Graphing My Confidence Interval Data

#1
Thanks for looking at this. I might be approaching this problem the wrong way because it seems like it would be simple to graph three lines for which I have data.

I have plotted a line that represents a mean species diversity estimate (Y axis is species count, X axis is individuals collected). I have the upper and lower bound confidence interval data that I would like to add to the line graph.

But, when I try to add the mean, upper bound and lower bound lines to the Y axis it ALWAYS gives me an index for the X axis (as if the x axis is not the same for all of them... they do vary a bit but they are very close). I could use some advice on how to produce the three lines together in SPSS v21.
 

bugman

Super Moderator
#2
Can you post your plot?

The issue maybe that lines graphs for catergorical values (which it sounds like your x-axis is) don't usually make sense. These (line graphs) are usually used for times (years seasons etc...) so you might run into trouble. Try plotting a scatter plot and adding your CI around the individual points. Im not famialr with SPSS but that method works in EXCEL.
 
#3
Bugman, thanks! My species are insects. These are the plots (attached) representing the # of species estimate mean and the upper and lower bound confidence intervals. The x axis variable is number of individuals collected. (Ultimately I want to split these the ten lines/sample areas into 2 or more plots because all 30 lines would be confusing on one.)
 

bugman

Super Moderator
#4
Ok, I see what you are doing (I think). But can you please explain a little bit better what you are trying to do so I can offer better advice.
Are you simply wanting to plot the mean and error for ten species?
Are there other factors involved (i.e. sites, years, season...) or are these results just a species inventory of a single site or sampling event.
Why have you opted for this line plot approach?
 
#5
Sorry for the confusion. Your questions are helping me to clarify, so thanks again.

bugman: Are you simply wanting to plot the mean and error for ten species?
No.

bugman: Are there other factors involved (i.e. sites, years, season...) or are these results just a species inventory of a single site or sampling event.

Each of the lines in the top graph represent a study site sampled over multiple years. The lines are a combination of real data and projections from a program called EstimateS.

bugman: Why have you opted for this line plot approach?

I am trying to produce a version of the attached with my data from 10 study sites. See attached "Rainbow" graph. The shaded areas of the image below are the confidence intervals.
 

bugman

Super Moderator
#6
Ok, that clarifies that for me.

I can only say (because I am not an SPSS user) to check the SPSS users manual and to first make sure that your dataframe is set up correctly.

Does EstimateS do graphics? I found the "rainbow plot" you are referring to in the EstimateS users manual. This is fairly straight forward to do in R, if you were inclined to give that a go.

Sorry to lead you down the garden path, I thought that if it were a straight forward issue, then I would have been able to help out in SPSS, but it looks like its a bit more complicated.
 
#7
Bugman, the garden path was useful and I appreciate your taking the time. I'll check my data frame in SPSS- I think there may be a way to layout the input data to get the result I want.

I would prefer to find a way to do this in SPSS since I'm more familiar with it...but yes, the Rainbow was produced in R so I might try it there.