Evaluation, statistics and OR

Getting a forecast wrong

As we get nearer the weekend, our fascination with the weather forecast grows.  Last week, we were promised a fine Friday and Saturday with rain and storms on the Sunday.  The promise of sunshine was a boost to the system - finally some nice weather!  Friday was a pleasant enough day - some long sunny intervals and slightly warmer temperatures, so when Saturday dawned uniformly grey and cold, we were devastated.  Seasonal Affected Disorder wasn't in it - we were downright depressed.

There is the potential for a similar scenario this week as well.  Meteo France are predicting a hot sunny Friday and Saturday and a hot Sunday though with the potential for some storms - the fact we were so badly let down last weekend makes this a cruel forecast.  Will they be right and we will have some summer at last, or will the clouds crawl in/down from the Channel/Atlantic and another depression (meteorological and mood-wise) sweep our household?

In hypothesis testing, there are type I and type II errors - the one being the chance of rejecting a hypothesis when it is in fact right, and the other being the chance of accepting a hypothesis when it is in fact wrong.  Last week, we believed the forecasters, thinking they had enough information to get it right, but they got it wrong (type II in my book).  This week we are suspicious and won't believe in nice weather until we have got just a little bit sun burnt, and have gone back to summer clothes from winter jumpers and long trousers (could this be a type I error on our part?).

In this instance, which error type is worse?  The disappointment last weekend at poor weather was immense and seemed to be almost physically painful, as we had set our hearts on one weather pattern as predicted by the weather boffins.  If we have the same disappointment this weekend, I will have to avoid the "contact us" bit of their website, or I could be in serious trouble.  But would it have been worse if they had predicted bad weather all along, giving gloom in anticipation, and they had been right?  Would I have felt so bad?  I don't think so.  If they had predicted poor weather and we had had a nice weekend, I would have been pleasantly surprised and felt happier about the world.

The nature of the forecast is therefore a factor in our feelings about whether type I or type II errors are more acceptable as far as the weather is concerned.  Of course this assumes that there is no cash or danger attached to getting the answer wrong.  In my days with British Gas, not having a reliable weather forecast to ensure enough gas in the system for diurnal requirements could cost money and even lives (and jobs for those who got the forecast wrong), and forecasting for the electricity grid system was even more sensitive than that, as it needed to respond to changes in demand within the hour.

Weather forecasting for the masses doesn't need that level of detail or accuracy and the weather is a chaotic and unpredictable system, so why do I get so upset about it?  Because in a year like 2007, the promise of sunshine and warmth is more than just a weather forecast, it is a promise of good things and normality and fun and enjoyment.  When someone promises you that, and it doesn't happen, you tend to take it personally, and that is when it hurts! 

Milk Market Failures

Is it me?  To quote a Terry Wogan/TOG catchphrase, but there is something weird going on in the European milk market and it seems farcical to me.

Today the French papers are reporting that milk product prices are due to rise between now and the end of the year, not because of the weather (perhaps the only part of the farming market not to be complaining about the weather at the moment) but due to a huge shortfall in production of milk in Europe.  It appears there is a Europe wide shortfall of a billion litres of milk, but in France, they are not using all the quota allocated for production.  In France the farmers are apparently turning away from milk production to cereal production which has a better profit margin.

This is not far off the situation in the UK where milk producers are going out of business as the price paid for milk at the farm gate is I gather a penny per litre below the cost of production.  Britain now imports milk from Ireland and Poland, where the costs of production are less or at least the farmers get paid the money owed to them by the relevant Department with responsibility for agriculture.

So now the relentless consumer race for the cheapest possible food from the supermarkets is going to turn round and bite the consumer on the bum.  By not being willing to pay a fair price for milk (and the culprit are: consumers, supermarkets, food producers), milk is now in short supply and the price will go up.

The list of culprits there is an interesting one, and to a certain extent is being investigated in the UK by the Office of Fair Trading.  The supermarkets are those with the overwhelming buying power that determines to a greater extent what is paid for products - while a litre of milk is about 50p in the supermarket, they only pay 17p to the farmer.  Those lovely buy one get one free offers, when applied to apples, mean that actually the farmer gets paid for only half of what he or she supplies, but has the cost of production of the total, so don't thank Tesco for your BOGOF offer, go out and thank a farmer directly.

Supermarket profits are going up, numbers of farmers are going down.  In the end the consumer pays, so make sure that you put the money you save on your supermarket bill now in a high interest savings account so that you can pay the extra for your milk products come Christmas.  Personally, I will keep buying my milk, cream and yoghurts directly from the farmer - it's about the same price, but the quality is better, the products are fresher and I know where they came from.

All change at the Department for Education and Skills

A new Prime Minister with a new Cabinet and a new way of looking at education and skills!  While Jacqui Smith as Home Secretary got the big headlines yesterday, the abolition of the Department for Education and Skills was hidden away rather, but will have a significant impact on the sector.  The BBC article "Brown shakes up education arena" focused on the wonders of Ed Balls and his new Department for Children, Schools and Families, once again leaving adult learning and skills as the Cinderella sector.

The new Department of Innovation, Universities and Skills covers the Higher Education sector, Science, Innovation and the skills sector, but the key delivery agents for skills in many areas - the Further Education Colleges - have more than one master, as their 16 - 19 remit sits with the Department for Children.  The Learning and Skills Council would appear to have some change to its remit, although not yet in the public domain as far as I can tell, as funding for 16 to 19 education will be allocated to schools and colleges via local authority education budgets (it is unclear whether schools and colleges also means private training providers).

For the LSC, the bulk of the budgets are for 16 to 19 provision so a significant transfer of power is to be expected with the movement of the funding role.  In some ways, my personal view is that this is probably right, as the LSC never really got to grips with or got support for its role in school 6th form funding, and a lot of the work that continues on academic routes at 16 to 19 detracts from the support needed to tackle those who chose not to continue in any form of education or training.  It would be nice if this meant that there is then a real concentration on the skills issues of adults and adult learning, however I think in reality there could just be a power struggle between local education authorities and the LSC on who determines policy and strategy for colleges.  FE Colleges will have to resist the temptation to deal with these two like a child with separated parents, playing one off the other to get more and better treats.

My real concern in all of this is that the internal politics and navel gazing that will ensue at the former Department and at the LSC and the other Skills organisations will mean that it will be too much like hard work to do anything about the Leitch report and its recommendations.  The report will be lost, while the sector jockeys for position, and once they return to it, it will be easy to say that it is out of date and needs refreshing, and we will wait another 2 or 3 years to see anything on improving the skills and competitiveness of the UK.

The benefits of business planning

It is a quiet week this week, as one project comes to an end and waiting for a phone call to get the next one kicked off.  Because of that, I have taken advantage of this lull in activity to do what I promised my bank manager I would do about 7 months ago and develop a business plan.

When I was in a group setting up the now disbanded Mercia Research & Strategy, we got a business plan together very early on in the process - in part to ensure that all partners understood what we were getting ourselves into, and in part to have something to show the bank.  But despite the fact that two of us at least had a planning background, we never systematically reviewed that plan.  It may be one reason why we drifted apart - we didn't take the time out to review and put down in writing our thoughts and ambitions and knowledge under a few key headings.

With Plantagenet Consulting, it has been a similar pattern.  In the initial phases, there was some stuff on the backs of envelopes or scrawled on my white board, and when I saw the bank, I blagged it a bit, along the lines of "you know me and my track record and my prudence with money".  This week I am setting time aside to review a number of key themes:

• What are my aims short term and long term?
• What are my targets?
• What are the markets I am targeting?
• What are my former, current and desired client bases?
• Who are my competitors?
• Who are my collaborators and partners?
• What do I do well?
• What are the services I can offer and how does that link with above?
• What are my marketing messages, and do they link with any of the above?

And that is just a start, as within those topics there are a whole other set of issues and linkages.  I have developed a template of headings, and I am in the process of filling those headings in very honestly.  I am spotting a whole range of actions and activities that have got to be done to make Plantagenet Consulting interesting to the markets that I want to be in, which will keep me well occupied in the long summer evenings, assuming the weather doesn't improve.

I am also identifying the training and self development that I need to do to get me where I want to be, the tools and skills that I need to provide a fully rounded service in the areas I want to work, and am starting to work on that self development.

This business planning lark might seem a bit luxurious and nebulous, but if I get it right and then follow through with those identified actions, I stand a good chance of meeting my first year targets for Plantagenet Consulting, and that means that it is time well spent.

If any other micro-businesses out there would like to try out my template and process, feel free to contact me via a comment on the blog or via the e-mail contact on the about link on the right of the screen.

The ability to speak the language of your host country?

Priceless!!  To quote a much hackneyed advert.

This is a topic that is doing the rounds in both the United Kingdom and France at the moment.  The reports on social inclusion of migrants, the call for ESOL lessons for migrants - both in the UK - and the demands by Sarkozy that immigrants to France learn French to an acceptable standard before they arrive are the talk of the moment in the media.  There are a number of issues here which are worth airing - so I will!

Firstly - the limits on who can get free ESOL training in the UK are so tight that the influx of migrant workers are effectively excluded from funding, and their employers will doubtless find their access to Train to Gain funding at the public expense for this also curtailed.  This is a major issue - providers expect that once the changes come into force in August 2007, demand will drop away, so that the capacity they have been trying to build will be wasted.

This brings in the second issue - the non-financial resources required to deliver the ESOL training.  Rooms, books, materials, teachers and assessments are stretched beyond what is reasonable and investment is required to meet the demand if funding rules aren't changed.

Then there is the issue of actually being able to get to any courses - little provision seems to be "roll on roll off" if I may use a ferry analogy while the fact that many migrant workers do long hours and anti-social hours means that actually getting to courses or being able to attend them regularly can be tricky.

But what of our own citizens that wish to emigrate elsewhere?  If the Sarkozy rule was applied to the British moving to France, how would we fare in getting the learning we need to prove an ability with French?  Well if my experience is anything to go by, not that well.  Realistically Beginners French or even GCSE French is not adequate for dealing with more than buying the morning baguette and croissants.  It is certainly not up to the standard required for sorting out your Internet access or exporting your car or explaining to the vet that your cat has been attacked by a chicken or eaten a lizard, or even signing on with a French GP.

So you need to look at AS/A2 French (A-level in old money).  Well, there aren't that many providers doing that as evening classes for those of us that work, the standard and requirements isn't much better than the much beloved O-level of 20 years ago, and still leaves you struggling when explaining your situation to the Social Security or Tax authorities in France.

So we can't train those that come in, as it costs too much, and we don't train those that want to leave, 'cos it is all a bit difficult and driven by the needs of 16 to 18 year olds to be entertained while they learn at 6th form.  We just haven't got our heads round languages and integration, have we?

Too much technology?

I have a favourite birthday card that I send to like minded chums, via the post, which has a nice sepia photo of an elegant lady from the 20s with a typewriter, and the caption reads "It was a hard day at the office - the computers went down and everyone had to learn to think again".  I am no technology Luddite - but the truth of that statement is frightening.  If my ADSL box goes down so I can't access the Internet and e-mail, or even worse I have to close down all computers as there is a thunderstorm near by, work stops.

Another situation where technology drives success was noted on a report on the BBC website about difficulties faced by community and voluntary organisations using the e-procurement system of the Learning and Skills Council.  While many organisations got through it fine, a small but significant number had problems with system crashes and lost work.  While the system saved an estimated 3 million sheets of paper, it has also caused heartache and stress to a number of organisations who didn't deserve it.

I fully applaud Rob Wye's unconditional apology on the system, but am a bit more worried by the attitude of the anonymous company that designed and implemented the system.  Their view was that any problems were entirely the fault of the users.

Now when I worked with similar systems from within the LSC, and I had community and voluntary organisations on at my team about the awfulness of the systems we were forcing them to use, one of my team would go out to see what was going wrong and do some training with the organisation that was having problems.  Or they would come back and say what the problem was - often dial up modems, dodgy telephony, or computers that were more than 2 years old.  While such issues might be the user's fault in some ways, the desire for bigger, better, faster technology and the need to upgrade every year or so is as much of a waste as the 3 million sheets of paper.  For a community or voluntary organisation, the cost of keeping up diverts resources from doing good.

When I worked at British Gas many years ago (here is the techie bit), I used to run multiple regression models with approximately 10 variables including an autoregressive term and using large data sets, and I could design them on a single sheet lotus worksheet (NOT a workbook), and I could run them on a PC with a 286 processor (remember those?).  It was a bit slow, but it could be done, and the results could be thought about, and the worksheet could be backed up on a floppy disk with 384Kb capacity.

Today I worked on a model for community and voluntary organisations to understand demand for learning, using computer power that is greater than that which powered an Apollo space rocket.  However when I looked at the quality of the data I was working with, an abacus might have been more practical! 

With all this technology around, you have to hope that from time to time the brain pitches in, looks at what you are doing, and says "that's garbage!"

Bad Timing!

After the lively debate on understanding and assessing risk, I am going to return today to another favourite theme of mine - migrant workers.  Or more precisely, the research undertaken into the migrant worker phenomenon in the UK.  This is sparked by two stories in the press yesterday - one on research undertaken for the Joseph Rowntree Foundation, and one on the needs of strawberry farmers.

The Joseph Rowntree Foundation story was about the publication of a study taking a longitudinal view of migrant workers and their desire to settle permanently in the UK.  The key finding being that after some time in the country and (with the final phase of field work being done in May 2004) with the Accession states joining the EU, workers were more likely to want to settle in Great Britain.

In a different story, the National Farmers' Union warned that the UK strawberry crop was under threat as there was a shortage of migrant workers in the sector.  This was blamed on rising standards of living in Poland and other Eastern European nations.

The juxtaposition of these two stories was fascinating to me - the research story trying to highlight the difficulty of integration and learning the language and getting information on immigration status faced by the migrant workers.  While the BBC also tried to introduce a little controversy into the story by questioning whether migrant workers weren't making life more difficult for UK marginal workers.  At the same time, there are 2,400 jobs picking strawberries going begging - jobs which the British seem unwilling to take on, and now the "cheap labour" of migrant workers are also staying away from.

The Joseph Rowntree Foundation Research was excellent, but took too long to come to light - the researchers from Oxford and Sussex Universities did their last field work (according to the story) in 2004, but other studies by companies such as Mercia Research have done work more recently and reported more quickly.  The findings are not new enough to advance the debate adequately, although the aspects that they highlight of limited use of English language learning, integration into British society and the emergence of enclaves are incredibly important.  Hence my view that there is bad timing going on here.

Another bit of bad timing was that this story was picked up at the same time as the plight of seasonal food producers, but no one tried to link the two.  Well apart from me and in a very small way.

A final  thought - it seems to me there will always be the jobs that people don't want to do - fruit picking, shallot planting, waiting tables, cleaning and sweeping, working in food processing factories.  In the UK, we have managed to help everyone aspire to not doing these jobs (even if we haven't succeeded in giving them the skills to do other jobs), so once again we need to look to other countries to supply the labour to undertake these roles.  However the NFU is finding that the sources of uncomplaining labour in Eastern Europe are now also setting their sights higher - where do we look for the next wave?  Or do we accept that sometimes people need to have jobs not careers?

Acceptable Risk Revisited

Last month, in a previous post called Acceptable Risk? I commented on a story from the BBC website looking at teaching union fears of the health risks of wi-fi.  I hadn't realised it was a puff piece, giving a lead into a Panorama programme to be broadcast today on "Wi-fi: a warning signal".  Licence payers' money was spent on this investigation which found (suitable pause for effect here) that wi-fi radiation levels at one school were three times the level of mobile phone mast radiation - they were also 600 times below government safety limits.

A number of experts from British institutions are quoted as saying that there is no risk and that this is a non-story.  The intensity of wi-fi radiation is 100,000 times less than that of a domestic microwave (where is the Panorama programme on let's ban the microwave oven?).  My favourite part of the report is a section from Professor Sperrin, an expert in the field of medical physics, who points out that it is impossible to prove a negative or to prove that something has absolutely no effect.  The implication is that Panorama, Jeremy Vine and other innumerate institutions will keep pushing this one until someone takes them seriously.

Other forms of electromagnetic radiation have been found to be harmful - the example quoted is of someone sitting in a wi-fi hot spot in full sun on a beach.  The significant danger here is of sunstroke and sun burn - get into the shade being the health and safety advice.  Other worries about wireless laptop computers come from the dangers of heat sources (the battery in particular) and the effect that prolonged exposure could have to the "soft parts of the body", where you might place a laptop.  Think about it chaps!!

On the same day, a survey for the NSPCC found that violence is a day to day problem for young people aged 11 to 16.  Of those questioned, 42% had been hit, punched or kicked at secondary school - a pretty tangible health risk.  A good three quarters had faced some form of bullying in school, but less than half felt there was adequate support to help them cope with it.

Jeremy Vine - here surely is something that is worth picking up on!  It would be a cheaper programme to fund by the BBC and a whole lot more relevant to the majority of young people than some non-existent risk that Panorama had to go to Sweden and to the USA to find some nebulous possible evidence for.  It would highlight something that needs action and a bit of government spend - but perhaps the public school boys among the powers that be think that it is a good part of the school tradition and never did them any harm.

So for the numerate amongst us, let us act to stop or reduce the actual threat, and leave the fantasy threat to the science fiction writers!  Or is there mileage in a study on the stupidity of the media - the chances (or risk) of getting funding for that are probably too low to be worth the effort of applying.

The Workforce of 2020 and Leitch

The pages of the BBC's education news on the web include a small and overlooked debate between Chris Humphries - Skills guru and Director General of City and Guilds - and the spokespeople of the DfES on the ultimate meaning of the Leitch report.

Chris is concerned that the new demographic time bomb (low birth rates in the 1990s) and the economically inactive will combine to create a "workforce black hole" in 2020.  He feels the demographic issue was missed by Leitch, giving a fundamental flaw.  DfES are naturally defending Leitch and the money and time that work took, but also point out that forecasting the size of the workforce of the future is difficult, while asserting that Leitch felt the inflow of young people would not be enough to meet future skills needs.

Both sides have good points to make and neither is essentially wrong.  Chris is right to say that we need to be engaging with the economically inactive - while unemployment has gone down significantly in the past 10 years, the numbers employed and economically inactive have both gone up.  It is the inactive group that are hard to engage with, that don't want to improve their skills, that don't seem to want to put themselves out for their own good or for that of society.  Don't get me started on what they might be doing with their time instead!!

Chris also has a slightly vested interest in highlighting the changes in the numbers of young people - City and Guilds provide an awful lot of qualifications to an awful lot of young people, and as the BBC has also pointed out recently, a substantial part of a college budget can be eaten up in examination fees to the likes of City and Guilds.

Leitch in many ways started from the premise that a lot of attention was being paid to the skills and education of young people, but what about the rest of the population?  A core part of his fundamental argument was that the majority of the working age population of 2020 (and excluding for the moment issues of changes in pensionable age) were ALREADY in the working age bracket, hence the need to focus on their skills, rather than that of youth.  This is not contrary to what Chris Humphries is saying.

Of course the whole argument could be a way of bringing the issue back to centre stage and developing media, academic and policy debate about how to engage with the economically inactive.  If that is the case, to this interested observer, I think they have missed out with the timings as Scottish election results, the Northern Ireland government and the climax of the northern hemisphere football season (not to mention the Lord's Test next week) are all rather more fascinating currently.

Acceptable risk?

According to the BBC today, one of the teachers' unions, PAT, is demanding research into the risks posed to the health of them and their pupils by wi-fi networks in schools.  So far the Health Protection Agency will only comment that wi-fi devices are "of very low power, much lower than mobile phones", and while the Chair of the HPA is being asked to condemn wi-fi, he is not prepared to do so - rather he has not taken a position on it.

I have just finished reading Managing Green Issues for a Review in the Journal of the OR Society.  Tom Curtin, the author, points out the difficulty of categorical denial in such cases - although most of his examples are to do with bigger environmental causes, the issue is the same.  Could any scientist or statistician say categorically that they were 100% certain that there was no health risk from wi-fi?

No we couldn't!  And that is because that categorical 100% certainty is impossible as there could be a bizarre set of circumstances that could cause a problem that could ultimately lead to harm.  In this way, campaigners are clever as they don't ask a comparative question.  If the question was posed as to whether wi-fi was safer than conventional Ethernet or similar networks, I suspect the answer would be yes.  Even with the best ducting, the cable will cause a hazard to something - I get my Ethernet cable tied up in the vacuum cleaner from time to time.  There is an identifiable risk - but one that teachers seem to be quite happy to live with.

It seems to me that one clear demonstration of the lack of numerical skills in the nation is the lack of understanding of acceptable risk and probability.  If something is potentially nice - a lottery win for example - people will put up with odds of 1 in 14 million, or indeed think they are a certain thing to win.  A similar very small risk is of death on a railway train that has an accident.  After the Hatfield accident a lot of people went back to their cars, as trains were now perceived as dangerous.  Rail travel remains the safest form of transport (I will not comment on comfort!), but people prefer the much more dangerous automobile.

Personally I am all for a little more acceptable risk in life - deal with real and clear risks, yes, but as we learn from mistakes and dangers, don't try to give everything a 100% certified safe certificate.  If a drug had to be 100% totally safe for example, we would not have access to Paracetamol, Aspirin, Ibuprofen or the contraceptive pill.  Take a few minutes to imagine your life, let alone modern society without those!