The Campus Analytics Suite is a set of tools that provide data to assist with increase graduation rates by predicting when students are at risk and providing daily updates to this data.
Product Page:
https://www.infinitecampus.com/products/campus-analytics-suiteCampus Early Warning
Campus Early Warning uses powerful statistical algorithms to measure how attendance, behavior, academics, home and school stability interact to predict graduation. GRAD Scores help Student Services respond with interventions targeted at students who would benefit the most. Privacy is our highest priority, so the system saves risk correlations that cannot be connected back to individual past students.
Overview video:
https://www.infinitecampus.com/video/early-warningWhat is Early Warning?
Early Warning is a tool that automatically discovers statistical relationships between educational records and enrollment outcomes. Early Warning uses these evidence-based risk factors to estimate each student's likelihood of a positive enrollment outcome (matriculation or graduation) versus a negative enrollment outcome, such as dropping out of school. See the full documentation below.
Early Warning provides school administrators, counselors, principals, etc., with a dashboard to quickly see a student's persistence toward graduation (successful completion of high school) and the likelihood of being promoted to the next grade level. School personnel can compare one student's likelihood to their peers and more effectively intervene on behalf of a student to increase the probability of promotion. Early Warning can be used with students in grades K-12.
Early Warning data is displayed for the current school year only, as determined by the year marked Active in the School Years editor. That year must be selected in the Campus toolbar. When a future or past school year is selected, a message indicates data cannot be displayed.
The following videos provide additional information on Early Warning:
The predictive analysis that Campus employs to generate a student's GRAD score is ever-evolving and constantly updated as new data is entered into Campus. A student's GRAD score includes (but is not limited to) the following categories of data:
|
|
|
|
|
|
|
|
|
|
Early Warning learns statistical relationships between educational records and enrollment outcomes. Machine learning uses computers to find evidence-based predictors of student outcomes using a large amount of anonymous data. To estimate a student’s likelihood of a particular enrollment outcome – such as dropping out of school – the computer system compares her record to past records of similar students.
The advantages of a machine learning system for Early Warning include:
The machine-learned statistical model has better predictive power than heuristics published in dropout prevention literature.
The student information system can detect changes in educational data faster than staff members can and update GRAD scores accordingly.
Early Warning Definitions and Calculations
Historical data for all students is used to calculate the scores. The application of that score to different categories and different types of scores varies. The following table lists the definitions and calculations of the terms used in Early Warning.
Term | Description | Interpretation of Score |
|---|---|---|
GRAD Score | The acronym for Graduation Related Analytic Data. It represents the score that summarizes a student’s educational record with a single number indicating their likelihood of promotion to the next grade level. It measures factors predictive of dropping out and factors indicative of a student's persistence to the next grade level or graduation.
A GRAD score is not provided for a student when there is no active primary enrollment record on the date the scores were calculated or when the student is new and has too little information available to calculate a GRAD score. | A GRAD score is between 50 and 150. A score closer to 150 indicates the student is more likely to be promoted to the next grade level. A score closer to 50 indicates the student has more risk of not being promoted to the next grade level. |
Category Impact Score | This score uses all of a student's data but imputes perfect values into a category subset of columns that are intended to forecast how much an intervention in a category would change the student's GRAD score based on historical data for that student and other students in the same situation. | An Impact score is between 0 and 100. An Impact Score that is closer to zero requires less need for intervention. |
Independent Category Outcome | This score uses a single category subset of columns of a student's original data (e.g., just the attendance-related columns), intended to explain the GRAD score and changes in a predictable manner as a student's category-related data changes. Positive behavior events are not included in this count and are not considered by Early Warning calculations. | An Independent Category Outcome Score is between 50 and 150. A score closer to 150 indicates the student has less risk of not being promoted to the next grade level; A score closer to 50 indicates the student has more risk of not being promoted to the next grade level. |
Score Distribution
The Score Distribution displays the selected school's score distribution for all students on a grid, with students on the left (Y-Axis) and scores across the bottom (X-Axis). The student distribution changes based on the entered tier definitions, selected schools, selected grade levels, and selected Score types.
Score Types
The score types include. but are not limited to the following definitions. Only one score type can be selected at a time. When a specific score type is selected, the list of students refreshes and displays results and color-coding based on that selected type.
Changing the values for the score, student count, percentage, and risk levels changes the list of students.
Score Type | Definition |
|---|---|
GRAD Score | This score considers the student's education record and the calculated determination of the likelihood of completing the grade level and being promoted to the next grade level. Note that when Early Warning is used in a district, the GRAD Score is also displayed on the Student Graduation and Student Profile tools. |
Attendance | This score looks at the student's total attendance data (absences and tardies), total attendance percentage, and whether the student is chronically absent. |
Behavior | This score looks at the student's total number of behavior incidents, out-of-school suspensions, and in-school suspensions. Positive Behavior events are not considered. |
Curriculum | This score looks at the student's transcript information and grades (including in-progress). |
Stability | This score looks at the total number of years the student has been enrolled at the selected school, how long the student has been enrolled in the district, how long the student has lived at the current address, the student's overall number of addresses, the enrollment status, and the number of portal logins. |
Tier Definitions
Enter the desired values for Score, Student Count, and Percentile for each High, Medium, and Low threshold. You can modify the percentile values by typing in the fields or using the up/down arrows. The tier settings return to the default values when the tool is refreshed.
A value closer to 50 indicates the student is at a higher risk for non-promotion and is displayed in red.
A value closer to 150 is at a lower risk of non-promotion and is displayed in green.
A value that falls between 50 and 150 tier definition boundary lines indicates the student is at medium risk for non-promotion and is displayed in yellow.
Or, use the default entries, which are a maximum percentile of 5% of the student population (high, more risk), 21% of the student population (medium, moderate risk), and 100% of the student population (low, less risk).
Intervention in some areas may benefit some students who have a high-risk score in a certain category, while other students may not benefit from such programs. Certain events may have caused a change in a student's score, and an intervention may not be the best way to affect change for that student (natural disasters, etc.).
When one score is changed for one of the tier definitions, every value adjusts accordingly. For example, when a counselor is viewing the 12th-grade students at the high school and changes the Score values from 50, 100, and 150 to 76, 110, and 150, the student count changes to match the entered Score values, as does the Percentile values.
If the entered values do not match the results, data is overridden to list the students that do match the entered values. For example, if ten students have a GRAD score of 50, and 20 students have a GRAD score of 55, and the Student Count is set at 15, the ten students with a GRAD score of 50 are listed.
dasd