Analytics Advanced Topics

Interpret Data Information

The predictive model refreshes at the beginning of each term. The Data Information page allows admin users only to verify term start and end dates, confirm that terms have switched in our systems, and ensure that data from each of your systems is being refreshed throughout the term. Contact your Support if term start and end dates are incorrect or if you have any questions about the data refresh schedule for your institution. Data freshness is subject to your institution-specific Service Level Agreement, which specifies how often data is processed.

How to verify that term switch has occurred

1. Confirm that the start date for the current term has passed and is correct.
2. Check the Data Freshness timestamp. This timestamp shows the most recent date and time of data ingestion across your internal systems.
3. If the Data Freshness timestamp shows a date after the term’s start date, the term has switched. Student lists will show active students for the current term with updated persistence predictions.

Data Timestamps — The timestamps at the top of the Data Information page show the most recent instance of data ingestion across your different systems. Hover over any timestamp to see these definitions.

• Data Freshness: The most recent date and time of data ingestion across your internal systems. If one of your systems (Learning Management System, for example) has been ingested more recently, the Data Freshness date will reflect the date and time when the complete data set was refreshed. Student prediction scores are generated based on this date, which ensures we are generating predictions on the most current and complete data set.
• Most Recent Data Processing: The most recent date and time the Civitas platform successfully processed your institutional data.
• Institutional Data Extract: The most recent date and time your institution pulled data from your internal systems for extraction by the Civitas platform. This timestamp will only be populated if your institution provides a date in the dataset you provide for us to ingest.
• Civitas Data Extract: The most recent date and time the Civitas platform ingested data from your internal systems.
• Most Recent Enrollment Record: The most recent date and time an enrollment record for a student was pulled from your institution’s Student Information System. This helps ensure regular student enrollment record updates. These updates could include data such as new transfer credit details, new future course registrations, or course withdrawals.
• Future Enrollment Record: The latest date of future enrollment pulled from the Student Information System. This indicates how far into the future we have a record of a student enrolling.
• Most Recent LMS Activity: The most recent date and time we have raw activity data pulled from the Learning Management System for any student. This helps ensure we are receiving the most up-to-date activity data for a student.

Current Active Term(s) — The list of Current Active Term(s) provides details for terms currently in progress at your institution:

• The Term Name shows the name for the term as stored in your systems.
• The Academic Level of students included in the term data is indicated with a U (for undergraduate) or a G (for graduate).
• Start and end dates for each term reflect the dates we have on record for your institution.
• The Census Date reflects the day the add/drop period ends for the term. This date is important because it signals when course enrollments are stable. If your institution has provided a census date, the student prediction trend will appear after the first Civitas data workflow runs following this date. If your institution has not provided this date, the Census Date field will read “Not Available” and the prediction trend will switch after the 14th day of the term by default.
• Current Day in Term shows how many days have elapsed since the start date for the term.

Upcoming Active Term(s) — The list of Upcoming Active Term(s) provides the same details for upcoming terms at your institution. The Start Date displayed for each upcoming term is the date Analytics will switch to reflect updated data for the new term. This data will include new active student lists with up-to-date persistence predictions and refreshed Powerful Predictors. If any of the upcoming start dates are incorrect, contact Support.

Powerful Predictor Term(s) — The list of Powerful Predictor Term(s) shows all of the terms included in the historical data set used for determining Powerful Predictors, or the institution-specific variables that are most predictive of persistence. You’ll see a Sample Day listed for each term (usually day 14), which shows the day in the term when the data was sampled to include in the historical data set and generate Powerful Predictors. As terms end, you will notice them move to this list as they get added to the historical data set. The student data from these terms will be used to determine the predictive variables surfaced as Powerful Predictors in future terms.

Customize student lists for additional analysis

You can identify students with shared characteristics and analyze specific behaviors that affect students’ likelihood to persist at your institution. Downloadable student lists make it easy to target these students with outreach campaigns or perform additional analysis in other tools.

Previously, student lists included data points related to the Powerful Predictors you were investigating. If you are exploring how GPA aligns with students’ persistence predictions, it may be relevant to review additional information such as major or number of accumulated credits. Now, these details can be added to any student list, allowing you to collect meaningful student information in one place.

Add additional Powerful Predictor columns to student lists to:

• Include up to five data points for each student in a single student list
• Segment a student group for more tailored outreach
• Sort the list by any data point and observe behavior trends in combination

Add Powerful Predictors to a student list — Exporting any student list will generate a file containing student data and personally identifiable information (PII). Lists must be handled in compliance with institutional policies and applicable laws.

1. Access a student list.
   • Click the Active Students number in the top right or go through the Prediction Distribution wheel or any Powerful Predictor chart.
2. At the top right, click the Add Columns button.
   • The drop-down will contain a list of all Powerful Predictors for the filtered student group. The top ten predictors are listed first, with all other Powerful Predictors listed beneath the top ten in alphabetical order.
3. Click the checkbox to the left of any predictor to add it to the list.
   • If you accessed the student list from a Powerful Predictor chart or Paired Predictor Plot, you can’t remove predictors that were included in the list by default.
4. Search for any predictor by typing into the search box above the list.
   • The list will update as you type to include only predictors meeting the search criteria.
5. When you are done choosing additional data points, click the blue Add Columns button.
   • The student list will refresh to include the selected predictors. If one of the columns is blank for a student, that student does not have data available for that predictor. This could mean that the student does not have a value for the predictor or that your institution has not made this data available to Civitas.
6. Download a .csv file of the list by clicking the orange Export Student List button.
   • The file will include all of the selected data points and can be opened in Excel and other data visualization applications. 
   • Previously, student lists were downloaded as a .tsv file, which did not open in Excel by default.

Filter by persistence prediction score

Filters enable you to narrow down your student population based on specific characteristics. Each filter category has options that identify various groups of your student population. Adding the ability to filter by Prediction Score will allow you to easily identify students most at risk of not persisting.

You can filter for one or more prediction groups (e.g. Low, Moderate, Very High) to:

• See updated Powerful Predictors, highlighting the specific factors that have had the greatest influence on whether historical students in the prediction group persisted or did not persist
• Download a student list including multiple persistence prediction groups

Interpret Powerful Predictors

Powerful Predictors are the variables that are most predictive of persistence at your institution. The list of Powerful Predictors updates dynamically as filters are added. This allows you to see what factors had the greatest influence on whether students with certain characteristics persisted or did not persist.

For example, review the Powerful Predictors for the Moderate prediction group to better understand what factors have historically contributed to students in this group being more or less likely to persist. These factors provide insight into what behaviors make students with Moderate predictions who persist different from students with Moderate predictions who do not persist. Powerful Predictors do not show how students with Moderate prediction scores are different from students in other prediction groups or any other student group.

Take action by downloading a list of current students in one or more persistence prediction groups and use your knowledge of what factors influence persistence for these students to inform your outreach.

Download a student list

1. From the Prediction Score filter, add one or more prediction groups (e.g. Very Low and Low). The Prediction Distribution in the top right will refresh to include only the prediction groups you have selected.
2. Click the blue Active Students link above Active Filters, which lists the filters you’ve added. This link will indicate how many students are included in the filtered student group (e.g. 2,025 of 8,007 Active Students).
3. The student list will include all students who are currently enrolled in at least one course with a persistence prediction score in the selected group(s).
4. Download the student list as a .csv file by clicking the Export Student List button at the top right of the list.

Add unenrolled students to student lists

Most institutions want to outreach to students who are not currently enrolled in the next term. Whether encouraging students who were enrolled in the Spring to re-enroll in the Fall during the Summer term or trying to win back students who skipped the last couple of terms, it’s important to know who has recently attended your institution but is not currently enrolled. 

You can include students who are not enrolled in the current term in student lists. Add details about these students to include them in nudge campaigns and understand context about their previous enrollment behaviors and intention to return for future terms. Targeting these students with outreach will support your teams’ efforts to win back these students during summer and other minor terms. It’s important to understand how this addition affects what you see in Analytics. Review the following details to see how to add these students to your student lists and how to interpret the insights found throughout the application. 

To access a student list:

1. Apply filters for a student group of interest.
2. Click the Active Students link in the upper right corner of the Overview page to view a student list including all students meeting the filter criteria who are enrolled in at least one course for the current term.
3. To see a student list containing a subset of these students, access a student list from the Prediction Distribution chart or any Powerful Predictor chart.
4. By default, you’ll see contact and enrollment details for students enrolled in the current term.

To add students not enrolled in the current term:

1. Click the ‘Show Students Not Enrolled in Current Term’ button located above the student list to add all students who were enrolled in a past term within the last calendar year. These students’ details will appear on a gray background to distinguish them from currently enrolled students.
2. For all students, including those who are currently enrolled, see the details about their history:
   • Enrolled Current Term: See whether the student is enrolled in the current term.
   • Persistence Prediction: For currently enrolled students, you will continue to see the most up-to-date prediction score. For students who were last enrolled in the most recent past term, see the prediction score they had as of the last data workflow at the end of the term. For students who were enrolled in an earlier past term, the prediction score will not be displayed.
   • Last Enrolled Term: Review the most recent term in which this student was enrolled. If the student is currently enrolled, the current term will be listed.
   • Next Enrolled Term: Check the next term the student has enrolled in courses. If the student is not registered for any upcoming terms, this field will be empty.
3. Click any column heading to sort the list by the values in that column.
   • For example, click the ‘Enrolled Current Term’ heading to separate students who are currently enrolled from those who are not.

To customize the student list and export for additional analysis:

1. After sorting, click the ‘Add Columns’ button to see these students’ values for one or more Powerful Predictors.
   • Try adding Credits Earned (Cumulative) to identify the students who may be close to finishing their degree.
2. Click the ‘Export Student List’ button to download the list as a .csv to begin an outreach campaign or perform additional analysis.
   • Tip: Be sure to export the list before navigating to another page as your list changes will not be saved.

Students who are not enrolled in the current term can be added only to student lists and will not appear or affect the student counts found in other parts of the application:

• The Active Students number you see at the top right will not change. Active Students includes all students enrolled in at least one course during the current term. However, the number of students included in the student list will update when students not enrolled in the current term are added. Find the updated number above the student list.
• Filter counts will only include currently enrolled students. If you are interested in sending a nudge to a certain filtered group, add these filters and navigate to the student list. From there, add students not enrolled this term and only students who met the filter criteria during their last enrolled term will be added.
• Powerful Predictors are still valid for exploration and analysis, as they are built using the last year or more of data from all types (i.e. seasons) of terms. The insights about factors that set students who persisted apart from students who did not persist for a filtered student group will not be specific to students enrolled in the current term. Instead, Powerful Predictors should reveal details about persistence trends over the past few terms.
• Adding columns to a student list with students who are not currently enrolled may result in some odd values or no value in some cases. The value you see for each student will reflect the data that was captured on the last day the data was refreshed during the term he was last enrolled. For example, if you’ve added a column to show GPA (Prior Term) and a student’s last enrolled term was Fall 2016, you’ll see his GPA from Spring or Summer 2016.

Inspiration & Intervention Factors

Inspiration and Intervention factors are the individual behaviors, qualities, and circumstances that most strongly influence a student’s likelihood to persist. These give you a good sense of the kind of attention a particular student needs, but until now, you could only see these details on a student-by-student basis. You can filter for students with shared Inspiration or Intervention factors to quickly identify groups of students with the same who would benefit the most from a targeted outreach.

Using the Inspiration and Intervention filters, you can tailor your outreach efforts to students not just with shared qualities, but with shared predictive qualities. Rather than filtering for all students with a low GPA, for instance, Inspiration and Intervention filters identify the specific students whose low GPAs are actively impacting their likelihood to persist or not persist.

Was this article helpful?

0 0

How can we improve this article?

Please submit the reason for your vote so that we can improve the article.

Submit