Contact Student Affairs Strategic Initiatives

Email:
SASI​@nau.edu

More resources

Data Jargon

Understanding data jargon is crucial to anyone who is sharing and review data set results. The ability to discern this jargon will ensure that you are able to have a comprehensive conversation about your data. The following definition of terms will help you to understand when you should use data terms, and what terms would best describe your dataset.

Statistics & Analysis Jargon Accordion Closed
  • Significant Results – ONLY to be used when running statistics and talking about confidence intervals and/or p-values.
    • Examples: XYZ shows a 3% lift in persistence and is statistically significant.
      • A higher persistence in XYZ students because of our program is significant.
      • There was a significant difference from the NAU populate and the XYZ group.
  • Lift/Difference/Percentage Difference – Used to describe the difference or gap between two groups. This gap can or cannot be statistically significant.
    • Example: XYZ shows a 3% lift in persistence for ABC students which is (or is not) statistically significant.
    •  Students who enrolled in XYZ averaged a 5% increase(decrease) in persistence.
  • Percentage Points – This is the overall score. This is NOT the difference or lift between the groups.
    • Example: Students who took XYZ course had a persistence rate of 65%.
  • Confidence Interval (CI): The likelihood that you are correct. This is how confident you are in your data. 90 or 95% is the general default for CI’s.
    • Example: I am 95% confident that the lift in the students who took NAU 100 was 4%.
  • P-value: The likelihood that you are wrong. This is not yet in a percentage; general form will look like 0.04.
    • Example: XYZ shows a 3% lift in persistence and is statistically significant with a p-value of 0.02.
  • Trend: the increase, decrease, or no change in a set of variables. This does not have to be statistically significant.
    • Example: There is a positive trend in persistence when XYZ is met.
  • Correlation: how much, or if, a variable is dependent on another. NOTE: this can be just a trend and if you are not stating that it is statistically significant you can infer this from graphs or numbers.
    • Example: There is a positive correlation between the XYZ program and the student’s retention.
  • Population: The entire data set. The entire NAU student body is a population.
    • Example: The average GPA for the population is X.XX while the average GPA for the XYZ group is Y.YY.
  • Data Set: The collection of data the will be used to run an analysis.
    • Example: Our data set contains only student in our program.