Monday 22 April 2013

Three home educating myths




I have before examined in detail some of the odder myths associated with British home education, but  there are many new readers here lately and for their benefit  I shall today  outline three popular misconceptions about home education.

1. Paula Rothermel found that home educated children did better academically than those at school.

This old chestnut is still doing the rounds; it cropped up last month in a comment on the Times Educational Supplement forum. Briefly, the legend is that Paula Rothermel conducted extensive research and found that home educated children did fantastically well academically. This is awfully misleading for two reasons. First, because the samples used were so tiny as to be meaningless. For instance one of the claims was that 94% of six year-olds were in the top band when tested for reading ability. This was in comparison with the 16% of schooled children who fell into this band. Very impressive, until you read the fine print and learn that this research relates to just seventeen children. Even less impressive when you learn that Rothermel did not test their literacy herself, but posted the tests out and relied upon the parents to conduct them! For some reason, Rothermel's tiny, self-selected samples, which rely almost entirely upon parental assessment, are still being touted around by home educators.

2. Many famous people were home educated.

Claiming that people like Einstein or the Wright brothers were home educated is intended to show people that being home educated can create great scientists, inventors and writers. The trick here is in how you define ‘home educated’. On one of the home education lists Dietrich Bonhoeffer, the German theologian, was recently described as being home educated. In fact, like most German children before World War I, he started school at the age of seven. What people do is use the British school starting age of five and then assume that any child in Europe who does not begin school at that age must be home educated! Sometimes, people are plucked at random and the assertion made that they were home educated, even though they really attended school like everybody else. Einstein and the Wright brothers are in this category as are most of the other names one sees in the lists of famous people who were home educated. Even more curious is the inclusion in such lists, of people who could not have attended school. The Home Education UK site has Joan of Arc down as being home educated. No girls at all were at school in France during her lifetime.

3. You don’t need GCSEs to get into college or university.

This one has proved an absolute killer for many home educating parents; particularly those who favour autonomous education. They avoid entering their kids for examinations and then when they are fifteen or sixteen, expect them to get into college. It very rarely comes off, except for subjects like  art, textiles, design, photography, drama and so on, where you can sometimes get in on a portfolio or audition. It is almost unheard of for a college to allow a child without GCSEs to study academic A levels like physics. Even on the rare occasions when this does happen, getting to university can be gravely handicapped by the lack of GCSEs. Because there are often many applicants for university places, all with the same A levels, a kind of tie-break sometimes takes place. If you have two young people, both with three A levels at A in the same subject, then a good way of deciding is to look at their GCSEs. The one who has ten GCSEs at A or A* will generally be preferred over the one with no GCSEs at all.

I shall be covering other misconceptions and myths over the next few weeks, because no matter how often these ideas are exposed as nonsense, a new generation of home educating parents is always ready to come forward and be led into error by people who really ought to know better.

68 comments:

  1. Not only do GCSE grades get taken into account, in some places the type of GCSE is important so Cambridge IGCSE trumps EdExcel. I emailed the places I could see my daughter being interested in and asked them what they were looking for for their courses, and was very much left with the impression that they'd prefer traditional progression thank you very much. The other thing that was, oddly, a potential problem was that, because of the cost of exams and my views about them, she's only doing the academic subjects, so she'll need to prepare a portfolio to show she's a rounded person and not a hothoused geek.

    That said, she's science based, and arts courses have very different attitudes so I guess it depends on what sort of child you have.

    And thanks for making the point that school for all is a comparatively recent invention. As I understand it, it was mass education to provide appropriately trained labour as we moved from rural to industrial. Before that, it was for the few and often linked to religion, which is another parallel with HE!

    Atb
    Anne

    ReplyDelete
  2. Simon wrote:
    "the samples used were so tiny as to be meaningless. For instance one of the claims was that 94% of six year-olds were in the top band when tested for reading ability. This was in comparison with the 16% of schooled children who fell into this band. Very impressive, until you read the fine print and learn that this research relates to just seventeen children."

    How big do you think the sample would have to be for this to result to be meaningful?

    ReplyDelete
    Replies
    1. With such a (proposed) large difference in ability, you wouldn't need a huge sample, though more than 17 for sure. Perhaps 60-100 off the top of my head without bothering to consult a statistical manual.

      The real issue is that the home-educated children were not tested in the same manner as the school educated children. This makes any statistical test invalid. Also, sampling should not be down to the experimenter's whim, which is what appears to be the case here.

      Even if this is fulfilled, the result wouldn't become terribly convincing until there are enough studies for a metastudy to be performed.

      Delete
    2. I'm looking for an answer from Simon - with his name on it - because he said, explicitly, that the sample size is too small, and so - regardless of any diversionary issues - he must know.

      Delete
    3. If I told you that you had only one bone in your body, would you need to know the exact number of bones in the human body to see it was obviously incorrect?

      Delete
    4. Simon said that "the samples used were so tiny as to be meaningless" and then gave an example; I would like to see his quantitative justification.

      "He who refuses to do arithmetic is doomed to talk nonsense." - John McCarthy

      Delete
  3. "I would like to see his quantitative justification."

    Don't be silly. Simon is a political animal rather than a deep thinker, When the truth is inconvenient or too difficult for him to figure-out, he'll brazen it out with a convenient lie and a lot of peripheral excuses.

    ReplyDelete
    Replies
    1. "Simon is a political animal rather than a deep thinker, When the truth is inconvenient or too difficult for him to figure-out, he'll brazen it out with a convenient lie and a lot of peripheral excuses."

      From what I have read of his blog so far that seems to be very true.

      Delete
  4. Sorry to rain on your parade but how can 14 yr olds get in to college with out gsces? Because they don't need them. Please do your research

    ReplyDelete
    Replies
    1. You need to read what he said more carefully.

      Delete
  5. Worn out old Webb says It very rarely comes off, except for subjects like art, textiles, design, photography, drama and so on, where you can sometimes get in on a portfolio or audition. It is almost unheard of for a college to allow a child without GCSEs to study academic A levels like physics.


    Peter studying A level Physics yet had no GCSE when he went to apply to the college to study it.The college interviewed Peter and looked at what he had done at home and offered him a place to study A level Physics common sense was used by the teachers after all good teachers want students to be on the right course its no good a student doing to easy a course if its clear he should be on a higher one next year Peter taking A level Maths and A level computing.
    When we went to the college i though it may be problem but far from it the college was very helpful and wanted to help Peter do the right course what a refreshing change if only Hampshire LEA could be like this

    ReplyDelete
  6. Anonymous22 April 2013 09:47
    Simon said that "the samples used were so tiny as to be meaningless" and then gave an example; I would like to see his quantitative justification.

    This is the kind of stupid childish comment Simon has commented on before. He said that - One of the most tiresome things about this blog is the way that some trifling remark of mine will be seized upon and analysed to death; those commenting being quite unable to let things drop.
    Anyone with a slight knowledge of statistics will know that a sample of 17 self-assessed responses are not really statistically valid.

    ReplyDelete
    Replies
    1. You're way out of your depth (Simon?). I'm quite confident of my experience and expertise in the use of statistics and I know bullshit when I see it.

      Delete
    2. Cheshire cat>
      "Anyone with a slight knowledge of statistics will know that a sample of 17 self-assessed responses are not really statistically valid."

      Whereas anyone with a substantial knowledge of statistics will know how to deal with the problem.

      Delete
  7. 17 self assessed children cited as evidence from the thousands of HE children there are?? If an LA used that kind of flawed research to push through visits or monitoring you would be up in arms. Get real

    ReplyDelete
    Replies
    1. You misunderstand the point being made.

      Delete
  8. Cheshire Cat wrote:

    "Anyone with a slight knowledge of statistics will know that a sample of 17 self-assessed responses are not really statistically valid."

    Regardless of the way the reading tests were performed and assessed, Simon said that the sample was too small; that's not a trifling remark and he should expect to be questioned about it.

    ReplyDelete
    Replies
    1. No, there really are stupid questions.

      Delete
  9. 'You're way out of your depth (Simon?). I'm quite confident of my experience and expertise in the use of statistics and I know bullshit when I see it.'

    Possibly. These tests, including the thirty five PIPS, are often used to support generalisations about the academic achievement of home educated children in this country. For a number of reasons, this cannot be done. To begin with, this sample was self-selected. In other words, Rothermel sent out five thousand questionairres. 80% of those receiving them did not respond and from the 20% who did, a small number of children were selected for testing. This in itself has the effect of skewing the results. We must ask ourselves which parents are most likely to want their children's academic ability assessed; those whose children are doing well or those whose children are doing poorly?

    The tests themselves were conducted not by the researcher, but by the parents. In a sample as small as this, it would only take a couple of careless or dishonest parents to distort the results. This is the case with a sample of seventeen, but would not be so serious with a sample of a hundred being tested under controlled conditions. We know that in at least one case, the reading test was left laying around the kitchen for the child to see for an extended period. Both the methodology and the size of the sample make it unwise to draw too many conclusions from this work.

    '(Simon?).'

    Please tell me that your are not suggesting that I am commenting anonymously on my own blog! I have been racking my brains to think why anybody would do such a thing.

    ReplyDelete
    Replies
    1. '80% of those receiving them did not respond and from the 20% who did, a small number of children were selected for testing. This in itself has the effect of skewing the results. We must ask ourselves which parents are most likely to want their children's academic ability assessed; those whose children are doing well or those whose children are doing poorly? '

      Yes. I was one who didn't continue that far because I knew my child was 'behind' at the time.

      Everyone who was involved at the time knew what was going on here; we just chose to turn a blind eye.

      Delete
  10. 'Simon said that the sample was too small; that's not a trifling remark and he should expect to be questioned about it.'

    It is not a trifling remark, Cheshire Cat and I am quite happy to be questioned about it. When you have a group of seventeen subjects, then small outliers can distort things badly. When you have a hundred, or better still a thousand, this does not happen; wild results are smoothed out.

    ReplyDelete
    Replies
    1. Actually, if the results are pronounced as reported here, around a hundred would be sufficient, but a thousand doesn't hurt, of course. The real problem here is the method (namely, lack thereof) of unbiased sampling. What the author has done in her paper is she has found a hypothesis, rather than a result. I wish I was in a field where I could publish hypotheses; I'd be prolific.

      Delete
  11. Cheshire 'Worn Out Webb' Cat23 April 2013 at 05:37

    I agree Simon although the reason I suggested it was trifling is that I didnt think that any reasonably intelligent person would see 17 self selected responses that were completed by parents as genuinley valid. Therefore I wondered why would they want an expalnation? perhaps to goad you or wind you up. Also how could anyone mistake me for you? I'm not sure who should be the most offended........

    ReplyDelete
  12. 'I'm not sure who should be the most offended........'

    Yes, I agree Cheshire Cat. I can imagine few things more mortifying than being taken for me! This is not the first time that this has happened. So convinced have some of these people become that I am a lone voice, that anybody who agrees with me on here is suspected of being my sock-puppet. I don't think that you are a figment of my imagination, but who can tell? Even more alarming, perhaps I am a figment of your imagination!

    ReplyDelete
    Replies
    1. Funnily enough, when I read this, and having seen some of Cheshire Cat's barkings on previous threads, a picture sprang to mind of Webb and Cheshire as -respectively- Jekyll and Hyde.

      Delete
    2. 'Funnily enough, when I read this, and having seen some of Cheshire Cat's barkings on previous threads, a picture sprang to mind of Webb and Cheshire as -respectively- Jekyll and Hyde. '

      Remember readers, if you cannot think of anything intelligent to say, then you can always fall back on making personal remarks about individuals; saying that somebody 'barks', for example. This will save you the trouble of having to think about the subject which is being debated.

      Delete
    3. "Remember readers, if you cannot think of anything intelligent to say, then you can always fall back on making personal remarks about individuals"

      such as Simon's comments about Alison Sauer and her choice of clothes, or... well any one the many disparaging and irrelevant personal remarks that he has posted in this blog over the years.

      Delete
    4. "such as Simon's comments about Alison Sauer and her choice of clothes, or... well any one the many disparaging and irrelevant personal remarks that he has posted in this blog over the years."

      +1

      Delete
  13. Cheshire Cat wrote:

    "17 self assessed children cited as evidence from the thousands of HE children there are?? If an LA used that kind of flawed research to push through visits or monitoring you would be up in arms. Get real"

    You seem to be unable to comprehend and decouple the different issues - sample size versus assessment methods.

    Why do you suppose that that my criticism of Simon's claims about Rothermel's statistics meant that I oppose visits or monitoring? I think they're entirely appropriate in some cases.

    I am generally sceptical about statistical results, at least until I've had a chance to to analyse them; I am even more sceptical about assertions - such as Simon's or yours - that are based on gut feeling, ignorance and prejudice, rather than evidence and calculation.

    ReplyDelete
  14. Simon wrote:

    "When you have a group of seventeen subjects, then small outliers can distort things badly. When you have a hundred, or better still a thousand, this does not happen; wild results are smoothed out."

    I asked the question about sample size in the first place and it's clear that you do not know what you are talking about.

    Let the null hypothesis be that Rothermel's sample of 17 was drawn from a parent distribution that is the same as the rest of the population. We can calculate or simulate the probability that a given result - or one that is more extreme - can occur through random chance under the null hypothesis. A generally accepted convention is that a probability, p<0.05 constitutes a significant result (most definitely not proof, and I'm not being as formal as I'd like but I want this to be understood).

    A quick Monte Carlo simulation indicates that Rothermel's result has p<0.03 and both a simulation and Binomial distribution calculation indicate that p<0.05 if seven or more of the 17 six-year-olds were in the top band.

    Therefore, at least ten of the 16 reported to be in the top band would have to have been misreported - i.e., exaggerated or overestimated - to nullify the significance of the result.

    One can find all sorts or problems in Rothermel's work, but your remarks about sample size are based on ignorance and inexperience rather than any mathematical substance. "Common sense" arguments about sample size are frequently wrong; bigger samples aren't always better.

    ReplyDelete
    Replies
    1. You're right you're not being formal; this is just bollocks. It's some deranged form of Bayesian reasoning.

      Delete
    2. "You're right you're not being formal; this is just bollocks. It's some deranged form of Bayesian reasoning."

      I can see you're hurt; perhaps you should try to calm down and explain why you think it's "deranged". Perhaps you could start by explaining "bollocks" in this context.

      Delete
    3. "It's some deranged form of Bayesian reasoning."

      And I'm not sure why you think it's Bayesian - deranged or otherwise. Simon's point didn't really warrant much effort. But perhaps you could suggest a suitable prior - other than "bollocks".

      Delete
  15. 'Let the null hypothesis be that Rothermel's sample of 17 was drawn from a parent distribution that is the same as the rest of the population.'

    If this were so, then you would be quite correct. It is not the case. The sample was skewed firstly by the fact that an initial 80% of those asked to participate declined to do so. This immediately raises the suspicion that those who did respond were more sure of their children's academic ability. The sample was further skewed when those who did respond were whittled down. Finally, a very small number were invited to take part in further tests. Some declined at this point, which further refined the sample.

    If we had assumed that the seventeen children were randomly picked from the general population, then of course you would be quite right and two or three large variations could not possibly push the number in the top band from 16% all the way up to 94%! This goes without saying. However, this was a sample that had already been filtered by self selection until the average ability in literacy was likely to be much higher. Under such circumstances, it would only take a few atypical results to bump it up further.

    ReplyDelete
    Replies
    1. The null hypothesis assumption is a working assumption that we use to calculate a probability that the results we see occur by random chance. When we see low p-values then that may be an indication that this sample is different. The question is, why?

      You have consistently and persistently argued that the sample was too small; now that you have been shown to be wrong on this point, you are changing tack. You're begging the question in (the true sense) by assuming that the selection effects will give rise to the result seen by Rothermel.

      Instead of bandying words about, perhaps you'd care to estimate these effects and quantify their impact on the result?

      Delete
    2. "The sample was skewed firstly by the fact that an initial 80% of those asked to participate declined to do so. This immediately raises the suspicion that those who did respond were more sure of their children's academic ability."

      Not really, since the questionnaires were sent out to every member of EO, many of whom were probably not even home educating or their children were the wrong age for the study. Also, if I'm remembering correctly, bundles were given to other people, LA staff for instance, and may have ended up in a bin without reaching home educating families at all. The number of questionnaires given out in this instance is largely irrelevant unless you know how many reached currently home educating families with children of the right age.

      Delete
    3. As I said earlier: Simon is a political animal rather than a deep thinker. When the truth is inconvenient or too difficult for him to figure-out, he'll brazen it out with a convenient lie and a lot of peripheral excuses.

      Delete
  16. Anon wrote:
    "the questionnaires were sent out to every member of EO, many of whom were probably not even home educating or their children were the wrong age for the study. Also, if I'm remembering correctly, bundles were given to other people, LA staff for instance, and may have ended up in a bin without reaching home educating families at all."

    Very interesting; Simon didn't mention any of that.

    ReplyDelete
  17. 'Very interesting; Simon didn't mention any of that. '

    I didn't mention it because I wrote only a short paragraph on one aspect of the results of the tests. In February 1997, Rothermel sent out 2500 questionnaires to members of education Otherwise. A year later, she sent out a similar number, again only to members of Education Otherwise. Two hundred were also sent to religious groups and local authorities. The idea that bundles were given to local authority staff who might then have put them in the bin is absurd. Rothermel only gave them to groups who actually wanted to hand them out.

    ReplyDelete
  18. 'the questionnaires were sent out to every member of EO, many of whom were probably not even home educating or their children were the wrong age for the study'

    What would be the 'wrong age' for the study? You seem to be saying that many people who are not home educating their children still pay an annual subscription to remain members of Education Otherwise.

    ReplyDelete
    Replies
    1. Yes, that is what I'm suggesting. I was a member for a couple of years before my eldest reached compulsory school age, and more recently I was a member for a year after my youngest passed compulsory school age (my membership wasn't continuous). I know others who have done the same or have been members whilst their children were at school.

      Rothermel herself states, "Also, membership numbers of 'EO' did not indicate whether families were practising home-educators. Families may have had grown children, pre-school aged children or children moving in and out of school; they may also have been people with an interest only."

      Delete
    2. Another reason for the relatively low response rate was that the questionnaire was 'very long and difficult'. This was one of the reasons given by non-responders. Many also returned the questionnaire too late to be included; over a 1000 questionnaires were returned, but only 419 were included in the analysis.

      Delete
    3. 'Another reason for the relatively low response rate was that the questionnaire was 'very long and difficult'. This was one of the reasons given by non-responders. Many also returned the questionnaire too late to be included; over a 1000 questionnaires were returned, but only 419 were included in the analysis.'

      Yes, I was aware of this. This too skewed the results. Those parents who completed the survey were likely to be those with high levels of what is known as 'document literacy'. This means that they belonged to families where close study of texts and a lot of reading were more common. Those who were discouraged, were in the opposite case. This too would help select children from literate and well read families and automatically increase the number of subjects with good literacy skills.

      Delete
    4. "Yes, I was aware of this. This too skewed the results. Those parents who completed the survey were likely to be those with high levels of what is known as 'document literacy'. This means that they belonged to families where close study of texts and a lot of reading were more common. Those who were discouraged, were in the opposite case. This too would help select children from literate and well read families and automatically increase the number of subjects with good literacy skills."

      There is nothing automatic about this; if there is a correlation, it certainly isn't 100%.

      I, for example, might be considered quite highly literate, having read early and very widely, and authored a wide range of material including a thesis, peer-reviewed papers in academic journals, reports, press and marketing material. However, I detest forms and defer form-filling or frequently don't bother unless it's absolutely essential or involves a large amount of out-of-pocket expenses (and I don't claim the small stuff). I know I'm not alone in this.

      On the other hand, I know people who can fill in a claim form - or anything else of no obvious advantage - in short order, but who don't touch books, can barely string a sentence together, and probably read no further than the first few pages of a red-top.

      Delete
  19. 'When the truth is inconvenient or too difficult for him to figure-out, he'll brazen it out with a convenient lie '

    You have made this comment twice. Perhaps you could point out the convenient lie?

    ReplyDelete
    Replies
    1. "the samples used were so tiny as to be meaningless."

      Or are you pleading incompetence on this one now?

      Delete
    2. '"the samples used were so tiny as to be meaningless."

      Or are you pleading incompetence on this one now?'

      Well, neither! This is not me brazening anything out; it was what I said in the original post. I got the impression that you were suggesting that I resorted to lies to get myself out of tricky positions and tight spots. This was not how it was with the post, which was not made in response to anything. You seem now to be saying that I am some kind of pathological case, who lies voluntarily and purely for the sake of it. Would this do justice to your point of view?

      Delete
    3. "You seem now to be saying that I am some kind of pathological case, who lies voluntarily and purely for the sake of it. Would this do justice to your point of view?"

      That's a possibility, although I think it's more likely in this case that you simply made the assertion without really having a clue about what you're taking about. So it was incompetence to begin with, then it becomes a lie as you try to sustain an untenable point. Or perhaps you're really too stupid to appreciate your position; it's difficult to know what's going on inside your head.

      Delete
    4. taking->talking

      Delete
  20. 'So it was incompetence to begin with, then it becomes a lie as you try to sustain an untenable point.'

    Forgive me, but I don't really follow your reasoning. You claimed at first that I resort to lies when I was forced, as you put it, 'to brazen it out'. When I asked for an example of this behaviour, you quoted a sentence from my original post. This was clearly neither a lie nor an attempt to brazen anything out.

    You seem now to be saying that the original statement was false and that I gave it mistakenly. It then became a lie when I defended it. Is this what you are asserting; that in the first instance I accidently, perhaps through sheer stupidity, made a false statement and then wittingly elaborated upon it? This would, I suppose, if true, make me both incompetent and a liar!

    ReplyDelete
    Replies
    1. "...This would, I suppose, if true, make me both incompetent and a liar!"

      That sums it up nicely. As I said, you're a political animal, and I think you try to avoid explicit, outright lies, but when you're trying to peddle something that involves arguing beyond your ability, you dig a hole for yourself. I think the nickname "Tangled Webb" has been used in reference to you, either on this blog or in reference to it.

      Delete
  21. Firstly I find it difficult to follow who said what on here as people seem incapable of choosing a name to use to help readers follow the thread.

    It was said that at least ten of the 16 reported to be in the top band would have to have been misreported - i.e., exaggerated or overestimated - to nullify the significance of the result. Does that mean if we know 9 of them to be exaggerated or overestimated that the results would be valid? And would you be happy to accept them?

    People can do as many Monte Carlo simulations and Binomial distribution calculations as they like. I stand by what I said that a sample of 17 children whose results have been completed by their parents is not valid research.

    Somebody once said It is one thing to be clever and another to be wise, there are people on here who prove that to be true.

    ReplyDelete
    Replies
    1. Go for it, Cheshire Cat! Never let reality stand in the way of a good old fashioned, "staan's to reason, don'it" style of non-argument - especially when you can throw in some trite words about cleverness versus wisdom.

      Delete
    2. "I stand by what I said that a sample of 17 children whose results have been completed by their parents is not valid research."

      Again, you are conflating the arguments. The point being made is that - contrary to Simon's claim - the sample of 17 is not necessarily too small. The issues of self-selection and assessment are different ones.

      "Does that mean if we know 9 of them to be exaggerated or overestimated that the results would be valid? And would you be happy to accept them?"

      You're asking for yes-or-no answers when there are only probabilities (and not even probabilities of what you're trying to determine). A threshold of p=0.5 is just a convention that people specify in advance for acceptance or rejection, but doesn't necessarily mean that one should believe a result one way or another, particularly in critical cases.

      In summary, Simon's point about sample size is spurious; there are selection effects and concerns about methods - but these aren't necessarily as well understood or clear-cut as Simon claims - and the result has a degree of significance that cannot readily be dismissed. Does it prove anything? Of course not. Should it be rejected out of hand? No. If anyone cares enough, it needs a better study, and this is the way many lines of research progress.

      Personally, I don't care very much, but I object to the aggresive dismissal of the work by those who want to advance their own agenda but are too ignorant or lazy to make an intelligent case.

      Delete
    3. Cheshire Cat said:
      "Somebody once said It is one thing to be clever and another to be wise, there are people on here who prove that to be true. "

      Sounds more like Kung fu Panda than Cheshire Cat.

      The truly wise don't come near this blog, but yes, there may be some clever, some wise, some clever and wise... then there's Simon and Cheshire Cat.

      Delete
    4. In fairness to Cheshire Cat, I don't think s/he is a bad sort, just simply misreading the arguments and disturbed by a result that I presented which is outside of their "common sense" experience.

      The wise reaction would be for CC to say, "that's interesting", question it by all means, and go away and try to understand it. That's the kind of thing we should be teaching our kids to do.

      Delete
    5. I think CC's a he, and he didn't have that reaction, did he; instead he reacted like a bombastic idiot. Unfortunately, he's also a home educator, so he looks like a proof-point for inspection.

      Delete
    6. 'he reacted like a bombastic idiot. Unfortunately, he's also a home educator, so he looks like a proof-point for inspection.'

      That should solve the unemployment problem then. You'd need to recruit a LOT of inspectors to get round every bombastic idiot who home educates.

      And I speak as one myself!

      Delete
  22. I shall explain the problem with small, skewed samples once more and then that will be it. Suppose that I wish to see what the average IQ is for people in Birmingham. If I use only a very small sample, say two people, then if one is very atypical; this will distort my results dramatically. If one of my sample had an IQ of 100 and the other had an IQ of 160, then I would conclude that the average IQ of people in Birmingham was 130. This would be less likely if I had a sample of three; two of whom had an IQ of 100 and one with an IQ of 160. In this case, I would conclude that the average IQ was 120. The larger the group; the less that one or two oddities will produce a distorted result.

    We must also make sure that our sampling is random. If I stood outside a hall where a meeting of MENSA was taking place, then this too would distort my results. (Don't laugh, this actually happened with a survey on a bacon snack. Those making it were surprised how unpopular it seemed to be in a survey. This was until they discovered that the survey had been conducted in the Golders Green district of London; an area with a high concentration of Jews.)

    These two things, skewing the original sample and then using too small a selection, can cause weird conclusions to be drawn.

    ReplyDelete
    Replies
    1. Your example does nothing for your case, Simon.

      When stating the mean, one should also give an error on the mean. In your IQ example, the standard deviation is 42, so one would state the result as 130±42. (That's 130 plus-or-minus 42, for any browser that doesn't render the character properly).

      That does NOT mean that the correct answer is definitely in the range 88-172; rather, with certain caveats (normal distribution etc), we believe that a result for this mean and sampling would be found in this range 68% of the time.

      If you look for certainty from statistics, you're likely to be bitterly disappointed; if you try to understand what they're saying, they can be a powerful guide. Quite often, you must be prepared to do further work - involving calculation - when you read a statistical result, in order to understand it properly. Only then will you be in a strong position to criticise the work, otherwise you may find yourself in a hole.

      Simply waving a shroud about sample size and skewed data isn't good enough; I can find issues with Rothermel's work (and I've no doubt that she and her peers are well aware of them), but attacking it from your naive and simplistic position is foolish.

      Delete
    2. Yet another example of Simon's complete lack of understanding of statistics.

      Delete
    3. It's always amusing watching Simon trying to teach granny to suck eggs. He ends-up with a mouth full of broken shell and egg all over his face.

      Delete
  23. "Perhaps you could start by explaining "bollocks" in this context."

    First, let's assume your analysis is sane; you've just rejected the null hypothesis, which is the opposite of what you seem to claim. Also, you can never accept the null hypothesis under Fisher. Second, where on Earth did you get the probability required for a binomial test? Empirically derived? Cute. Made up in you head? Lunacy. Thirdly, you're answering the wrong question in the first place: http://www.surveysystem.com/sscalc.htm is something like what is required.

    ReplyDelete
    Replies
    1. "First, let's assume your analysis is sane; you've just rejected the null hypothesis, which is the opposite of what you seem to claim."

      HUH?! This makes no sense.

      "Also, you can never accept the null hypothesis under Fisher. Second, where on Earth did you get the probability required for a binomial test?"

      Religious battles aside, look at the information provided by Simon and think.

      "Made up in you head?"

      No.

      "Thirdly, you're answering the wrong question in the first place: is something like what is required."

      Oh dear; someone pushing a product. I'll stick to R or julia - or plenty of other reputable products if I have to pay.

      Delete
  24. None of the three myths that have been mentioned hold any sway with me and my decision to Home Educate.

    The fact that my children's schools where failing to Educate them was more of a concern of mine.

    ReplyDelete
  25. Just a few things. I am a male.

    Am I the Kung Fu Panda? No Do I bark? No

    am I a bombastic Idiot? You must decide (Cue the anonymous agreements).

    Are Webb and Cheshire Cat the same person? No

    I wasnt sure why we were described as Jekyll and Hide but if we get bored of it can we swap roles?

    However Anonymous said Unfortunately, he's also a home educator, so he looks like a proof-point for inspection.' If you had read some of my previous points, the ones where I bark presumably, you will have seen that I am not a home educator. I have never said I was or pretended to be one.

    Again though I am perplexed at people’s inability to choose a name and stick to it. It is difficult to know if those who think I am an idiot think I’m a panda etc…

    ReplyDelete
  26. "I am not a home educator."

    Thanks Cheshire Cat, that's good to know.

    ReplyDelete
  27. Actually Dietrich Bonhoeffer was educated at home, as were his 7 siblings. Their mother was a qualified teacher and she taught him at home up to the point where he successfully took his Abitur exams. The Abitur was and still is the highest German secondary school qualification and the prerequisite for matriculation at a University.

    It was not uncommon for well off families such as the Bonhoeffers to keep their children at home and teach them there. Other parents may have simply employed a tutor rather than do the teaching themselves. When the Nazis came to power, they put an end to this as part of their attempt to immerse the population as a whole in Nazi ideology. Obviously they had a keen interest in ensuring that children were not being brought up under the influence of an ideology other than their own. One way of preventing this and minimising the effects of parental indoctrination was to make schooling compulsory and for all schooling to be at state-controlled schools only. Since that point, home education has been illegal in Germany. Prior to the Nazi period, it was as common in Germany as it would have been in the UK.

    ReplyDelete