Obesity and height and weight

Phase B:

        I will now need to separate my sample into year groups. Although I could have done age, it would have been much more difficult, and a month or two off does not tend to affect someone’s BMI a great deal. I will now investigate how age (carried out by year group) affects a person’s BMI. From section A all I am told is that the middle 50% lie in between roughly 17 and 22. From my own knowledge, I know that different pupils will have different BMI’s for different reasons. For example, males even if they are the same height as a female, tend to be heavier. Also a person’s BMI is changed by their age. Hence the reason that I will now investigate the trends of how BMI changes with ages.

Here are some graphs to show the results that I have found.

 

        

As can be seen from the box and whisker plots, the medians seem to be more or less in line; they are all near enough the same. However, to exclude year 7 and year 11, there is a general trend that as we go up through the year groups the interquartile range gets smaller. The BMIs of the all the pupils seem to get very cramped together, which seems to suggest that their dimensions may be more similar. In the box and whisker plot I have also included th box and whisker pot for the whole school.. There is very little change in the median and the skew. A we can see, the skew is to the left, which means that the values are not evenly spread. It is particularly difficult to come up with an accurate conclusion for this, however, the evidence seems to be reliable enough for me to conclude that your age does not really affect the normal (middle and medium) BMI, however people seem to have a larger range of BMIs. For example, the range for year 7 is smaller than the range of values for year 8 which is also smaller than year 9. Year 10 is the one year that does not seem to fit the trend. I can only come up with one reason to explain this. In my view, this is because as we grow our BMIs generally increase, however, year 10 is the age when the growth spurt starts to kick in. However, many of the year 10s will not have started their growth spurt and have stayed the same as the year before (I must be careful not to try and compare them to this years year nines, as they are a completely different set of data. So, to conclude my theory, it could be that not everyone has started their growth spurt in year 10, which leaves the average 10 as the average 9 was. In year 11, however, almost everybody has started the major growth spurt, which has made up for the lacking in year 10 and carried out the expected result. It is almost perfectly in line with years 7, 8 and 9. If I draw a line through the maximum limbs or whiskers of the box plot for every year I get an almost perfect line. I also notice upon viewing the table that most of the year tens that were included in my sample, are girls. Therefore, I would like to investigate whether gender has any affect.

Phase C:

I felt that Box and Whisker diagrams are more useful for comparing data than cumulative frequency graphs. It clearly diplays where the middle 50% of the sample lie and where the middle, lowest and highest values are, hence enabling us to determine the skew of the data. Below are box and whisker plots comparing boys to girls within each year group. I am doing this as I would like to investigate ho gender affects a persons BMI. I will investigate gender within years as I this, although I have smaller samples, would mean that I am able to see, instead of a representative of all years, how the trend changes from year to year.

From looking ath the box and whisker plots, it is immediately clear to the eye that females have an almost constant median BMI. It ranges from 18.5 to 19.05 over al year groups. From my own knowledge, one of the main things that I can put this down to would be that as girls grow, they seem to keep a similar ratio between their height and their weight. Boys however fill out and become more built. This can be seen from the graphs as the boys medians seems to constantly increase as we go through the years. Throughout the years, the girls seem to have a fairly similar interquartile range. The boys seem to have a similar interquartile range except for year 11. I only had 4 pieces of data for this category, so I may have chosen 1 or 2 severely overweight people which would have influenced my results.

 The two values that I have highlighted, are ones that I feel have changed the median and upper quartile, and made the interquartile range too high.

So, to exclude year 11, there is a trend that there is a substantial change in the BMI for the boys. However for the girls it tends to remain fairly similar.

The council will use this information to try to improve the borough’s health and use the median values as rough guidelines for the ideal weight of a child of a particular age and gender.

I think that this project could have been improved if I had taken a larger sample and had it stratified within year groups rather than a stratified sample of the the whole school. I probably should have also analysed only the middle 50% of all the data, thereby eliminating any particularly strange and anomalous BMIs.

Appendix

Here is where I have included all the data I was given and my sample, which I then split down further into year groups and then gender within year groups.

Here is all the data after I eliminated any information that I considered to be irrelevant to my investigation, such as IQ and favourite subject.

Below is all the data after I  split it up by year group:

I have used colours to differentiate between year groups but as can be seen the genders are still mixed.

Here it is clear that it has been split own into genders within years.

Sami azzam 10E