Skip to content
Tutorials Center
My learning

Topics

Lesson 16 of 39 ยท Descriptive Statistics

SPSS Descriptives: Mean, SD, Skewness, Kurtosis and Z-Scores

Use Analyze > Descriptive Statistics > Descriptives in SPSS: options, standard error, skewness and kurtosis formulas, saving standardized z-score variables, listwise N and syntax.

Descriptives is the lean cousin of Frequencies: it handles only numeric variables, and it prints one compact row per variable instead of long tables. It also has a feature nothing else has so conveniently: it can save z-scores as new variables. In this lesson you learn what each statistic means, how SPSS computes the less obvious ones (standard error, skewness, kurtosis) and how to use standardized scores.

The dialog

Analyze > Descriptive Statistics > Descriptives...

  1. Move numeric variables into Variable(s).
  2. Optionally tick Save standardized values as variables (creates a new variable with the prefix Z, such as Zage).
  3. Click Options... to pick statistics:
    • Mean, Sum;
    • Dispersion: Std. deviation, Variance, Range, Minimum, Maximum, S.E. mean;
    • Distribution: Kurtosis, Skewness;
    • Display Order: Variable list, Alphabetic, Ascending means, Descending means.
  4. OK.

The syntax:

TEXT
DESCRIPTIVES VARIABLES=age income visits
  /SAVE
  /STATISTICS=MEAN STDDEV MIN MAX SEMEAN SKEWNESS KURTOSIS.

The output is the Descriptive Statistics table: one row per variable (N, Minimum, Maximum, Mean, Std. Deviation and any extra columns), plus Valid N (listwise). Each row's N counts that variable's valid cases; the last row counts cases valid on all of them.

What each statistic means

Statistic Meaning
Mean Arithmetic average
Sum Total of the values
Std. deviation Typical distance from the mean, in the units of the variable (sample formula, divide by n - 1)
Variance Standard deviation squared
Range Maximum minus minimum
S.E. mean Standard error of the mean: standard deviation divided by the square root of N; how much the sample mean would vary between samples
Skewness Asymmetry: 0 symmetric, positive = long right tail, negative = long left tail
Kurtosis Tail heaviness relative to a normal curve (SPSS reports excess kurtosis, so a normal distribution scores about 0)

SPSS prints a Std. Error next to Skewness and Kurtosis. The ratio statistic / standard error gives a rough z test: values beyond about plus or minus 1.96 (or 3.29 for large samples) suggest that the distribution differs from normal. With big samples this flags harmless deviations, so look at the plot as well. See explore, outliers and normality.

Python Stand-in: the Descriptives statistics, formula by formula

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The positive skewness and large kurtosis come from the single extreme value. In SPSS the numbers are the bias-corrected versions shown here, which differ from the simple "population" formulas some textbooks use.

Z-scores

A z-score (standard score) says how many standard deviations a value is from the mean:

z = (x - mean) / sd

A z of 0 is average, +1 is one standard deviation above, -2 is two below. Z-scores put different variables on the same scale and appear in:

  • outlier screening (|z| > 3);
  • comparing scores from different tests ("Is 72 on test A better than 65 on test B?");
  • combining variables with different units into one index;
  • standardizing predictors in regression.

With Save standardized values as variables, SPSS adds Zage, Zincome, ... computed with the sample standard deviation of the valid cases. The new variables have mean 0 and standard deviation 1. SPSS labels them "Zscore: Age in years". You can compute them yourself in syntax:

TEXT
COMPUTE z_score = (score - 62.4) / 11.8.

but typing the constants by hand is error-prone; use the saved variables or AGGREGATE to attach the mean and SD. With Split File on, SPSS standardizes within each group, which is useful ("relative to peers") but can surprise you.

Python Stand-in: compare two tests with z-scores

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Both scores sit above average; the z-scores show which is further above its own test's average, a fairer comparison than the raw points.

Descriptives vs Frequencies vs Explore

Need Use
Quick table for many numeric variables Descriptives
Counts, percentages, percentiles, charts Frequencies
Confidence interval, outliers, normality tests, by group Explore
Statistics by categories of another variable Compare Means > Means (compare means)
Z-scores Descriptives (save standardized values)

Missing values

Descriptives uses listwise deletion only for the final row Valid N (listwise); each variable's own row uses its own valid cases ("analysis by analysis"). The Descriptives Options dialog does not offer a pairwise or listwise switch for the means. If you need statistics on the same cases for every variable, filter to the complete cases first (see select, split and weight).

Good habits

  • Read N first. If it is lower than expected, find out why.
  • Compare the range with what is possible (Minimum and Maximum are your first data check).
  • Use mean and SD only for approximately symmetric scale variables; otherwise add the median and IQR from Frequencies.
  • Do not describe a skewed variable using only the mean.
  • Report the SD with the mean ("M = 41.2, SD = 7.9").

Common mistakes

  • Reading a Std. Error as a Std. Deviation.
  • Running Descriptives on a nominal code and interpreting the mean.
  • Standardizing after a Split File and forgetting the z-scores are within groups.
  • Treating |z| > 3 as proof of an error; it is only a signal to look.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Which statistic describes how much a sample mean would vary from sample to sample?

    The standard error of the mean is the standard deviation divided by the square root of N.
  2. 2. A positive skewness value means:

    Positive skew means a long right tail, usually with the mean above the median.
  3. 3. What does Save standardized values as variables create?

    Each new variable holds the z-score of the original, with mean 0 and standard deviation 1.
  4. 4. A student scores z = -1.5 on a test. What does that mean?

    A z-score is measured in standard deviations from the mean.

Summary

  • Descriptives gives a compact numeric summary: N, min, max, mean, SD and optionally SE, skewness and kurtosis.
  • SPSS uses sample (n - 1) formulas and reports excess kurtosis, with standard errors for skewness and kurtosis.
  • Saving standardized values creates z-score variables (mean 0, SD 1).
  • Each row's N is per variable; Valid N (listwise) counts complete cases.
  • Use Frequencies for percentiles and charts, Explore for outliers and normality.

Next: Explore: Outliers, Confidence Intervals and Normality