13.7.1. Chi-Square Test

Used to test the goodness-of-fit of the adjustment:

$\chi^2 = \sum_{i} \frac{v_i^2}{\sigma_i^2}$

Compared with a critical value to assess if the residuals are within expected limits.

The Chi-Square Test is a statistical method used to determine how well the observed data from an adjustment fits the expected data. In simpler terms, it checks if the differences we see between what we observed and what we expected (residuals) are reasonable given the uncertainty (variability) in our measurements. The formula reveals that we sum the squares of these residuals divided by their variances. The result, $\chi^2$, is then compared to a critical value from the Chi-Square distribution table, which tells us whether our results fall within an acceptable range. If $\chi^2$ exceeds the critical value, it indicates that there are significant differences between observed and expected values, suggesting the adjustment might not be valid.

Imagine you're a teacher trying to see if the grades of your students fit the grading curve you set at the beginning of the term. If you predicted that most students would score around a B, and instead, a lot scored F's and A+'s, you might question your grading curve. You could apply a Chi-Square Test to see how far off your actual grades are from what you expected based on your predictions. If the results show a significant difference, it suggests that maybe your grading curve needs to be adjusted.

Residuals are the differences between observed values and the values adjusted based on the model or assumptions. The critical value is derived from statistical tables and varies based on the degrees of freedom in the data.

Residuals represent the errors or deviations from the expected values in your model. They provide insight into how well the model explains the data. The critical value used in the Chi-Square Test is obtained from a distribution table that accounts for degrees of freedom, which is essentially a measure of the number of independent pieces of information available in the data. This allows you to make informed decisions about the accuracy and reliability of the adjustment you've made.

Think of residuals as the feedback you receive after giving a presentation. If you planned to hit key points but missed some entirely, the comments from your audience (residuals) will highlight these gaps. The critical value is like a benchmark or threshold you use to decide if the feedback is significant enough to warrant changing your presentation style in the future.

--