Measure Flashcards
process mapping
a graphical representation of a process, showing the sequence of tasks using a modified version of standard flow-charting symbols
steps of process mapping
- select a process to be mapped
- define the process
- map alternative paths
- map intersection points
- use map to improve the process
flow charts
- graphically displays the inputs, actions, and outputs of a given system
- symbols standardized by the ANSI standards
operational definition
- an operation definition is a precise description that tells how to get a value for the characteristic you are trying to measure
- it describes what something is and how to measure it
- it should remove ambiguity so that everyone has the same understanding
normality assumption
- many statistical techniques assume data normality (t-tests, Z-tests, ANOVA, many others)
- 2 approaches to testing normality: graphical and statistical
what if the data is not normal?
- do nothing (if the curve fits the region of interest you might be okay to proceed even if p<0.05 might occur even if the lack of fit is not practical importance)
- transform the data to make it normal
- use averages
- fit another distribution
- use a Non-Parametric technique
Why use NP statistics
- no assumptions regarding population distributions
- the qualifiers on the conclusions are less restrictive
- apply to numbers on scales other than interval or ratio
Why no use NP statistics all the time?
- if assumptions are valid, parametric tests are more powerful
- more people are familiar with parametric tests
- parametric tests produce models that are more useful in many real-world solutions
classic interpretations of Cp and Cpk
- if Cp (Pp) > Cpk (Ppk), the process is not centered at its target value. if they are approximately equal, then the process is centered
- if Cp or Cpk < 1, process is incapable
- if Cp or Cpk are between 1 and 1.33, process is barely capable
if Cp or Cpk > 1.33, process is capable
for six sigma quality, Cp and Cpk = ?
- Cp = 2
- Cpk = 1.5
What is z.bench?
the combination of the two tail areas into one and thier corresponding Z or sigma value
dpu
defects per unit (average) = number of defects / number of units
dpmu
defects per million units = (dpu) * (10^6)
dpo
defects per total opportunities (i) = number of defects / total # opportunities = dpu / opportunities per unit
dpmo
defects per million opportunities = dpmu / # opportunities per unit = (dpo)8(10^6)