Transcript week9_descriptives

Descriptive Statistics &
SPSS introduction
Nan Yu
COMM 420.8
Basic steps in conducting research
1.
2.
3.
4.
Literature review
Raise questions or make hypotheses
Design of a study
Collect evidence (qualitative evidence or
quantitative evidence)
5. Analyze your evidence (data)
6. Draw conclusion from the analysis
Tasks for the next part of your group project
Collect evidence
Analyze your data
Results and Conclusion
Prepare for the presentation
Data Analysis
Requirement
Understand statistics
Be able to use statistical software
Recommendation
Take other stat classes or data analysis classes
if you are truly interested in this area.
This class will review very basic skills and
principles about data analysis
A review – levels of measurement
Nominal/Categorical
e.g. Gender, Race, Favorite Music Genre
Ordinal
In a typical month, how many movies do you
rent on video?
None
1-2
3-5
5-20
More than 20
Interval/Ratio
I believe in love at first sight.
Strongly Disagree
1
2
Neutral
3
4
Strongly Agree
5
On a typical day, how many hours of
television do you watch?_______
How many siblings do you have?______
Text Book p.327, Table 14.1
Descriptive statistics
--reduce and simplify the number to
interpret the results
Inferential statistics
--make a judgment of what you observe in
the sample can be generalized to the
population from which the sample was
drawn
Text Book p.327. Table 14.1
Nonparametric
Nominal, Ordinal
Parametric (continuous)
Interval, Ratio
Text Book p.327. Table 14.1
 Univariate
One variable
 Bivariate
Two variables
 Multivariate
Three or more variables
Note: Sometimes scholars combine bivariate and
multivariate so the distinction would be one variable
(univariate) versus multiple variables (multivariate)
Descriptive Stats
What to describe?
What is the “center” of the data?
How the data vary? (variability)
Measures of central tendency
Mean
Mode
Median
Mean
Another name of average
If describing a population, denoted as μ,
the Greek letter “mu.”
If describing a sample, denoted as, X
called “x-bar.”
“Balance point” or the “value center” of the
distribution.
 Only use for interval/ratio.
Calculating the Mean
Imagine that we have a population of 5 objects
with heights of 2, 4, 6, 8, and 10 inches.
X1 = 2
X2 = 4
X3 = 6
X4 = 8
X5 = 10
The mean of the population is:
X 1  X 2  X 3  X 4  X 5 30
X

6
N
5
Sum of all the numbers divided by the number of
observations contributing to that sum
Mode
Most frequently occurring score in a
distribution.
One data set can have many modes.
Appropriate for all types of data, but
most useful for categorical data.
Mode
Hometown
Number of students
Pennsylvania
New York
New Jersey
Ohio
Maryland
Other States
What is the mode in this case?
Median
Middle score in a distribution (50%
above, 50% below)
Can use with ordinal and interval/ratio,
but not with nominal
Median
E.g., Scores:
4
9
2
2
Rank order them:
1
2
2
4
1
8
10
7
8
9
Median
9
9
7
10
Summary
Choice of descriptive measures depends entirely on the
level of measurement for a particular variable.
A normal distribution
Symmetric, bell-shaped curve.
Most values fall around the mean, but some values are
smaller, and some are larger.
mean
median
mode
But distributions are not always normal
(p.345)
Skewness: How far the peak is from the center of the distribution
What happens if the distributions
are not normal?
Impact of skewness on measures of Central Tendency
Mode
Median
Mean
Right (Negative) Skew
Mode
Median
Mean
Left (Positive) Skew
Distribution Kurtosis (p.345)
Normal (bell-shaped)
peaked
flat
Dispersion
The mean, mode and median are not enough to understand the
distribution of a variable.
Variability
How variable are the scores in a
distribution?
Range
Standard Deviation
Variance
Range
Standard Deviation
A measurement of variability that indicates
how much all of the scores in the
distribution typically deviate from the mean.
Calculating Standard Deviation
Variance
What is the standard deviation for this
distribution?
SPSS
Statistical Package for the Social Science
A list of other available statistical software,
p.332
Please download the file “Class Survey”
from ANGEL (week 9) and save it on your
desktop
Introduction to SPSS
Opening a data file
STEP 1
Double click on the
“class survey.sav” file.
STEP 2
Select "Edit," then "Options"
STEP 3
Select "Display names" and "File," then
Click "OK"
Introduction to SPSS
Variables
Cases
(People)
Variable Labels
Place cursor
over variable
name
To See Value Labels
Click Value Label button
Sorting Cases
Click Data -> Sort Cases
Select Variable, Move to Right
Box, Click on OK when ready.
Data Editor in Variable View
Variables Listed
Down
Information About
Each Variable Across
Values
Frequencies/Descriptive Statistics
Requesting frequencies
Analyze -> Descriptive Statistics
-> Frequencies
Select the variables you
want to analyze, and use
the arrow keys to move them
into the "Variables" box.
Click "Statistics."
Select the statistics
you want,
and then click
"Continue"
and then "OK."
Output Window
You can double-click
on the chart and edit
the appearance…
You can copy and
paste the chart into
other applications...
Notice the
new window...
Graphs
A. Select "Graphs" then
"Histogram"
B. Put “tvhours" in the
variable box.
C. Select "Display normal curve"
D. Select "OK"
In Your Output Window
You can double-click
on the graph and
edit the appearance.
You can copy and
paste the graph into
other applications.
In Your Output Window
As new procedures are
run, they are appended to
the output and added to
the outline on the left of
the screen.
You can navigate around
your output by selecting
from the outline menu.
You can also delete any
output by selecting it and
then clicking the delete
key.
Entering Your Own Data
A. Open Data (Type in Data)
B. Go to “Variable View”
and name and label variables
Variable Names Listed Down
(No spaces)
Values Dialog Box:
Click “Add” after
each new value
label that you
include.
Label for the Variables
Labels for the Values
C. Enter data in “Data View”
Cell Editor
Case
Cell
1. Click on a given cell where you want to add data.
2. Type the value you want to include in that cell. The value will
appear in the cell editor.
3. The TAB key will take you to the next variable in the same case.
4. The ENTER (return) key will take you to the next case.
D. Save your data after each case
Oh-my-gosh!
I think I forgot
to save the data!
I lost all my data
because I forgot
to save it.
Ha ha ha.
Too bad for them.
My data is safe!
Saving Your Data and Output
Extensions Used for File Types
Date Files: .sav
e.g., "mydata.sav"
Output Files: .spo
e.g., "myoutput.spo"
Saving Your Data File
A. From within the Data Editor, select
"File" then "Save as"
B. Select your floppy disk.
C. Type in the name of your data file
(with the extension .sav).
D. Select "Save."
Note: After you have saved your data
for the first time, you can save it
periodically by selecting "Save" rather
than "Save as."
Saving Your Output File
A. From within the Output window,
select "File" then "Save as"
B. Select your floppy disk.
C. Type in the name of your output
(with the extension .spo).
D. Select "Save."
Note: After you have saved your
output for the first time, you can
save it periodically by selecting
"Save" rather than "Save as."
Z-scores
If two measures were used to measure the
liking toward “Nan Yu”
not likable 1 2 3 4 5 likable
not nice 1 2 3 4 5 6 7 nice
The raw scores are not comparable because
they are measured differently.
So we need to standardize the scores…
What we do is…
Translating each individual score. The
transformed scores will necessarily have a
mean of zero and a standard deviation of one.
The standard score indicates how many
standard deviations an observation is
above or below the mean.
-3
-2
-1
0
1
2
3
Requesting Z-scores from SPSS
Open Class Survey.sav
Go to Analyze  Descriptive Statistics 
Descriptive
Z-scores in SPSS
A. Put “tvpm" in the
variable box.
B. Check “Save
standardize values as
variables, can click OK
Z-scores in SPSS
Practice
Request z-scores for “tvsit.”
Sort “Ztvsit” in a descending order.
In-Class Demo
Run frequencies on “wwwbuy”
Request mean, median, mode, standard
deviation, range, and skewness, and
historgram with normal curve.
Does it look at a normal distribution? Base
on the reported mean, median and mode,
explain the skewness of the distribution of
“wwwbuy.”
In-class Demo answer
You can compare your SPSS output with
the one on ANGEL (class demo
answer.spo)
No, it doesn’t look at a normal curve. The
mean is higher the median and the mode.
So it is a positive (left) skew.