“The government is very keen on amassing statistics. They will collect them, raise them to the nth power, take the cube root, and prepare wonderful diagrams. But you must never forget that every one of these figures comes in the first instance from the village watchman, who just puts down what he pleases.”
Sir Josiah Stamp,
Research methods & techniques (see http://krypton.mnsu.edu/~tony/courses/609Design.html%20)
· Control/randomization
· Cross-sectional & longitudinal design
· Quasi-experimental design
· Operational definition
· Variables: independent (x), dependent (y), intervening
· Relationships: linear, curvilinear, bivariate, multivariate
· Spurious relationships
· Validity: internal (control), external (representativeness)
· Reliability (replicability)
Collecting Data
· Nominal (categories)
· Ordinal (rank order)
· Interval & ratio
Analyzing Information
Descriptive Statistics (see http://krypton.mnsu.edu/~tony/courses/609StatsI.html%20)
· Parameter (scale factor) vs. statistic (summary measure)
· Range & variance
· Measures of central tendency:
o Mean, median, mode
o Standard deviation & coefficient of variation (sd/mn)
· Kurtosis (platykurtic [flat], leptokurtic [peaked], & skew)
Significance Testing
· Null hypothesis
· Significance level
· Type I (reject true null) & Type II (accept false null) error
· One-tail/two-tail test
· Sample size
· Standard error of mean (confidence interval)
· T-test of difference of means
· Central limit theorem
· Significance vs. measures of association
Nonparamteric Statistics
· Chi-square
· Rank-order correlation (Spearman’s r)
Parametric Statistics (see http://krypton.mnsu.edu/~tony/courses/609StatsII.html )
·
· General linear model (regression line)
· Homoscedasticity & independence
· Scattergram
· Correlation (y=a+bx+u)
· Multiple regression (y=a+b1x1+b2x2…+u)
· Partial correlation coefficient & beta weight
· Analysis of variance
· Factor analysis
· Logit/analysis of covariance
· Data transformations & z-scores
Choosing Inductive Statistics:
X
categorical interval
interval Y |
ANOVA |
R, r |
categorical |
Chi-Square |
Logit/ANCOVA |
Teri Renner, Statistics Unraveled (Washington, DC: ICMA, 1988)
(see http://krypton.mnsu.edu/~tony/courses/609Pop.html )
· Simple
o Exponential
o Modified exponential
o Gompertz
· Composite
· P = P + NI + NM
· Net Migration (NM) = inmigration – outmigration
· Natural Increase (NI) = births – deaths
· Births = Women of childbearing age * net reproduction rate
· Net reproduction rate = fertility – infant mortality
· Replacement rate
· Cohort
· Survival rate
· Basic & nonbasic employment
· Basic (export) multiplier
· Methods
o Survey
o Location quotient
o Minimum requirements
o assumptions
· Location Quotient
o LQ = local share / national share
· National share
· Industry shift
· Employment = Total Output / Output per Worker
· Allocate base employment
· Allocate basic housing
· Allocate service employment
· Allocate service housing
Michael Greenberg, et alii, Local Population and Employment Projection Techniques (Rutgers NJ: CUPR Press, 1978)
(see http://krypton.mnsu.edu/~tony/courses/609Benefit.html )
· All costs & benefits are explicit, finite, and quantifiable
· Comparable projects
· Optimality
· “do-nothing” option
· No complementarities
· Market prices
· Dead-weight loss
· Intangibles
· Differential impacts (use weights)
· Externalities (sum weighted effects)
· Project constraints (linear programming)
· Indivisibilities (treat as a constraint)
· Cost vs. “negative benefit”
· Time-value of money
· Discounting (PV = FV / (1+i)^n)
·
· Inflation
(see http://krypton.mnsu.edu/~tony/courses/609Fiscal.html )
· Case Study Method
1. Most accurate
2. Most expensive
· Average Cost Methods (Per Capita Multiplier; Service Standards)
1. Steady growth (+/-)
2. Past predicts the future
3. Usually for residential projects
· Marginal Cost Methods (Comparable City; Employment Anticipation)
1. Large project—past is no guide
2. Significant unused capacity or no remaining marginal capacity
J.A. Schofield, Cost Benefit Analysis in Urban & Regional Planning (London: Allen & Unwin, 1987)
Edith Stokey & Richard Zeckhauser, A Primer for Policy Analysis (NY: WW Norton, 1978)
(see http://krypton.mnsu.edu/~tony/courses/609CIP.html )
Essentially, a political process of trying to determine who will get what when and how:
· Generate a list of projects (and their costs)
· Rank the projects
o Criteria
o Weights
· Allocate available funds
o By fund
o Debt capacity
· Capital reserve account
(see http://krypton.mnsu.edu/~tony/courses/609PERT.html )
Revise PERT estimates to arrive it minimum time, neither wasting nor sparing costs.
· All activities have distinct beginning and end points
· Time estimates are well defined
· Resources must be able to be shifted from one activity to another
· Linear programming may be used, if cost is direct function of time and there are no indivisibilities in activities
· No discounting, indirect costs, or time benefits
Krueckeberg & Silvers, Urban Planning Analysis: Methods & Models (NY: Wiley, 1974)
§ Determine organizational goals
§ Convert goals into performance objectives for individuals
§ Measure performance against stated results or standards
§ “Horizontal comprehensive”—compare alternatives to determine which best contributes to larger program objectives
§ Organize spending by program rather than by function
§ Extend programs far enough into future to determine spending implications (justify costs against established needs)
§ Subject all programs to explicit quantitative analysis (measure cost of meeting needs)
§ Starts from assumption of “zero funding” for each year
§ “Vertical comprehensive”—compare alternative funding levels for same program
§ Shifts burden of proof to each manager to justify why money should be spent
Last revised: