Food for the Brain: Child Survey: Review Part 10

Professor Patrick Holford of Teesside University (and also Head of Science and Education at Biocare) and Drew Fobbester are joint researchers and authors of the Food for the Brain Child Survey, September 2007 (pdf). We have previously examined the overview of the literature in this report and found that it was misleading and irrelevant. Unfortunately, the reporting and analysis of the survey data look to be even worse: there are frequent errors in the graphs and even more frequent misinterpretations. Because of these mistakes, it looks as if few of the claims are accurate.

For Part 8 of this review we looked at the alleged benefits of eating dark green leafy vegetables. For this post, we look at the claims for fried/and or takeaway food (FTF): however, Holford Watch can not find evidence to support any of the important claims. It is very difficult to comment about the claims made for the deletrious impact of FTF on children because there are so many mis-labellings and mistakes in the report that the reader can not be confident that the correct data are being offered. E.g., Section 7.2 (pg. 19) is supposedly about Food Group Consumption and Behaviour – V Good Rating; however, the label for the figure reads, “Increase (Green +ve) or reduction (Red -ve) in very good health ratings comparing high consumers with low consumers of each food group”. In addition to that, we learn:

Children who eat a lot of fried/and or takeaway food are less than half as likely to be well behaved. [pg. 19]

Ignoring those errors, Fried and/or Takeaway Food appears as the worst of the top 3 negative foods in terms of impact for both Overall Health and Behaviour but not (despite the publicity) for academic performance.

Children who eat a lot of fried/and or takeaway food…are half as likely to be in very good health. [pg. 18]

If the reader wishes to co-operate with the authors, the reader overlooks the use of the phrase “a lot” and concentrates on comparing and contrasting the highest and lowest consumers of FTF. The first set of graphs for Fried Food, Section 9.1 (pg. 22) look intimidating.[1] The doom-laden summary statements reads:

Children who eat fried or takeaway food most days are nearly three times more likely to be badly behaved than children who never eat these foods. [pg. 22]

FFTB Child Survey 2007 Section 9.1 Fried Food and/or Takeaway Food or Food Cooked in Heated Fat

Just to add to the jeremiad about the impact of FTF on children, Section 11 discusses overall health and looks at the

four most common symptoms of behaviour or performance problems, and their correlations with aspects of diet. [pg. 34]

The section is a remarkable indictment about FTF which is implicated in all four of these symptoms/problems and occupies the top spot out of the three most negative foods for three of these.

  • Section 11.2 (pg. 35) looks at “concentration and poor attention span”; FTF is first out of three for “most negative” of the “significant diet correlations with this symptom”
  • Section 11.3 (pg. 35) looks at irritability, mood swings and emotional outbursts; FTF is first out of three for “most negative” of the “significant diet correlations with this symptom”
  • Section 11.4 (pg. 36) looks at “difficulty sleeping and restless sleep”; FTF is third out of three for “most negative” of the “significant diet correlations with this symptom”
  • Section 11.5 (pg. 36) looks at “slow to get going in the morning”; FTF is first out of three for “most negative” of the “significant diet correlations” for this observation.

The table of food impacts in Section 4.3 (pg. 12) makes grim reading for FFT so it is understandable that in the overall diet recommendations of Section 4.4 (pg. 13), the authors recommend that:

children should eliminate completely or, at least keep to an absolute minimum, all of the following ‘foods’.

  • Fried food and takeaways…

However, given this grim litany of the impacts of FTF, it is a little surprising that when the authors detail the average diet of children, FTF are not mentioned. It is an exercise for the reader to consult the Section 8 (pg. 21) food frequency tables and learn that 32% are said to never eat FTF and an addtional 48% eat it once per week. So, it would seem that 80% of children are already observing this restriction even without have a guideline to follow.

So, given that we are discussing a reported 20% of children who eat FTF more than 1 a week, just how strong is this evidence for them? After all, FTF are said to be strongly associated with undesirable symptoms and problems, and overall ‘half as likely to be in very good health’ and ‘three times as likely to be (rated as) badly behaved’.

Again, it is a question of whether you prefer to look at comparisons of percentages which can be very misleading when you are comparing groups of different sizes, or you look at absolute numbers. For all of the above assessments, the group with the highest rate of consumption represents 2% of the children (or 204 for a sample of 10,222; 63 for a sample of 3139 as per SAT scores).

If we reverse-engineer the numbers, we learn that for the 10,222 sample:[2] 3271 children ‘never’ eat FTF and 204 eat it ‘most days’. Looking at the charts for Section 9.1 (pg. 22) reproduced above, we see that:

  • For behaviour, 16% of the children in the ‘never’ category are rated as Poor/V. Poor for behaviour (red) and 21% are rated as Very Good (green). There are 687 (green) children and 523 (red) children
  • For the ‘most days’ category, 44% of children are rated as Poor/V. Poor for behaviour (red) and 6% are rated as Very Good (green). There are 12 (green) children and 90 (red) children
  • There are several ways of performing comparisons but it should be apparent that there are many more children who are rated Poor/V Poor for behaviour who never eat FTF than those who eat it ‘most days’; 523 v. 90. It is inappropriate to attempt to derive any significance from comparing % or absolute numbers here because of the disproportionate size of the groups
  • For academic performance, 16% of the children in the ‘never’ category are rated as Poor/V. Poor for academic performance (red) and 19% are rated as Very Good (green). There are 621 (green) children and 523 (red) children
  • For the ‘most days’ category, 40% of children are rated as Poor/V. Poor for academic performance (red) and 14% are rated as Very Good (green). There are 29 (green) children and 82 (red) children
  • There are several ways of performing comparisons but it should be apparent that there are many more children who are rated Poor/V Poor for academic performance who never eat FTF than those who eat it ‘most days’; 523 v. 82

Assuming that the authors made similar comparisons to derive their observations for the association of FTF with various health symptoms and performance problems, then the evidence for those claims does not look robust. There may be many reasons to deprecate FTF but the authors of this report have not produced any reliable evidence for their recommendation. If research from other groups should indicate that restricting FTF is helpful, then it is perhaps fortunate that 80% of those who completed this survey indicate that children and families are independently choosing to restrict their intake without the help of any guidelines based on poor number-handling.

Holford Watch does not pretend to understand the relationship between the above numbers and the claims made in the Section 4.3 (pg. 12) food impact table.[3]

For overall health, behaviour and academic performance it shows the increase or decrease in propensity to have a very good score between high consumers of a food and low/non consumers of the food. [pg.12]

The numbers quoted in the table are as follows:

Extract from FFTB food impact table

Holford Watch notes that the authors derive -71% for the calculation for table 4.3 (pg. 12) and the claimed impact on behaviour. Initially[3] the baseline number for comparison was not clear to us and that is because Holford and Fobbester have done something with the numbers that is so surprising that it is difficult to describe. Remember that if the reader compares and contrasts the high consumption (‘most days’) group with the ‘never’ group, then there are children who are rated very good or poor/v. poor for (say) behaviour in both groups. Further, the high-consumption group represents 2% of the children and the ‘never’ group, 32%. For FTF in Section 9.1 (pg. 22) as reproduced above, the percentages reported for these very different sizes of frequency of consumption groups are 21% (green) and 16% (red) for the ‘never’ group; for the ‘most days’ group, 6% are (green) children and 44% (red) children. In raw numbers, there are 687 (green) children and 523 (red) children in the ‘never’ category; for the ‘most days’ group, there are 12 (green) children and 90 (red) children.

Holford Watch believes that Holford and Fobbester derived a -71% for the calculation for table 4.3 (pg. 12) and the claimed impact on behaviour by expressing the percentage number of green children from the ‘most days’ group over the number of green children from the ‘never’ group and then subtracting to estimate the negative impact on behaviour. Thus:

6/21 = 0.29
1-0.29 = 71

The same sort of calculation produces the same results as Holford and Fobbester claim in the food impact table 4.3 (pg. 12). Remember that the authors are comparing children with very good ratings for behaviour from one group with those from another. Holford Watch has no idea why the authors think that this is a useful calculation or in what way it is supposed to support their claims. We leave it as an exercise to the reader to marry up this calculation and the many statements scattered through the report and quoted above with a sample given below:

For overall health, behaviour and academic performance it shows the increase or decrease in propensity to have a very good score between high consumers of a food and low/non consumers of the food. [Re: food impact table Section 4.3, pg.12]

Children who eat a lot of fried/and or takeaway food are less than half as likely to be well behaved. [pg. 19]

Children who eat fried or takeaway food most days are nearly three times more likely to be badly behaved than children who never eat these foods. [pg. 22]

Under the circumstances, Holford Watch can not confirm that there is sufficient quality or quality of evidence to support the report’s findings and claims for FTF; it looks as if the claims are exaggerated at best or wrong. The data are questionable and the analyses of that data are inappropriate at best in some instances and wrong in others. We can not find that there is any reason to have confidence in the authors or their findings. There may well be some disadvantages to eating FTF on a regular/frequent basis, but Holford and Fobbester do not have the evidence to support their assertions.

We could have interpreted the data in several different ways for overall health, academic performance and behaviour but there are so many serious flaws that it really was not worth further effort. There is a good chance that both your grandmother and mother tried to persuade you to eat more greens and less fried food; they probably told you that the former would make you more ‘regular’ and the latter piece of advice would give you fewer spots. The FFTB is making similar recommendations but predicting different outcomes. Holford and Fobbester have dressed up their recommendations in scienciness and presented the reader with lots of distracting figures and graphs, but it doesn’t mean that there is any greater quality of evidence for their recommendations or foundations for their claims than those from your grandmother or mother.

Notes

[1] We have discussed the accurate interpretation of these line graphs and the misleading nature of the chosen display in Review Part 8.
[2] For completeness, in case these charts are based on the sample for whom there are SAT scores, there are 3139 children: 1004 children ‘never’ eat FTF and 63 eat it ‘most days’. Looking at the charts for Section 9.1 (pg. 22) reproduced above, we see that:

  • For behaviour, 16% of the children in the ‘never’ category are rated as Poor/V. Poor for behaviour (red) and 21% are rated as Very Good (green). There are 211 (green) children and 161 (red) children
  • For the ‘most days’ category, 44% of children are rated as Poor/V. Poor for behaviour (red) and 6% are rated as Very Good (green). There are 4 (green) children and 28 (red) children
  • There are several ways of performing comparisons but it should be apparent that there are many more children who are rated Poor/V Poor for behaviour who never eat FTF than those who eat it ‘most days’; 161 v. 28. It is inappropriate to attempt to derive any significance from comparing % or absolute numbers here because of the disproportionate size of the groups
  • For academic performance, 16% of the children in the ‘never’ category are rated as Poor/V. Poor for academic performance (red) and 19% are rated as Very Good (green). There are 191 (green) children and 161 (red) children
  • For the ‘most days’ category, 40% of children are rated as Poor/V. Poor for academic performance (red) and 14% are rated as Very Good (green). There are 9 (green) children and 25 (red) children.

[3] We were so baffled as to what the authors intended that our first attempt at this section was as follows.
Holford Watch does not pretend to understand the relationship between the above numbers and the claims made in the Section 4.3 (pg. 12) food impact table.

For overall health, behaviour and academic performance it shows the increase or decrease in propensity to have a very good score between high consumers of a food and low/non consumers of the food. [pg.12]

The numbers quoted in the table are as follows:

Extract from FFTB food impact table

However, the baseline number for comparison is not clear to us. Despite many different calculations, we can not guess how the authors derived those numbers for the relative impact. If the reader compares and contrasts the high consumption (‘most days’) group with the ‘never’ group, then there are children who are rated very good or poor/v. poor for behaviour in both groups. Further, the high-consumption group represents 2% of the children and the never group, 32%. E.g., even if the behaviour ratings from Section 6.2 (pg. 17) were generalised across all of the consumption frequency groups (irrespective of their disparate sizes) then we learn that 15.7% of the children should have a very good behaviour rating, and 16.6% of children should have a poor behaviour rating and 4.4% a very poor behaviour rating. So, for FTF in Section 9.1 (pg. 22) as reproduced above, because the poor/v. poor groups have been merged, we would expect to see relative percentages of 15.7% for children with a very good behaviour rating (green), and 21% of children with a poor/v.poor behaviour rating (red). The actual percentages reported for these very different sizes of frequency of consumption groups are 21% (green) and 16% (red) for the ‘never’ group; for the ‘most days’ group, 6% are (green) children and 44% (red) children. In raw numbers, there are 687 (green) children and 523 (red) children in the ‘never’ category rather than the 513 (green) and 687 (red) children predicted by generalising from Section 6.2. Similarly, for the ‘most days’ group, there the prediction was for 32 (green) children and 43 (red) children rather than the 12 (green) children and 90 (red) children. Holford Watch does not understand how the authors derived a -71% for the calculation for table 4.3 (pg. 12) and the claimed impact on behaviour.

Further Reading

Food for the Brain Child Survey 2007: The Promotion
Holford Watch looks at the literature review:
Food for the Brain Child Survey 2007: Review Part 1
Food for the Brain Child Survey 2007: Review Part 2
Food for the Brain Child Survey 2007: Review Part 3
Food for the Brain Child Survey 2007: Review Part 4
Food for the Brain Child Survey 2007: Review Part 5

Holford Watch appeals for help to Professor Holford and two members of the Scientific Advisory Board who approved this report and then looks at the data and analyses:
Food for the Brain Child Survey 2007: Review Part 7
Food for the Brain Child Survey 2007: Review Part 8
Why Don’t Food for the Brain Report Their Survey Results on Supplement Pills Survey: Review Part 9
Food for the Brain Child Survey 2007: Review Part 10

Advertisements

4 Comments

Filed under children, education, Food for the brain, health, Holford, patrick holford

4 responses to “Food for the Brain: Child Survey: Review Part 10

  1. Claire

    It seems the purpose was hype…sorry, ‘hypothesis-generating’, explained here:
    http://dcscience.net/?p=218

  2. tifosi246

    I’m currently reading Extraordinary Popular Delusions and The Madness of Crowds by Charles Mackay (originally published in 1841). One chapter lists various petitions for share issues in companies related to the South Sea Bubble, one of which is:

    For carrying on an undertaking of great advantage; but nobody to know what it is.

    Apparently in just 6 hours, the originator of this scheme sold 1000 shares at £2 each (in 1720’s money).
    The obscure way this FFTB report has written, I can’t help but think its authors are taking a leaf from the stock book of this sage of the South Sea Bubble. Hopefully, your merciless thoroughness here will prevent government or other policymakers paying up their £2 worth in 2008 money.

  3. Pingback: Our Original Questions to Patrick Holford About the Food for the Brain Child Survey 2007 « Holford Watch: Patrick Holford, nutritionism and bad science

  4. Pingback: The Economist: The End of a Childhood Illusion « Holford Watch: Patrick Holford, nutritionism and bad science

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s