15.6% of drivers (about 1 in 6.5 drivers) are doing more than 500 miles per day

https://www.suffolk.gov.uk/asset-librar ... y-2025.pdf

Was just thinking about the TPTP survey again while looking at those numbers from the Rhonnda about inspection failures and the like, and thus the TPTP percentages usually quoted to one hundreth of one percent, therefore a tad spurious in terms of accuracy. For example 0.82%, or 37.54% in the table about app used depending on driver age.

(And another slightly strange thing is the inconsistency throughout the document as regards how they round those percentages - sometimes the likes of the figures above would be rounded to 0.8% and 37.5% respectively, or to 1% and 38% respectively rounding up even further, but with no obvious reason for this inconsistency of approach.)

On the other hand, anyone else notice that in the 'average UK driver' section on pages 8 and 9, there are 12 different regions, but in every one of them the average insurance is either £1,500 or £1,600?

Which at least demonstrates - assuming a degree of integrity to the figures - that insurance doesn't vary much throughout the whole of the UK

Then again, particularly in view of the lack of variance demonstrated in the figures, wouldn't those be the ones where a greater degree of accuracy would be useful for regional comparisons - for example, £1,562 or £1,647 (say) would presumably both be reported as £1,600 in the survey, but there's obviously a reasonable difference between the two figures that simply doesn't come out if they're both rounded to £1,600.

Of course, if the top region had been £1,700, and the bottom region £900, say, then rounding in that way would have been more acceptable because of the greater gap between the top and bottom figures.

But since they're all reported as either £1,500 or £1,600 then surely a less rounded figure would be more instructive. And the way they're reported is totally inconsistent with the way many of the other figures in the survey are presented.