This site is built around the idea that the most important use of statistics is to support scientific reasoning. If this is the goal then statistics should not be taught simply as a menu of procedures to use ritualistically. Statistical procedures, or the software in which they are implemented, are not able to do the thinking required of good science.

Similarly, there are different ways to do statistics. Some ways can get you results quickly, but you might not really understand what you did. Other ways may require more time and effort, but also promote a better understanding. The biggest barrier and least appealing aspect of statistics for most people is the math. Statistics is math, so there is really no way to avoid it. You have to know the math to understand statistics. But let’s be clear on what we mean by math. Math is more than computational procedures. Fundamentally, math is formal reasoning, and it is this aspect of math that is most important to understand if you want to master statistics. We can get a computer to do the computational aspects of math for us, but, at least for now, computers will not replace good reasoning. But, a good statistical programming language can allow us to pass off the tedious computations to the computer while still being very close to the reasoning needed to understand statistical methods.

This site demonstrates quantitative methods commonly used in the social sciences, using statistical programming. There are two primary goals for developing these demonstrations.

**Learning to code**- First, to help you learn to use a computer to do simple and complex mathematical and statistical computations. That means programming with statistical software. On this site I use R, SPSS, and Mplus.**Coding to learn**- Second, using your statistical programming skills to build an intuitive understanding of theoretical knowledge emphasized in most formal statistical training.

Speaking of formal statistical training, this site is aimed at supplementing – but not replacing – the more traditional modes of learning statistics. Statistical concepts can be complex and are not always intuitive. So, mastering the basics of statistics requires active practice using the concepts and procedures taught in the classroom. Indeed, one reason for creating this site it to use as a resource supplementing my own teaching of statistics. So, while lectures are helpful, they are just the beginning. We have to apply what we learn to become proficient with statistical methods. With statistics, like many things in life, we learn by doing.

This is the third tutorial in a series that demonstrates how to us full information maximum likelihood (FIML) estimation using the R package lavaan. In this post, I demonstrate two methods of using auxiliary variable in a regression model with FIML. I am using data and examples from Craig Ender’s website Applied Missing Data. The purpose of these posts is to make the examples on Craig’s website, which uses Mplus, available to those who prefer to use lavaan

When I have to prepare data for Mplus, I use the MplusAutomation package in R. Its great! I import the SPSS data file into R with the foreign package. Then I use the prepareMplusData() function to create a .dat file for use in Mplus. This function also creates basic Mplus code that can pasted into Mplus or a text file used to prepare Mplus code files. MplusAutomation has many other great features and I highly recommend it for those who use Mplus and R.

This tutorial demonstrates how to use full information maximum likelihood (FIML) estimation to deal with missing data in a regression model using lavaan.
Import Data In this post I use FIML to deal with missing data in a multiple regression framework. First, I import the data from a text file named ‘employee.dat’. You can download a zip file of the data from Applied Missing Data website. I also have a github page for these examples here.

FIML for Missing Data in Lavaan Full information maximum likelihood (FIML) is a modern statistical technique for handling missing data. If you are not familiar with FIML, I would recommend the book entitled Applied Missing Data Analysis by Craig Enders. The book is both thorough and accessible, and a good place to start for those not familiar with the ins and outs of modern missing data techniques.
The purpose of the FIML in Lavaan series of posts and the related git repository is to take some of the examples related to FIML estimation within a regression framework from the Applied Missing Data website, and translate them into code for the R package lavaan.

If you are a visitor to this site who is not taking one of my courses, you can use the site as you would any other site you find on the internet. If you are currently, or have previously enrolled in one of my courses, and want to find information relevant to that course, use the Courses tab to find your course, click on it to see the related posts.

Select a course for further information.

.courseERMA 8340 page

Measurement Error

Removing Barriers

ERMA 7300 page

ERMA 7310 page

ERMA 7310 page

Current and future research projects.

.valenciaERMA 8340 page

Measurement Error

Removing Barriers

ERMA 7300 page

ERMA 7310 page

ERMA 7310 page

This is a list of courses I teach( or plan to teach) at Auburn University and that are relevant for this site:

- ERMA 7200 Basic Methods in Educational Resarch
- ERMA 7300 Design and Analysis in Education I
- ERMA 7310 Design and Analysis in Education II
- ERMA 8320 Design and Analysis in Education III
- ERMA 8340 Introduction to Structural Equation Modeling

- wmm0017@auburn.edu
- 334 844 3806
- 4064 Haley Center Auburn AL, 36489, USA