Statistical Analysis with SPSS: A Beginner's Guide

Introduction

SPSS (IBM Statistical Package for the Social Sciences) is one of the most widely used statistical software packages in academic research. Yet many students find it intimidating. This guide walks you through the fundamentals: setting up your data, running analyses, and interpreting results.

Getting Started: Setting Up Your Data

The Data View

Your raw data goes here—each row is a case (participant, observation), each column is a variable.

Example:

The Variable View

Define your variables here:

• Name: Short variable name (no spaces)
• Type: Numeric, string, date, etc.
• Label: Full description of the variable
• Values: For categorical variables, define codes (1=Male, 2=Female)
• Measure: Scale (continuous), ordinal, or nominal (categorical)

Descriptive Statistics

Start here. Understand your data before running inferential tests.

Accessing Descriptive Statistics

Menu: Analyze → Descriptive Statistics → Frequencies

For each variable, you'll get:

• Mean (average)
• Standard deviation (spread around mean)
• Minimum and maximum values
• Quartiles

Reading Your Output

Important numbers:

Rule of thumb: If skewness or kurtosis > 2 or < -2, your data is not normally distributed (matters for some tests).

Visualizing Data

Menu: Graphs → Chart Builder

Create histograms, bar charts, and scatterplots to visualize patterns.

Correlation Analysis

Measures the relationship between two continuous variables.

Running a Correlation

Menu: Analyze → Correlate → Bivariate

Select your variables and choose Pearson correlation.

Interpreting Correlation Output

Correlation coefficient (r) ranges from -1 to +1:

• r = 1.0: Perfect positive correlation (as one increases, so does the other)
• r = 0.7: Strong positive correlation
• r = 0.5: Moderate positive correlation
• r = 0: No correlation
• r = -0.5: Moderate negative correlation
• r = -1.0: Perfect negative correlation

Example: If anxiety and test score correlation is r = -.45, p = .02:

"There is a moderate negative correlation between anxiety and test performance (r = -.45, p = .02). Students with higher anxiety tend to score lower on tests. This relationship is statistically significant."

T-Tests: Comparing Two Groups

Use when comparing means between two groups (e.g., treatment vs. control).

Types of T-Tests

Independent t-test: Different people in each group

• Example: Do men and women differ on leadership scores?

Paired t-test: Same people measured twice

• Example: Do participants improve from pre-test to post-test?

Running an Independent T-Test

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

Select your dependent variable (the outcome you're measuring) and grouping variable (the groups to compare).

Interpreting T-Test Output

Key numbers:

Example: If t(48) = 2.34, p = .023, Mean Diff = 3.5:

"Men and women significantly differed on leadership scores (t(48) = 2.34, p = .023). On average, men scored 3.5 points higher than women."

The P-Value

• p < .05: Result is statistically significant (less than 5% chance it occurred by random chance)
• p > .05: Result is not significant (could easily have occurred by chance)

ANOVA: Comparing Multiple Groups

When comparing 3+ groups, use ANOVA (Analysis of Variance).

Running One-Way ANOVA

Menu: Analyze → Compare Means → One-Way ANOVA

Select dependent variable and factor (grouping variable).

Key Output

Example: F(2, 87) = 4.32, p = .016

"There were significant differences in test scores across the three teaching methods (F(2, 87) = 4.32, p = .016)."

Post-hoc tests (if significant): Use Tukey HSD to see which specific groups differ.

Chi-Square Test: Categorical Data

When analyzing categorical variables (e.g., "Did students pass or fail?").

Running Chi-Square

Menu: Analyze → Descriptive Statistics → Crosstabs

Check the "Chi-square" option.

Output

Example: χ²(1) = 5.23, p = .022

"There was a significant association between gender and college major choice (χ²(1) = 5.23, p = .022)."

Reporting Results

Use this format for any statistical test:

Descriptive + Test + P-value + Effect:

"Students who received the intervention (M = 87.3, SD = 5.2) scored significantly higher than control students (M = 82.1, SD = 6.8) on the post-test, t(98) = 3.45, p = .001."

Common Mistakes

1. Using wrong test – Check assumptions before choosing (normal distribution? equal groups?)
2. Ignoring assumptions – SPSS will run analyses even if your data violates assumptions
3. Confusing correlation and causation – Relationship ≠ cause and effect
4. Over-interpreting p-values – p < .05 means "probably not by chance," not "this is true"
5. Forgetting to report effect size – Significance depends on sample size; effect size shows practical importance

Resources

• SPSS Help Files (built into software)
• YouTube tutorials for specific tests
• Your instructor (ask before running analyses you're unsure about)

Conclusion

SPSS is a powerful tool for quantitative research. Start with descriptive statistics to understand your data, then move to inferential tests to answer your research questions. Always check assumptions, interpret results in context, and remember: statistics inform decisions, but don't make them.

Need help with SPSS analysis? EssayBoffin's statistics tutors can guide you through setup, analysis, and interpretation. Get statistics support