Using SPSS

Numeric Variable

numbers

String Variable

series of letters

Date Variable

Date options

Label of a Variable

string of text to identify in more detail what a variable represents

Describe Nominal or Ordinal

Analyze —> Descriptive —> Frequency

How many of a Nominal or Ordinal

frequency of X

What percent are X

valid percent of X

Describe Scale variables

analyze —> descriptive —> explore

Mean age or Mean BMI

just put the value

Best measure of Central Tendency for X

look at boxplot and see for asterisk
if none use mean value

Analyze Scale based on Nominal

analyze → descriptive → explore

can be asked for mean or best measure of central tendency

Analyze Scale based on Nominal central tendency

outlier but no asterisk
mean value answer

Analyze Scale based on Nominal variation

since normal
mean ± standard deviation

Best way to Describe normal distribution

mean and SD

Best way to Describe abnormal distribution

median and range

New Variables from Existing

transform —> compute variable —> add variable

New Variable based on Dates

transform —> date and time wizard —> calculate with dates and times
calculate number of time units —> add variables latest —>interest unit —> label right

Recode into different variables

transform —> recode into different —> add variable —>

range just before and just after test value

Confidence Interval for 1 variable

analyze → descriptive → explore

Transform Data for abnormal

transform → create variable → label → lg10(V) → ok

only do for the scale variable

Running Runs test

analyze → nonparametric tests → legacy designs →

use original variable and given value

recode string into nominal

transform —> automate recode —>

Test Paired-Test assumption

compute variable → post-pre→ label and ok

assess using explore

Do Paired t-test

compare means → paired samples t-test → post 1 pre 2

Difference in Pre and Post

2 scale diff→ log10(difference) → explore → outlier do WXN

Do Wilcoxon-Signed rank test

non-parametric tests → legacy dialogs → 2 related samples → post 1 pre 2

Independent Samples t-test

analyze → describe → explore

compare means → independent samples t-test → set test and define group → continue

Running Mann-whitney Test

nonparametric tests → legacy dialogs → 2 independent samples → set test and define group → continue

Convert scale to nominal

If log data has no outliers, what data do you use

use the log data in the test

Anova

Compare means → one way → post hoc if overall significant → Bonferroni or tomhanes —> option’s —> descriptions→ homogenity

Kruskil

analyze → nonparametric → legacy → k independent samples

explore to get the min and max

Compare medians

graphs → chart builder → drag to center bar → bot left picture → scale then nominal → change to median