Introduction Some investigative statistical tools are described below, with associated graphics. Click on graphic for larger image.
-
Graphs - visually explore relationships among variables, look for outliers, confirm and visualize intuitively-understood relationships.
Learn about making effective graphs to visualize relationships in our
Introduction to Minitab and Graphical Analysis and
Introduction to STATISTICA workshops.
-
Design of Experiments (DOE) - change many or several factors that affect process responses, in a minimum number of runs; simultaneously optimize several responses; build empirical models to predict response behaviour as a function of process factors.
Learn about DOE in our
Design of Experiments workshop, or in our
Minitab workshops:
Factorial Experimental Design,
Response Surface Designs,
Mixture Designs,
DOE in Practice.
-
Statistical Process Control (SPC) - determine process stability; identify when process signals an out-of-control situation; use charts such as
- Variable Control Charts: X-bar, Range, I, MR
- Attribute Control Charts: c, u, p. np
- One Point Signal Control charts: EWMA, Cusum charts
When process is stable, assess capability using Process Capability Indices such as Cp, K, Cpk, Cpm, Cr, Pp, Pr.
Learn about SPC in
Process Control and Capability workshop, or in
Minitab workshops:
Quality Analysis for Manufacturing Quality, and
Quality Analysis for Service Quality
workshops.
-
Measurement System Analysis (MSA) - assess measurement instrument stability (state of statistical control), determine and quantify MSA metrics:
- gage R&R (Repeatability and Reproducibility),
- precision, accuracy,
- % study variation,
- % P / T (Precisicion / Tolerance ratio),
- discrimination,
- bias, linearity of spread and bias;
Use MSA metrics to decide if measurement system is good enough for it intended use, or needs improvement, and where it needs improvement.
Learn about MSA in
Minitab workshops:
Quality Analysis for Manufacturing Quality and
Quality Analysis for Service Quality workshops.
-
Source of Variation Study (SOV) - quantify and rank contributions of factors to variations in responses; see how a Measurement Gage R&R study is a special case of an SOV study. Use Minitab Multi-Vari charts to visualize sources of variation, from SOV studies, Gage R&R studies, and also factorial designed experiments.
|