View
228
Download
3
Embed Size (px)
Citation preview
A Simple Guide toUsing SPSS© for Windows
By: Julia Hartman
2
Using SPSS© for Windows
• Introduction• Data procedures• Statistical procedures• Syntax files• Editing output
3
Introduction
4
Steps for Analyzing Data
• Enter the data• Select the procedure and options• Select the variables• Run the procedure• Examine the output
5
Common Operations - Menu Options
• In the menu, click Statistics
• Choose Summarize
• Click Frequencies
6
Common Operations - Variables Dialog BoxThis type of dialog box is used for many procedures.
Variables are selected from the list on the left.
Click the arrowto move them to theappropriate box on the right.
7
Using SPSS© for Windows -Data Procedures
• Ways to Enter Data• Entering Data Directly
– Defining variables– Entering data
• Viewing Data • Recoding Variables• Computing New Variables• Selecting Cases
8
Ways to Enter the Data
• SPSS datafile
• Import data– Database file– Spreadsheet file– ASCII text file
• Enter data directly with Data Editor
9
Entering Data Directly-Define the Variables
10
Entering Data Directly-Define the Variable
• Name• Type and size• Labels• Missing values
11
Define the Variable - Name
• Name the variable– No more than 8
characters– Each name unique– Must begin with a
letter– Certain characters
not allowed– Not case sensitive
12
Define the Variable - Type
• Define the variable type.
• Define the variable width.
• Define the number of decimal places.
13
Define the Variable - Labels
• Labels will be displayed in the output.
• Variable Label– can be more descriptive
than variable name
• Value Labels– useful for files that use numeric
codes
14
Define the Variable - Missing Values
• Missing values are used to define user-specified missing information.– No response
– Refused to answer
– Data entry mistakes
15
Define the Variable - Column Format
Column Format is used to define column width
and alignment in theData Editor window
16
Entering Data Directly
• Each row is a case(e.g., survey form).
• Enter the value for each variable.
• Press <Tab> key or right arrow key to move to next variable.
17
Entering Data Directly
• Leave blank or use user-defined missing value if no answer.
• Press <Enter> key to move to next case.
18
Change the View - Value LabelsData entered as numeric codes can be displayed as value labels.
• In the menu, click View
• Click Value Labels
19
Recode Procedure
Recode is used to• to change the values of
an existing variable• to create a new
variable based on the values an existing variable
20
Recode into New Variable
• In the menu, click Transform.
• Select Recode.• Click
Into Different Variable(s)
21
Recode into New Variable
• Select and move variable(s) over.
• Name and label new variable.
• Click Old and New Values
22
Recode into New Variable
For each value of the existing variable
• Repeat for each value or range of values
• Click Continue
• Enter the old value• Enter the new value• Click Add
23
Recode into New Variable
• Click Change
• Click OK
24
Define Labels for New Variable
• In the Data menu, click Define Variable.
• Click Labels.
• Enter value labels for the new variable.
• Click Continue.
25
Compute Procedure
Compute is used to create a new variable.
• In the menu, click Transform.
• Click Compute.
26
Compute Procedure
• Name the newvariable.
• Click Type&Label to define the characteristics of the new variable.
27
Compute Procedure
• Label the newvariable.
• Enter the variable type.
28
Compute Procedure
• Enter the numeric expression thatwill determine the values of the new variable.
• Click OK.
29
Select Cases
For a subset of the datafile, use Select Cases.
• In the menu, click Data.• Click
Select Cases...
30
Select Cases - Alcohol drinkers only
To select only those cases which meet certain criteria, choose the If option.
31
Select Cases - Alcohol drinkers only
• Enter the expression thatwill determine which variables will be selected.
• Click Continue.
32
Select Cases - Alcohol drinkers only
When you’ve finished specifying selection criteria, click OK.
33
Using SPSS© for Windows -Statistical Procedures• Summarizing Data
– Frequencies– Crosstabs (Chi Square)
• Comparing Means– T-Tests– One-Way Analysis of Variance
• Nonparametric Tests– Wilcoxon Signed Ranks– Mann-Whitney U– Kruskal-Wallis
34
Frequencies
• In the menu, click Statistics
• Choose Summarize
• Click Frequencies
35
Frequencies
• Select and move the variables.
• Click Statistics.
36
Frequencies
• Choose the appropriate statistics.
• Click Continue.
37
Frequencies - Charts
For histograms or other charts, click Charts.
38
Frequencies
Choose the type of chart and click Continue
39
Frequencies
To select the format of the table(s), click Format.
40
Frequencies
Choose the format and click
Continue
41
Frequencies
Click OK to run the Frequencies
procedure.
42
Frequencies - Format optionOrganize Output by Variables
43
Frequencies - Format OptionCompare Variables
44
Frequencies - Distribution TableApgar 1 minute score
8 1.2 1.9 1.9
4 .6 .9 2.8
2 .3 .5 3.3
3 .5 .7 4.0
11 1.7 2.6 6.6
9 1.4 2.1 8.7
13 2.0 3.1 11.8
38 5.7 9.0 20.8
170 25.6 40.1 60.8
145 21.8 34.2 95.0
21 3.2 5.0 100.0
424 63.9 100.0
240 36.1
664 100.0
0
1
2
3
4
5
6
7
8
9
10
Total
Valid
SystemMissing
Total
Frequency PercentValid
PercentCumulative
Percent
45
Frequencies - Histogram
Apgar 1 minute score
10.08.06.04.02.00.0
Apgar 1 minute scoreF
req
ue
ncy
300
200
100
0
Std. Dev = 1.83
Mean = 7.8
N = 424.00
46
Crosstabs
• In the menu, click on Statistics
• Choose Summarize
• Click Crosstabs
47
Crosstabs• Move the outcome
variable(s) to the Row(s) box.
• Move the predictor variable(s) to the Column(s) box.
• Click Statistics.
48
Crosstabs
• Select the appropriate statistics.
• Click Continue.
49
Crosstabs
To select the counts, percentages, and residuals to be displayed in each cell, click Cells.
50
Crosstabs
• Select the information to be displayed in each cell.
• Click Continue.
51
Crosstabs
To run the Crosstabs procedure, click OK.
52
Crosstabs - Output
Case Processing Summary
45
95.7%
2
4.3%
47
100.0%
N
Percent
N
Percent
N
Percent
CasesValid
Missing
Total
Birthweight category *Smokes cigarettes
53
Crosstabs - OutputBirthweight category * Smokes cigarettes Crosstabulation
0 8 8
3.2 4.8 8.0
.0% 100.0% 100.0%
.0% 29.6% 17.8%
.0% 17.8% 17.8%
18 19 37
14.8 22.2 37.0
48.6% 51.4% 100.0%
100.0% 70.4% 82.2%
40.0% 42.2% 82.2%
18 27 45
18.0 27.0 45.0
40.0% 60.0% 100.0%
100.0% 100.0% 100.0%
40.0% 60.0% 100.0%
Count
Expected Count
% within Birthweightcategory
% within Smokescigarettes
% of Total
Count
Expected Count
% within Birthweightcategory
% within Smokescigarettes
% of Total
Count
Expected Count
% within Birthweightcategory
% within Smokescigarettes
% of Total
Low birthweight
Normal birthweight
Birthweightcategory
Total
no yes
Smokes cigarettes
Total
54
Independent Samples T-Test
• In the menu, click Statistics.
• Choose Compare Means.
• Click Independent Samples T-Test.
55
Independent Samples T-Test
• Select and move– Test Variable(s)
– Grouping Variable
• Click Define Groups.
56
Independent Samples T-Test
• Enter the values for the groups.
• Click Continue.
57
Independent Samples T-Test
Click OK to run the procedure.
58
Independent Samples T-Test - Output
Group Statistics
14 8.14 .95 .25
20 6.70 2.98 .67
Smokes cigarettes?no
yes
Apgar 1 minute scoreN Mean
Std.Deviation
Std. ErrorMean
Independent Samples Test
11.107 .002 1.746 32 .090 1.44 .83 -.24 3.13
2.026 24.180 .054 1.44 .71 -2.61E-02 2.91
Equal variancesassumed
Equal variances notassumed
Apgar 1 minute scoreF Sig.
Levene's Test forEquality of Variances
t dfSig.
(2-tailed)Mean
DifferenceStd. ErrorDifference Lower Upper
95% ConfidenceInterval of the
Difference
t-test for Equality of Means
59
One-Way Analysis of Variance
• In the menu, click on Statistics.
• Choose Compare Means.
• Click One-Way Analysis of Variance.
60
One-Way Analysis of Variance
• Move the dependent variable(s) to the Dependent List box.
• Move the grouping variable(s) to the Factor box.
• For comparison tests, click Post Hoc.
61
One-Way Analysis of Variance
• Select the appropriate Post Hoc comparisons.
• Click Continue.
62
One-Way Analysis of Variance
Click Options for• Descriptive statistics• Homogeneity of
variance• Mean plots• Missing values options
63
One-Way Analysis of Variance
Make appropriate selections, then click
Continue.
64
One-Way Analysis of Variance
To run the One-way ANOVA
procedure, click OK.
65
Test of Homogeneity of Variances
BTWT
1.384 3 647 .246
LeveneStatistic df1 df2 Sig.
One-Way Analysis of Variance - Output
Descriptives
BTWT
90 2888.72 599.25 63.17 2763.21 3014.23 670 3856
254 3031.24 602.39 37.80 2956.80 3105.68 706 4820
84 3050.98 632.61 69.02 2913.69 3188.26 506 4224
223 3264.89 525.97 35.22 3195.48 3334.30 1049 4734
651 3094.12 594.67 23.31 3048.36 3139.89 506 4820
Underweight
Normal
Overweight
Obese
Total
N MeanStd.
Deviation Std. ErrorLowerBound
UpperBound
95% Confidence Intervalfor Mean
Minimum Maximum
66
ANOVA
BTWT
11460978 3 3820325.9 11.318 .000
2.18E+08 647 337556.868
2.30E+08 650
Between Groups
Within Groups
Total
Sum ofSquares df
MeanSquare F Sig.
One-Way Analysis of Variance - Output
67
One-Way Analysis of Variance - Output
Multiple Comparisons
Dependent Variable: BTWT
Tukey HSD
-142.52 71.271 .188 -325.62 40.58
-162.25 88.143 .254 -388.70 64.19
-376.17* 72.556 .000 -562.57 -189.77
142.52 71.271 .188 -40.58 325.62
-19.74 73.127 .993 -207.60 168.13
-233.65* 53.317 .000 -370.62 -96.68
162.25 88.143 .254 -64.19 388.70
19.74 73.127 .993 -168.13 207.60
-213.92* 74.379 .021 -405.00 -22.83
376.17* 72.556 .000 189.77 562.57
233.65* 53.317 .000 96.68 370.62
213.92* 74.379 .021 22.83 405.00
(J) BMI CategoryNormal
Overweight
Obese
Underweight
Overweight
Obese
Underweight
Normal
Obese
Underweight
Normal
Overweight
(I) BMI CategoryUnderweight
Normal
Overweight
Obese
MeanDifference
(I-J) Std. Error Sig.LowerBound
UpperBound
95% Confidence Interval
The mean difference is significant at the .05 level.*.
68
Wilcoxon Signed Ranks Test
• In the menu, click Statistics
• Choose Nonparametric Tests
• Click 2 Related Samples
69
Wilcoxon Signed Ranks Test
• Move selected variable pairs to the Test Pair(s) List box.
• Choose the statistical test(s).
• Click Options...
70
Wilcoxon Signed Ranks Test
Check Descriptives box for descriptive
statistics.
71
Wilcoxon Signed Ranks Test
Click OK to run the procedure.
72
Wilcoxon Signed Ranks TestRanks
0a .00 .00
27b 14.00 378.00
7c
34
Negative Ranks
Positive Ranks
Ties
Total
Apgar 5 minute score -Apgar 1 minute score
N Mean RankSum ofRanks
Apgar 5 minute score < Apgar 1 minute scorea.
Apgar 5 minute score > Apgar 1 minute scoreb.
Apgar 1 minute score = Apgar 5 minute scorec. Test Statisticsb
-4.631a
.000
Z
Asymp. Sig. (2-tailed)
Apgar 5minutescore -Apgar 1minutescore
Based on negative ranks.a.
Wilcoxon Signed Ranks Testb.
73
Mann-Whitney U Test
• In the menu, click Statistics
• Choose Nonparametric Tests
• Click 2 Independent Samples
74
Mann-Whitney U Test
• Select and move– Test Variable(s)
– GroupingVariable
• Click Define Groups.
75
Mann-Whitney U Test
• Enter the values for the groups.
• Click Continue.
76
Mann-Whitney U Test
• Click Options.
• After changing options, click Continue.
• Click OK to run the procedure.
77
Mann-Whitney U Test - OutputDescriptive Statistics
34 7.29 2.44 0 9
47 .60 .50 0 1
Apgar 1 minute score
Smokes cigarettes?
N MeanStd.
Deviation Minimum Maximum
Ranks
14 19.89 278.50
20 15.82 316.50
34
Smokes cigarettes?no
yes
Total
Apgar 1 minute scoreN Mean Rank
Sum ofRanks
Test Statisticsb
106.500
316.500
-1.238
.216
.245a
Mann-Whitney U
Wilcoxon W
Z
Asymp. Sig. (2-tailed)
Exact Sig. [2*(1-tailedSig.)]
Apgar 1minutescore
Not corrected for ties.a.
Grouping Variable: Smokes cigarettes?b.
78
Kruskal-Wallis Test
• In the menu, click Statistics
• Choose Nonparametric Tests
• Click K Independent Samples
79
Kruskal-Wallis Test
• Move the dependent variable(s) to the Test Variable List box.
• Move the grouping variable(s) to the Grouping Variable box.
• Click Define Range.
80
Kruskal-Wallis Test
• Enter the minimum and maximum values for the Grouping Variable.
• Click Continue.
81
Kruskal-Wallis Test
• Check the box for the Kruskal-Wallis H.
• Click OK to run the procedure.
82
Kruskal-Wallis Test - OutputRanks
201 242.52
340 347.55
108 407.53
649
Age group<20
20-29
30+
Total
BMIN Mean Rank
Test Statisticsa,b
64.743
2
.000
Chi-Square
df
Asymp. Sig.
BMI
Kruskal Wallis Testa.
Grouping Variable: Age groupb.
83
Using SPSS© for Windows -Editing the Output
• Pivot Tables• Scatterplots• Charts
84
Scatterplot
• In the menu, click on Graphs.
• Choose Scatter….
85
Scatterplot
• Choose the appropriate type of plot.
• Click Define.
86
Scatterplot
• Select and move the variables for the X and Y axes to the appropriate box.
• Click OK to run the procedure.
87
BMI
70605040302010
BT
WT
5000
4000
3000
2000
1000
0
Scatterplot - Output
Regression line must be added.
88
Edit the Scatterplot
In the Output Window• Click the chart object to select it.
• In the menu, click Edit. • Choose SPSS Chart Object.
• Click Open.
89
Scatterplot
The Chart Window
will open.
90
Edit the Scatterplot
In the Chart Window• In the menu, click Chart.
• Click Options.
91
Edit the Scatterplot
• Check the Total box.• Click OK.
92
Scatterplot - Output
Regression line is added.
BMI
70605040302010
BT
WT
5000
4000
3000
2000
1000
0
93
Syntax Files
A Syntax File is a text file that contains SPSS commands.
94
Uses for Syntax Files
• Some commands and options are available only by using command syntax.
• Syntax files let you save a program so you can repeat it or edit it to use different variables.
• Syntax files let you run a program using SPSS Production Facility.
95
Copying to a Syntax File
The Paste button copies the
commands for a procedure into a
syntax file.
96
Editing Syntax Files
A syntax file can be edited in the Syntax Editor window.
97
Opening a New Syntax File
A new Syntax File can be opened from the File menu.
98
Running Syntax Files
A syntax file can be edited in the Syntax Editor window.
99
SPSS KeyWords newsletter:http://www.spss.com/keywords/
Chi Squarehttp://vflylab.calstatela.edu/edesktop/VirtApps/VflyLab/ChiSqBg.html