# Multivariate analysis

#### biologist13

##### New Member
Hello,
I need a basic answer regarding the data matrix used in multivariate analysis.
It contains data of sample spots, which are categorical variables, species, and their abundance on a given scale (1-5), at each given location. What kind of data are species in this case - also categorical, but their abundance is ordinal?
What are your experiences regarding missing values - if I don't have a species present on one location, do I give it a zero or make another number, since PAST won't do PCA analysis with missing data?

Thank you in advance for your answers.

#### GretaGarbo

##### Human
If your dependent variable is abundance, e.g. the number of snails at a sample spot. Then you can count the number of snails 0, 1, 2, 3, ...

An zero is also a number, not a missing value. (And the number of snails is a ratio variable). The number of snail can be classified on a 1 to 5 scale. A zero number of snails would then be classified as "1" on the 1 to 5 scale.