Contact and Bio

Jeffrey L. Harding, Ph.D.
jeffrey.l.harding@gmail.com
PH: 615.713.0473
LinkedIn

A public-school classroom teacher turned researcher and creator. Accountability Research Specialist at Georgia Department of Education analyzing and reporting on millions of administrative student- and school-level data points. Former Senior Director of Cognia Innovation Lab. Seasoned presenter, facilitator, and trainer of internal and external constituents. Published author of empirical research.

Introduction

Hello, and thank you for taking the time to review these application materials. Given the nature of this particular position, I felt it might be helpful to present (as an alternative) all of my applications in the dynamic HTML I created using the R Markdown languages. Of course, a standard CV and Cover Letter have been provided as well if you prefer. It is my hope, though, that this document communicates that same passion and demonstrates my qualifications for this position at least as well as the traditional application materials and, if all goes well, elevates those arguments through this interactive format. Please note that all data used in this report come from either publicly available datasets or those included (and available free of charge) in the R and R studio applications.

Résumé

Please use the tabs below to navigate to the relevant parts of my résumé. Again, you will find all of the same information here that you would in the PDF version of my CV in my application. However, you may find this format easier to navigate.

Education

Ph.D., Higher Education (Policy), University of Georgia, Athens, GA
Additional Concentrations in Quantitative Methods and K-12 Policy
Dissertation: Timely Information and STEM Pipeline Outcomes

Committee:

M.Ed., Higher Education Administration, Vanderbilt University
Additional Concentration in Institutional Advancement

B.A., English Middle Tennessee State University
Minors in Secondary Education and Spanish

Current Employment

Senior Research and Data Scientist (Nov. 2021 - Present)
Georgia Department of Education

  • Lead research and analysis projects using a range of methodologies, including econometric, psychometric, and machine learning approaches
  • Develop and oversee projects leveraging AI and ML, including generative AI
  • Query and analyze large scale, administrative datasets using ORACLE/SQL
  • Prepare and present analyses of findings for audiences including state deputy superintendents and myriad district officials, (e.g., superintendents, directors, managers, etc…)
  • Liaise with district and school officials to provide support implementing rigorous data-driven initiatives
  • Suggest and implement research projects related to state agency initiatives

Adjunct Instructor (2021 - Current)
Logan University, Doctor of Education Program
Quantitative Research Methods Course

  • Record and post weekly teaching segments to provide online, asynchronous learners a positive experience with the instructor of record
  • Maintain course learning environment in Canvas LMS
  • Update course materials as needed for students’ asynchronous course engagment
  • Monitor weekly discussion board and facilitate engagement as needed
  • Mentor and advise students in the DHPE program

Employment History

Accountability Research Specialist (2019 - 2021)
Georgia Department of Education

  • Query and analyze large scale, administrative datasets for routine and ad hoc state education reports
  • Liaise with district and school officials to provide support interpreting state accountability measures
  • Advise Director of Accountability regarding implementation of policy changes
  • Provide quality checking of Georgia accountability data and reports
  • Suggest and implement research projects related to state accountability measures and practices
  • Support Assessment department in review of state standardized test forms

Senior Director, Innovation Lab (July 2019 - December 2019)
Cognia (Formerly AdvancED|Measured Progress)

  • Oversee and lead long- and short-term projects
  • Develop new products and services for emerging business opportunities
  • Liaise with internal and external constituents throughout development process
  • Train internal and external stakeholders regarding existing and emerging products and processes
  • Provide strategic guidance and inform prioritization of organization goals in light of competing interests and constraints
  • Identify new opportunities for expanding offerings of services and products
  • Project management for various review and product creation purposes
  • Evaluate schools and school programs for accreditation and certification purposes
  • Serve as lead evaluator on domestic and international school accreditation reviews
  • Contribute to the execution of pilot studies for testing new products and processes
  • Contribute to the research efforts of the organization
  • Analyze and/or review analysis of data for research projects and program evaluation
  • Write, revise, and edit reports for accreditation and STEM certification purposes
  • Assist Chief Solutions Architect with creation and oversight of department budget

Director, Research & Development (2018 - 2019)
Cognia

  • Oversee and lead long- and short-term projects
  • Evaluate schools and school programs for accreditation and certification purposes
  • Liaise with internal and external constituents throughout development process
  • Serve as lead evaluator on domestic and international school accreditation reviews
  • Design and lead pilot studies for testing new products and processes
  • Contribute to the research efforts of the organization
  • Analyze and/or review analysis of data for research projects and program evaluation
  • Seek out and apply for relevant grant opportunities germane to the work of the research division
  • Develop new products and services for emerging business opportunities
  • Assist Senior Vice President of Development with creation and oversight of department budget
  • Identify new opportunities for expanding offerings of services and products
  • Project management for various review and product creation purposes
  • Analyze and present trend data from accreditation and STEM certification processes
  • Write, revise, and edit reports for accreditation and STEM certification purposes

Senior Specialist, Research & Development (2017 - 2018)
Cognia

  • Lead various aspects of long- and short-term projects
  • Serve as lead evaluator on domestic and international school accreditation reviews
  • Design pilot studies for testing new products and processes
  • Develop new products and services for emerging business opportunities
  • Identify new opportunities for expanding offerings of services and products
  • Project management for various review and product creation purposes
  • Evaluate schools for accreditation and certification purposes
  • Create and manage dataset for analysis of certification processes
  • Analyze and present trend data from accreditation and certification processes
  • Write, revise, and edit reports for certification purposes

Specialist, Research & Development (2015 - 2017)
Cognia

  • Develop new products and services for emerging business opportunities
  • Identify new opportunities for expanding offerings of services and products
  • Project management for various review and product creation purposes
  • Evaluate schools for accreditation and certification purposes
  • Create and manage dataset for analysis of certification processes
  • Analyze and present trend data from certification processes
  • Write, revise, and edit reports for certification purposes

Graduate Research Assistant (2012 - 2015)
UGA Institute of Higher Education, University System of Georgia

  • Served as research assistant to the Executive Vice Chancellor and Chief Academic Officer of the University System of Georgia in Atlanta, Georgia and to senior and junior faculty members at the Institute of Higher Education
  • Compile reviews of relevant research
  • Collect and analyze data
  • Consult on special projects and committees as determined by the executive vice-chancellor and other supervising professors

Middle School Classroom Teacher (2004 - 2011)
Rutherford County (TN) School System

  • Create and execute daily lesson plans based on latest research and technology
  • Design, implement, and facilitate various professional development seminars
  • Maintain and update daily records of student attendance and progress
  • Supervise and instruct classes of up to thirty students
  • Collect various fees and maintain/balance bookkeeping records
  • Organize and direct various off-campus field trips
  • Communicate student progress with parents and administration
  • Maintain accurate records per TMSAA athletic and school board regulations
  • Perform routine cleaning and maintenance of classroom and resources

Publications and Presentations

Publications

Peer-Reviewed:

González Canché,M.S., Lee, J.C., Harding, J.L., Turk, J.M., Bae, J.Y., & Zhang, C. (2023). Post-baccalaureate federal loans de-subsidization: Impacts on compositional attributes, extensive and intensive borrowing margins, and anticipatory effects. The Journal of Higher Education. DOI: 10.1080/00221546.2023.2187176

Harding, J., Parker, M., Toutkhoushian, R. (2017). Deciding about college: How soon is soon enough. Teachers College Record, 119,(4), 1-40.

Non-Peer Reviewed:
Harding, J. (2019). The bullying problem in our schools. Retrieved from https://go.cognia.org/the-bullying-problem-in-our-schools

Harding, J. (2019). So, did they learn anything? Strengthening performance-based assessments in the classroom. The Source, Winter. Retrieved from https://source.cognia.org/issue-article/so-did-they-learn-anything/

Harding, J. (2018). Did we do that? Evaluating the impact of turnaround efforts. The Source, Spring/Summer. Retrieved from https://source.cognia.org/issue-article/did-we-do-evaluating-impact-turnaround-efforts/

Practitioner Presentations

Keynote Addresses:

Harding, J. (2018). Building and growing STEM education models: Lessons learned from the field. Keynote address given to attendees at the 2018 AdvancED Midwest Regional Conference in Kansas City, Missouri to ~200 attendees comprised of teachers, school and district leaders.

Harding, J. (2018). The who, what, when, where, why and how of Student Engagement. Keynote address given to attendees at the 2018 AdvancED North Dakota State Conference in Bismarck, North Dakota to ~400 attendees comprised of teachers, school and district leaders, and representatives from the North Dakota Department of Education.

Practitioner Conference Presentations

Harding, J. (2018). Program evaluation in K-12 educational settings. Conference session created for and presented at AdvancED North Dakota State Conference in Bismarck, North Dakota to ~50 attendees comprised of teachers, school and district leaders, and representatives from the North Dakota Department of Education.

Harding, J. (2015-2018). Numerous AdvancED conference training sessions covering a range of topics related to AdvancED-specific and related education content presented in Georgia, North Carolina, Tennessee, Missouri, Wyoming, New York, New Jersey, Illinois, Arkansas, Cairo (Egypt) and Dubai (United Arab Emirates.

Workshop Sessions

Harding, J. (2016). Everyday data: Utilizing your school’s data to its full potential. Full-day training session created for and presented to the principals and leaders of the Diocese of Trenton Schools in Point Pleasant Beach, NJ.

Harding, J. (2016). Leveraging Instructional Design for STEM Educational Models. Full-day training session created for and presented to the principals and leaders of the Academy of Greatness and Excellence in Teaneck, NJ.

Academic Presentations

Stollberg, R., Harding, J., & Gonzalez Canché, M. (2016). Virtuous or Vicious Circles in Student Loan Debt: Measuring the Effects of Debt Level on Early Career Salary and Early Career Salary on Debt Repayment. Paper presented at the Association for Education Finance and Policy annual conference, Denver, CO.

Harding, J., Parker, M.C., & Toutkoushian, R.K. (2014). Deciding about college: How soon is soon enough? Paper presented at the Southern Economic Association annual conference, Atlanta, GA.

Stollberg, R., Harding, J., & Gonzalez Canché, M. (2014). Buy now, pay later: Post-graduate earnings as a function of loan debt. Paper presented at the annual conference for the Association for the Study of Higher education, Washington, D.C.

Harding, J. , Parker, M., & Toutkousian, R. (2014). Deciding about college: How soon is soon enough? Paper presented at the annual conference for the Association for Education Finance and Policy, San Antonio, TX.

Harding, J., Gonzalez Canché, M., Stollberg, R., Quarles, D., & Brajkovic, L. (2014). Fitting the mold: Using text network analysis to model university peer groupings. Paper presented at the annual International Sunbelt Social Network Conference, St. Pete Beach, FL.

Harding, J. and Gonzalez Canché, M. (2013). A wolf in sheep’s clothing? The influence of proprietary institutions on student loan outcomes. Paper presented at the annual conference for the Association for the Study of Higher Education, St. Louis, MO.

Harding, J. and Julseth D. (2012). Social media and service-learning: Using technology to enhance reflection and communication. Paper presented at the annual Gulf South Summit on Service-Learning in Higher Education, Hattiesburg, MS.

Skills

Research Methodology

  • Descriptive Statistical Analysis
  • Experimental Design
  • Hypothesis testing (t-Test, ANOVA, Chi-Squared analyses)
  • Linear Regression
  • Regression for Categorical Variables
    • (Multinomial) Logit & Probit
    • Ordered Logit/Probit
  • Causal Inference & Quasi-Experimental Design
    • Difference-in Differences
    • Regression Discontinuity
    • Instrumental Variables/Two-Stage Least Squares
    • Propensity Score Matching/Weighting
  • Panel Data Analysis
  • Bayesian Statistics

Data Management, Analysis, and Visualization Skill

  • Intermediate R user (Tidy-verse, ggplot2, etc..)
  • Intermediate SQL user (Oracle, Postgre SQL)
  • Experience querying administrative relational databases containing millions of records, requiring multiple joins
  • Entry-Level STATA user
  • Data Camp Courses Completed: Statistician in R Skill Track
    • Introduction to Statistics in R
    • Foundations of Probability in R
    • Introduction to Regression in R
    • Intermediate Regression in R
    • Generalized Linear Models in R
    • Modeling with Data in the Tidyverse
    • Sampling in R
    • Hypothesis Testing in R
    • Experimental Design in R
    • A/B Testing in R
    • Dealing with Missing Data in R
    • Handling Missing Data with Imputations in R
    • Analyzing Survey Data in R
    • Survey and Measurement Development in R
    • Hierarchical and Mixed Effects Models in R
  • Data Camp Courses Completed: Data Scientist with R Skill Track
    • Introduction to R
    • Intermediate R
    • Introduction to the Tidyverse
    • Data Manipulation with dplyr
    • Joining Data with dplyr
    • Introduction to Data Vizualization with ggplot2
    • Intermediate Data Vizualization with ggplot2
    • Reporting with R Markdown
    • Introduction to Importing Data in R
    • Intermediate Importing Data in R
    • Cleaning Data in R
    • Working with Dates and Times in R
    • Introduction to Writing Functions in R
    • Exploratory Data Analysis in R
    • Exploratory Data Analysis in R (Case Study)
    • Introduction to Statistics in R
    • Introduction to Regression in R
    • Intermediate Regression in R
    • Supervised Learning in R: Classification
  • Data Camp SQL Courses Completed
    • Introduction to Oracle SQL
    • Introduction to SQL
    • Introduction to Relational Databases in SQL
    • Functions for Manipulating Data in Postgre SQL
    • Joining Data in SQL
    • Intermediate SQL
    • PostgreSQL Summary Stats and Window Functions

Data Visualization Examples

Transparency and Replication

Wages By Degree Type

This visualization, while basic in appearance, is meant to demonstrate the ways in which we can leverage publicly available datasets to explore areas of interest. Data for this particular visual come from the website of the Governor’s Office for Student Achievement in Georgia. By using public data (when possible) and by providing the code used in performing analyses, we enable others to review and critique (or confirm!) our work. It’s also good practice to annotate code to let others know what steps you’ve taken, especially since must of us have our own coding conventions. Of course, having both annotated and plain code allows for transparency and readability.

Plot

Code

The code provided here can be copy/pasted into your own R terminal session and then executed as written to render the figure on the first tab. The ‘pacman’ package may ask to be installed (if it is not found on your machine). Once you have installed the pacman package, it will automatically check for (and install and load) any other packages you may need to produce the figure.

if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, readr, ggplot2,RColorBrewer )

url <- "https://gosa.georgia.gov/sites/gosa.georgia.gov/files/related_files/site_page/Higher%20Learning%20and%20Earnings.zip"
download.file(url, 'wagedata.zip')
unzip('wagedata.zip')
wage_data <- read.csv("GHLE/Earnings by Degree Type.csv")

wage_data %>%
  filter(percentile == '50') %>% 
  mutate(`Degree Type` = factor(degree, 
        levels = c("Certificate", "Associate's", "Bachelor's","Master's",
                   "Education Specialist", "PhD", "Professional")),
        Year = as.factor(year)) %>% 
  ggplot(aes(x = `Degree Type`, y = wages , fill = Year)) +
  geom_bar(stat = 'identity', position = 'dodge') +
  scale_fill_brewer('Years After Graduation', palette = 'Paired') +
  theme(axis.text.x = element_text(angle = 45, vjust = .5),
        plot.title = element_text(hjust = 0.5),
        plot.subtitle = element_text(hjust = 0.5)) +
  scale_x_discrete(name = '') +
  labs(title = 'Median Wages in Georgia by Degree Type',
       subtitle = 'One and five years after graduation',
       y = 'Median Wages')

Annotated Code

### Checking to see whether this R environment has the necessary packages installed and loaded. The pacman package, if not already installed, should ask the user permission to install and run. From there, pacman will take care of the other packages needed for the plot below. 
if (!require("pacman")) install.packages("pacman")
pacman::p_load(dplyr, readr, ggplot2,RColorBrewer )

### Pointing R to the URL where the publicly available data are stored. R then downloads and unzips the compressed file (into the working directory)/. The final line reads in the .csv file necessary for creating this figure and assigns it to the object "wage_data". 
url <- "https://gosa.georgia.gov/sites/gosa.georgia.gov/files/related_files/site_page/Higher%20Learning%20and%20Earnings.zip"
download.file(url, 'wagedata.zip')
unzip('wagedata.zip')
wage_data <- read.csv("GHLE/Earnings by Degree Type.csv")


wage_data %>%       # Intializng the object
  filter(percentile == '50') %>% # Filtering for the median values
  mutate(`Degree Type` = factor(degree, #Renaming 'degree' and converting to a factor
        levels = c("Certificate", "Associate's", "Bachelor's","Master's",
                   "Education Specialist", "PhD", "Professional")),
        Year = as.factor(year)) %>% # Assigning year to a factor variable for plotting
  ggplot(aes(x = `Degree Type`, y = wages , fill = Year)) + # Mapping the Data layer
  geom_bar(stat = 'identity', position = 'dodge') + #Building the geom layer
  scale_fill_brewer('Years After Graduation', palette = 'Paired') + # Changing color and Legend name
  theme(axis.text.x = element_text(angle = 45, vjust = .5), # Tilting Labels for readability
        plot.title = element_text(hjust = 0.5), # Ajdusting title for readability
        plot.subtitle = element_text(hjust = 0.5)) + # Ajdusting subtitle for readability
  scale_x_discrete(name = '') + #Removing the label from the x axis
  labs(title = 'Median Wages in Georgia by Degree Type', #Adding labels for (sub)title and y axis
       subtitle = 'One and five years after graduation',
       y = 'Median Wages')

Simulating Data and Exploring with Shiny Apps

Overview

Student Assessments During a Pandemic

Many models are used for predicting results. Predictions are important, especially when we can later test the accuracy of those predictions. Sometimes, though, we are missing data that we may never be able to capture. For instance, consider a state where a large percent of students who would be expected to take the end of grade test in a normal year did not do so in the spring of 2021 due to circumstances created by the Covid-19 Pandemic. It is interesting, and possibly instructive, to simulate what scores might look like given some different assumptions. For this example, we will hold constant the idea that all students in a given subgroup (e.g., race, gender) in a given setting (system/county and school) that did not take the test would have behaved similarly and according to a normal distribution. We then use information on the number of participants and the mean achievement of tested (4th grade) students in each content area (found at this link here on the Georgia Department of Education Website) to build a simulated dataset of test scores for the number of students who took the test. We then want to simulate the scores of students who did not take the test. We can add some variation in to the mix by allowing different mean scale scores and standard deviations according to our priors. For instance, we may believe that students who did not take the test participated in a 100% virtual mode of instruction during the 2020 - 2021 school year. We may further believe that given the abrupt change to this model (as opposed to traditional face-to-face), these students would have been less prepared for the test and would have performed worse, (lower mean scores), and with much more variation (larger standard deviation). Alternative, we might believe that these students were from families that chose to keep their children home rather than send them into school for a test. These families may have had more resources for childcare or more opportunities to stay home and work closely with their children due to parents having the flexibility to work remotely. Either case might lead to varying results. The shiny app on the next tab allows for the user to explore any number of possibilities.

Using & Interpreting the App

Mean Adjustment Scalar

By default, this slider is set to 1. This means that you expect that the non-participants, had they taken the test, would have received the same scale scores on average as those who actually took the test. The more representative you believe the participants were of non-participants, the closer you should set this to 1. If you know something about the population of non-participants that leads you to believe they would have performed worse, on average, than those who actually took the test, move the slider toward .5, at which point you are simulating scores based on the assumption that non-participants would have earned a score on average that was half the scale score of participants (an unlikely outcome, but theoretically possible). Alternatively, if you believe non-participants would have performed better on average, then adjust the slider to the right. At the far end, 1.5 means these students would have performed 1.5 times (150%) as well as those who took the test.

Standard Deviation Adjustment Scalar

By default, this slider is set to 1. This means that you expect that the non-participants, had they taken the test, would have had the same dispersion/variation in their scores as those who took the test. The more representative you believe the participants were of non-participants, the closer you should set this to 1 (the default standard deviation for this simulation is an arbitrary 37.5 scale score points). If you believe that the non-participants were a different group than those who took the test (e.g., more high achieving) and would have had less variation in their scores, move the slider toward the left, where a value of .5 simulates scores based on a standard deviation for non-participants that is half what we ‘observed’ for participants. If you know something about the population of non-participants that leads you to believe they would have performed in ways that were not as predictable or as consistent as those who took the test, move the slider to the right toward 3, which simulates scores assuming the standard deviation would have been 3 times as large for non-participants.

Number of Bins Slider

This slider only exists to adjust the visualized shape of the distribution. Increasing the number of bins adds more detail. Decreasing the number of bins places more students in each bar of the histogram, possibly at the expense of visualizing nuance.

It is important to note that this particular histogram is filled (color) based on Performance Category designation. Because these categories shift around a cut-score, certain places on the histogram will display as two bars side-by-side. In reality, the count in that bin should be interpreted as if those two bars were “stacked” on top of one another.

Application

Click on the image below to open the shiny app related to this project.