Thanks Shandy for your question!  I believe you and I met at the Boulder Summit last year, I hope you're doing well.  

StatPREP does not directly tackle topics related to the ethical aspects of data science such as data interrogation.  We do, however, focus on a more broad concept of using real data in introductory statistics.  We encourage faculty to use larger real data sets instead of fabricated small data sets in their classroom. 

This would be a great topic for our MAA Connect board!  If you're interested please join our StatPREP Hub Community at MAA connect (https://connect.maa.org/home)! For assistance doing so, try looking at our getting started YouTube video https://www.youtube.com/watch?v=Pgt4cf-eB3c or document https://drive.google.com/file/d/1Cy67ziOwS4D5XpZDiGmPsP0su6Yf6Ny0/view

Thanks for watching our video Leigh.  I agree, it's a great question!  I think that with so much data being released in the media especially now, it's important for our students to think about how data is being represented.  

Yes, the ethics and legal dimensions of the data we collect and how it may be used are important to grapple with in our liberty-loving society. We are seeing in real-time in this epidemic how various leaders use or refuse the meaning of data and models based on them.  We recently submitted a paper on legal and ethical aspects related to an analytical technique used for SARS-CoV-2 that may predict COVID-19 in a second wave. Here's part of the abstract:  "Scientists have observed that molecular markers for COVID-19 can be detected in wastewater of infected communities both during an outbreak and, in some cases, before the first case is confirmed.... This scientific breakthrough may lead to many useful potential applications for tracking disease, intensifying testing, initiating social distancing or quarantines, and even lifting restrictions once a cessation of infection is detected and confirmed. ... Even if reliability and efficacy are established, limits on sample and data collection, use, and sharing, must also be considered to prevent undermining privacy and autonomy in order to implement these public health strategies consistent with legal and ethical considerations."

Thank you Lisa for watching the video! With Statistics especially continued support is important as the vast majority of community college statistics instructors do not have a background in statistics but rather in math.    

Thank you for this insight.  While I agree with the notion that substantially the conversation is less about specifics, I would argue that it is a way to draw students into the world of statistics and to help understand the importance of interpreting data accurately/science communication.  In far too many circumstances we have seen gross misinterpretations of data to drive socioeconomic and sociopolitical motives.  A small unit embedded into a course may provide additional context for students to see the relevance of statistics.

Stephen, Thank you for your question. StatPREP has presented in six different metropolitan areas to teachers at many different institutions. We haven't looked into the gain of performance of students since as PIs we do not have access to that information at all the institutions. We also are not sure how the teachers are using the materials in the classes. We ask that the participants use at least one activity in a semester, so looking at student gain may be hard to assess without more control. But that is the hope that there will be student gain in knowledge.

Since each participant can decide how they will utilize the material, having students create their own data sets and integrating them into their other STEM disciplines is beyond our control. However, I will suggest that instructors may want to try that. We have created Little Apps for the instructors to use in their classes and the new generation of these Little Apps have the capability for the students to upload their own datasets. It would be easy for instructors to encourage this. 

Guest lecturers may not be feasible, since everyone is at a different institution, but that is an interesting idea. Maybe the PIs can be involved in this. 

Thank you for the great ideas. 

Kathryn: Thanks so much to you and your team for helping instructors “make statistics about data” and Stephen: Thanks for your question about student outcomes.

I hope it is ok if I share some findings on this front. We have collected data on students from a wide range of academic settings whose instructors have made data projects central to their courses. That is, students choose their own research question and teaching happens in the service of helping them to answer those questions. Results have shown that a data-driven approach 1) attracts more under-represented students compared to a traditional math statistics course; 2) is associated with students being more likely to report increased confidence in working with data and increased interest in pursuing advanced statistics coursework compared to a traditional statistics course and 3) is associated with students actually taking an additional data oriented course in the future (not just saying they are interested) compared to both a traditional statistics course and a statistics course taught through a psychology department. For these and related findings see: https://passiondrivenstatistics.com/publications/

From the StatPREP video, the importance of supporting teachers to make this change in their classrooms is clear. StatPREP is not just inspiring teachers, but also providing the kind of real support that allows them to make a data-driven curriculum happen. Warmest wishes for your continued success!

Sarah, thank you for watching our video. The goal of StatPREP is to encourage teachers of introductory statistics to integrate data into their classes. Especially data that we see now with multiple variables. We don't have specific activities for explaining to students randomness and noise in the data, but each Little App that we have developed allow the user to take a new sample with a push of a button. Students are exposed to randomness and noise with each push of this button. Even though we don't have an activity that focuses on this topic, each activity exposes students to this concept.

As for your ideas of Open Science and research transparency, can you tell me more? These may be beyond the scope of the introductory statistics course, but I am not familiar with these terms. I would like to know more about both. Once I have a better understanding of these terms I can let you know if I think these are topics to discuss in an introductory statistics course. 

Thank you again.

I really like the Little Apps because of the low barrier for student access and engagement. 

I personally do not yet use R in my classes, but probably will in the future. I do not use calculators. I've used textbooks that have some decent pedagogical technology directly linked to the text - but that is changing as I'm exploring OERs. 

Some feedback I've personally received from colleagues in different departments tells me that many different programs are used across our disciplines. I'm interested in what programs are being used in different disciplines and in different settings (e.g. K-12) within the STEMforall community. 

Hi Wendy! AP teachers weren't initially part of the target group, but I can see them using the material from our website http://statprep.org/ , webinars, and applets easily. In fact, there's a nice alignment of our Little Apps to the free Open Intro textbook, which is similar to the AP-aligned statistics book that I teach with at Portland CC. 

Thanks for viewing, Feng! Those are great questions! One measure of success is the return rate of instructors the next year, and their level of excitement and use of the apps. We hold monthly webinars and use MAA connect to support the faculty during the rest of the year. There haven't been any formal assessment of the outcomes of the usage of the apps, although that could be a direction that some faculty could chose to tackle in the future. I am one of the faculty participants in StatPREP, so I'm hoping that one of the named grant writers could help better answer your question. :)

I agree! Getting to student level impacts is an important direction for both evaluation and research. 

I am also a faculty participant and I'd like to share a bit from a Scholarship of Teaching and Learning (SoTL) project I completed in relation to my StatPREP experience. The SoTL project was supported by my institution as part of the Anoka-Ramsey CC Scholars Program. 

My SoTL inquiry was a teaching experiment focused on a two week learning unit on exploratory data analysis and descriptive statistics (think - beginning of semester stuff). For the unit, I used some NHANES data for instruction, class activities and assignments / assessments. I used a pre-post design and assessed student learning with the Comprehensive Assessment of Outcomes in Statistics (CAOS) tool (delMas, Garfield, Ooms & Chance, 2007), and assessed student confidence with a knowledge survey that I created (Nuhfer & Knipp, 2003). 

Students (n = 28) took the CAOS assessment in the week prior to learning unit as a pretest and 12 weeks following the unit as a posttest. The Knowledge survey was completed on the first and last days of the unit. 

Both learning gains and increases in student confidence were demonstrated. Although the learning gains were mixed and modest (I looked at the item level focusing on CAOS items whose learning descriptors aligned with the unit) - but I was excited to see them sustained over a 12 week period. 

I was able to share the outcomes of this project at the 2019 JMM meetings. 

Thanks for the response, Jennifer. I agree that you can look at the return rate of instructors and some other outcomes as you mentioned (e.g., level of excitement assuming you have a reliable and valid instrument to measure that) to measure the impact of StatPREP program. 

Thanks for the detailed explanation on the evaluation of student level impact, Megan. As you described, a within-group pre-post design would give you the ability to interpret a correlational relationship between a treatment/intervention under investigation (e.g., StatPREP program) and the outcomes to be affected. If feasible, adding a control group (if not a randomized control trial) with/without matching would make it a stronger study. This applies to the teacher-level impact evaluation as Jennifer describe as well.

Hi Jeffrey!

To add to what Megan said, here's our webinar from last month where we run through some of the Little Apps. 

Here's a description of the webinar from our website:

The third anniversary of the StatPREP Little Apps is approaching. To celebrate, we’re launching a brand new set of apps with a new look and new functionality. Among the changes are a dramatically increased choice of data sets (including those keyed to textbooks often used by StatPREP instructors) and the ability to upload your own CSV files, a feature that lets you freeze a display and show it side-by-side with the current display. The original apps will continue to be available at their current web addresses, but we encourage you to switch to the new ones and try out their new capabilities.

Maisha,

Thank you for your question. StatPREP offers workshops in the summer. This summer workshops were cancelled, but we will be offering some next summer. One will be in the New York City area. If your faculty can't come to a workshop, we offer webinars almost every moth. There are recordings of previous webinars on our website, statprep.org. There are also links to the Little Apps and activities to the Little Apps under resources on our statprep.org site. The activities are for teachers to use in their classes for student group work, but teachers can also use the material to learn about data. Also under resources are some tutorials for both students and instructors to learn data techniques. Please let me know if you need anything else.