Correlation & Regression
A statistic measuring the strength and direction of a relationship
Correlation
The first step where the issue is clearly identified
Clarify and validate the problem
A structured analysis designed to identify problems and propose solutions
Special study
A number that describes a population’s characteristic
Parameter
Analysis used to forecast future events using current and past data
Predictive analysis
Reporting all findings, even unfavorable ones, builds this
Mission trust and ethical integrity
An r-value showing a perfect positive relationship
+1.0
The step that investigates underlying causes
Determine root cause
The number of essential questions that every study must answe
Five
A measure calculated from a sample to estimate the population’s parameter
Extending an existing trend line into future time periods
Extrapolation
Differences that naturally occur between data points
Variation
An r-value showing no linear relationship
0
Creating targeted solutions for verified causes
Develop countermeasures
The question that defines the purpose of a study
“Why are we doing this study?”
Reason analysts rely on samples instead of entire populations
Efficiency and resource savings
The long-term directional movement in time-series data
Secular trend
Why small samples can mislead conclusions
They may not represent the population
A relationship where one variable increases as the other decreases
Negative correlation
Checking whether fixes worked and are sustainable
Confirm results and process
The section that defines a study’s scope and constraints
Assumptions and limitations
Sampling entire units such as squadrons.
Cluster sampling
Why mechanical extrapolation is less accurate
Subjective human interpretation
A measure describing how peaked or flat a distribution is
Kurtosis
The analysis that predicts one variable based on another
Regression analysis
A questioning tool that uncovers the true root cause by repeatedly asking “Why.”
5 Whys
The condition that triggers the start of a study
Identified issue or performance shortfall
Bias caused by unintentionally selecting easier-to-reach data
Unintentional bias
Difference between regression and extrapolation
Regression = model-based; Extrapolation = visual extension
A chart that distinguishes between random and assignable causes
Control chart
The mathematical equation used to express a regression line
Y = a + bX
The step ensuring improvements are standardized
Standardize successful processes
An analysis used to identify the causes behind performance issues
Root cause analysis
Sampling method ensuring each subgroup is properly represented
Stratified sampling
Why correlation should be checked before regression
To confirm a relationship exists
Ignoring ethics in reporting leads to this consequence
Poor decisions and reduced readiness
The measure showing how far actual points fall from the regression line
Standard error of the estimate
The most common mistake during problem solving
Misidentifying the root cause
The overall objective of any completed study
Improve processes and mission effectiveness
Method to reduce sampling error
Increase sample size and randomness
The advantage of regression for analysts
Quantitative prediction of performance
Using factual data to improve efficiency and planning
Data-driven decision making