# Comparing snake diversity across primary, secondary and degraded sites - Shannon diversity index?

#### StatsUnknown

##### New Member
Hi Everyone - hopefully you can help me with the following!

I am looking to ascertain whether there is a difference in snake diversity across primary, secondary and degraded sites, encompassing a protected reserve. I have a total of ten survey sites; two in primary, three in secondary and five in degraded forested areas. An example of the raw data is included below.

Sites 1 and 2 (primary forest)
Sites 3, 4 and 9 (secondary forest)
Sites 5, 6, 7, 8 and 10 (degraded forest)

Please could you let me know if the following steps would be an appropriate means in which to analyse the data:
1). Calculate the Shannon diversity index (H) for each of the ten sites.

2). Compare the Shannon diversity indices of primary, secondary and degraded sites using a one-way ANOVA.

3). If a significant difference is calculated, use a post-hoc test (unsure which one?) to identify between which types of forest (primary, secondary and degraded) the significance lies.

4). Do you recommend that I also look at calculating evenness? I believe however (rightly or wrongly), that the Shannon diversity index already takes evenness into account, so I am not sure how necessary this is.

5). Is there anything else you suggest I do with the data, from which I can make meaningful conclusions?

It is also worth noting that there have been 17 snake species historically documented in the proximal surveying area, however only 10 species were sighted during the surveying period. Am I right to include zero values for the seven unsighted species, or do I exclude these species entirely? I have so far left them in, as can be seen in the data table.

I look forward to hearing your thoughts!

Cheers!!

#### Karabiner

##### TS Contributor
Does that mean you want to 3 types of sites with regard to Shannon index?
With n1=2, n2=3 and n3=5, your sample sizes are too small for a meaningful analysis
of variance, AFAICS.

With kind regards

Karabiner

#### StatsUnknown

##### New Member
Does that mean you want to 3 types of sites with regard to Shannon index?
With n1=2, n2=3 and n3=5, your sample sizes are too small for a meaningful analysis
of variance, AFAICS.

With kind regards

Karabiner
Thanks for your prompt reply! I would like to compare three types of sites (primary, secondary and degraded). I realise that I have an uneven number of each type of site. It was also not possible to have more sites (especially primary and secondary), owing to the size and protection status of the reserve.

Many thanks!

#### Karabiner

##### TS Contributor
I do not critisize that you have been able to obtain a small sample only.
Nevertheless with n=2 in one group and n=3 in another group, and n(total) = 10
you cannot perform a meaningful oneway ANOVA, at least to my knowledge.
For example, one cannot check distributional assumptions. H-test would
certainly work, but power would still be very low.

With kind regards

Karabiner

#### StatsUnknown

##### New Member
I do not critisize that you have been able to obtain a small sample only.
Nevertheless with n=2 in one group and n=3 in another group, and n(total) = 10
you cannot perform a meaningful oneway ANOVA, at least to my knowledge.
For example, one cannot check distributional assumptions. H-test would
certainly work, but power would still be very low.

With kind regards

Karabiner
I don't take your comments to be critical, I certainly appreciate and thank you for your feedback! I realise that the stats I could use would be limited but I was hoping that I could use something.

Many thanks!