SPSS Cheat Sheet
Bootstrapping
1. Analyze > compare means > independent samples t-test
2. Add your grouping variables (for grouping variables define groups)
3. Click bootstrap
4. Check perform bootstrapping option
5. # of sample 2000-5000
6. Level of CI at 95% + check bias correct
7. Continue, paste + run!
Interpretation of output
Table 1: specific actions bootstrap
o 5000 samples
o Bias correct
Table 2: t-test group statistic
o Tells us how many candies in each group
o CI for mean + SD
o Mean, SD, standard error mean
Table 3: independent samples t-test
o Levene’s test > if p-value above .05 assume variances are equal
o T-value (Sig. 2 tailed) above .05 no statistical difference
Table 4: bootstrap results
o Mean difference in average weight
o CI for bootstrap > can have negative/positive differences
o Look at right row equal variances assumed/not assumed
Exact Test
1. Look for categorical variable to use
2. Analyze > non-parametric tests > legacy dialogue
3. Choose test (in this example, chi-squared test)
4. Click exact button + select exact option
5. Set test time (can be 5 minutes)
6. Paste + run!
Crosstab Exact Test
, 1. Analyze > descriptive stats > crosstabs
2. Add variable to row + columns
3. Select exact approach
4. Select statistic
5. Choose chi-square test
6. Select your measure of association
7. Cells > select percentages & columns
8. Paste + run!
Interpretation of output
Table 1: case
o Tells us the number of cases
Table 2: cross-tabulation
o Tells you the proportions + associations
Table 3: chi-square test
o Look at Fisher > If p-value is below .05 is a statistically significant association
between the values
Table 4: measures
o Tells us the association
Confidence Interval
1. Finding mean (use Bootstrap)
Analyze > descriptive statistics > explore/frequencies
2. One sample t-test
Analyze > compare means > one sample t-test > insert test variables > options
> CI% at 95
3. One-way ANOVA = F-test
Analyze > compare means > one-way ANOVA > post-hoc > ‘Bonferroni’ >
add significance level
4. Two-way ANOVA
General > linear > univariate > add variables > click post-hoc > ‘Bonferroni’ >
but no significance level > options > add significance level
5. Linear regression
Analyze > regression > linear > stats > click on CI
6. Correlations
Analyze > compare > bivariate > add variables > must bootstrap > we get CI
7. Frequencies
2
Bootstrapping
1. Analyze > compare means > independent samples t-test
2. Add your grouping variables (for grouping variables define groups)
3. Click bootstrap
4. Check perform bootstrapping option
5. # of sample 2000-5000
6. Level of CI at 95% + check bias correct
7. Continue, paste + run!
Interpretation of output
Table 1: specific actions bootstrap
o 5000 samples
o Bias correct
Table 2: t-test group statistic
o Tells us how many candies in each group
o CI for mean + SD
o Mean, SD, standard error mean
Table 3: independent samples t-test
o Levene’s test > if p-value above .05 assume variances are equal
o T-value (Sig. 2 tailed) above .05 no statistical difference
Table 4: bootstrap results
o Mean difference in average weight
o CI for bootstrap > can have negative/positive differences
o Look at right row equal variances assumed/not assumed
Exact Test
1. Look for categorical variable to use
2. Analyze > non-parametric tests > legacy dialogue
3. Choose test (in this example, chi-squared test)
4. Click exact button + select exact option
5. Set test time (can be 5 minutes)
6. Paste + run!
Crosstab Exact Test
, 1. Analyze > descriptive stats > crosstabs
2. Add variable to row + columns
3. Select exact approach
4. Select statistic
5. Choose chi-square test
6. Select your measure of association
7. Cells > select percentages & columns
8. Paste + run!
Interpretation of output
Table 1: case
o Tells us the number of cases
Table 2: cross-tabulation
o Tells you the proportions + associations
Table 3: chi-square test
o Look at Fisher > If p-value is below .05 is a statistically significant association
between the values
Table 4: measures
o Tells us the association
Confidence Interval
1. Finding mean (use Bootstrap)
Analyze > descriptive statistics > explore/frequencies
2. One sample t-test
Analyze > compare means > one sample t-test > insert test variables > options
> CI% at 95
3. One-way ANOVA = F-test
Analyze > compare means > one-way ANOVA > post-hoc > ‘Bonferroni’ >
add significance level
4. Two-way ANOVA
General > linear > univariate > add variables > click post-hoc > ‘Bonferroni’ >
but no significance level > options > add significance level
5. Linear regression
Analyze > regression > linear > stats > click on CI
6. Correlations
Analyze > compare > bivariate > add variables > must bootstrap > we get CI
7. Frequencies
2
