The Analysis Institute - Making Statistics Make Sense

Using AMOS 16.0 to Run Regression Models with Missing Data

The effects of missing data on statistical results range from slight loss of power to devastating bias that renders results compeletely inaccurate.

And many of the traditional solutions to missing data create more problems than they solve.

But Maximum Likelihood approaches to Missing Data give unbiased results with no loss of power in many circumstances.

Maximum Likelihood is quite easy to implement--it's much easier than Multiple Imputation, which can take weeks to do well. You may already be using it for your analysis.

But unfortunately, it's only possible for certain kinds of regression models and sometimes requires specialized software.

One great, and easy-to-use software package that implements Full Information Maximum Likelihood is AMOS 16.0. AMOS is actually a Structural Equation modeling software package, but can be used to run regression models.

One really nice thing about AMOS is that it reads SPSS files and comes bundled with many versions of SPSS, including the GradPak.

If you are a researchers who uses SPSS (or any other package--it's not necessary to use SPSS) who needs to run a regression model on a data set with more than 5% missing data, this product is for you. You get two items:

Video: Using AMOS 16.0 to Run Regression Models with Missing Data

The video will show you step-by-step how to implement Full Information Maximum Likelihood (ML) approaches to Missing Data for Linear Regression models in AMOS 16.

The video is great for seeing the steps actually implemented--seeing where each button is, exactly which boxes to check, and an annotated explanation of why you need to do various steps.

Along with the video, you also get the 12-page ebook: How to Use Full Information Maximum Likelihood in AMOS to Analyze Regression Models with Missing Data

The ebook cuts it down to the essentials. Once you've gotten your bearings watching the video, the ebook will take you through each of the steps, with screenshots. Use it as a checklist to make sure you've done each step correctly.

The information in both the video and ebook also apply to version 4.0 of AMOS. There are slight differences in the menus, which are discussed in the video.

the video & ebook for $19.00

If you are new to using Maximum Likelihood for missing data, I suggest you purchase this in combination with the Maximum Likelihood Approaches to Missing Data that Eliminate Bias, Retain Power, and Produce Accurate P-Values video. This workshop excerpt explains all the maximum likelihood approaches to missing data, their assumptions, and when each one works.

both products for $27.00

The effects of missing data on statistical results range from slight loss of power to devastating bias that renders results compeletely inaccurate.

And many of the traditional solutions to missing data create more problems than they solve.

But Maximum Likelihood approaches to Missing Data give unbiased results with no loss of power in many circumstances.

Maximum Likelihood is quite easy to implement--it's much easier than Multiple Imputation, which can take weeks to do well. You may already be using it for your analysis.

But unfortunately, it's only possible for certain kinds of regression models and sometimes requires specialized software.

One great, and easy-to-use software package that implements Full Information Maximum Likelihood is AMOS 16.0. AMOS is actually a Structural Equation modeling software package, but can be used to run regression models.

One really nice thing about AMOS is that it reads SPSS files and comes bundled with many versions of SPSS, including the GradPak.

If you are a researchers who uses SPSS (or any other package--it's not necessary to use SPSS) who needs to run a regression model on a data set with more than 5% missing data, this product is for you. You get two items:

Video: Using AMOS 16.0 to Run Regression Models with Missing Data

The video will show you step-by-step how to implement Full Information Maximum Likelihood (ML) approaches to Missing Data for Linear Regression models in AMOS 16.

The video is great for seeing the steps actually implemented--seeing where each button is, exactly which boxes to check, and an annotated explanation of why you need to do various steps.

Along with the video, you also get the 12-page ebook: How to Use Full Information Maximum Likelihood in AMOS to Analyze Regression Models with Missing Data

The ebook cuts it down to the essentials. Once you've gotten your bearings watching the video, the ebook will take you through each of the steps, with screenshots. Use it as a checklist to make sure you've done each step correctly.

The information in both the video and ebook also apply to version 4.0 of AMOS. There are slight differences in the menus, which are discussed in the video.

the video & ebook for $19.00

If you are new to using Maximum Likelihood for missing data, I suggest you purchase this in combination with the Maximum Likelihood Approaches to Missing Data that Eliminate Bias, Retain Power, and Produce Accurate P-Values video. This workshop excerpt explains all the maximum likelihood approaches to missing data, their assumptions, and when each one works.