Protocol Online logo
Top : New Forum Archives (2009-): : General Biology Discussion

Bacterial Growth Curves and Error Bars dilemma - (Jan/11/2014 )


I need some help with statistics and excel.

Basically, I have to plot a graph Absorbance (at 550 nm) vs time (hours) for E.coli. I measured absorbance value every hour for 4 hours. So I have 4 values for every hour, then I'm calculating the mean (for every single hour of course) and plot a graph. And then my assignment says I have to do error bars and sadly statistics is my weakness. I do not understand why do I need them and I do not know how to obtain them- I mean there's an option 'Y error bars' when I click on the graph and then I can choose 'error category' such as constant value, percentage, standard error, standard deviation etc, and also 'error indicator' like positive and negative, positive, negative. My guess is: standard error as 'error category' and positive and negative in 'error indicator' part.

Could anyone tell me how should I do it (if there's any formula or anything like that)? and also shed some light on why do I need them? I really have no clue what they represent and if they have any importance.


Thanks in advance


Error bars are very very important in biology - think about it like this:  You have done your growth curve experiment and  get a curve out of the end (no replicates and hence no error bars).  Now you change a variable (say component of medium) and repeat the experiment and get a curve out which is slightly different, but is slightly lower than the original curve.  Just looking at the two you would say that they are different, but how do you quantify how different?  If you have the error bars you might well find that the two curves lie within each-other's error bars, indicating that in fact the two curves aren't different.


There are certainly formulae you can use - but I recommend getting a basic statistics book and having a read (here's a good free guide).  If you carry on in biology, a basic knowledge of statistics is very useful.


Note: I'm also moving this to homework!


The error bar options in excel are cumbersome. At the risk of telling you how to do this before you understand why: You can calculate the standard deviations/errors alongside your means. You can then enter this range of cells into the “constant value” category.


Hopefully the ‘why’ will become apparent if you imagine your lecturer asking you how accurate do you think those means are?


Thank you for your help.

And yes, I wasn't really sure if this topic falls into biostatistics category.