7 replies to this topic

[yobou]

Dear All
when presenting data as (% of control) how can we add error bars in this case? I mean the error bars would represent what? If I make the error bars represent standard deviation, they are not visible on the chart because its value is small compared to the scale used. So what should be done in these cases?
thanks

Edited by yobou, 26 February 2013 - 10:53 PM.

[pcrman]

Error bars represent deviation (SD or SEM), but cannot represent % of control. Can you give some example data?

[yobou]

I want to send you an excel file. could you give your email

[pcrman]

you can PM me as attachment. thanks.

[yobou]

plz check ur PM

[doxorubicin]

The easy way:
When you convert to "% of control", you multiply all values by:  100/[control value].  Just multiply all of your standard errors by this same exact value.  It's just changing the scale of your graph, like changing from inches to cms.

[hobglobin]

doxorubicin, on 27 February 2013 - 03:21 AM, said:

The easy way:
When you convert to "% of control", you multiply all values by:  100/[control value].  Just multiply all of your standard errors by this same exact value.  It's just changing the scale of your graph, like changing from inches to cms.

but this won't solve the problem what the error bars finally show...

[pcrman]

From the data yobou passed me I know he(she) has done one experiment which includes a control, other 3 treatments with 3 repeats in each treatment. In this case, he can do a plot of the actual measurement with error bars. But if he wants to plot as value relative to control, he needs to repeat his experiment at least another time, so that he can SD for the %of control values.