SPSS Cheat Sheet
Basic Descriptive Statistics
Reverse Scoring
Transform > Recode into different variables
To ensure respondents are reading through each question and not rushing through
Nominal/Ordinal Data
Analyse > Descriptive Statistics > Frequencies
Report percentages
Scale Data using Descriptives
Analyse > Descriptive Statistics > Descriptives
Report mean and standard deviation
Scale Data using Explore
Analyse > Descriptive Statistics > Explore
Report mean and standard deviation
Testing for Normality in SPSS
Analyse > Descriptive Statistics > Explore
Statistics > Descriptives (95%) > Outliers
Plots > Histogram > Normality plots with tests
H0: Data is normally distributed
H1: Data is not normally distributed
Three ways to test:
1. Hypothesis testing (Kolmogorov-Smirnov or Shapiro-Wilks)
2. Descriptives (Skewness or Kurtosis)
3. Histograms
Sample > 50 = Kolmogorov-Smirnov
Sample < 50 = Shapiro-Wilks
Skewness within range -1 and 1
Kurtosis within range -1.5 and 1.5
1
, Crosstabs, Reliability and Validity
Crosstabs
Analyse > Descriptive Statistics > Crosstabs
Does “nominal variable x” influence “nominal variable y”?
Is there an association between “nominal variable x” and “nominal variable y”?
Measures associations between 2 nominal variables
H0: There is no association between x and y
H1: There is an association between x and y
Look at Pearson’s Chi-squared statistics
Item Reliability
Analyse > Scale > Reliability Analysis
[Put in individual items not summated; remember they need to be reverse scored]
Statistics > Scale if deleted item
Reliability is analysed to determine how a scale performs when used multiple times.
It enables us to determine whether the items in the scale consistently measure their
respective construct.
Cronbach’s alpha (scale) should always be above 0.7
Summated Scales
Transform > Compute Variable
Statistical & Mean (must be reverse scored first!)
Use the means = necessary to preserve the measure of the scale; has to accurately
represent the construct
Scale Validity
Analyse > Dimension Reduction > Factor
Extraction > Principle Components
Rotation > Varimax
Validity is analysed to determine how items load onto factors. It enables us to
determine whether items are correctly associated with their respective constructs
thus ensuring that the correct constructs are being used in the study. % of Variance
& Rotated Component Matrix
Factor Analysis:
Exploratory – to search for latent variables when developing new scales.
2
Basic Descriptive Statistics
Reverse Scoring
Transform > Recode into different variables
To ensure respondents are reading through each question and not rushing through
Nominal/Ordinal Data
Analyse > Descriptive Statistics > Frequencies
Report percentages
Scale Data using Descriptives
Analyse > Descriptive Statistics > Descriptives
Report mean and standard deviation
Scale Data using Explore
Analyse > Descriptive Statistics > Explore
Report mean and standard deviation
Testing for Normality in SPSS
Analyse > Descriptive Statistics > Explore
Statistics > Descriptives (95%) > Outliers
Plots > Histogram > Normality plots with tests
H0: Data is normally distributed
H1: Data is not normally distributed
Three ways to test:
1. Hypothesis testing (Kolmogorov-Smirnov or Shapiro-Wilks)
2. Descriptives (Skewness or Kurtosis)
3. Histograms
Sample > 50 = Kolmogorov-Smirnov
Sample < 50 = Shapiro-Wilks
Skewness within range -1 and 1
Kurtosis within range -1.5 and 1.5
1
, Crosstabs, Reliability and Validity
Crosstabs
Analyse > Descriptive Statistics > Crosstabs
Does “nominal variable x” influence “nominal variable y”?
Is there an association between “nominal variable x” and “nominal variable y”?
Measures associations between 2 nominal variables
H0: There is no association between x and y
H1: There is an association between x and y
Look at Pearson’s Chi-squared statistics
Item Reliability
Analyse > Scale > Reliability Analysis
[Put in individual items not summated; remember they need to be reverse scored]
Statistics > Scale if deleted item
Reliability is analysed to determine how a scale performs when used multiple times.
It enables us to determine whether the items in the scale consistently measure their
respective construct.
Cronbach’s alpha (scale) should always be above 0.7
Summated Scales
Transform > Compute Variable
Statistical & Mean (must be reverse scored first!)
Use the means = necessary to preserve the measure of the scale; has to accurately
represent the construct
Scale Validity
Analyse > Dimension Reduction > Factor
Extraction > Principle Components
Rotation > Varimax
Validity is analysed to determine how items load onto factors. It enables us to
determine whether items are correctly associated with their respective constructs
thus ensuring that the correct constructs are being used in the study. % of Variance
& Rotated Component Matrix
Factor Analysis:
Exploratory – to search for latent variables when developing new scales.
2
