Hi there,
I am currently analyzing 6 PCB congeners (Polychlorinated Biphenyls) in the river sediments. I am using SPSS for data analysis. I don't quite understand how to explore normality considering one group of concentration for each PCB congener when, in fact, I have two seasons, 9 sampling points and 3 repetitive samples for each sampling points. I am considering Shapiro Wilk W test for normality. My goal is to present summery statistics and show the differences among the concentrations of PCBs in the sampling points and seasonal variations by using one-way ANOVA.
Being a novice in this field, any kind of help is highly appreciable. Thanks in advance.
With an n=9 it doesn't matter much. I'd probably just do graphical comparisons. If you really want a hypothesis test try Shapiro Will. Best wishes David Booth Don't forget normal probability plots
David Eugene Booth Thanks for your kind response. I am asking whether taking all pollutant concentration data (that is 6 individual PCB congeners) in dependent list in SPSS for Shapiro wilk test while not inputting seasons (that is 2 seasons) and sampling points (9 sampling points) in factor list, or taking all pollutant concentration data in dependent list and seasons or sampling sites in factor list in SPSS. In the data sheet, 6 individual PCB congeners and season name are the variables. I am giving a png file to show how does it look. I am looking forward to it.
Assuming that I understand you I would just do the standard residual analysis for each of your regression models as discussed in Kutner applied linear Statisical models which is available in the z-library for download. Best wishes, David Booth
- Badges
- Science topic
- Similar topics
- Chemistry, Organic
- Organic