Understanding the GLM output

[KaydenSawyer]

Not sure if this is the right sub for this, so let me know if I should crosspost elsewhere.

I'm working on an analysis where I'm looking at a binary variable and comparing it to multiple ordinal/continuous variables. I used a series of ANOVAs using the proc glm command. In the output for each comparison, I have 3 results tables: Type I SS. Type III SS, and Least Squares Means. The first two tables give me the degrees of freedom, Type X SS, mean square, f value, and p value. The Least Squares Means table just has the lsmean and p value.

I'm trying to figure out which of these I want to use to report my results (or a combination?). I figured out the difference between the Type I SS and Type III SS and I'm pretty confident I want to use the Type III because I have a couple variables I'm trying to control for. What I'm not sure is if I want to use the Type III SS table or the Least Squares Means table, or both? I'm having trouble figuring out exactly what Least Squares Means is and why that table gives me so much less data. Any help would be appreciated.

[PaigeMiller]

You can use both the Type III hypothesis tests/sum-of-squares and the Least Squares Means table. They are all pieces that help you understand the big picture.

The Least Squares Means (in layman's terms) is the mean at each level of your independent variables, after removing the effect of the other variables. When the design is balanced ... in other words the exact same number of observations in each cell of the design ... then the means equal the least squares means. But if there is imbalance in the number of observations in each cell, then you need the least squares means which take into account the effects of each variable and mathematically balances the unequal number of observations in each cell.

[PGStats]

It is not clear from your statement which variable is the dependent (or outcome) variable in your model. Giving us the proc code would help clear things up.