Flashcards in Study Unit 6 Deck (13)

Loading flashcards...

1

## relevant range

### range from highest to smallest measures in set for a linear relationship; can't project beyond it

2

## High-low method

###
generate regression line based on highest and lowest observations

criticism- values may be abnormal

high $600 for 1300 hours

low $400 for 800 hours

increase $200 for 500 hours

variable cost=200/500=$.4

for low month VC= $.4*800=320

400 is total and 320 is VC, 80 is FC

y=80+.4x

3

## Perfect direct vs perfect inverse vs strong direct

###
direct r=1

inverse r= -1

strong direct r=.7

4

## r=0

### doesn't mean there isn't a relationship; relationship can't be expressed as a linear equation

5

## coefficient of determination

###
measure of fit between independent and dependent variables

-proportion of total variation in dependent variable that is accounted for by the indep variable

-closer r^2 is to 1 the more useful the indep var (x) is to predicting y

r^2=.64; 64% of new car sales explained by changes in income

6

## standard error

###
how well linear equation represents the data

vertical distance between the data points in a scatter diagram and the regression line

-closer the data points are to line, lower the std error

7

## learning curve analysis

###
reflects increased rate ppl perform tasks as they gain experience

-time req becomes shorter

-expressed as % of reduced time to complete task for each doubling of cumulative production

-80% l.c. means doubling will reduce cumulative avg unti completion time by 20%

8

## expected value

###
used by DM that is risk neutral

associating dollar amount w/ each possible outcome of probability distribution; outcome yielding the highest expected value is the optimal alternative

-decision alt is under mgr control

-state of nature is future event whose outcome mgr predicts

-payoff is financial result of comb of mgr decision and actual state of nature

calculated by multiplying probability of outcome by payoff and summing products

criticism- repetitive trials, decision involve 1 event

9

## sensitivity analysis

###
how sensitive expected value calculations are to the accuracy of the initial estimates

-useful to determine if should spend adtl resources

-used in cap budgeting, small changes of interest rates or payoff amnts can make difference in profitability

10

## simulation analysis

###
refinement of std probability theory

computer generates examples of results based on assumptions

-costly

11

## monte carlo technique

###
simulation to generate indiv values for a random variable

perf of quant model under uncertainty investigated by random selection of values for variables in model & calculating the soln

12

## Delphi technique

###
opinions from experts, summarize opinions, feeds summaries back to experst w/o revealing participants

-repeat until opinions converge to optimal solution

13