AICP Review - Quantitative

Methods in Urban Planning

David Schoen, AICP

Department of Urban Planning
Ball State University

[email protected]
765-285-5871
Presentation Goals
z Vocabulary
z BasicStatistical Methods
z Shotgun approach
– Brief overviews
Vocabulary
z Constant
– An unchanging value (pi, or π: 3.1417)
z Variable
– A value that can take on different
quantities
z Levels of Measurement
– Nominal, Ordinal, Interval, Ratio
Vocabulary
z Levels of Measurement
– Nominal
• Non-ordered categories (redhead, blonde)
– Ordinal
• Ordered categories (Excellent, moderate, poor)
– Interval
• Equal intervals, arbitrary, 0 (temperature)
– Ratio
• Zero based, used for proportions
Vocabulary
z Histogram
– A graphic depiction of the
distribution of a variable
z Frequency Distribution
– The pattern or shape
formed by the group of
measurements in a
distribution
Vocabulary
z Ogive Curve
– A frequency distribution curve
in which the frequencies are
cumulative
z Scatter Plot
– Also known as scattergram or
scatter diagram. A two
dimensional graph representing
a set of bi-variate data
Vocabulary
z Sample
– A subset of a population
• Random (probabilistic)
• Stratified (separate groups)
• Systematic (every kth element)
• Cluster (correct mix group)
Population

Sample
Vocabulary
z Parametric
– Assumes that the data is drawn from a
normal distribution
– the data is measured on an interval
scale
– parametric tests make use of
parameters such as the mean and
standard deviation
– T-test, ANOVA, regression, correlation
Vocabulary
z Non-parametric
– Not normally distributed data
– Nominal and ordinal data
– chi2
Vocabulary – Central Tendency
z Mode
– The most frequently occurring value of
a group of values
z Mean
– is the sum of all values divided by the
number of observations
− x1 + x2 + x3 + ...xn
x=
n
Vocabulary – Central Tendency
z Median
– the 50th percentile, such that half the
values are above the median and half
the values are below
Vocabulary – Dispersion
z Range
– the maximum value minus the
minimum value of a data set n

∑ i
( x − x ) 2

z Variance s2 = i =1
n
– Population
n
– Sample ∑ (x − x)
i
2

s2 = i =1
z Standard Deviation n −1
– the square root of the variance
n

∑ (x − x)
i
2

s= i =1
n
Statistics – Basic Concepts
z Normal Distribution
– A theoretical frequency distribution for
a set of variable data, usually
represented by a bell-shaped curve
symmetrical about the mean
Statistics – Basic Concepts
z Cross-tabs
– Tables showing the joint distribution of
two or more categorical variables
• e.g., cell counts, cell percentages, expected
values, and residuals
– Various measures of association,
between the variables, can be obtained
– 2-way and 3-way
Statistics – Basic Concepts
z Cross-tabs
– 2-way
Sex * Town Government Crosstabulation

Town Government
Excellent Good Average Fair Poor Total
Sex Male Count 13 25 21 9 10 78
% within Sex 16.7% 32.1% 26.9% 11.5% 12.8% 100.0%
% within Town
48.1% 45.5% 38.2% 39.1% 47.6% 43.1%
Government
% of Total 7.2% 13.8% 11.6% 5.0% 5.5% 43.1%
Female Count 14 30 34 14 11 103
% within Sex 13.6% 29.1% 33.0% 13.6% 10.7% 100.0%
% within Town
51.9% 54.5% 61.8% 60.9% 52.4% 56.9%
Government
% of Total 7.7% 16.6% 18.8% 7.7% 6.1% 56.9%
Total Count 27 55 55 23 21 181
% within Sex 14.9% 30.4% 30.4% 12.7% 11.6% 100.0%
% within Town
100.0% 100.0% 100.0% 100.0% 100.0% 100.0%
Government
% of Total 14.9% 30.4% 30.4% 12.7% 11.6% 100.0%
 Statistics – Basic Concepts
z Cross-tabs
– 3 way, used to test a 2-way cross-tab
Sex * Town Government * Age group Crosstabulation

Town Government
Age group Excellent Good Average Fair Poor Total
65 or over Sex Male Count 1 11 1 2 2 17
% within Sex 5.9% 64.7% 5.9% 11.8% 11.8% 100.0%
% within Town Sex * Town Government * Age group Crosstabulation
33.3% 50.0% 25.0% 40.0% 66.7% 45.9%
Government
Town Government
% of Total 2.7% 29.7% 2.7% 5.4% 5.4% 45.9%
Age group Excellent Good Average Fair Poor Total
Female Count 2 11 3 19 to 29 3 1 20
Sex Male Count 1 3 3 3 1 11
% within Sex 10.0% 55.0% 15.0% 15.0% 5.0% 100.0%
% within Sex 9.1% 27.3% 27.3% 27.3% 9.1% 100.0%
% within Town
66.7% 50.0% 75.0% 60.0% 33.3% % within Town
54.1%
Government 50.0% 50.0% 23.1% 60.0% 50.0% 39.3%
Government
% of Total 5.4% 29.7% 8.1% 8.1% 2.7% 54.1%
% of Total 3.6% 10.7% 10.7% 10.7% 3.6% 39.3%
Total Count 3 22 4 5 3 37
Female Count 1 3 10 2 1 17
% within Sex 8.1% 59.5% 10.8% 13.5% 8.1% 100.0%
% within Town
% within Sex 5.9% 17.6% 58.8% 11.8% 5.9% 100.0%
100.0% 100.0% 100.0% 100.0% 100.0% 100.0%
% within Town
Government
50.0% 50.0% 76.9% 40.0% 50.0% 60.7%
% of Total 8.1% 59.5% 10.8% 13.5% 8.1% Government
100.0%
% of Total 3.6% 10.7% 35.7% 7.1% 3.6% 60.7%
Total Count 2 6 13 5 2 28
% within Sex 7.1% 21.4% 46.4% 17.9% 7.1% 100.0%
% within Town
100.0% 100.0% 100.0% 100.0% 100.0% 100.0%
Government
% of Total 7.1% 21.4% 46.4% 17.9% 7.1% 100.0%
Statistics – Basic Concepts
z Z-score
– A measure of the distance, in standard deviation
units, from the mean
– Used to determine probability if something
would, or would not, happen
– Unit-less value!

x−x
z=
s
Statistics – Basic Concepts
z Hypotheses
– Research hypothesis
• H1 – what we really want to prove
– Null hypothesis
• H0 – what we statistically use, and really want
to prove wrong
– It’s easier to statistically prove
something wrong, thus the null
hypothesis
Statistics – Basic Concepts
z Hypotheses
– Type 1 error
• Reject null hypothesis when it is actually
true
– Type 2 error
• Fail to reject null hypothesis when it is
actually false
Statistical Tests
z Chi2 (χ2)
– Test for a relationship (dependence)
between two nominal or ordinal based
variables.
– Examines the joint probabilities between
the two variables
Statistical Tests
z Chi2 (χ2)
Statistical Tests
z Spearman Rank Correlation
– Correlate two ranked data sets

6∑i =1 Di2
N

rs = 1 −
N ( N 2 − 1)
Statistical Tests
Spearman Rank Correlation example

Sum of d2 is 74

rs = 1- (6(74) / 990)) = 1 – 444 / 990 = .55

Z = .55 (√10-1) = 1.65
1.65 < 1.96 … cannot reject hypothesis of independence
Crime and poverty not related

Use Z score to test significance (the probability that an observed outcome is

not due to chance alone )
Z = rs / σr where σr = 1 / √N-1
Z* = 1.96 at .05
Statistical Tests
z t-test
– A t-test allows us to compare the means of two
groups and determine how likely the difference
between the two means occurred by chance
– The calculations for a t-test requires three
pieces of information:
• the difference between the means (mean difference)
• the standard deviation for each group
• and the number of subjects in each group
Statistical Tests
z There are three types of t-tests and each is calculated
slightly differently.

– A correlated (or paired) t-test is concerned with the difference

between the average scores of a single sample of individuals
who is assessed at two different times (such as before
treatment and after treatment) or on two different measures. It
can also compare average scores of samples of individuals
who are paired in some way (such as siblings, mothers and
daughters, persons who are matched in terms of a particular
characteristics).

– An independent t-test compares the averages of two samples

that are selected independently of each other (the subjects in
the two groups are not the same people). There are two types
of independent t-tests: equal variance and unequal variance.
Statistical Tests
z Analysis of Variance (ANOVA)
– Extension of the t-test
– Identify relationship between two
variables
• Variable X is nominal scale (groups)
• Variable Y is interval scale
– Assume that by grouping one variable
we will get a better estimate of Y
Statistical Tests
z Analysis of Variance (ANOVA)
– If relationships exists, we can get better
estimate of Y using group means
Statistical Tests
z Correlation
– A correlation is a relationship
between two variables.
– The data can be represented by the
ordered pairs (x, y) where x is the
independent, or explanatory, variable
and y is the dependent, or response,
variable
Statistical Tests
z A correlation coefficient indicates the
extent to which two variables are related
– Known as Pearson product-moment coefficient
z It can range from -1.0 to +1.0
z A positive correlation coefficient
indicates a positive relationship, a
negative coefficient indicates an inverse
relationship
z Correlation CANNOT be equated with
causality.
Statistical Tests
Ranges of correlation
Statistical Tests
z Correlation Coefficient

Sx & Sy is std. deviation

– Squaring r gives us R2
• Calculates how much one variable explains the other
– R2 is an extremely useful measure
• Example
• r = -.44102 = 0.194, thus var1 accounts for 19.4% of the
variability of var2. That leaves 80.6% to be explained…
- But still does not indicate causality!
Statistical Tests
z Regression
– Explore the nature of relationships
between variables
– Simple (two variables) and multiple
regression
– Always one dependent and 1 or more
independent variables
Statistical Tests
z Regression
z Issues:
– What is the association between Y and X?
– How can changes in Y be explained by
changes in X?
– What are the functional relationships
between Y and X?
A functional relationship is symbolically written as:
Y = f(X)
Statistical Tests
Simple Regression

Y = b0 + b1X
b0 is the intercept,
b1 is the slope.
Statistical Tests
z Multiple Regression
– Just an extension of simple linear regression!

Y = b0 + b1X
Y = b0 + b1X1 + b2X2 + … bnXn
Planning Models - Population
z Population Projections
– Historical
• Historical trends
– Component (Cohort Survival)
• Birth and death rates by cohorts
• Migration can be included
– Ratios
• Use larger unit (state, region) to ratio with
Planning Models - Population
z Population Projections – terms
– Estimates
• Current population levels
– Projections
• Future population levels
– Forecasts
• Selected projection (judgmental)
– Migration
• Migration = Pop2000 – Natural increase1990-2000
Planning Models - Population
z Population Projections - Historical
– Linear
Pt + n = Pt + b ( n )

m +1

∑(P − P
t =2
t )
t −1

b=
m
Planning Models - Population
z Population Projections - Historical
– Exponential
• r is rate of change

Pt + n = Pt (1 + r ) n

1 Pt − ( pt − 1)
r= ∑
m Pt − 1
Planning Models - Population
z Population Projections - Historical
– Modified Exponential
• Ceiling to growth
• V is % of unused capacity

Pt + 1 = K − [( K − Pt )(v ) n ]

1 K − Pt
v= ∑
m K − Pt − 1
Planning Models - Population
z Population Projections – Historical
– Gompertz
• Similar to modified exponential but more rigorous
Planning Models - Population
z Population Projections – Historical
– Ratio your community to another larger known
quantity

Concept

Rearranging terms
Planning Models - Population
z Population Projections – Cohort
– Population increases with natural increase
and migration
– Pt+n = Pt + N + m
– Pop2006 = Pop2005 + (B-D) + net migration
Planning Models - Population
z Population Projections – Cohort
– Assumptions
• New births: women 15-44 years of age
• Births divided by some ratio (m/f)
• Cohorts could be anything
– 0-1
– 1-5
– 5-9
– 10-14
– 15…
Planning Models - Location
z Gravity Models
– Based on attractiveness of a location as a
destination and difficulty getting there

F – attraction
K – regional index
P – mass PA PB
d - distance FAB =K 2
d
Planning Models - Location
z Gravity Models
– Unconstrained model for generating trips
between two locations

T – total trips from zone i to j

H iα S βj
Tij = K H – trip makers
d ijλ S – Opportunities
d – Cost of travel
K, α,β,λ – exogenous parameters
Planning Models - Location
z Gravity Models
– Constrained Model
• Limit of disposable income, available land etc.

S βj
d ijλ Probability of a trip between ones i and j
p ji =
S βj
∑ dλ
ij
Planning Models - Location
z Gravity Models
– Residential Location
Planning Models - Location
z Gravity Models
– Lowry Model
• Classic application of a gravity model
• Assigns secondary employment and
population to zones
• Driving force is location of basic
employment
Planning Models - Location
z Gravity Models
– Lowry Model
• Step 1: Calculate population (N) dependent on basic
employment
• Step 2: Calculate non-basic employment (E) dependent
on population
• Step 3: Calculate population (N) dependent on non-
basic employment
• Step 4: Repeat steps 3 and 4 until minimum increase is
reached
Ei Ni
d ij2 d ij3
Nj = E sj =
Ei Ni
∑ ∑ d2 ∑ ∑ d3
ij ij
Planning Models - Economics
z Basic sector
– Those economic sectors that produce
goods for export from the region
– Brings in new $ to the region
z Nonbasic sector
– Those economic sectors that produce
goods for local consumption
Planning Models - Economics
z NAICS
– The North American Industry Classification System
(NAICS) has replaced the U.S. Standard Industrial
Classification (SIC) system
– 6 digit code hierarchy
– All industry is assigned to a consistent set of
categories

– http://www.naics.com/index.html
Planning Models - Economics
z Location Quotient (LQ)
– Estimates if an industry is a basic
industry
– Compares local industry employment to
national employment for the same sector
– If LQ is > 1.0 then industry is an exporter
– If LQ is < 1.0 then imports are implied
Planning Models - Economics
z Location Quotient (LQ)
– Simple ratios of employment

LQ =
E / ERj R > 1
E /E Nj N
<

ERj – regional employment for industry j

ENj – national employment for industry j
ER – Total regional employment
EN – Total national employment
Planning Models - Economics
z Minimum Requirements Approach
– Estimates if an industry is a basic
industry
– Compares employment to a similar
region rather than the nation
• Similar region output is just sufficient to
satisfy local requirements (i.e. minimum)
– If region produces more than minimum
than it is an exporter industry
Planning Models - Economics
z Employment Projections
– Constant Share approach
• Assumes the local region maintains a
constant share of employment to a larger
region
• Larger regions typically have employment
projections
– Simple matter to ratio down to local economy
– R is regional growth rate
Planning Models - Economics
z Employment Projections
– Constant Share approach

e = (1 + R )e
i
t'
i
t −t '
i
t

E ti ' − E ti
R ti − t ' =
E ti
Planning Models - Economics
z Employment Projections
– Shift-Share technique
• Assumes local economy does not keep
constant share of larger region
• Assumes local sectors grow faster, or more
slowly, so share changes
• Hence, employment shifts into, or out of, a
region
Planning Models - Economics
z Employment Projections
– Shift-Share technique

e = (1 + R
i
t'
i
t −t ' + s )e
i
t −t '
i
t

sti−t ' = rit −t ' − Rit −t '

Planning Models - Economics
z Employment Projections
– Economic Base Model
• Assumes basic sector is the prime cause of local
economic growth
• Assumes changes in basic sector can be converted into
changes for total economy by applying a base multiplier
ratio
• Changes in basic employment can be obtained from
projections

eTt
BM = t Ratio of the total employment
bT to the total basic employment
Planning Models - Economics
z Input-Output Analysis
– A technique used in economics for tracing
resources and products within an economy
– The quantities of input and output for a given
time period, are entered into an input-output
matrix
• one can then analyze what happens within, and
across, various sectors of an economy where growth
and decline takes place and what effects various
subsidies may have
Planning Models - Economics
z Input-Output Analysis
– Example input-output matrix