Using SSD for R in the Classroom: Analyzing Baseline Data

Well, the summer semester is officially more than halfway over here in steamy NYC!

What does that mean for students in our Practice Research and Evaluation courses?  It means that our students are hard at work on their midterm papers describing their baselines from their own projects. To continue learning about this, we taught our students how to use several functions in SSD for R to visually and statistically analyze their data.

We showed the students how to produce and interpret one and two standard deviation band graphs in order to assess the stability of their baseline data and look for outliers.  Then we taught the students how to evaluate their baselines for two problems that could impact how they compare their baseline and intervention phases, which we will begin focusing on next week.  The first problem we discussed was trending while the second was autocorrelation.

Here's an example of a graph that our students were able to produce for testing for a trend by using the Aregres() function:

Autocorrelation is, perhaps, the most difficult concept we teach in this course.  Most of our students have never heard of the notion of serial dependence before, and it is a fairly complicated concept to teach.  The other issue with teaching students about autocorrelation is that it is not detectable visually.  Our students MUST learn to interpret some statistical output!

To help our students, we have several videos posted on our website and YouTube on all of these topics.

Their mid-term papers are due next week, and we're pretty sure that our students are going to do really well with these!

Using SSD for R in the Classroom: Analyzing the Baseline

We just finished Week 3 of our 7 week summer semester here in sunny (finally) NYC, and we are spending more and more time in Practice Research and Evaluation course using SSD for R.

Now that our students know how to navigate SSD for R, build input files, and create basic graphs, we have moved on to teaching the students how to analyze their baseline data.  We have taught them how to create one- and two-standard deviation band graphs to help them identify "typical" behavior and to look for outliers.  We have shown them how to retrieve a host of descriptive statistics (sample size, mean, trimmed mean, standard deviation, coefficient of variation, quantiles) and produce a box plot depicting the variation in their data.

Today, we began talking about analyzing baseline data visually for trends and then to calculate a regression line to determine if any trend they may see could be statistically significant.  One of the classes also began to talk about that niggling little problem of autocorrelation, which needs to be evaluated statistically since there is no way to visually determine if autocorrelation is problematic.

All this, of course, is a precursor to teaching the students how compare their baseline and intervention phases for their final projects.

As we are working through these lessons, we are referring the students to drafted chapters of the manuscript for our upcoming book.  We are hoping that student feedback will help us make this book as useful and user-friendly as possible.

For more information on SSD for R, visit our website or email us!

Using SSD for R in the Classroom: Unintentional Visual Analysis

Well, our Social Work Practice and Evaluation Research class just completed its second week.  So far, we have been devoting the first part of each class to teaching the students the fundamentals of SSD for R which we think has been extremely helpful.  The students clearly seem more comfortable navigating RStudio than they did last week, and they seem a bit more computer savvy, in general.

So far, the goals for using SSD for R have simply been pragmatic - learning how to use the various functions.  During our first class this week, we entered intervention data into our Excel spreadsheets and then imported all of this into SSD for R.  The lesson then focused on creating a simple line graph that showed multiple phases (both baseline and intervention).  We showed the students how to add a vertical line between phases and then label each phase. We did this for each of the two behaviors that we were tracking for our hypothetical client.  We also showed them a handy-dandy searchable PDF that is located on the ssdanalysis.com website  and is useful for identifying each of the functions based on what is trying to be created.

Today, the second class for the week, we continued the lesson based upon a student's request - he wanted to know how to view two graphs simultaneously.  This was a great jumping off point for a discussion about the impact of an intervention on multiple indicators, so this was our hands-on lesson for today.  We showed the students how to use the plotnum() and ABplotm() functions to build multiple graphs in one window.  We then practiced using data from our hypothetical client.

Here's what the students produced:

This was SUPPOSED to be the end of the lesson for today; however, it wasn't.  One student observed how the behaviors moved in tandem, particularly during the baseline.  Then, another student noted how variable the data was during the baseline and how could you make comparisons between the baseline and the intervention, if the baseline data weren't stable - what was it actually representing?  Good point!

While we couldn't address this issue entirely today, this provides a great entrée into subsequent lessons on the baseline.

For more information on SSD for R, visit our website or e-mail us.

Using SSD for R in the classroom

Here at Wurzweiler School of Social Work, we have a summer program designed for long-distance students.  Some live in the US, but many come from other countries.  Our students spend seven weeks living and studying for their MSW degree in New York City over the course of three consecutive summers.

During their third summer, our students are required to take a course in Social Work Practice and Evaluation Research.  This class teaches students to empirically evaluate their clients' progress so the students' can adapt their work according to the needs of each client.  During the course of the semester, students are taught how to do this regardless of whether their clients are individuals, couples, families, groups, or communities.This semester, we will be using SSD for R in this class for the students' projects, which will be a semester-long evaluation of one of their clients.

We thought it would be interesting to use this blog to share with you how we are using SSD for R in this course and what our experiences are.

We just finished the second of fourteen classes of the semester and we are already using SSD for R! During the first class, we had the students become oriented to the package by opening RStudio and giving them a "tour" of the interface.  We explained how to install SSD for R (only once) and how to require it (every session).  We also encouraged all the students to follow the instructions on our website to download R, RStudio, and SSD for R on their own computers at home.  We showed them how to view our tutorial videos on our website, in case they needed extra help.

I can't tell you how delighted we were this morning when we found out that all the students who attempted to install everything, did so successfully!  This is great because, like many social workers, many of our students are a little reticent about computer usage, in general.  I think this bodes well for the semester, and we are hoping that the students feel that they can use the skills we teach them in class at their own agencies.

Today's lesson was also interesting. We started out by creating a fictional client whose problems we plan to measure.  We showed the students how to create an Excel file to track client data and save it as a .csv file.  We then opened the file in SSD for R and created simple line graphs for each of the behaviors we were tracking. A couple of students had minor difficulties.  These included people forgetting to attach the file they had opened or in mistakes in entering the command for the simple line graphs.  However, between the instructor and our trusty teaching assistant, each student was able to create two different line graphs.  We concluded this lesson by showing the students how to cut and paste their graphs into Word.

We are already hearing sighs of relief.

For more information about SSD for R, visit our website.  You can also reach us by e-mail.