Page 446 - Six Sigma Demystified

426 Six SigMa DemystifieD

70. Some useful purposes for model transformations include
A. to calculate parameters that are not easily measured directly.
b. to understand the effect of parameter settings on process variation.
c. to stabilize the variance to improve parameter estimates.
d. All the above are true.
71. When transforming data to stabilize the variance,
A. we can plot the log of the standard deviation at each experimental condition
against the log of the mean of the experimental condition to determine a suitable
transform function.
b. we should verify that the transformed data have a stabilized variance.
c. we will use the transformed response in place of the original response when
estimating significance of factors or the model.
d. All the above are true.

72. Positive correlation implies that
A. the dependent variable improves as the independent variable increases.
b. the dependent variable decreases as the independent variable increases.
c. the dependent variable increases as the independent variable increases.
d. the independent variable decreases as the dependent variable increases.
73. Strong correlation implies that
A. the dependent variable improves as the independent variable increases.
b. there is little error between the predicted response and the actual response as the
dependent variable increases.
c. the dependent variable increases rapidly as the independent variable increases.
d. All the above are true.

74. In linear correlation analysis, if the slope of the line is low, then
A. the dependent variable is not well predicted by the model.
b. there is weak correlation between the variables.
c. as the independent variable changes, there is a small change in the dependent
variable.
d. All the above are true.

75. If the cycle time of a process is predicted by cycle time = 5.25 × (number of
items) + 4.3, with a correlation coefficient R of 0.8, then it is fair to say that
A. we can predict cycle time with no error.
b. only the number of items influences the predicted cycle time.
c. cycle time is definitely influenced by the number of items.
d. None of the above are true.

76. If the correlation coefficient R is 0.9 for a simple linear regression, then
A. 90 percent of the variation in y is explained by the regression model.
b. 81 percent of the variation in y is explained by the regression model.
c. 90 percent of the variation in y is explained by the variation in x.
d. approximately 95 percent of the error is explained by the variation in x.