- I.Q
- Average number of hours of T.V watched per week
- Number of siblings
- Key Stage 2 results in English, Maths and Science
As there are 1183 pupils who filled in this questionnaire I shall use other fields, to help sort my data into smaller, more manageable samples as 1183 is far too larger number to work with. The other fields I will use are year group and gender as I intend to split all the pupils into their 5 different year groups (7, 8, 9, 10 and 11) and then into boys and girls within the year groups. This will give me 10 groups. I shall then take a systematic sample from each of the 10 groups. To determine my nth term, used in my systematic sampling I shall roll a dice and use this number for my selections, for example if I rolled a 4 I would take every 4th pupil’s data until I reach the total amount of data required from each group. I have decided to take a total of 15 pieces of data from each group, leaving me with a total of 150 samples, which is just over 10% of the total number of pupils at the school. This way I shall get an equal spread of boys and girls from each year group. If I were to take a percentage from each year group, or even from each of my 10 groups, I would end up with more pupils from larger groups, leading my results to be bias towards a certain year group or gender, as each year group is a different size with unequal amounts of each gender.
Once I have followed my sampling methods, cut my results down to my final 150 and remove unneeded fields, which I will display on a separate spreadsheet, I shall then represent and compare my data in different graphs to make it simpler to interpret its relation to my hypothesis. The first thing I am going to do is copy my data into MS Access so that it is easier to sort, making it easier to create my graphs. The first graph I am going to create is a cumulative frequency curve, showing I.Q. This will enable me to calculate the average I.Q. To create this graph I will create a table in Microsoft Excel by splitting I.Q into the following 7 groups:
0 < x 20
20 < x 40
40 < x W 60
60 < x 80
80 < x 100
100 < x 120
120 < x 140
I am using these groups as the lowest I.Q from my sample is 11 and the highest is 134. I will then find the frequency of pupils from my sample that belong in each group. From finding the frequency for each group I can then calculate the cumulative frequency and plot it onto my graph, by hand, creating a cumulative frequency curve. Then I will go on to calculate the median, lower quartile, upper quartile and inter-quartile range, which I will display in a box plot.
I am also going to produce a bar chart representing the average number of siblings and the average, average number of hours of T.V watched per week within the same groups of I.Q as used in my cumulative frequency graph. First I will do a table, using Microsoft Excel, to show theses 2 averages in each I.Q group. I will then turn this data into one bar chart using Excel. I hope to see some sort of negative trend for both averages on my graphs if my hypothesis is to be correct.
I am then going to see how I.Q affects key stage 2 results. I have decided the best way to do this would be to once again split my I.Q into groups and find out what percentage people who received level 4 and above for all subjects at key stage 2 belong in each group of I.Q. I am using people who got level 4 and above as level 4 was average for key stage 2 and level 5 was above average. This information will then be put into a pie chart. This diagram will show me whether more people with a higher I.Q received average and above than people with a lower I.Q. if my hypothesis is to be correct about key stage 2 results I expect to see a higher percentage for the higher I.Q groups and a lower percentage for the lower I.Q groups. To make this pie chart I will produce a table with the following groups:
0 < x 30
0 < x 60
60 < x 90
90 < x 120
120 < x 150
I have used larger ranges for each group this time to try and ensure that I get a more reasonable percentage to deal with rather than having a group with say 0% or decimal percentages. Then I will find how many pupils in my sample received level 4 and above for every subject and out of these I will find the number of pupils belonging to each I.Q group. For example there may be 75 pupils all together who receive level 4 for all subjects and 5 of these pupils could have an I.Q between 0 and 30. Once I have recorded this I will convert each I.Q groups frequency into a percentage. I will use Microsoft Excel to show this data in a pie chart, as it will be more accurate than drawing one by hand.
I will also produce a time series graph for key stage 2 results. I will do this by creating 2 tables in Excel one for above average I.Q and the other for below average I.Q. I will find out the average I.Q from my cumulative frequency graph. Each table will show how many people received what level (0, 1, 2, 3, 4, 5 and 6) for each subject at key stage 2. Both of these tables will be plotted on the same axis but in different colours. This graph will show me whether people with below average I.Q received the higher or lower marks and the same for above average I.Q. If my hypothesis is correct I expect to see more people with above average I.Q receiving the higher marks, for example level 4 and up, and more people with below average I.Q receiving the lower marks, for example level 4 and below. Subject and level will be plotted along the X-axis and number of pupils plotted along the Y-axis.
The last graphs I am going to create are 2 scatter diagrams using Microsoft Excel. I shall compare I.Q to number of siblings on one graph and I.Q with average number of hours of T.V watched per week on another. On one I will plot I.Q along my X-axis and number of siblings up the Y-axis. I will then compare I.Q and number of siblings on this scatter graph. I expect to see a negative correlation for this comparison if my hypothesis and assumptions are to be correct. On my other scatter graph I shall once again plot I.Q on the X-axis and then average number of hours of T.V watched per week up the Y-axis. I will then compare I.Q and average number of hours of T.V watched per week on this graph. Once again I expect to see a negative correlation if my hypothesis is to be correct. On my 2 scatter graphs I shall draw a line of best fit, if possible, for each comparison and see if there is any correlation and whether it is the correlation I was expecting to see to make my hypothesis correct.
I will draw conclusions from all my graphs and comparisons, comment on any patterns I can see and say what parts if any of my hypothesis were correct and how I can tell this. Finally I will evaluate my strategy and make sensible suggestions about how I could improve the process.