Σ2

Aim

The aim of this investigation is to see if the predicted grades given by subject teachers at the time of Yellis, has more correlation to grads achieved by students in there gsce’s then there Yellis predictive grades

Yellis is a test taken in year 9 of schooling; there are 2 parts to the test, a mathematic and vocabulary styles paper, each student has to complete in exam conditions.

The subject teachers predicted grads are based on departmental meetings where all teachers of the student discuses the progress of the child. They take in to consideration the students natural ability, work rate in class, behaviour, quality of work, and test results, these aspect are taken over many weeks so a rounder picture is built up of the student of which a prediction can be based,

Yellis only take in to account the mark achieved on the test, ranks the results all around the country, then assign grand to each percentage, eg top 20% are predicted A.

The population that I will be sampling the data from, is Stoke Dameral community College, but there may be implication outside he given population, as the trend may be evident over the whole of the examined students, but I can’t say that it will be just with data from the Stoke Dameral as there may be different methods of producing predicted grades. The reason that I have chosen stoke Dameral as the parent population is because the data was available to me very easily, and I have spoken to the heads of department and they do not give false predicted grades do e.g. one’s that will give a misrepresentative grade to stop student getting complacent.

I would like to investigate the correlation between the achieved GCSE results and both the predictions by yellis and staff, I will be looking at Mathematics and English language because yellis test specify these two skills so it is more likely to have less deviation, then to a subject such as art which is far from the original type of skill tested, I will also look at the results for Science double award because this is a subject that uses both skills, so it should be a truer representation to the results achieved.

There are over 220 students in the class of 2001 that took double award Science, mathematics, and English language

I will use a form of random quota sampling using a stratified ???? method, I will start by selecting every fifth student, starting from the sixth piece of data. These are in alphabetical surname order. For privacy reasons there have been replaced with ascending numerical values. Taking 50 students as the sample population is the quota part of the sampling. They also need to have taken all of the exams and have staffs predicted grades.

I will plot predictive grades (by Yellis and staff) against grades achieved, I will look at the distribution of the data and I will adopt an appropriate technique, to calculate the correlation coefficient, depending on the correlation on the scatter graphs.

What am I looking for and why.

In the scatter diagrams I will be looking for a relationship between the two variables (x,y), this relationship is know as the correlation.