How to Use SPSS for Data Analysis: Step-by-Step Examples

Home / Blog / How to Use SPSS for Data Analysis: Step-by-Step Examples

In Brief

SPSS (Statistical Package for the Social Sciences) is the most widely used software for quantitative data analysis in academic research. To use SPSS for your thesis or dissertation, you follow a structured workflow: prepare your data, define variables, run descriptive statistics, perform inferential tests (t-tests, ANOVA, correlation, regression), and interpret the results using APA reporting standards. This guide walks you through each step with concrete examples, so you can analyze your data correctly without needing a statistics degree.

What Is SPSS and Why Do Researchers Use It?

SPSS, developed by IBM, is a point-and-click statistics software package designed for data analysis, data management, and data communication. It was originally created for the social sciences but is now used across virtually every discipline that relies on quantitative research — including education, psychology, health sciences, business, and nursing.

For graduate students, SPSS matters because:

It handles large datasets — unlike Excel, which struggles with thousands of rows
It offers every common statistical test — from basic frequencies to multivariate regression
It produces APA-ready output — tables and statistics formatted for dissertation results sections
It’s the most taught software — most graduate programs include SPSS in their coursework

If you’re writing a dissertation with quantitative data — survey results, experimental measurements, or structured questionnaire data — SPSS is likely the tool your department expects you to use.

A Simple Workflow: 7 Steps to Analyze Your Data in SPSS

Before running any test, it helps to understand the overall process. Here’s the complete SPSS data analysis workflow:

Prepare your data (clean, code, and import)
Define your variables (Variable View)
Run descriptive statistics (mean, median, standard deviation)
Check assumptions (normality, reliability, outliers)
Choose and run the right statistical test
Interpret the output
Write up results in APA format

Let’s walk through each step with real examples.

Step 1: Prepare and Import Your Data

Getting Data Into SPSS

Most researchers collect data in Excel, Google Forms, or SurveyMonkey. SPSS accepts several formats: .xlsx, .csv, .txt. To import your data:

Go to File → Open → Data
Select your file (change “Files of type” to Excel if needed)
Check Read variable names from first row — this ensures your column headers become your variable names
Click OK

When the Data Editor opens, you’ll see your spreadsheet laid out in rows (participants) and columns (questions or variables). Each row represents one person; each column holds one variable.

Tip: Organize Before You Import

Remove blank rows at the top or bottom
Ensure every column has a clear header (no spaces, use underscores instead)
Make sure numeric variables are set as numbers, not text
Save the file before importing — sometimes SPSS locks files in use

Step 2: Define Your Variables in Variable View

Switch from Data View (the spreadsheet) to Variable View (the attribute table). Each variable has several important fields:

Field	What It Does	Example
Name	Internal variable identifier (no spaces, max 64 chars)	`Q1`, `Age`, `Gender`, `Satisfaction`
Type	Data type — Numeric or String	`Numeric`
Width	Maximum characters	`5`
Decimals	Decimal places for numeric types	`0` for Age, `2` for Likert scales
Label	Full descriptive name (appears in charts and output)	`Gender of participant`
Values	Codes for categorical variables	`1 = Male, 2 = Female`
Type	Variable measurement level — Nominal, Ordinal, Scale	`Nominal`
Measure	Measurement level (same as Type in newer versions)	`Nominal`

Why Variable View Matters

Many students skip Variable View and jump straight to analysis. This is a mistake. Setting up your variables correctly means your output tables will have proper labels, your charts will show readable categories, and your APA write-up will be accurate.

For categorical variables (like gender, ethnicity, or treatment group), always define value labels. For example, if your “Education Level” variable uses 1–5, define those values in Variable View. Without this, SPSS will show just “1” in every table instead of “High School.”

Step 3: Run Descriptive Statistics

Descriptive statistics give you the basic numbers — means, standard deviations, ranges, and frequencies — that describe your dataset.

How to Run Them

Go to Analyze → Descriptive Statistics → Frequencies
Move your variables into the right panel
Click Statistics — select Mean, Median, Std. Deviation, Minimum, Maximum
Click Continue, then OK

What the Output Means

You’ll get a table showing:

N — how many respondents answered this question
Mean — the average score
Std. Deviation — how spread out the scores are
Minimum and Maximum — the range

For example, if you asked 200 students to rate their satisfaction on a 1–5 Likert scale, your output might show:

Satisfaction	N	Mean	Std. Dev.	Minimum	Maximum
Satisfaction	200	3.74	1.12	1	5

This tells you most students are moderately satisfied (3.74), with some variation (±1.12).

What to recommend: Always run Frequencies first for categorical variables (gender, education, treatment group) and Descriptives for continuous variables (age, scores, ratings). Don’t skip this step — descriptive stats are your reality check before you start hypothesis testing.

Step 4: Check Assumptions

Before running inferential tests, you need to verify that your data meets the necessary assumptions. Skipping this step can produce invalid results.

Testing Internal Consistency (Cronbach’s Alpha)

If you use a multi-item scale (like a 5-question survey measuring “workplace stress”), test reliability with:

Analyze → Scale → Reliability Analysis
Move your scale items into the “Items” box
Click OK

Interpretation: A Cronbach’s alpha above 0.70 is acceptable; above 0.80 is good; above 0.90 is excellent. If your alpha is below 0.70, consider removing items that don’t correlate well with the rest.

Checking for Normality

SPSS doesn’t have a built-in normality test in the menu, but you can use:

Analyze → Descriptive Statistics → Explore
Add your variable to “Dependent List”
Click Plot — check Normality prob plots
Click OK

Look at the Shapiro-Wilk test in the output table. If p > 0.05, your data is roughly normal. If p < 0.05, your data is significantly non-normal.

Practical tip: Many dissertation datasets are only slightly non-normal. In those cases, most statistical tests (t-tests, ANOVA) are robust enough to proceed anyway. Only worry if your data is wildly skewed or has extreme outliers.

Identifying Outliers

Outliers — extreme scores that distort results — can sneak into your data. In the Explore dialog above, also check Boxplot under “Plots.” Points outside the whiskers are potential outliers. You can then decide whether to keep them (they might be real) or remove them (they might be data-entry errors).

Step 5: Choose and Run the Right Statistical Test

This is the heart of your analysis. Here’s a decision framework to help you pick the right test:

Research Question	Statistical Test	SPSS Menu Path
Compare two groups (e.g., males vs. females on GPA)	Independent-samples t-test	Analyze → Compare Means → Independent-Samples T Test
Compare two paired measures (e.g., pre-test vs. post-test)	Paired-samples t-test	Analyze → Compare Means → Paired-Samples T Test
Compare three or more groups (e.g., different teaching methods)	One-way ANOVA	Analyze → Compare Means → One-Way ANOVA
Compare many variables across groups (e.g., ANOVA with multiple dependent variables)	Multivariate ANOVA (MANOVA)	Analyze → General Linear Model → Multivariate
Check relationship between two continuous variables	Pearson correlation	Analyze → Correlate → Bivariate
Predict an outcome from one or more predictors	Linear regression	Analyze → Regression → Linear
Compare proportions between groups	Chi-square test	Analyze → Descriptive Statistics → Crosstabs

Example 1: Independent-Samples t-Test

Suppose you want to know if male and female students differ in their average course grades.

Analyze → Compare Means → Independent-Samples T Test
Put the outcome variable (e.g., GPA) in the Dependent Variable box
Put the grouping variable (e.g., Gender) in the Grouping Variable box — define your groups (1 = Male, 2 = Female)
Click OK

Reading the output: Look at the Levene’s Test for Equality of Variances. If p > 0.05, assume equal variances. Then look at the t value, df (degrees of freedom), and Sig. (2-sided) — this is your p-value.

A significant result (p < 0.05) means the groups differ meaningfully. If p = 0.032, male and female students have statistically different average GPAs.

Example 2: Pearson Correlation

Suppose you want to see if hours of study time correlate with exam scores.

Analyze → Correlate → Bivariate
Move both variables into the variables box
Select Pearson as the correlation coefficient
Click OK

Reading the output: The r value tells you the strength (0.0 = no relationship, 1.0 = perfect relationship). The Sig. value tells you if it’s significant. A correlation of r = 0.65, p < 0.001 means a strong, significant positive relationship.

Example 3: Linear Regression

Suppose you want to know which factors predict GPA: study hours, attendance, and class participation.

Analyze → Regression → Linear
Put GPA in the Dependent box
Put the three predictors in the Independent(s) box
Click OK

Reading the output: The Model Summary table shows R² (how much of the GPA variation your model explains). The Coefficients table tells you which predictors are significant. For example, if “study hours” has a Beta of 0.42 with p < 0.001, it’s a strong predictor of GPA.

Decision-oriented advice: For most dissertations, regression is the most complex test you’ll run. Start with correlation to understand your variables, then move to regression to build your prediction model. Don’t force a regression model with too many predictors — keep it to a reasonable number (roughly 1 predictor per 10–20 cases) to avoid overfitting.

Step 6: Interpret Your Output

SPSS generates an Output Viewer window full of tables. Here’s what to look for in each common output:

t-Test Output

Levene’s Test: Determines whether variances are equal
t statistic: The test value
df: Degrees of freedom
Sig. (2-tailed): The p-value (look for < 0.05)
95% CI: Confidence interval for the difference

ANOVA Output

F value: The ratio of between-group to within-group variance
Sig.: The p-value
Eta squared (η²): Effect size (0.01 = small, 0.06 = medium, 0.14 = large)

Correlation Output

r: Correlation coefficient
Sig. (2-tailed): p-value for significance
Sample size (N)

Regression Output

R²: Proportion of variance explained
F (model): Overall model significance
Beta: Standardized coefficient (predictive strength)
t and Sig.: Significance of each predictor

Common mistake students make: Reporting only the p-value without the effect size. APA format requires reporting exact p-values and effect sizes (Cohen’s d for t-tests, η² for ANOVA, r for correlation). This tells readers not just whether something is significant, but how strong the effect is.

Step 7: Write Results in APA Format

APA style has specific formatting rules for statistical reporting. Here are examples for each common test:

Reporting a t-Test

An independent-samples t-test was conducted to examine differences in course grades between male and female students. The results indicated a statistically significant difference, t(48) = 2.45, p = .019, d = 0.35, with male students reporting higher average grades (M = 3.42, SD = 0.52) than female students (M = 3.18, SD = 0.58).

Reporting a Correlation

A Pearson correlation was conducted to assess the relationship between hours of study per week and course grade. The results showed a strong positive correlation, r(198) = 0.65, p < .001.

Reporting a Regression

A linear regression was conducted to predict GPA from study hours, attendance, and class participation. The model explained 43.2% of the variance in GPA, F(3, 96) = 24.31, p < .001, R² = 0.432. Study hours (β = 0.42, p < .001) and attendance (β = 0.28, p = .003) were significant predictors, while class participation was not (β = 0.11, p = .18).

APA formatting rules:

Italicize statistical symbols (t, F, p, M, SD)
Report exact p-values (not p < .05, but p = .023)
Use three decimal places for p-values (except when p < .001, which is written as p < .001)
Report effect sizes with every test
Round statistical values to two decimal places (except p-values, which get three)

Common Mistakes Students Make When Using SPSS

1. Skipping Variable View Setup

As mentioned above, skipping Variable View means unreadable output tables. This is the most common beginner mistake and is completely avoidable.

2. Using the Wrong Statistical Test

Using a chi-square when you need a t-test, or vice versa, invalidates your results. Use the decision framework above to pick the right test.

3. Ignoring Effect Sizes

A statistically significant result (p < .05) might still be so small as to be practically meaningless. Always report effect sizes.

4. Forgetting to Define Group Codes

When running a t-test or ANOVA, SPSS needs to know which numbers represent which groups. If you forget to define groups, SPSS will use all your data, producing an error or misleading results.

5. Misinterpreting Correlation as Causation

Correlation shows association, not cause-and-effect. Never write “X causes Y” — write “X is associated with Y.”

6. Not Checking for Missing Data

If half your participants skipped a question, SPSS will exclude them from all analyses (listwise deletion). Check for missing values before you start.

SPSS vs. Other Statistical Software: A Quick Comparison

Feature	SPSS	R	Excel
Interface	Point-and-click menus	Command-line/script	Spreadsheet
Learning curve	Easy	Steep	Easy
Statistical coverage	Broad (standard tests)	Extremely broad (every test)	Limited
Cost	Paid (IBM license)	Free	Often included
APA-ready output	Yes	Manual formatting	Manual formatting
Data handling	Good for 10,000+ rows	Excellent for any size	Struggles above 10,000 rows
Best for	Students needing quick results	Researchers needing advanced methods	Basic descriptive stats

Recommendation: If your program requires SPSS, use it. It’s the safest choice for most dissertation committees because it’s familiar and produces publication-ready output. If you have the time and willingness to learn, R is free and more powerful — but only if you’re comfortable with scripting. For most graduate students, SPSS strikes the right balance between power and usability.

When to Get Help with SPSS

There are three common scenarios where students should consider external support:

Your analysis plan is unclear — If you can’t translate your research questions into statistical tests, seek help before you start collecting data.
Your data is messy — If you have hundreds of missing values, coded incorrectly, or unusual distributions, professional help can save months of frustration.
Your committee expects advanced methods — If you need structural equation modeling, multilevel modeling, or factor analysis beyond SPSS’s menu options, consider additional tools (AMOS, Mplus) or expert consultation.

Summary: Your SPSS Data Analysis Checklist

[ ] Import your data from Excel or CSV
[ ] Set up Variable View with names, labels, and value codes
[ ] Run Frequencies and Descriptives to understand your data
[ ] Check reliability (Cronbach’s alpha ≥ 0.70)
[ ] Check assumptions (normality, outliers)
[ ] Choose your test using the decision framework above
[ ] Run the test and interpret the output tables
[ ] Write results in APA format with exact p-values and effect sizes

Related Guides

Statistical Analysis for Dissertations — A broader introduction to statistical analysis methods
Writing Academic Tone and Style in Dissertations — APA formatting guidelines for dissertation writing

Next Steps

Analyzing your data is one of the most challenging parts of your dissertation. If you need help with SPSS setup, statistical interpretation, or APA formatting of your results, our professional writing team can assist you. Visit our Order page to get started, or contact us for a free consultation.