Transform data for factorial two-way ANOVA/General Linear Model??

Hi, I'm not very statistically minded but I would really appreciate some guidance. I am needing to do several stat tests for my MSc dissertation.

I have one dependent variable (species count) and one independent variable (year), but two species. I have constructed a basic count against year graph for each species. I have then constructed a residual plot to determine linearity. I have concluded that the data of both species is non-linear.

I have also checked for normality (Shapiro-Wilks test), and have concluded that neither species data is not normally distributed and therefore further non-parametric tests are required.

I am wanting to determine whether there are any associations between the two species and believe a factorial two-way ANOVA is the most appropriate test, however as my data is non-parametric would I have to transform my data first? I am using SPSS and when constructing a factorial two-way ANOVA you do so in the same way you would construct a General Linear Model. Apologies if this is an obvious question but is a factorial two-way ANOVA the same as a GLM?

Any advice would be most appreciated, I can upload the data if anyone is able to help!

Many thanks in advance!


TS Contributor
I am wanting to determine whether there are any associations between the two species
I am not sure what you mean by this.

With 1 factor "year" and 1 factor "species", and count data,
you could crosstabulate year versus species and perform a
Chi² test of independence. But this will give you only information
whether there is an association between year and species. If
you want to know something else, please be more specific.

With kind regards


however as my data is non-parametric
there is no such thing as nonparametric data.
There are nonparametric tests, though.
Hi thank you both very much for your replies! I think potentially I am trying to be too ambitious and perhaps a General Linear Model is not the most suitable for my research question. Trying to keep is simple!

SO taking a step back, I have two species, the common frog and marsh frog and I have count data on both species since 1980. The marsh frog is non-native and I am trying to ascertain whether since the introduction of marsh frog they are negatively impacting/outcompeting the common frog. If they were outcompeting you would expect to see the population counts of the marsh frog increase whilst the population of common frogs decrease. I want to statistically test this "mode/hypothesisl"!

How is it best to go about this, as mentioned before both species are non-normally distributed so I'm assuming non-parametric tests. I don't really know where to start, maybe a Mann Whitney U test or Kruskal-Wallis test?

(Sorry I cant actually upload the actual data to a public forum, I have not got permission for this but please find attached fabricated data!)

Many thanks once again!