I have mean biomass from a number of plots in each forest types. I want to know if there is any significant difference in mean biomass among these forest types or not. Which test is more appropriate in this matter? Thank you.
One way analysis of variance at either 95 or 99 per cent confidence level. If p
value is less than 0.05 or 0.01 depending on the confidence level you will use, you proceed with a posthoc test (preferably the Least Significant Difference test) to separate the means.
Is necessary to apply an normality and hemogeneity variance test on the variable biomass before Anova test. If the biomass variable is not normal and the variance is not homegeneous you can use an nonparametric test Kruskal Wallis o Wilcoxon test. Another way is to transform the biomass variable until reaching it is normal and hemogeneity variance.
Interesting question. My advice, check for normality and homogeneity first. If your data are homogeneous and normally distributed you can use one-way ANOVA with the Tukey test. If the data is not normally distributed, you can use square root or logarithm transformation. See also the below links:
ANOVA is the appropriate statistical test in this case. But before using ANOVA, you must check for normality of your data set. If your data set is normally distributed, fine. If otherwise, you can apply the nonparametric counterpart of ANOVA (Kruskal Wallis test) or transform your variables to convert non-normally distributed variables to normally distributed variables.
Ideally, you are suppose to use one-way ANOVA (with forest types as the factor), after testing for normality and homogenous of variance; otherwise, non-parametric. Currently, I'm working on similar research, i.e. estimating biomass of three forest management types (two Community Forest Areas, two Govt. Forest Reserves and two Sacred Groves). Thus, I want to know if there is significant different within and between the forest management types. In my own case, I intend to use two-way ANOVA (after testing for normality and homogenous of variance)-two factors: location and forest Mgt. types i.e. within forest types (all the six locations of forest); and between forest Mgt. types (i.e. between the 3 forest Mgt. types).