Last class

1
URBP 204 A Class 4

• Last class
• Descriptive Statistics
• THIS CLASS
• Excel and SPSS Tutorial (Tutorial 1)
• Census Overview
• Normal Distribution
• Hypothesis Testing

Note the class notes summarize Salkind (2004)
Chapters 6 to 8
2
Census Overview
Source U.S. Census
3
Five Wounds Brookwood Terrace Project
Census tract numbers 5014.00, 50515.01 and
5015.02 lie fully within the FWBT area while
part of census tract number 5036.01 falls within
the FWBT area.
Source Mike Reilly, Instructor URBP 179, Fall
2004
4
Census Overview Go to www.census.gov Go to
American Fact finder Go to Maps Go to
Reference Maps Go to Select a boundary
grouping Select 2000 Census Tracts and Blocks
Select CA as State and 95112 as Zip Code Go
to Reposition on Select A street address or
zip code Type E St James and N 26TH St under
Street Address San Jose
under City Choose CA under state Zoom out to
2.8 miles across
How to obtain files for socio-economic, housing
characteristics ? Datasets - Decennial Census
SF1 - Demographic (DP-1) and Housing (QT-H1).
Also explore SF-2 and SF-3.
5

• Normal Distribution
• Features
• Mean, median, and mode are the same
• Symmetrical about the mean
• Tails are asymptotic- tails tend to reach x-
  axis, but never touch
• Example with Mean 100
• Standard Deviation 10
• Between mean and 1 std. dev (between 90 and 110)
  34.13 cases (on either side) - total 68.26
• Between mean and 2 std. dev (between 80 and 120)
  47.72 cases (on either side) - total 95.44
• Between mean and 3 std. dev (between 70 and 130)
  49.87 cases (on either side) - total 99.74
• Go to the following link for a dynamic example
• http//www-stat.stanford.edu/naras/jsm/NormalDens
  ity/NormalDensity.html

Note the class notes summarize Salkind (2004)
Chapters 6 to 8
6
Z Score How to compare different
distributions? Z (X- X) / s z z score X
raw score X mean s sample standard
deviation Example Variable Height Other
Examples accidents per month crimes reported
per year permits issued per year Raw score 78
inches (6 6 feet) Mean 72 inches (6 feet) s
6 inches z (78-72) / 6 6 / 6 1 That means
78 inches is 1 std. dev. above the mean score is
1 z score above the mean. If raw score 66
inches z (66-72) / 6 -6 / 6 -1 That means 66
inches is 1 std. dev. below the mean score is 1
z score below the mean.
Note the class notes summarize Salkind (2004)
Chapters 6 to 8
7
Z score continued Larger the z score, further
away from the mean is the raw score. 84.13
(50 34/13) of all scores fall below z score
of 1. For raw score of 78 inches, when the mean
72 inches and s 6 inches Probability of the
score greater than 78 inches is 16 (100-84).
Note the class notes summarize Salkind (2004)
Chapters 6 to 8
8
Hypothesis Does watching good examples of high
density residential development have an effect on
a persons attitude towards high density
residential development as measured by the
density attitude scale? - Research question Null
hypothesis no effect no relationship between
variables H0 ?before ?after Research
Hypothesis effect there is relationship between
variables H1 Xbefore Xafter Non-
directional research hypothesis or H1 Xbefore
Xafter Directional research
hypothesis

Note the class notes summarize Salkind (2004)
Chapters 6 to 8
9
Null and research hypothesis Null hypothesis
equality Research hypothesis inequality Null
hypothesis population Research hypothesis
sample Null hypothesis indirectly tested (as
dont survey the entire population) Research
hypothesis directly tested (as the entire sample
is included in the analysis) Null hypothesis
Greek symbols Research hypothesis Roman
symbols
Note the class notes summarize Salkind (2004)
Chapters 6 to 8
10

• Characteristics of a good hypothesis
• Declarative, clear and forceful not a question
• Presents an expected relationship between
  variables
• Reflects the theory / literature
• Brief and to the point
• Testable

Note the class notes summarize Salkind (2004)
Chapters 6 to 8
11
Statistical Significance Research question Does
watching good examples of high density
residential development have an effect on a
persons attitude towards high density
residential development as measured by the
density attitude scale? Research
hypothesis Watching good examples of high density
residential development has positive effect on a
persons attitude towards high density
residential development as measured by the
density attitude scale. How can we be sure that
the increase is not by chance but only because of
watching the movie? - Level of risk associated
with being incorrect Significance level risk
associated with not being 100 confident that
what you observe in an experiment is due to the
treatment or what was being tested.. (Salkind,
p.143)
Note the class notes summarize Salkind (2004)
Chapters 6 to 8
12

• 4 Possibilities
• Null hypothesis is true and you fail to reject
  (Salkind book accept) it.
• Great! There was actually no difference in
  scores in the entire population and you also
  found none in your sample.
• Null hypothesis is true but you rejected it
  Type 1 error
• If level of significance 0.05 5 chance that
  you will reject the null hypothesis when it is
  true.
• Made a mistake! Actually there was no
  difference between before and after score in the
  entire population but you found a difference in
  your sample.
• Null hypothesis is false and you fail to reject
  (Salkind book- accept) it - Type 2 Error. Accept
  equality when inequality exists. As sample size
  increases the probability of type 2 error
  decreases. Other reason sample not
  representative of the population - reliability
  issue.
• Made a mistake! Actually there was a
  difference in scores in the entire population,
  but you did not find any difference in your
  sample!
• Null hypothesis is false and you reject it
• Great! There was a difference in scores in
  the entire population and you found a difference
  in your sample too.

Note the class notes summarize Salkind (2004)
Chapters 6 to 8