SPSS Tutorial 1 What is SPSS

1
SPSS Tutorial 1What is SPSS?

• SPSS is a computer program for organizing,
  analyzing and graphing statistical data.
• SPSS stands for Statistical Package for the
  Social Sciences.
• However, it can be used in any area of
  application.

2
SPSS Tutorial 1Getting started

• Start the program
• First window is the data editor window.
• As the name suggests, this is where you enter
  your data.
• You can type data in directly.
• Try this.
• For the purposes of this tutorial, we will load
  an existing database.
• File-gtOpen-gtData
• Open folders Tutorial\sample_files
• Open file demo.sav or demo
• The data from demo will appear in the data
  editor.

3
SPSS Tutorial 1Viewing data with labels

• Ultimately, all SPSS numeric variable are encoded
  with numbers.
• However, it is sometimes useful to associate
  meaningful labels to these numbers.
• Try toggling between the two modes.
• View-gtValue Labels
• Labels are easier to read, but number codes may
  be easier during data entry.
• In the Labels mode, try clicking on a label.
• A menu arrow appears.
• See what other values the variable can have.
• Change it, and then change it back.

4
SPSS Tutorial 1Analyzing data

• Data means nothing without some analysis.
• Analysis functions can be found under the Analyze
  menu!
• Frequency table example
• Analyze-gtDescriptive Statistics-gtFrequencies
• In the Frequencies dialog box, notice the list of
  variables on the left.
• Try running your cursor over the variable names
  to get a more complete name to pop up.
• The variable name is in square brackets.
• The long version is called the label.

5
SPSS Tutorial 1Analyzing data

• A ruler indicates a numeric variable.
• A graph indicates an ordinal scale.
• An two circles and an A indicates a nominal
  string.
• Some types of analyses are only appropriate for
  numeric variables.
• Select some variables to analyze.
• Click each then move them to the box on the right
  using the arrow in the middle.
• Select Gender and Income category for analysis.

6
SPSS Tutorial 1Analyzing data, continued

• Click OK
• A viewer window will appear.
• On the left is an outline pane.
• The outline pane allows you to jump to any part
  of the output.
• Try clicking on Income category under Frequency
  Table.
• This will take you to the Income category
  frequency table.
• Shown is the number of people in each category,
  by absolute numbers and in percent.
• Valid percent handles the case where some data is
  missing.
• We will discuss Cumulative Percent in chapter 2.

7
SPSS Tutorial 1Creating charts

• Charts are sometimes created by an analysis
  procedure.
• You can also create them directly, using the
  Graphs menu.
• Lets create a chart to show the relationship
  between the cell phone and PDA ownership.
• Graphs-gtBar
• Choose a chart type, say Clustered.
• Click Define.
• Transfer Wireless service into the Category Axis
  box.
• Transfer Owns PDA into the Define Clusters box.
• Click OK to produce the chart.
• Scroll down to see chart or use outline pane by
  clicking Graph.
• Notice that people with cell phones are more
  likely to own PDAs.

8
SPSS Tutorial 1Help

• Do not hesitate to use the Help menu as needed.
• Check it out.
• Topics
• Tutorial
• Case Studies
• Statistics Coach
• Etc.
• Also, most dialog boxes have their own Help
  button.
• Pull down Graphs-gtBar
• Notice the Help button.
• Check it out.

9
SPSS Tutorial 1Data entry

• Pull down Window-gtdemo.save
• You can type data directly into the data window
  or import it.
• Import sources
• SPSS data files ( like demo.sav )
• Excel spreadsheets
• Microsoft Access database files
• Text files
• Data in an SPSS data file is organized into cases
  (the rows) and variables (the columns).
• A case refers to a single measurement or instance
  of a random variable.
• See data view.
• In the demo example, a case is a single person
  responding to a survey.
• Variables represent each question asked in the
  survey.

10
SPSS Tutorial 1Data entry, continued

• SPSS data files have a .sav suffix.
• How to control the display of file suffixes (in
  Windows, not SPSS)
• With any desktop window open
• Tools-gtFolder Options
• Click View tab.
• Select or deselect Hide Extensions
• We have already opened an SPSS data file, so we
  know how to do it.
• Excel example
• File-gtOpen-gtData
• At the bottom of the dialog box, select
  Excel(.xls) file type.
• Open demo.xls
• In an Excel file, where is SPPS going to find the
  variable names?
• This is not built into Excel.
• Select the check box for Read variable names from
  the first row of data if desired.
• The entire worksheet is imported unless a range
  of cells is specified in Range.

11
SPSS Tutorial 1Data entry, continued

• In our case, we will read the entire spreadsheet.
• Click OK.
• Click No to save contents question.
• Excel data appears in data window.
• Notice that the structure is not as detailed as
  with the SPSS file.
• Which databases can exchange data with SPSS?
• Answer Databases that use Open Database
  Connectivity.
• See help tutorial and other help facilities for
  more information on databases.

12
SPSS Tutorial 1Data entry, continued

• Text files, delimited by commas or tabs, can be
  imported.
• Use a tab or comma between each variable (column)
  and
• Use a carriage return to separate cases (rows)
• Go to Windows Start-gtMy Computer
• Open folder containing the demo file.
• It should be in the programs folder.
• Open demo.txt. It should open in Notepad or
  some such program.
• Notice the delimited format.
• Return to SPSS

13
SPSS Tutorial 1Data entry, continued

• SPSS text example
• File-gtRead Text Data
• Select file type Text(.txt)
• Open demo.txt
• You are now in the Text Import Wizard
• The Text Import Wizard allows to specify how your
  text data is to be interpreted.
• Select No for predefined format
• Click Next
• We are tab delimited, so select Delimited
• The top row in our file has variable names, so
  select Yes for variable names.
• Click Next.

14
SPSS Tutorial 1Data entry, continued

• Data begins on line 2.
• Click Each line represents a case.
• Click All the cases.
• Check the data preview to see if things look
  right.
• Select Tab as the only delimiter.
• None for text qualifier.
• The next dialog box permits you to edit any names
  that have been truncated by SPSS
• You can also define data types.
• Text files do not have data type information.
• Select Income in the Data Preview window
• Select Dollar from the Data Format list.
• Now Income is in dollars.
• Select No, No in the next dialog box and Finish.

15
SPSS Tutorial 1Introduction to the Data Editor

• The Data Editor has two views
• Data view - columns represent variable, rows
  cases
• Variable view - each row is a variable, each
  column is a variable attribute
• The tabs at the bottom allow you to switch back
  and forth.
• Try it.
• Thinking in terms of a survey, each case is a
  respondent and each variable is a question.
• Data view is for entering and editing data.

16
SPSS Tutorial 1Defining variables in Variable
View

• Click Variable View tab.
• Lets define three variables
• Age
• Marital status
• Income
• Type these names in the name column.
• However, note that length is limited and spaces
  are not allowed.
• Notice that defaults are given for the variable
  attributes.
• Type is numeric.

17
SPSS Tutorial 1Entering data in the Data View

• Click Data View tab.
• Notice that our variable names now appear at
  column heads
• Age
• Marital status
• Income
• Enter 2 cases
• 55 1 72000
• 53 0 153000

18
SPSS Tutorial 1Changing variable attributes

• We prefer integer formats for age and marital.
• Return to Variable view.
• Set the number of decimals to 0 for age and
  marital in the Decimals column.
• See the change in data view.

19
SPSS Tutorial 1Changing variable attributes

• Add a new variable.
• Go to variable view.
• Type sex in the name column.
• Notice that the default type is Numeric.
• We want to designate sex as m or f (string type).
• Select the cell in the Type column, sex row.
• Click the button in the right half of the cell.
• Select String as the Variable Type.
• Click OK.
• In the same way, change type of income to dollar.
• Dollar types permit a choice of format.
• Choose ,,
• Click OK.

20
SPSS Tutorial 1Variable descriptive labels

• Each name is short.
• However, for use in statistical reports and
  charts, we want a more descriptive name or label.
• Go to variable view.
• In the Label column for age, type Respondents
  Age.
• In the Label column for marital row, type Marital
  Status.
• In the Label column for income row, type
  Household.
• In the Label column for sex row, type Gender.

21
SPSS Tutorial 1Value labels

• By default, SPSS stores most values as numbers or
  numeric codes.
• It is nice to give these numeric codes meaningful
  labels.
• For example, 0 and 1 are not very meaningful
  values for marital status.
• Adding value labels can fix this problem.
• Go to Variable View.
• Select the Values column, marital row cell.
• Click the button in the right side of the cell.
• The Value Labels window pops up.

22
SPSS Tutorial 1Value labels

• Type 0 in the Value field.
• Type Single in the Value Label field.
• Click Add button.
• In the same way give the 1 code a label of
  Married.
• Click OK.
• Go to Data View to see how this affects the
  married variable.
• Toggle View-gtValue Labels on and off, to see how
  numeric codes are labeled.
• Repeat this process for the sex variable, using
  the labels Female and Male for code F and M.
• Enter sex data for each case.
• Note that string codes are case sensitive (m?M).

23
SPSS Tutorial 1Missing data

• In research, it commonly happens that data is
  missing.
• It is important to handle these situations
  carefully.
• SPSS converts empty cells into system missing,
  designated by .
• It is important to note the specific reasons why
  data is missing.
• Example data missing because
• Respondent refused to answer.
• Respondent found question objectionable.

24
SPSS Tutorial 1Missing data

• Go to Variable View.
• Click the age row, Missing column cell.
• Click the button in the right half of the cell.
• The Missing Numbers dialog box pops up.
• You can select up to 3 discrete missing values or
  a range of missing values plus one discrete
  value.
• Select Discrete missing values.
• Type 999 into the first cell.
• Leave other 2 empty.
• Click OK.

25
SPSS Tutorial 1Missing data

• 999 is now a code for a particular type of
  missing data.
• Select (age,Values) cell.
• Then right button in cell.
• Assign No Response to the code 999.

26
SPSS Tutorial 1Missing data

• Missing strings are not designated as system
  missing.
• The are interpreted as an empty string.
• Format missing data for a string.
• Open the Missing Values dialog box for the sex
  variable.
• Select Discrete missing values.
• Type NR in the first text box, for no response.
• Dont forget that NR is different from nr.
• Since NR means no response, give NR a label
  of No Response using the Values column.

27
SPSS Tutorial 1Reusing variable formats

• If you have multiple variables with the same
  format, that format can be reused.
• In variable view, create a new variable called
  agewed.
• In the Label column, type Age Married.
• Click the Values cell in the age row.
• Edit-gtCopy.
• Click Values cell in age row.
• Edit-gtPaste.
• Since age and agewed are both ages, it makes
  sense that they share the same Values format.

28
SPSS Tutorial 1Reusing variable formats

• You can reuse all the attributes from a variable.
• Click the row number for the marital row.
• Copy using the Edit menu.
• Select the row number of the first empty row.
• Edit-gtPaste.

29
SPSS Tutorial 1Defining variable properties for
categorical variables

• A categorical variable is a variable on a nominal
  scale.
• It is important for SPSS to know what kind of
  scale to use for each variable.
• SPSS will use different formulas for different
  scale types.
• Open demo.sav.
• Data-gtDefine Variable Properties.
• Move all the variables starting with Owns into
  the Variables to Scan box.
• The default type of variable is scale (i.e.
  ratio), indicated by the little yellow rulers.
• The selected variables should be nominal
  (categorical).

30
SPSS Tutorial 1Defining variable properties for
categorical variables

• Click Continue.
• In the Scanned Variable List, select ownpc.
• Set the Measurement Level to Nominal.

31
SPSS Tutorial 1Working with output

• File-gtOpen-gtOutput.
• Open viewertut.spo.
• This is the output Viewer.
• The outline pane is on the left.
• The contents pane is on the right.
• You can access various parts of the output by
  either scrolling or by clicking an item in the
  outline.
• An open book icon in the outline pane indicates
  that part of the output is open.
• You can open or close items in the contents pane
  by double clicking on the items icon in the
  outline pane.
• Try this with Marital Status.

32
SPSS Tutorial 1Working with output

• You can also hide or display any section of the
  output by using the /- signs in the output pane.
• Try this with the first Frequencies section in
  the outline pane.
• Collapse and expand the section.
• Notice the effect in the contents pane.

33
SPSS Tutorial 1Pivot tables

• The results of most SPSS computations are
  displayed as pivot tables.
• Statistical terms appear in the pivot tables.
• Definitions of these terms can be accessed
  directly in the Viewer.
• Click on Owns PDAGenderInternet Crosstabulation
  in the outline pane.
• Double click the table.
• Notice the mysterious term, Expected count.
• Left click it, then
• Right click it.
• A popup menu will appear.

34
SPSS Tutorial 1Pivot tables

• Select Whats this? from the menu.
• The definition appears.
• The Viewer permits you to transpose, reformat and
  otherwise rearrange data in the pivot tables.
• We will not cover this now because we wont be
  using this capability much in this course.
• However, if you need to manipulate these pivot
  tables for inclusion in a paper say, refer to
  Help-gtTutorials Working with output for details.

35
Chapter 2A

• Frequency tables, graphs, and distributions

36
Unsorted data
Salaries
37
Sorted data
Salaries

• Trends become apparent
• Range
• How many of each
• We can make how many of each explicit with a
  frequency table.

38
Frequency table

• For any given salary, we can see how many people
  earn that much.

39
Relative frequency table

• Say 3 people earn 55,000
• Is that a lot of people?
• We want a relative measure.
• Relative to what?
• Why not 1?
• rf f / Total

40
Cumulative relative frequency table

• Add what you have accumulated so far to the next
  relative frequency.

41
Cumulative relative frequency table

• Add what you have accumulated so far to the next
  relative frequency.

42
Cumulative relative frequency table

• Add what you have accumulated so far to the next
  relative frequency.

43
Cumulative relative frequency table

• Add what you have accumulated so far to the next
  relative frequency.

44
Cumulative relative frequency table

• And so on.
• Should eventually accumulate to 1

45
Cumulative percentage frequency table

• Shows the rank of a salary relative to others.
• Percentile.
• The person earning 60,000 is earning as much or
  more than 75 of the other people in the sample.
• How to calculate Multiply crf by 100.

46
Exercises

• Page 31 2 ac, 5 a,b,c
• You may skip cf columns

47
Grouped frequency distributions

• What if we had slightly different data?
• 33,013 is almost the same as 33,014.
• Does it make sense to count them in separate
  bins?
• Meanwhile, there is a large gap between 70,000
  and 200,000.
• How can we represent these aspects of the data?

48
Grouped frequency distributions

• Answer Grouped Frequency Distributions
• Place the data into equally sized (20,000) bins.
• Notice that the 33,000 salaries are lumped
  together with similar salaries.
• The 200,000 salary is really outstanding.

49
Grouped frequency distributions

• What size bin is the right size?
• Different size bins will show different aspects
  of the data.
• Try different sizes.
• Explore.

50
Histograms

• The mode is the peak value
• Histograms can be unimodal, bimodal, or
  multimodal
• Some cases may be ambiguous, depending on the bin
  structure

51
Distributions

• Histograms are a kind of distribution
• Distributions from small samples tend to have
  large deviations from the true population
  distribution
• Not smooth
• Theoretical distributions are usually assumed to
  be from an infinite population and so are
  perfectly smooth

52
Distributions

• By dividing each count by the total sample size,
  one gets a relative frequency distribution
• The vertical axis now becomes a probability
• The sum of all probabilities is 1
• A relative frequency distribution from an
  infinite population is called a probability
  density function
• The area under the curve is 1

53
Exercises

• Page 43 1 a b (use bin size of 5), 2 a b