Unfortunately, Microsoft Excel 2003 does not have the capability to run regressions on non-numerical data. Regression analysis requires numerical data in order to calculate the relationship between variables. However, there are a few workarounds you can try:

1. Convert non-numerical data to numerical data: If you have categorical data such as gender or race, you can assign numerical values to each category. For example, you can assign 1 to male and 2 to female, or 1 to white, 2 to black, and so on. This will allow you to run a regression analysis, but keep in mind that the numerical values you assign do not have any inherent meaning.
   
2. Use a different statistical softwa There are many statistical software packages available that can handle non-numerical data, such as R or SAS. These programs have more advanced regression analysis capabilities and can handle a wider range of data types.
   
3. Consider a different type of analysis: If you cannot convert your non-numerical data to numerical data, you may need to consider a different type of analysis that is more appropriate for your data. For example, you could use a chi-squared test to analyze the relationship between two categorical variables.