Analyzing the Setup
Imagine you are standing in a classroom. The teacher has just handed back the math exams, and the results are, well, not great. To lift everyone's spirits, the teacher makes a generous decision: every single student will receive a grace mark of 10.
We have a set of original marks, x1,x2,…,xn. These marks form a distribution—a beautiful, bell-shaped curve representing the class's performance. The center of this curve is the mean, defined as:
The Change of Origin
When the teacher adds 10 to every student's score, we create a new set of marks, yi, where yi=xi+10. In the world of statistics, this is known as a Change of Origin.
Visually, imagine the entire bell curve picking itself up and sliding 10 units to the right on the number line. The shape of the curve—its height, its width, and its spread—remains identical. It has simply moved to a new neighborhood.
Why the Mean Follows
It is intuitive that the mean, which represents the center of the data, must move with the data. If every student gets 10 more marks, the average score of the class must also increase by 10.
Mathematically, the new mean yˉ is calculated as:
By splitting the summation, we get:
This simplifies beautifully to yˉ=xˉ+10. The center has shifted, just as we expected.
The Invariance of Variance
Now, we arrive at the heart of the mystery: the variance. Variance, denoted as σ2 or Var(X), is defined as the average of the squared deviations from the mean:
This formula captures the 'spread' of the data. It measures how far, on average, each student's score is from the class average.
Let's calculate the variance of our new, shifted marks, Var(Y)=n∑(yi−yˉ)2. If we substitute our expressions for yi and yˉ, something magical happens.
The deviation term becomes:
(yi−yˉ)=(xi+10)−(xˉ+10)
When we open those brackets, the +10 and −10 cancel out perfectly. We are left with (xi−xˉ).
The Elegant Conclusion
Because the deviation of the new data is exactly the same as the deviation of the old data, the squared deviations are also identical. When we sum them up and divide by n, we find that:
The spread of the curve did not change because the relative distances between the students' marks remained constant. Shifting the entire class by 10 marks didn't make the class performance more or less consistent; it just made everyone look a little better.
And that, my friend, is why variance is independent of the change of origin. It is a fundamental property of dispersion that remains untouched, no matter how much you shift the starting point.