Hi all,

The hypothesis of my study is that different factors have an affect on the

formation of a certain geological formation. These formations are

categorized to 5 different types. The factors affecting the formations are

“sediment type” (1= clay, 2= sand….22 altogether) and “depth of the sample”

(from 9.9 m to 65 m). I have 88 samples. It is important to analyze the

different geological formations separately and this leads to rather small

sample sizes (7,11,12,13 and 45 samples in each of the 5 geological

formation categories).

I am confused which test in SPSS to use to get valid results. The variables

are categorical nominal (geological formations and sediment types) and scale

(depth). So far I have considered/tried different tests for all of the

samples together (88 samples). I tried using chi-square test since I have

nominal categorical variables but I realized that the data is not valid for

that test (48 cells (96,0%) have expected count less than 5. The minimum

expected count is ,04). For what I understand Fisher’s exact test is also

not possible..

I have considered One-way Manova but I’m pretty sure that the data I have

does not meet the requirements for that test either.. (For instance I don’t

have two or more variables at the interval or ratio level). SPSS version 22.

I am new to this and I would really appreciate any help.

Best Regards,

Irene

--

Sent from:

http://spssx-discussion.1045642.n5.nabble.com/=====================

To manage your subscription to SPSSX-L, send a message to

[hidden email] (not to SPSSX-L), with no body text except the

command. To leave the list, send the command

SIGNOFF SPSSX-L

For a list of commands to manage subscriptions, send the command

INFO REFCARD