Thread by @goldstone_tony

1/ Short 🧵 on lies, damned lies, & statistics

I've seen some rather misleading use of statistics on the BBC / Sky today, its not clear whether they dont understand the differences, or they are trying to mislead

Please read all & share / RT

2/ First of all lets rewind back a couple of months to @CommonsHealth from May of this year with @BrineMP chair of that committee disputing the 35% pay restoration figure (juniors) with the excellent @ERunswickBMA

@CommonsHealth @BrineMP @ERunswickBMA 3/ The claim made by @BrineMP was wrong on many levels (as pointed out by Emma)
- conflating pay erosion with figure to needed restore (listen to her excellent explanation)
- different inflation measures (RPI vs CPI vs CPIH)
- different dates (i.e. 2008 start vs 2010)

@CommonsHealth @BrineMP @ERunswickBMA 4/ But its the rather dismissive quote "the 35% figure has been has been heavily disputed by the @IFS, who I assume you respect, a more accurate figure would be X,Y" - if @IFS want to criticize the BMA or anyone else, then at least try and make the comparison more valuable

@CommonsHealth @BrineMP @ERunswickBMA @ifs 5/ For example one of the key differences between the BMA and @IFS were the start dates. The BMA like almost every other union had started from 08/09 as that is broadly when austerity started. Missing that initial loss is *HIGHLY* significant & precludes useful comparison

@CommonsHealth @BrineMP @ERunswickBMA @ifs 6/ When I first asked (the excellent) @benzaranko from @IFS why he started his calculation in Aug 2010 he very frankly admitted he used that start date as that was when the dataset he had used started

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko 7/ But others including @FT, the BMA and myself have also used the same datasets from NHS Digital, but have been able to work back to March 09 (covering 08/09)- which would provide a better comparator to allow other non-experts / journalists to compare more fairly

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT 8/ To be entirely fair to @BenZaranko he said at the time he agrees the methods are not comparable (contrary to their comparison for example by @BrineMP in the select committee)👇

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT 9/ So this lunchtime Ben was on BBC news and presented this chart - showing pay erosion for juniors at 15%-18.9% & consultants at 17.2% (from Aug 10 - Mar 23). I cant read the small print but I know Ben's preferred inflation measure is "CPIH"

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT 10/ CPIH is a measure of inflation from the ONS (their preferred one) although it is almost always ignored by the media (Ben did say he shared both CPI and CPIH btw - Im guessing CPI erosion would have been higher, so interesting choice by BBC who always ignore CPIH usually!

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT 11/ Ang again in an interview this morning, trying to compare directly different figures (like the BMA pay erosion based on RPI, and net pay, from 2008) directly with Bens figure of 17.2% which is gross pay, different time period & CPIH.

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT 12/ So when journalists are trying to compare figures for example as they did this morning using most recent IFS analysis, they are just not directly comparable (as apparently Ben told the BBC)

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT 13/ So its really important if journos want to compare figures, they are are clear they comparing apples with oranges. Ben did say he pointed this out to the BBC, but Im not sure this was fairly reflected in the reporting.

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT 14/ I suspect there is a media undertone that the BMA are trying to be dishonest, when I suspect its the media either not understanding the subtle differences, or some examples deliberately misrepresenting them. This example from @skynews "research"

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 15/ The graphic is blatantly wrong, BMA have never produced a comparison to 2010 as far as I know (happy to be proven wrong), so they are comparing apples with oranges, and they are not even faithfully reporting what the apple was (different start dates, RPI vs CPI, net v gross)

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 16/ And no offence to the Nuffield methodology but thats the least relevant, not only is the time period different, it also is only based on estimated earnings for 22/23, now known

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 17/ But coming back to today's @IFS data I would make 3 observations

(1) There is, in my personal opinion, a valid debate about use of CPI vs CPIH vs RPI. Personally in my own charts, I prefer to use CPI (as per the @FT charts).

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 18/ But Im not as dismissive as Ben and some others of RPI. I think there are valid concerns including that RPI is used by government when it suits them (i.e. student loans which are now often >£100k for medics), and elsewhere). RPI also includes things like mortgage interest.

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 19/ For those interested there is a reasonable discussion of the three main methods here - I would highlight that "Some stakeholders are concerned that the Government is intentionally sidelining the RPI in favour of a measure that shows lower inflation"

incomesdataresearch.co.uk/resources/news…

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 20/ (2) I accept the Ben has stuck to Aug 2010 because he wants to include other NHS groups in his data to show comparisons WITHIN the NHS, but in my opinion that is at the expense of provide a valid comparison to groups in pay dispute to allow a more valid comparison

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 21/ (3) Whilsts Ben's comparisons to other NHS groups is excellent, and his chart still puts doctors at the bottom of the pack, I do think its valid to look at other groups outside the NHS (and comparator professions).

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 22/ I know I wont win the inflation argument with Ben who will stick with his preferred CPIH, but I thought it would be useful to update my charts to give a better comparison covering points (2) and (3) above

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 23/ So this my pay chart (CPI left); (CPIH right)

Fin & Bus. ⬆️ 4.8% CPI, 8.3% CPIH
Prof Sci & Tech ⬆️2.5% CPI , 6.1% CPIH
All workers, All sectors ⬇️-3.5%, -0.2% CPIH
Nurses ⬇️⬇️ -12.8%, -9.8% CPIH
Junior Doctors ⬇️⬇️-24.2%, 21.6% CPIH
Consultants ⬇️⬇️-23.9%, 21.2% CPIH

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 24/ Looking at each profession (CPI left);(CPIH right)

NHS = BRUTAL Real-terms Cuts - £5-40k/ each & every year DOWN

Whole economy = NOT so much - £1k/yr down - £2k/yr UP

Please *ANY* politician tell me this is "FAIR" or REASONABLE 👇

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 25/ And finally gross loss in real terms (CPI left);(CPIH right)

Again please someone tell striking workers this is REASONABLE 👇
and inflation is "hurting" us all, and we should show "restraint"

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 26/ To summarise to journalists, please try & be more careful. Dont compare apples with oranges.
Look at
- start & end date
- inflation measure
- gross vs net pay
- consider comparators professions & dont use misleading generalisations we are all affected equally

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 27/ The bottom line is that pay has been destroyed in the NHS whether measure of inflation is used. Ministers may gaslight that strikes are causing waiting lists, but they are not. Waiting lists are a #PoliticalCHOICE

RT if you want this/next government to #FixPay #RETENTION

@CommonsHealth @BrineMP @ERunswickBMA @ifs @BenZaranko @FT @SkyNews 28/ Addendum 1 - I understand that in relation to tweet 16 the Nuffield analysis is no longer based on predicted 22/23 pay data, but now known 22/23 pay data. The criticism remains that this is a different time period, deliberately misrepresented by @SkyNews

@CommonsHealth @BrineMP @ERunswickBMA @BenZaranko @FT @SkyNews 29/ Addendum 2
Much more balanced reporting here IMHO including a comparator to all workers, correctly starting when pay erosion started from @catherine5news