Get AI summaries of any video or article — Sign up free
Running, Interpreting, and Reporting Descriptive Statistics using SPSS thumbnail

Running, Interpreting, and Reporting Descriptive Statistics using SPSS

Research With Fawad·
5 min read

Based on Research With Fawad's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.

TL;DR

Use **Analyze → Descriptive Statistics → Frequencies** for categorical variables like gender, and report counts plus appropriate distribution summaries.

Briefing

Descriptive statistics in SPSS hinge on matching the analysis tool to the type of variable—categorical data gets Frequencies, while continuous data gets Descriptives. In the session, the workflow starts under **Analyze → Descriptive Statistics**, where **Frequencies** is used to summarize discrete categories like gender. After selecting the variable (e.g., Gender), the analysis can report counts and distribution measures such as mode, minimum, and maximum. The output also distinguishes between **percentage** and **valid percentage**: valid percentage ignores missing cases, while percentage reflects them. That difference matters when reporting results, because a table can show 100% valid percentage even when some responses are missing—an example is given where removing cases changes percentage but leaves valid percentage at 100%.

Reporting the results follows a consistent AP-style approach: copy the SPSS table, add a table number and italicized caption/title, and then present the descriptive statistics for each variable before moving on to hypothesis testing or model evaluation. For categorical variables, the session demonstrates how to report both **n** and percentages (e.g., male vs. female counts and their percentages) and how to state the total sample size.

For continuous variables, the process shifts to **Analyze → Descriptive Statistics → Descriptives**. Here, the key outputs are **mean**, **standard deviation**, and related dispersion statistics such as minimum, maximum, variance, range, skewness, and kurtosis. The session illustrates how to translate SPSS output into thesis or paper language—for instance, reporting the average age of students along with standard deviation.

A more complex reporting need arises when a study uses a multi-item construct measured by indicators. For a construct like **CSR (corporate social responsibility)** measured with eight items, the session shows how to summarize item-level descriptives and also compute an **overall composite mean score**. In SPSS, that composite is created via **Transform → Compute Variable**, using the mean of CSR1 through CSR8. Once computed, the overall mean and standard deviation provide a single interpretive summary of respondents’ perceptions, which the session links to the response scale (e.g., a mean near 4 suggests agreement on a strongly disagree to strongly agree scale). It also emphasizes that item interpretation should be grounded in the questionnaire wording—such as explaining what CSR1 represents by checking the instrument.

Finally, the session covers descriptive comparisons across groups and cross-tabulation. To report CSR scores by country (Pakistan, China, Italy), it uses **Analyze → Compare Means → Means**, treating **country** as the factor and requesting mean, standard deviation, and sample size. For categorical relationships, it uses **Analyze → Descriptive Statistics → Crosstabs**, placing **country** in rows and **gender** in columns to produce counts by cell and a bar chart. The session closes by tying these outputs to where they belong in a write-up: demographic profiles at the start of a paper, or descriptive statistics right before evaluating measurement and structural models in thesis work.

Cornell Notes

SPSS descriptive statistics depend on variable type: use **Frequencies** for categorical variables (e.g., gender) and **Descriptives** for continuous variables (e.g., age). Frequencies output counts plus measures like mode, minimum, and maximum, and it distinguishes **percentage** from **valid percentage**—valid percentage excludes missing data. For multi-item constructs such as **CSR** measured by CSR1–CSR8, the session recommends reporting item-level descriptives and creating a composite score using **Transform → Compute Variable** to compute the mean across items. To compare descriptive results across groups like countries, use **Compare Means → Means** and report mean and standard deviation (often with sample size). For relationships between categorical variables, use **Crosstabs** with rows and columns to produce cell counts and charts.

When should SPSS use Frequencies versus Descriptives?

Use **Frequencies** for categorical variables with discrete values (e.g., gender: male/female; country: Pakistan/China/Italy/Spain). Use **Descriptives** for continuous variables where mean and dispersion are meaningful (e.g., age or an overall CSR score). In the session’s workflow, gender is analyzed with Frequencies, while age is analyzed with Descriptives.

What’s the practical difference between percentage and valid percentage in SPSS output?

**Valid percentage** excludes missing data, so it can sum to 100% even when some responses are missing. **Percentage** includes missing cases in the denominator, so it can be less than 100% when missing values exist. The session demonstrates this by rerunning frequencies after removing cases: cumulative percentage reaches 100% while percentage drops (e.g., 99%), reflecting missing-data handling.

How should a multi-item construct like CSR be reported descriptively?

First, run descriptives for the individual items (CSR1–CSR8) and report their means and dispersion. Then compute an overall composite score by averaging the items: **Transform → Compute Variable**, set the new variable as the mean of CSR1, CSR2, …, CSR8 (using a mean function). Report the composite mean and standard deviation to summarize overall perception.

How do you create and interpret an overall mean score for CSR in SPSS?

Create the composite via **Transform → Compute Variable** by computing the mean across CSR1–CSR8 (separating each indicator with commas inside the mean expression). Interpretation should match the questionnaire scale: if the scale runs from strongly disagree to strongly agree, an overall mean around 3.84 (close to 4) indicates respondents lean toward agreement. The session also notes that the meaning of specific items (e.g., CSR1) must be taken from the questionnaire.

How can descriptive statistics be compared across countries for a continuous score like CSR?

Use **Analyze → Compare Means → Means** with **country** as the factor list and CSR as the dependent variable. Select options such as mean, standard deviation, and number of cases. Reporting then typically lists each country’s sample size and the mean (with standard deviation in parentheses).

How does Crosstabs differ from Descriptives, and what does it produce?

**Crosstabs** produces a contingency table with rows and columns for categorical variables, showing counts in each cell (and optionally charts). In the session, country is placed in rows and gender in columns to show how many male and female respondents appear in each country, with a bar chart summarizing the cross-tabulation.

Review Questions

  1. If a categorical variable has missing responses, how would you decide whether to report percentage or valid percentage?
  2. What SPSS steps would you use to compute an overall CSR composite score from CSR1–CSR8, and what should guide the interpretation of its mean?
  3. How would you report descriptive statistics for age versus gender in an AP-style thesis or paper?

Key Points

  1. 1

    Use **Analyze → Descriptive Statistics → Frequencies** for categorical variables like gender, and report counts plus appropriate distribution summaries.

  2. 2

    Report **valid percentage** when you want results that ignore missing data; use **percentage** when you want missing cases reflected in the denominator.

  3. 3

    Use **Analyze → Descriptive Statistics → Descriptives** for continuous variables and report mean and standard deviation (plus min/max if relevant).

  4. 4

    For multi-item constructs (e.g., CSR measured by CSR1–CSR8), report both item-level descriptives and a composite mean created via **Transform → Compute Variable**.

  5. 5

    Interpret composite means using the questionnaire scale direction (e.g., a mean near the upper end suggests agreement).

  6. 6

    Compare descriptive statistics across groups like countries using **Analyze → Compare Means → Means**, reporting mean and standard deviation (often with sample size).

  7. 7

    Use **Analyze → Descriptive Statistics → Crosstabs** to show relationships between categorical variables via row/column contingency tables and charts.

Highlights

Valid percentage can hit 100% even when percentage is lower, because valid percentage excludes missing responses.
A construct like CSR can be summarized with an overall mean by averaging CSR1–CSR8 using **Transform → Compute Variable**.
Descriptive comparisons across countries for CSR are handled through **Compare Means → Means**, with mean and standard deviation reported per country.
Crosstabs turns categorical relationships into a contingency table (and bar chart) by placing one categorical variable on rows and another on columns.

Topics

Mentioned

  • SPSS
  • CSR