No it doesnt. The software will randomly select one of the level within the variable to be the reference group. You will have better control of what the reference group is if you transform it into a dummy variable.

@Lim : Not randomly. At least for any reasonable software. You just have to read the doc to know how it choose it. And change it if necessary. "Manual" coding is most often a waste of time with most softwares, except in some special cases.

@Mustafa Cansiz: it strongly depends on 1) the software you will use and 2) the question you're asking.

Yes, assuming you are using SPSS for multiple regression analysis, you will need to create dummy variables for nominal data. But when converting ordinal data (into dummy variables) you have to be very careful as it might distort your results. So, it depends on the type of analysis you do.

Here's a discussion about this in researchgate before,

https://www.researchgate.net/post/Dummy_variables_creation_for_a_regression_model

Hope this helps.

If you assign integer values to your ordinal categories (1, 2, 3,...) you are imposing the assumption that they are equally spaced. If you assign a dummy variable to each one, you will lose the constraint of their order (but allow the data to tell you whether it was a good assumption.) I prefer the second option.