turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- BI
- /
- Enterprise Guide
- /
- Understanding what drives a variable in Survey dat...

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-04-2014 10:29 PM

Hello,

I have some survey data which includes an effort score (interger 0-5). I'd like to understand from all other information in the survey data set what other factors (any of the variables) influence that score. I'm after advice on both:

- The correct statistical procedure here to use; and
- guidance on how to do this in SAS EG (including how to account for survey weights)

could you please recommend concise resources that would help me understand how to beat do this, and interpret the results?

thank you,

silvertreetops

Accepted Solutions

Solution

12-05-2014
11:02 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-05-2014 11:02 AM

I'll make an attempt at an answer, but first I'm required to utter this disclaimer: "I am not a statistician".

With that out of the way, I can say that I've *seen* people use Factor Analysis (PROC FACTOR) or Principle Component Analysis (PROC PRINCOMP) to achieve what you've described. Both of these are tasks in SAS Enterprise Guide. Here are descriptions from the online help:

**Factor analysis** performs a variety of common factor and component analyses and rotations. Input can be multivariate data, a correlation matrix, a covariance matrix, a factor pattern, or a matrix of scoring coefficients.

You might want to use this task to perform a common factor analysis on an annual employee review. For example, suppose that 103 police officers were rated by their supervisors on 14 scales (variables). You can conduct a common factor analysis on these variables to see what latent factors are operating behind these ratings.

**Principal component analysis** is a multivariate technique for examining relationships among several quantitative variables. You should use principal component analysis if you are interested in summarizing data and detecting linear relationships.

You might use the Principal Components task when you have too many variables to plot simultaneously. For example, suppose that you have the crime rates per 100,000 people in 7 categories for each of the 50 states in 1977. Since there are seven numeric variables, it is impossible to plot all the variables simultaneously. You can use principal component analysis to summarize the data in two or three dimensions and to help you visualize the data.

Both tasks are found in the **Tasks->Multivariate** menu.

Chris

All Replies

Solution

12-05-2014
11:02 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-05-2014 11:02 AM

I'll make an attempt at an answer, but first I'm required to utter this disclaimer: "I am not a statistician".

With that out of the way, I can say that I've *seen* people use Factor Analysis (PROC FACTOR) or Principle Component Analysis (PROC PRINCOMP) to achieve what you've described. Both of these are tasks in SAS Enterprise Guide. Here are descriptions from the online help:

**Factor analysis** performs a variety of common factor and component analyses and rotations. Input can be multivariate data, a correlation matrix, a covariance matrix, a factor pattern, or a matrix of scoring coefficients.

You might want to use this task to perform a common factor analysis on an annual employee review. For example, suppose that 103 police officers were rated by their supervisors on 14 scales (variables). You can conduct a common factor analysis on these variables to see what latent factors are operating behind these ratings.

**Principal component analysis** is a multivariate technique for examining relationships among several quantitative variables. You should use principal component analysis if you are interested in summarizing data and detecting linear relationships.

You might use the Principal Components task when you have too many variables to plot simultaneously. For example, suppose that you have the crime rates per 100,000 people in 7 categories for each of the 50 states in 1977. Since there are seven numeric variables, it is impossible to plot all the variables simultaneously. You can use principal component analysis to summarize the data in two or three dimensions and to help you visualize the data.

Both tasks are found in the **Tasks->Multivariate** menu.

Chris

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-10-2014 03:03 PM

Thanks Chris,

Apologies for the slow reply - I have a look at those, they look helpful