Turn on suggestions

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

Showing results for

- Home
- /
- Analytics
- /
- Stat Procs
- /
- What is best approach - DV Cont and IV Ordinal

Options

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

🔒 This topic is **solved** and **locked**.
Need further help from the community? Please
sign in and ask a **new** question.

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Posted 05-23-2018 05:37 PM
(1718 views)

Hi,

I have 15 questions and each question has 4 different statuses like 0 – ‘no action’ 1 – ‘started ‘2- ‘implemented’ and 3 – ‘completed’. I'd like a good way to measure the impact of each question on our in visitors (number of visitors), and if possible apply regression.

- It is possible to use dummy-encoding to the ordinal variable and use regression ( 3 INV variable – 1,2,3)
- It is possible to output each dummies Standardized Coefficients - Beta
- Also get from R-squared how different independent variables (dummies) are contributing to a dependent variable ( number visiting)

Please let me know if the above makes sense

Thanks in advance.

1 ACCEPTED SOLUTION

Accepted Solutions

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Thank you both ..

8 REPLIES 8

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Procedure intending to use what?

Many of the modeling / regression procedures will allow specifying categorical variables as Class variables and handle creating/using "dummy variables" in the background.

Output options tend to depend on which specific procedure is used.

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Thanks for the reply.

The intention is - which status (driver) are contributing more on the outcome ( i.e Driver 4 - Completed).

In that case we were thinking to show each RSQR and Beta for each status.

Thanks again.

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Is it ok to model /regression

by question;

Model Visit = Status1 (0,1) Status2 Status3 / STB output each R..

In ordinal variable

Please let me know

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

SAS has over 20 different regression procedures.

Choice relates to type of regression wanted, data types involved and possibly sample design (surveyreg and surveylogistic.

The model statements have different options depending on specific procedure chosen.

What do you mean by Status1 (0,1) in that model statement?

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

The status are ordinal VR :

Status like 0 – ‘no action’ 1 – ‘started ‘2- ‘implemented’ and 3 – ‘completed’..

All the 16 groups (participant) - They have to choose one of the scores for each question (total = 15 questions)

I was thinking to convert the ordinal data to dummy (0, 1) - But not sure.? I see before where categorical is coded 0 and 1.

Ideally would like to get an idea which group (group 1 to group 16) has impact on the continuous outcome (visit)

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

@Diimoo wrote:

The status are ordinal VR :

Status like 0 – ‘no action’ 1 – ‘started ‘2- ‘implemented’ and 3 – ‘completed’..

All the 16 groups (participant) - They have to choose one of the scores for each question (total = 15 questions)

I was thinking to convert the ordinal data to dummy (0, 1) - But not sure.? I see before where categorical is coded 0 and 1.

Ideally would like to get an idea which group (group 1 to group 16) has impact on the continuous outcome (visit)

As I explained in my above post, there is NO need to create dummy variables here. SAS will do the work for you.

--

Paige Miller

Paige Miller

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

@Diimoo wrote:

Hi,

I have 15 questions and each question has 4 different statuses like 0 – ‘no action’ 1 – ‘started ‘2- ‘implemented’ and 3 – ‘completed’. I'd like a good way to measure the impact of each question on our in visitors (number of visitors), and if possible apply regression.

- It is possible to use dummy-encoding to the ordinal variable and use regression ( 3 INV variable – 1,2,3)

If I am understanding properly, there's no need for YOU to create dummy variables here. In most modeling procedure in SAS, such as PROC GLM or PROC PLS, you can take the variable (let's call it by the name STATUS) with the 0,1,2,3 value and place it is a CLASS statement, and then SAS will create the dummies for you internally, so you don't have to.

In this case, your model becomes (because you have 15 different status variables)

CLASS STATUS1-STATUS15;

MODEL VISITS = STATUS1-STATUS15/options ... ;

It is possible to output each dummies Standardized Coefficients - Beta

Yes, you can get coefficients for each level of STATUS.

Also get from R-squared how different independent variables (dummies) are contributing to a dependent variable ( number visiting)

There is NO SUCH THING as an R-squared for individual terms in the model; the only R-squared number you can get is for the entire model fit.

--

Paige Miller

Paige Miller

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Thank you both ..

Build your skills. Make connections. Enjoy creative freedom. Maybe change the world. **Registration is now open through August 30th**. Visit the SAS Hackathon homepage.

What is ANOVA?

ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.

Find more tutorials on the SAS Users YouTube channel.