SPSS Cheat Sheet — Statistical Analysis, Syntax & Data Management | Dataplexa

SPSS Cheat Sheet

Data Entry · Variable Setup · Descriptives · t-Tests · ANOVA · Regression · Chi-Square · Syntax

Sheet 6 of 6 Stats & Math Intermediate Printable

SPSS Interface Overview

Core

View / Window	Purpose
Data View	Spreadsheet — rows = cases, columns = variables
Variable View	Define variable properties (name, type, labels, measure)
Output Viewer	Results tables, charts, and notes from every procedure
Syntax Editor	Write, save, and re-run SPSS command syntax
Chart Editor	Double-click any chart in Output to edit it

Tip: Always switch to Variable View first to set up variables correctly before entering data.

Variable View — Properties

Setup

Column	What to Set
Name	Short identifier, no spaces, starts with letter (e.g. age)
Type	Numeric / String / Date / Currency
Width / Decimals	Display width and decimal places
Label	Full descriptive name shown in output
Values	Code map for categoricals (1=Male, 2=Female)
Missing	User-defined missing values (e.g. 99)
Measure	Scale / Ordinal / Nominal

Rule: Set Measure level correctly — SPSS uses it to suggest appropriate charts and analyses.

Measurement Levels

Theory

Level	SPSS Label	Example	Allowed Stats
Nominal	Nominal	Gender, Blood Type	Mode, Frequencies, Chi-Square
Ordinal	Ordinal	Likert Scale, Rank	Median, IQR, Spearman r
Interval / Ratio	Scale	Age, Income, Score	Mean, SD, t-test, ANOVA, Pearson r

Memory Aid — NOIR

N · O · I · R

Nominal → Ordinal → Interval → Ratio (each level adds properties)

Descriptive Statistics

Analyze

Menu: Analyze → Descriptive Statistics → …

Procedure	Use When	Key Output
Frequencies	Nominal / Ordinal	Count, %, Bar chart, Pie chart
Descriptives	Scale (continuous)	N, Mean, SD, Min, Max, Skewness
Explore	Scale — normality check	Stem-leaf, Boxplot, S-W & K-S tests
Crosstabs	Two categoricals	Contingency table, Chi-Square, Phi/V

Syntax — Descriptives

DESCRIPTIVES VARIABLES=age score income
  /STATISTICS=MEAN STDDEV MIN MAX SKEWNESS KURTOSIS.

Frequencies & Crosstabs Syntax

Syntax

Frequencies with Bar Chart

FREQUENCIES VARIABLES=gender education
  /BARCHART PERCENT
  /ORDER=ANALYSIS.

Crosstabs with Chi-Square

CROSSTABS
  /TABLES=gender BY smoker
  /STATISTICS=CHISQ PHI
  /CELLS=COUNT ROW COLUMN EXPECTED.

Chi-Square assumption: Expected cell count ≥ 5 in all cells. Check the footnote in output.

Normality Testing

Assumptions

Explore — Normality Tests

EXAMINE VARIABLES=score
  /PLOT NORMPLOT
  /STATISTICS DESCRIPTIVES
  /CINTERVAL 95
  /NOTOTAL.

Test	Best For	Interpret
Shapiro-Wilk (S-W)	n < 50	p > .05 → normal
Kolmogorov-Smirnov (K-S)	n ≥ 50	p > .05 → normal
Skewness / Kurtosis	Any n	±1.0 acceptable; ±2.0 borderline
Q-Q Plot	Visual	Points near diagonal → normal

Independent Samples t-Test

Inference

Menu: Analyze → Compare Means → Independent-Samples T Test

Test Statistic

t = (x̄₁ − x̄₂) / SE_diff

SE_diff = √(s₁²/n₁ + s₂²/n₂) | df = n₁ + n₂ − 2

Syntax

T-TEST GROUPS=gender(1 2)
  /MISSING=ANALYSIS
  /VARIABLES=score
  /CRITERIA=CI(.95).

Levene's Test: Check first. If p < .05 → variances unequal → read "Equal variances not assumed" row.

Paired & One-Sample t-Test

Inference

Paired Samples t-Test Syntax

T-TEST PAIRS=pre WITH post (PAIRED)
  /CRITERIA=CI(.95)
  /MISSING=ANALYSIS.

One-Sample t-Test Syntax

T-TEST
  /TESTVAL=100
  /MISSING=ANALYSIS
  /VARIABLES=iq_score
  /CRITERIA=CI(.95).

Effect Size — Cohen's d

d = (x̄ − μ₀) / s

0.2 = small · 0.5 = medium · 0.8 = large

One-Way ANOVA

Inference

Menu: Analyze → Compare Means → One-Way ANOVA

F Statistic

F = MS_between / MS_within

MS = SS / df | SS_total = SS_between + SS_within

Syntax with Post Hoc

ONEWAY score BY group
  /STATISTICS DESCRIPTIVES HOMOGENEITY
  /PLOT MEANS
  /MISSING ANALYSIS
  /POSTHOC=TUKEY BONFERRONI
     ALPHA(.05).

Assumption: Levene's test for homogeneity of variance. If violated, use Welch's ANOVA or Games-Howell post hoc.

Factorial (Two-Way) ANOVA

Inference

Menu: Analyze → General Linear Model → Univariate

GLM Univariate Syntax

UNIANOVA score BY gender treatment
  /METHOD=SSTYPE(3)
  /INTERCEPT=INCLUDE
  /PRINT=ETASQ HOMOGENEITY DESCRIPTIVE
  /CRITERIA=ALPHA(.05)
  /DESIGN=gender treatment gender*treatment.

Source	Meaning
gender	Main effect of gender
treatment	Main effect of treatment
gender*treatment	Interaction effect

η² (Eta squared): .01 = small · .06 = medium · .14 = large effect

Simple Linear Regression

Regression

Menu: Analyze → Regression → Linear

Regression Equation

Ŷ = b₀ + b₁X

b₀ = intercept · b₁ = slope (unstandardized) · β = standardized

Syntax

REGRESSION
  /MISSING LISTWISE
  /STATISTICS COEFF OUTS R ANOVA
  /DEPENDENT salary
  /METHOD=ENTER experience
  /SCATTERPLOT=(*ZRESID, *ZPRED)
  /RESIDUALS=DURBIN.

Output	Meaning
R²	Proportion of variance in Y explained by X
Adj R²	R² adjusted for number of predictors
F (ANOVA)	Overall model significance
β (Beta)	Standardized coefficient — comparable across predictors
Durbin-Watson	≈ 2 = no autocorrelation; <1 or >3 = problem

Multiple Regression

Regression

Multiple Regression Equation

Ŷ = b₀ + b₁X₁ + b₂X₂ + … + bₖXₖ

k = number of predictors · VIF < 10 (no multicollinearity)

Enter Method — Multiple Predictors

REGRESSION
  /MISSING LISTWISE
  /STATISTICS COEFF OUTS R ANOVA COLLIN TOL
  /DEPENDENT salary
  /METHOD=ENTER age education experience.

Method	When to Use
ENTER	Theory-driven — enter all predictors simultaneously
STEPWISE	Exploratory — auto selects predictors by F criterion
FORWARD	Adds predictors one at a time (best contribution)
BACKWARD	Starts full, removes least significant predictors

Multicollinearity: Check Tolerance (> 0.1) and VIF (< 10) in Coefficients table.

Correlation Analysis

Bivariate

Pearson r

r = Σ[(xᵢ−x̄)(yᵢ−ȳ)] / √[Σ(xᵢ−x̄)² · Σ(yᵢ−ȳ)²]

Range: −1 to +1 · |r|: .1=small · .3=medium · .5=large

Pearson & Spearman Syntax

/* Pearson (Scale data) */
CORRELATIONS
  /VARIABLES=age salary score
  /PRINT=TWOTAIL SIG FULL
  /MISSING=PAIRWISE.

/* Spearman (Ordinal or non-normal) */
NONPAR CORR
  /VARIABLES=rank satisfaction
  /PRINT=SPEARMAN TWOTAIL SIG
  /MISSING=PAIRWISE.

Note: Correlation ≠ Causation. Always report r and p-value together.

Chi-Square Tests

Non-Parametric

Chi-Square Statistic

χ² = Σ [(O − E)² / E]

O = observed · E = expected · df = (r−1)(c−1)

Goodness-of-Fit Syntax

NPAR TESTS
  /CHISQUARE=category
  /EXPECTED=EQUAL.  /* or specify proportions */

Test	Variables	Null Hypothesis
Goodness-of-Fit	1 categorical	Observed = expected distribution
Independence	2 categorical	Variables are independent

Association Measure	Use When
Phi (φ)	2×2 table
Cramér's V	Larger tables (r×c)
Lambda	Asymmetric nominal associations

Non-Parametric Alternatives

Non-Parametric

Parametric Test	Non-Parametric Alternative	SPSS Procedure
Independent t-test	Mann-Whitney U	Nonparametric → Legacy → 2 Ind. Samples
Paired t-test	Wilcoxon Signed-Rank	Nonparametric → Legacy → 2 Related Samples
One-Way ANOVA	Kruskal-Wallis H	Nonparametric → Legacy → K Ind. Samples
Repeated Measures ANOVA	Friedman Test	Nonparametric → Legacy → K Related Samples
Pearson r	Spearman ρ	Nonparametric → Correlations → Spearman

Mann-Whitney U Syntax

NPAR TESTS
  /M-W=score BY group(1 2)
  /MISSING ANALYSIS.

Data Transformation

Transform

Compute New Variable

COMPUTE bmi = weight / (height * height).
EXECUTE.

/* Log transformation for skewed data */
COMPUTE log_income = LN(income).
EXECUTE.

Recode Into Different Variable

RECODE age
  (18 THRU 35=1)(36 THRU 55=2)
  (56 THRU HIGHEST=3)
  INTO age_group.
VARIABLE LABELS age_group 'Age Group'.
VALUE LABELS age_group 1 'Young'
  2 'Middle' 3 'Senior'.
EXECUTE.

Data Management Commands

Manage

Select Cases

/* Keep only females */
SELECT IF (gender = 2).
EXECUTE.

/* Reset — all cases */
FILTER OFF.
USE ALL.
EXECUTE.

Sort & Split File

SORT CASES BY department (A) salary (D).

/* Analyze by group separately */
SPLIT FILE SEPARATE BY department.
SPLIT FILE OFF.  /* Reset */

Aggregate — Mean per Group

AGGREGATE
  /OUTFILE=* MODE=ADDVARIABLES
  /BREAK=department
  /mean_salary=MEAN(salary).

Missing Data Handling

Data Quality

Strategy	When to Use	SPSS Option
Listwise deletion	MCAR, small % missing	MISSING=LISTWISE
Pairwise deletion	Correlation/regression, preserve n	MISSING=PAIRWISE
Mean substitution	Simple continuous replacement	Transform → Replace Missing
Multiple Imputation	MCAR / MAR, formal analyses	Analyze → Multiple Imputation

Replace Missing with Mean

RMV score_imp=SMEAN(score).
EXECUTE.

Caution: Mean substitution reduces variance and can distort correlations. Prefer Multiple Imputation for formal research.

Reliability — Cronbach's Alpha

Scale

Menu: Analyze → Scale → Reliability Analysis

Cronbach's Alpha Formula

α = (k / k−1) · [1 − (Σσᵢ²) / σ²ₜ]

k = items · σᵢ² = item variance · σ²ₜ = total scale variance

Reliability Syntax

RELIABILITY
  /VARIABLES=q1 q2 q3 q4 q5
  /SCALE('ALL VARIABLES') ALL
  /MODEL=ALPHA
  /STATISTICS=DESCRIPTIVE SCALE
  /SUMMARY=TOTAL.

α Value	Interpretation
≥ .90	Excellent
.80 – .89	Good
.70 – .79	Acceptable
.60 – .69	Questionable
< .60	Unacceptable

Factor Analysis (EFA)

Multivariate

Menu: Analyze → Dimension Reduction → Factor

Principal Components Syntax

FACTOR
  /VARIABLES q1 q2 q3 q4 q5 q6
  /MISSING LISTWISE
  /ANALYSIS q1 q2 q3 q4 q5 q6
  /PRINT KMO EXTRACTION ROTATION
  /CRITERIA MINEIGEN(1)
  /EXTRACTION PC
  /ROTATION VARIMAX
  /SAVE REG(ALL FAC).

Diagnostic	Acceptable Value
KMO	≥ 0.60 (sampling adequacy)
Bartlett's Test	p < .05 (correlations exist)
Communality	≥ 0.40 per item
Factor Loading	≥ \|0.40\| for inclusion
Eigenvalue Rule	> 1 (Kaiser criterion)

SPSS Syntax — Essentials & Quick Reference

Syntax

File Operations

/* Open data file */
GET FILE='/data/survey.sav'.

/* Save data file */
SAVE OUTFILE='/data/clean.sav'
  /COMPRESSED.

/* Import CSV */
GET DATA
  /TYPE=TXT
  /FILE='/data/data.csv'
  /DELIMITERS=","
  /FIRSTCASE=2.  /* skip header */

Variable & Value Labels

VARIABLE LABELS
  gender 'Participant Gender'
  score  'Test Score (0–100)'.

VALUE LABELS gender
  1 'Male'
  2 'Female'
  3 'Non-binary'.

MISSING VALUES score (99)
  gender (9).
EXECUTE.

IF / DO IF Logic

/* Conditional compute */
IF (score >= 50) pass=1.
IF (score < 50)  pass=0.
EXECUTE.

DO IF (group=1).
  COMPUTE label=1.
ELSE IF (group=2).
  COMPUTE label=2.
ELSE.
  COMPUTE label=9.
END IF.
EXECUTE.

Rule	Detail
Commands end with period	Every SPSS syntax command must end with a . (full stop)
Subcommands use /	Subcommands within a procedure start with /
Strings in quotes	Text values must be in single quotes: 'Male'
Case insensitive	Keywords can be upper or lower case — GET FILE = get file
Comments	/* comment / or start line with
EXECUTE	Required after transformation commands to force processing

Reading SPSS Output

Interpret

Table / Row	What to Report
Model Summary	R, R², Adjusted R², Std. Error of estimate
ANOVA table	F, df₁, df₂, p-value (model significance)
Coefficients	b, SE, β, t, p, 95% CI — one row per predictor
Group Statistics	n, Mean, SD per group (t-test)
Levene's Test	p > .05 → equal variances assumed
Post Hoc Tests	Pairwise p values — which groups differ

APA Format: t(df) = value, p = .xxx, d = value · F(df₁, df₂) = value, p = .xxx, η² = value

Selecting the Right Test

Decision

Research Question	DV Level	Use
Compare 2 independent groups	Scale	Independent t-test
Compare 2 related measurements	Scale	Paired t-test
Compare 3+ independent groups	Scale	One-Way ANOVA
2 factors + interaction	Scale	Factorial ANOVA
Predict continuous outcome	Scale	Linear Regression
Predict binary outcome	Nominal	Logistic Regression
Association between 2 categoricals	Nominal	Chi-Square
Linear relationship strength	Scale–Scale	Pearson Correlation
Rank-based relationship	Ordinal	Spearman Correlation

Common SPSS Errors & Fixes

Troubleshoot

Error / Issue	Fix
Variable name has spaces	Use underscore: test_score
String variable in numeric analysis	Recode to numeric or change Type in Variable View
User-missing values included in mean	Define missing values in Variable View
Chi-Square warning: expected < 5	Merge cells, use Fisher's Exact Test (2×2), or collect more data
Syntax "End of command expected"	Check for missing period . at end of command
Output: "There are no valid cases"	All cases filtered or selected; check SELECT IF / FILTER

Paste Syntax: In any dialog, use the Paste button instead of OK to send syntax to the editor — great for reproducibility.

SPSS Mastery Checklist

Self-Assessment

Interface & Setup

Navigate Data View and Variable View

Set Name, Type, Label, Values, Missing, and Measure for each variable

Import CSV/Excel files and save .sav files

Use the Output Viewer to interpret results

Open, write, and run commands in Syntax Editor

Data Management

Compute new variables with formulas

Recode variables (same / different variable)

Select and filter cases using SELECT IF

Sort cases and use SPLIT FILE by group

Handle missing data (listwise, pairwise, imputation)

Descriptive & Normality

Run Frequencies for categorical variables

Run Descriptives for continuous variables (mean, SD)

Use Explore to check normality (S-W, K-S, Q-Q plots)

Interpret skewness and kurtosis values

Produce Crosstabs with row/column percentages

Inferential Tests

Conduct and interpret independent samples t-test

Conduct and interpret paired samples t-test

Run One-Way ANOVA with Tukey post hoc

Run Factorial ANOVA and interpret interactions

Choose a non-parametric alternative when needed

Correlation & Regression

Compute Pearson and Spearman correlations

Interpret r, R², and p-values

Run simple and multiple linear regression

Check assumptions (normality, homoscedasticity, multicollinearity)

Report β, t, p, and 95% CI in APA format

Advanced Procedures

Run Chi-Square independence test with Cramér's V

Assess scale reliability with Cronbach's Alpha

Conduct Exploratory Factor Analysis (EFA)

Use Multiple Imputation for missing data

Write reproducible syntax and save .sps files

🎉 You've completed the Statistics & Math series!

Explore more cheat sheets — Rust, JavaScript, Power BI, Tableau, and R Language are coming next.

Browse All Cheat Sheets →

← Back