While I have been loving playing with Datawrapper, it does not have a sankey option. So I went in search of a WordPress compatible application that makes sankey visualizations. The first I came across was Amcharts. It is a bit more complicated than Datawrapper, but I think I got it figured out. I still have a ton of options to figure out, but I believe I am able to make a chart appear. Since building the chart took so much time I am going to save the explanation of the data for when I update this post. Right now I just want to see how the visualization works.
Maryland Public School College Pipeline | Caroline Boice
As I figure out the program you need to click on the actual post or this link to see the chart. I am working on this issue.
After my last post about community college outcomes using data published by the Maryland Higher Education Commission (MHEC), I went in search of more data to get a better sense of the story. I want to know what happens to students that initially enroll in a Maryland community college. I found some data published by the Maryland Longitudinal Data System (MLDS) Center in 2016. It is almost exactly the data I am looking for, unfortunately, there was an unlabeled category that consisted of 25% of the cohort. With this unknown category being such a large part of the pie, I feel that the data is of limited use for really understanding the outcomes of community college enrollees. Despite my belief that the data may be limited, I am going to take a closer look at the data to see if I can learn anything from it.
It is also unclear the time frame after enrollment that the data covers. I know that the population is community college students that were enrolled during the 2008-2009 academic year. I also know that the data was published in 2016. It is unclear if the data set is all students; first-time, full-time students; or another population. I think that my next step may be to contact MLDS Center to clarify what data is being presented. There may have been additional information in the documentation that I have missed.
When I publish graphs in the future I want to be sure to be clear about the data that is presented. In education public policy I feel that this is especially important as the specifics matter. For example, a cohort of first-time full-time students is not the same as a cohort of all community college students.
If Data is Accurate: Few CC STudents Transfer; Fewer earn a BA
If this data is accurate, few community college students transfer to a Maryland four-year institution, only about 6% of the cohort. This is much less than the 12% of the first-time, full-time cohort in the data reported by MHEC. This leads me to believe that the transfer rate in this data source may be underestimated. I wish that I could trust that this data is fully accurate because, unlike the other data source, this data source reports the percent of students that earned a bachelor's from a Maryland four-year institution after transfer. It shows the specific community college to four-year institution pathway to a bachelor's degree that some experts recommend, especially to lower-income students as a method to save money. But it isn't a method to save money if students do not actually end up earning a degree. According to this data, 1% of the students (421 students) that begin at a community college earn a BA from a Maryland four-year institution.
If only 1% of a community college cohort earn a BA, then that pathway to a BA is broken. However, I need to dig into the data more, because other data sources report a slightly less bleak picture. According to data published by the Aspen Institute, which is shown below, nationwide 14% of community college students earn a BA within six years of transfer. However, like the MLDS data I discussed above, I am not sure of the student population, so I need to do more research into the data. Overall I need to better understand the data sources before I make any conclusions.
A Note About THis Data Presentation
As I have said in previous posts I have really been loving playing with data presentation using datawrapper. However, for the multiple pie charts shown above I ran into a few issues. Firstly, the website was being glitchy, not showing changes I had made to the graph and hiding the graph completely. Second, I was not able to make this type of graph look exactly how I wanted it to look. I did not want to show the percent for the grayed-out space and that was not an option because the categories are not the same. This is not surprising, I think that for that type of functionality I would need a graphic designer and not a web application. So while there are a few issues, I am still extremely happy with datawrapper. I am just trying to document for myself the limits of its functionality.
According to the Career and College Readiness and College Completion Act of 2013, it is the State’s goal that all degree-seeking students enrolled in a public community college earn an associate’s degree before leaving the community college or transferring to a public four-year higher education institution. Therefore the Maryland Higher Education Commission (MHEC) tracks community college students’ graduation and transfer outcomes. MHEC does not have the ability to follow students who transfer to out-of-state institutions. Therefore, the data only reflects transfers to institutions in Maryland.
For this data set, the analysis cohort is all first-time, full-time students entering Maryland community colleges within the fall term of a given year. Three successful outcome measures are tracked.
Graduated and transfered: Student graduated with an associate degree or lower-division certificate and transfered to a Maryland four-year college or university.
Graduated/Did not transfer: Student graduated from a Maryland community college with an associate degree or lower-division certificate and did not transfer to a Maryland four-year instituion.
Transferred to four-year college or univeristy without graduating: Student transferred to a Maryland four-year insitution, withou having completed an associate degree or lower-division certificate.
Now Most Community College Transfer Students Graduate Before Transfering
I decided to display the data as a line chart, but with different visual clues from MHEC’s original line chart. For my chart, I added shading between graduated and transferred and transferred without graduating. I like how it clearly shows that in 2014 the number of students who graduated before transferring surpassed those who transferred before graduating. Although the numbers had been trending in that way, the legislation likely had some impact.
I wish we had the data to analyze the outcome of both types of transfer students at four-year institutions. Past research suggests that the students that graduated community college first prior to transferring be more successful, but I wonder if this is still true. This might be a question to examine in the future.
64% of a Cohort does not transfer or graduate in four years
The line chart above does not capture the approximately 64% of the cohort that do not graduate or transfer within four years, so I decided to make an area chart. The area chart shows that the majority of students that begin as first-time, full-time community college students do not graduate or transfer within four years. Although when the number of first-time, full-time students decreased after peaking during the Great Recession, the percent of students that did not graduate or transfer has dropped to about 60% of the cohort.
This data set does not capture if these students are still persisting or if they dropped out. It is also possible that these students transferred to a two-year or four-year institution out of State. This data set only captures a subset of successful outcomes for community college students: graduation and in-state transfer.
This chart reminds me that while measuring the success of transfer students that did or did not graduate is interesting and important, there are a large number of students that entered community college as full-time students who have not graduated or transferred.
Maryland Community College Student Outcomes
This data only tells part of the story of community college student outcomes, but it is a good place to start. One of the biggest issues in education public policy is the lack of data that tells the whole story. Another issue is knowing what data is important to analyze. I plan to continue to look for data to analyze.
Student Performance On THird Grade Assessments 2019
Yesterday the Maryland State Board of Education presented data about student performance on the third grade Statewide tests in a discussion on learning loss and recovery prior to the COVID-19 pandemic. The new State Superintendent said that Maryland has one of the bigger performance gaps between students eligible for free/reduced meals (FARMS) and students not eligible for FARMS; however, I am not sure how that is measured. I was happy to see that they divided the data by students eligible for free/reduced meals and by race/ethnicity as that is not always done. The graphs show that regardless of race/ethnicity, family income matters when it comes to performance on third-grade assessments.
I would be curious to know if this data is available further divided by local school systems. If any school system is doing better than the others with regards to the performance gaps it might point to policies that are contributing to better student performance that may be able to be replicated in other local school systems.
Third Grade English Language Arts Performance
Student Math Performance
Data PResentation
For these graphs, I generally copied the presentation by the Maryland State Department of Education for this data for their presentation. I did not add axis markers because I did not release that they were missing until I published the graphs. I thought about making a range plot, but I didn't have the data for "all students" in a category, and for this data, I decided I preferred the bar graphs for this data. The bar graphs show that there is a performance gap between FARMS and non-FARMS eligible students for every race/ethnicity category.
I played with using low-income vs. non-low-income rather than FARMs, for the presentation of the data to a general audience, but I did not commit to the choice as you can see. This is a problem that I face frequently as a policy analyst, do I simplify data presentation for a general audience or do I provide all of the specifics so that fellow wonks can understand specifically what I am data presenting? This is a skill I want to work on in the future.
Today I looked at a report on the Pathways in Technology Early College High Schools (P-TECH) program. In six years or less, a P-TECH student can graduate with a high school diploma and a no-cost, two-year Associate of Applied Science (AAS) degree in a specified discipline. Each P-TECH program requires a partnership among a local school system, a local institution of higher education, and a local employer. The partnering company agrees to provide students with mentoring, paid summer internships, and first-in-line consideration for job openings with the school’s partnering company.
Maryland is currently phasing in nine P-TECH programs in six jurisdictions. The first P-TECH programs in Maryland were established at Carver Vocational-Technical High School and Paul Laurence Dunbar High School, both in Baltimore City. They had their first class of students earning graduate from high school in the 2019-2020 school year. According to data reported by the Maryland State Department of Education, 92.6% of Carver P-TECH students and 100% of Dunbar P-TECH students earner their high school degree. In addition, 48.1% of Carver P-TECH students also earned an AAS degree in four years. Additional students may earn their AAS in the fifth and sixth years of the program. According to the chart shown below, 31 P-TECH students at those two schools returned for the fifth year of the program in 2020-2021.
The report included information on Maryland P-TECH program enrollment for the 2020-2021 school year by grade and by school. Overall, there were 1,065 students enrolled in a P-TECH program. Grade 9 enrollment ranged from 25 students to 61 students. Enrollment is anticipated to grow as the programs continue to phase in. All currently operating programs will be fully phased in by the 2025-2026 school year.
The Pathways in Technology Early College High (P-TECH) School Act of 2017, states that it is the intent of the General Assembly that no additional P-TECH school shall be established other than those that receive a P-TECH Planning Grant in fiscal year 2017 or 2018 until the P-TECH Program is shown to be successful in preparing students for the workforce or for further postsecondary education. Thus, no new P-TECH schools funded by the State of Maryland will likely be established until an evaluation of the current programs is completed. However, the statutory language does not require a formal evaluation.
Since I used U.S. Census Bureau data for part of my discussion in my last post, I decided to quickly examine their educational attainment data. For this visualization, I used a doughnut chart. I briefly considered a chart with multiple doughnut charts to include the information desegrated by males and females, but it was not very interesting because the percentages for males and female educational attainment is remarkably similar. Women are one or two percentage points higher for the attainment of college degrees.
The Census Bureau does not publish the percentages for all options for educations by race. It just publishes the percentage of the population that has obtained a high school degree or higher and a bachelor’s degree or higher. I am not sure how best to show the data yet, but it shows a real difference in the attainment of bachelor’s degrees by race.
Unlike the Maryland Longitudinal System Center data I examined in my last post, this data reflects the population living in Maryland at the time of the survey, not just public high school graduates.
Using Datawrapper
A note about datawrapper, I had to update the chart because I forgot to uncheck the box that makes the top row the label row. It was easy to fix, but it is a reminder to check your data before publishing.
Some College, No Degree
What I find most interesting is what a large percent of the population is in the "some college, no degree" category. Seven percent of the population has earned an associate's degree, while 18 percent are in the "some college, no degree" category.
No High School Degree
Another fact that I find interesting is that nearly 10 percent of the population of Maryland age 25 and over do not have a high school diploma. From this data set I do not know if the people without a high school diploma are younger, older, or evenly distributed between younger and older people.
I have been testing the features of the Datawrapper data visualization tool. I started with maps, and still have mapping features to explore, but today I decided to try making a chart. In particular a data range chart. Like my other experiences with Datawrapper the range chart was easy to make. I just put the data into four columns, changed the percent data to just the number without the percent sign, saved the Excel workbook, and uploaded it to Datawrapper. With a few clicks and a bit of typing, I made the chart below which shows college enrollment for high school graduates.
Public Policy Questions about the CHart
Statewide 78% of public high school students enrolled in college either full-time or part-time as a degree-seeking or non-degree seeking at any point after high school graduation. The statewide average hides the variability in college enrollment in the State both by county and by family income. Only 44% of low-income high school graduates from Kent County enrolled in college, while 93% of non-low-income high school graduates from Howard County enrolled in a college. That is a difference of 49 percentage points!
All counties also have gaps in college enrollment between low-income and non-low-income high school graduates. In particular, Carroll and Queen Anne's counties have the largest gaps, 31 and 29 percentage points respectively. Talbot Couty stands out as having the smallest gap in college enrollment between low- and non-low-income high school graduates, 4 percentage points.
What I don't know from this data is what is a good level of college enrollment. Many studies have found that earning a bachelor's degree pays off financially for most people, but there are other pathways to financial and life fulfillment. This data captures some "non-traditional" education pathways such as certificate programs, so some "trades" are captured, but not all. Although I do not know what the college enrollment rate should be, I think that high school graduates should have equal opportunities to enroll in college. The huge variability in college enrollment rates might point to underlying factors that prevent some students from enrolling in college.
This data also does not tell me why high school graduates choose to enroll or not enroll in college. However, given that low-income students enrolled at a lower rate for all counties, money is likely a major factor. Other factors may be the distance to an affordable college, high school preparation, or community expectations.
Again I am using publically posted data from the Maryland Longitudinal Data Center. In their data set they did not post the percentage of public high school students that earned a college degree by age 25, I calculated it by dividing the number of high school graduates in a county by the number of students that earned a college degree from that county by age 25. From my understanding of the data this should work, but I haven’t done a deep analysis into the potential flaws of that process.
For this Datawrapper map, I added a legend caption to reflect the Statewide average. The caption appears directly above the legend. For now I this placement makes sense for a Statewide average. Otherwise, I used the same settings I have used with the other maps I have made thus far.
Public Policy Thoughts About the Data
According to this data, 40% of Maryland public high school students who graduated in 2011-2012 earned a college degree by age 25. There is not directly comparable data nationwide because Maryland Longitudinal Data Center only collects and publishes data about Maryland. The American Community Survey Data collects data about the educational attainment of individuals based on where they live, not where they graduated high school or where they were educated. According to that data 44% of Marylanders 25-44 years old, have earned a bachelor's degree or higher. Maryland is known as a State with a highly educated workforce. Is that because a large number of the State's high school graduates graduate college by age 25 or is it because educated workers move to the State? I do not know the answer, but I interested in exploring the data more.
What strikes me about the map is the difference in the percentage of high school graduates who earn a college degree by age 25 in Baltimore City, 16%, and Howard County, 60%. That is a huge difference, 44 percentage points. As someone who is familiar with Maryland I am not surprised by the difference, but the difference is striking. I want to dig deeper into the data. I want to see if low-income high school graduates from Howard County earn college degrees at a similar rate to high-income graduates in the county, or is the rate more similar to that of jurisdiction that have higher levels of poverty such as Baltimore City. I also want to look at the opposite for Baltimore City.
When looking back on the map showing percent college enrollees with a college degree by age 25, there is no jurisdiction that stands out as being radically different in this map, but I will need to dig deeper into the data.
If I can understand the potential reasons behind the data better I hope I will be able to give better policy advice.
Yesterday I made my first map using Datawrapper. I noted in the write-up of my experience that I could not figure out how to display the data as a percentage; I ended up displaying it as a decimal. I figured that it was possible because I had seen Datawrapper maps with percentages, so I was pretty sure I just needed to dig through the menus and options. Unsurprisingly, I quickly found a tutorial written by Datawrapper that explained how to customize a choropleth map, which included the information I needed to figure out how to display percentages.
Since the Datawrapper tutorial does not directly address displaying map data as a percentage, I will give a quick explanation to remember how to do it next time. Basically, it is a three-step process.
Upload the data striped of the percent sign as you want it displayed, not in decimile form. For example if you want a data point to be 8% upload the data as 8. This will allow your data to be displayed in the map. If you add the percent sign the data will not be displayed.
On the “visulize” step choose a percent number format (there are a few choices) from the legend menu. This will display the data as a percentage in the legend.
Use “Tooltips” to add the percent symbol after the variable code for the data. It is the second box.
As I suspected, it was not difficult to get the data to display as I wanted. It just took some digging into the menus and options.
Now for a look at the map.
As with the first map I made, I used data published by the Maryland Longitudinal Data Center. For their visualization of the data, you have to pull up each county individually because they have rich data on students from each county. I like seeing all of the counties at once on one map. I used the same color scheme as with my last map, with red being the lower percentages and green being the higher percentages. For this map, I allowed the lowest percentage to be the darkest red and the highest percentage to be the darkest green. I have not yet given much thought to if that is the best way to display the data.
Another piece of data that I want to explore adding to the map is the Statewide average. I know that the Statewide average is 51%, that is, Statewide, 51% of public school students who graduated in 2011-2012 earned a certificate, associate's, bachelor's, or master's degree or higher by age 25. At least among the students captured in the dataset.
It stands out that only 22% of students from Baltimore City who enroll in college earn a college degree by age 25. What the data does not tell me is why. Since I am a curious person, I plan to dig into the data more to look for why this might be. I might not find the whole story, but I hope to find some elements of the story. Dorchester and Somerset counties also have low levels of degree attainment for students who enroll in college. For future maps I want to look at total college degree attainment by high school graduates, college degree attainment by FARMS and non-FARMS students, and by FARMS percentage of the entire county. I also want to see how the college enrollment rate correlates with the college graduation rate. While examining this data, I want to see the capabilities Datawrapper has to display the data.
As a public policy professional, I consume and manipulate a ton of data. Unfortunately, as a government employee, I often do not have access to the latest data tools. There tend to be many layers of approval and cost restrictions. Therefore, I am really excited to give Datawrapper a try here on my own website. It offers an extensive free service that I want to explore with public data.
Today I was able to follow the easy instructions to install the Datawrapper Plugin on my WordPress-based website. Even without much website experience, the installation was easy. Then I decided to build a map in Datawrapper using data I have been looking at published by the Maryland Longitudinal Data Center for one of their data dashboards. I am really excited about their data, but I feel like I have to manipulate it myself to understand it. The map below is me just dipping my toe into the water to understand the data and how to use Datawrapper.
The map shows the percentage of public high school graduates from 2011-2012 that enrolled in college at any point after high school graduation. I was looking at 2011-2012 graduates because I am also interested in the number/percentage that graduated from college by age 25. I will likely examine the college graduation data in a future post.
It was really easy to upload the data into Datawrapper, I already had it in an Excel file. The map below was created in a few minutes. I did have an issue with getting the data to display using the percent symbol. I will have to see if there is a way to get the percent symbol to display using the program.
I am not sure about the choices I made regarding the scale and colors for the map. I want to make a bunch of maps and graphs to publish here on my website to explore how best to represent the data.
Overall I am really excited that I was able to quickly make a professional-looking map. I am looking forward to testing its capabilities and seeing if I can learn any insights about the data.
We use cookies to personalize content and ads, to provide social media features and to analyze our traffic. We also share information about your use of our site with our social media, advertising and analytics partners who may combine it with other information that you have provided to them or that they have collected from your use of their services. You consent to our cookies if you continue to use our website.
We use cookies to personalize content and ads, to provide social media features and to analyze our traffic. We also share information about your use of our site with our social media, advertising and analytics partners who may combine it with other information that you have provided to them or that they have collected from your use of their services. You consent to our cookies if you continue to use our website.
Necessary cookies help make a website usable by enabling basic functions like page navigation and access to secure areas of the website. The website cannot function properly without these cookies.
Preference cookies enable a website to remember information that changes the way the website behaves or looks, like your preferred language or the region that you are in.
Statistic cookies help website owners to understand how visitors interact with websites by collecting and reporting information anonymously.
Marketing cookies are used to track visitors across websites. The intention is to display ads that are relevant and engaging for the individual user and thereby more valuable for publishers and third-party advertisers.
Unclassified cookies are cookies that we are in the process of classifying, together with the providers of individual cookies.
