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
- /
- Re: I have FIML estmates for my missing data... now what?

Options

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

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

Posted 10-16-2019 05:52 PM
(364 views)

Hello! First post here, and I am super new to SAS, so please bear with me!

I am working on addressing some missing data in this data set I have. I decided to use the Full Information Estimation Likelihood (FIML) method to estimate the mean values of my missing data. I used the following code to yield the estimates:

proc calis data=[name of data set] method=fiml;

mstruct var=[listed pertinent variables here];

run;

Now that I have all the estimates I need for the missing data, I am unsure how to proceed with integrating them with the rest of my data. Is it as simple as using a series of "if then" statements (e.g., if [variable 1] = . then [variable 1] = [estimated mean]) to plug in the missing values? Or is it something more complex than that?

Thank you for your consideration!

4 REPLIES 4

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

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

Hello!

Thank you so much for sending those resources! I have reviewed them, but unfortunately they don't quite address my confusion. Previously, I had ran some analyses to calculate effect sizes from one set of variables to the other, but we later decided that we wanted to impute missing values rather than use listwise deletion. I have successfully run the proc calis code (with the FIML option), and I have output that tells me the mean and co-variance structures of the estimation, such as the estimate of the variables, their standard errors, t values, and p values. I am just confused on what to do next. Do I just re-run the previous effect size analyses, and will that take into account the estimated values I just got from my proc calis code? Or do I have to do something else to ensure the estimated values are accounted for in my effect size calculation?

I apologize if my description of the problem is confusing. I am doing the best I can with the knowledge I have. Thank you again for considering!

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

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

What is the final goal of your analysis? Are you trying to perform some regression? Estimate variance? Impute the data to construct a predictive model?

Briefly,

1. if you are interested in inferential statistics such as standard errors, confidence intervals, and p-values, then use PROC MI and PROC MIANALYZE to perform multiple imputations.

2. If your data are from a statistical survey, use PROC SURVEYIMPUTE instead.

3. If you are interested in predicted values for machine learning/data mining applications, some people use mean imputation.

Secure your spot at the must-attend AI and analytics event of 2024: SAS Innovate 2024! Get ready for a jam-packed agenda featuring workshops, super demos, breakout sessions, roundtables, inspiring keynotes and incredible networking events.

Register by March 1 to snag the Early Bird rate of just $695! Don't miss out on this exclusive offer.

** **

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.