Community Health Statistics
Reading:
Book Chapter 2
Week 2 Lecture PPT
Week 2 Lecture Video
central tendency- refers to “average”, representative score
no single measure will always produce a central AND representaive value in every situation
3 methods
mean- interval data (age, income, IQ)
median- interval data w/ outliers, ordinal data (likert scale)
mode- nominal data (gender, weight classification)
mean- mathematical average
mean of a sample= M
mean of a population= μ
sensitive to extreme scores
weighted mean- overall mean for a combined group
to calculate you need 2 values
overall sum of the scores combine group
total # of scores in the combined group
median- score that divides a data set exactly in half
2 types
odd # of scores
even # of scores
mode= frequently occurring value
when 2 values occur the same # of times there is a bimodal distribution
2 modes= bimodal
>2modes= multimodal
levels of measurement:
nominal→ ordinal→ interval→ ratio
nominal- numbers are used as names for diff categories
ex) gender, female=1 male=2
ordinal- stands for order, characteristics of things being measured are in a meaningful order
very bad, okay, great, excellent
interval- related to distance or interval between 2 scored o a scale
ex) 3-6 year old, 6-9 year old
ratio- strongest data, has an absolute zero on scale, level of measurement defined by the presense of an absolute zero on the scale
ex) age, weight
variability- reflects how scored differ from one another, way to describe the distribution of data collected
descriptive tool
computing the range, standard deviation, and variance
measures of variability
range
standard deviation
variance
low vs high variability
range- the distance from the biggest score to the smallest score
standard deviation- represents the average amount of variability set of scores
CH 3 Computing Correlation Coefficents
correlation between X (age) and Y (height)
examines the relationship between variables and values ranging from -1.00 to +1.00
bivariate correlation (2 variables)
correlation is only used for continous variables
Visual picture: The Scatterplot
Types of Correlation Coefficients
positive correlation
direct
when one variable increases, the other variable also increases
negative correlation
indirect
when one variable increases the other variable decreases
To remeber
absolute value indicates strength
Curvilinear relationships
not all correlations are reflected by straight line (parabola)
some relationship may not be linear
EX) age & memory
Correlation Matrix
shows relationship between more than 2 variables
same variables across columns and down the arrows
Coefficient of Determination
percentage of variance in one variable that’s accounted by the variance in the other variable
Partial Correlations
the relationship between 2 variables is explored, but the impact of a third variable is removed from the relationship between the 2
Association VS. Causality
correlation between 2 variables can’t determine casuality
Chapter 5 The Normal Curve
distributions- pattern of how a sample of scores are arranged across all possible values, showing the frequency or likelihood of each score
Probability
normal curve provides us w/ a basis of understanding the probablitlity associated w/ any possible outcome
Normal curve
normal curve AKA bell shaped curve
mean, median, and mode are equal to one another
right and left side of curve are symmetrical
tails are asymptotic (never tough the horizontal axis, forever ongoing)
Standard Scores
scores that are comparable b/c they are standardized in units of standard deviation
the z score
raw score subtract by the mean and divide by the standard deviation
Calculate 2 z scores
area between the curve
subtract smaller area from larger area
Hypothesis Testing & Z scores
if the probability value is less that 5% chance of occurring, and you have a significant result (p<.05)
Fat & Skinny Frequency Distributions
variability- same average different distributions
skewness- measure of lack of symmetry, one tail is longer than the other
kurtosis- how flat or peaked a distribution appears
Chapter 6 Hypothesis
“educated guess”
take our research question into something we can actually test
population- group to which you would like to generalize your findings
sample- smaller, representative group of the population used to do the research
sampling error- measure of how well a sample represents the population
null hypotheses- populations and are written in greek symbols
research hypothesis- refer to samples
non directional research hypothesis
reflects a difference, but the direction isn’t specified
use a two-tailed test
here’s a difference in physical fitness scores between those who lift weights and those that don’t lift weights
directional research hypothesis
reflects a difference, and the direction is specified
one tailed test
Significance Testing
your level is risk i always .05
Mid-term Review
CH 1
Desciptive statistics
mean, median, mode
inferential stats used to make inferences based on a smaller group of data
types of data
nominal= category
ordinal= order
interval= from 0
ratio= scale
parameter: characteristics of an entire population
statistics: characteristics of a sample
definitions of variability
range, variance, standard deviation
CH 3
correlations= relationship between X & Y
relationship between -1.00 to +1.00
bivariate correlation (2 variables)
Pearson product-moment correlation examines relationship between 2 continuous variables
Rule of thumb
Correlation Matrix
same variables across columns and down the rows
Coefficient of determination
percentage of variable in one variance thats accounted for by the variance in the other variable
computed by squaring the correlation r2
coefficent of non-determiannce
1-r2
level of significance is usually set at 0.5
anything less than .05 is considered statistically significant
Write up:
df= n-2
r(3)= 0.985, p<.05 or p=.002
partial correlations
the relationship between 2 variables is explored, but the impact of a 3rd variable is removed from the relationship between the two
confounding variables
Association VS Casuality
correlations between 2 variables can’t determine casuality
CH 4
reliability- the measure that’s consistent in measuring what is was designed to measure
validity- test, scale, instrument measures what it’s supposed to
test-retest reliability- when you want to know whether a test is reliable over time
internal consistency reliability: when you want to know if the items on a test asses one and only one dimension
crobach’s alpha- measure of reliability that measures internal consistency, rane 0-1
parrallel-forms reliability: when you want to know id several forms of a test are reliable or equivalent
test can be reliable but not valid
test cant be valid unless it’s reliable
the normal curve
bell-shaped curve
tails are asymptotic (get closer to horixaontal axis but never touch it)
the z-score
?
skewness-
kurtosis-
CH 6
hypothesis
null
research
2 types of research hypothesis
non-directional- reflects a difference, but direction is not specified, use a 2 tailed test
directional- reflects a difference, and the direction is specified, use the one-tailed test
CH 6 statistical significance
degree of risk you’re willing t take that you’ll reject a null hypothesis when its actually true
conventional levels are set between .01 and .05
larger smaple size easier to detet sidgificnae
know difference between sample size and significance
siginigicance?
meaning?