Descriptive and Inferential Statistics in SPSS
Ordinal Scale: Size of hometown or city
Size of home town or city
10,00 to 99,999
100,000 to 499,999
Paired Samples t-Test and ANOVA in SPSS
Example: AA Concepts Survey
Are there differences in preferences for Economy 4 seat hybrid and Standard 4 seat hybrid
ANALYZECOMPARE MEANSPAIRED SAMPLES t-TEST.
Select Preference: Eco
Hypothesis Testing: Non-Directional Hypothesis
For categorical data (or proportions for nominal or ordinal scales):
Write non-directional null (Ho) and alternative (Ha) hypotheses:
Ho: = Hstatistically, the population proportion, , and H are the same.
Formula for significance of the difference between two proportions/percentages
(nominal & ordinal scales):
1. Write the null and alternative hypotheses.
Ho: p1 = p2statistically, the two percentages are not significantly differen
ANOVA in SPSS
If sig value is less than or equal to 0.05, we reject the null and conclude that at least two group means are
different. The Duncans post-hoc test reveals the differences.
1. People in the age group of 19-24 years have the highest mean prefe
MODE & MEDIAN Categorical datanominal or ordinal scales
SPSS command sequence:
Points to think about:
1. How does the frequency distribution expla
Sample t-Test in SPSS
p is highlighted. Sp was previously calculated as 1.4 (please see page 5). Z-calculated = (26.7
25)/1.4 = 1.214286 or 1.214.
Since -1.96 1.214 +1.96, we fail to reject the null hypothesis.
For Metric data (interval and ratio scales)
One-Sample Test in SPSS
(i)The default confidene level in SPSS is 95%. This can be changed to the desired level by
clicking on options in the One-Sample Test window and typing in the desired level (e.g., 99%).
(ii) For large sample, t-test and z-tes
Inferencial Analysis in SPSS
Example from AAConcepts data:
We are 95% confident that the true % of the population that lives in a city of 1 million and
more population falls between 24.0% and 29.4%.
1. Compute frequency table for Size or town or city.
Computing Confidence Intervals in SPSS
1. Compute using: z = (p - ) / sp
For metric data (interval and ratio scales):
1. Computing Confidence Intervals:
ANALYZECOMPARE MEANSONE-SAMPLE t-TEST with test value 0.
Step 1: Click ANALYZE.
Step 2: Move mouse