Error bars on graph, skewed data, negative standard deviation

I have only had basic stats. I conducted some research where I have four different treatments and results for the data range from 0-6. I don't need any elaborate stats test, I just want to know how I should handle the data.

I summed each treatments results and got 162, 72, 75, 14. I turned these numbers into percentages and graphed them. But I am required to put error bars on the graph. So I calculated the standard deviation and when I placed the error bars on the graph, the SD is a negative value for the 5% data value and the data that I tested cannot be a negative value.

I need to report some type of descriptive statistic with this data. I also calculated the mean and for the treatment group with 5% was 0.17. The data value cannot be less than 1 because I'm researching eggs laid on a bean and a beetle can't lay 0.17 eggs.

Thanks you so much for your help!

• Please show how you got a negative standard deviation. I expect you don't mean exactly that. I doubt you should be summing your results; what is the question you're interested in finding out from the data? A mean can certainly be less than 1; there's nothing that says the mean has to be an observable value -- it frequently isn't. – Glen_b Apr 10 '15 at 3:36
• While not duplicates, you may find discussion at this question and possibly also this one good starting points for understanding issues in and possibly reframing your question. The first one deals with an error bar exceeding a known upper bound (rather than a lower one), but the discussion largely carries over. – Glen_b Apr 10 '15 at 3:42