topic Re: Statistics Methods in Statistical Procedures

Statistics Methods

collegestudent2 — Tue, 30 Nov 2021 16:04:57 GMT

I am wanting to perform an analysis on census information. I want to analyze the data using SAS to determine if there is a relationship between college education and income in the southern region of the united states. My data consists of the percent of the population with a bachelor's degree and the median annual income.

I believe the best procedures to use are bivariate analysis, specifically scatterplot, correlation, and linear regression, but is there any advice or suggestions for a better analysis method? Would ANOVA work?

Re: Statistics Methods

PaigeMiller — Tue, 30 Nov 2021 16:13:20 GMT

This doesn't seem like an ANOVA to me. ANOVA requires a categorical independent variable, you don't have that.

Re: Statistics Methods

sbxkoenk — Tue, 30 Nov 2021 17:38:02 GMT

Hello,

I would also consider a PROC LOESS for local regression and Scatter Plot Smoothing Using LOESS. The relationship might be non-linear.

Or fit a spline (PROC TRANSREG or EFFECT statement in a regression procedure).

Good luck,

Koen

Re: Statistics Methods

StatDave — Wed, 01 Dec 2021 20:47:01 GMT

If it is census data, that suggests it is likely to be data drawn using sample survey methods. To get proper tests for survey data, you should only use one of the procedures with SURVEY in its name. Then if you want to fit a regression model to predict income as a function of education (assuming that your income measure is approximately normal in distribution or can be transformed to be so), PROC SURVEYREG is probably the one to use. See the discussion and examples in the SURVEYREG documentation.