which statistical test should i use - (Aug/30/2012 )

which statistical test should i use?
I has two different treated cells by time couse.
I did western blot and I took picture by ECL and Film.
I have quantitated all bands on the film by Freeware ImageJ http://rsbweb.nih.gov/ij/ .

Now I have two series of digit numbers.

I have read statistics log time ago and I have forgotten most of them.

I just want to know which statistical test should i use?

Personally I think that CORRELATION (Pearson) by SPSS software is the best, but I am not sure yet.

In the papers I have seen using "one way anova with tukey test".
Which one do you recommend?

-----
Babak

I am not an expert on statistics but I think that the test depends on what do you want to know. I mean, correlations usually are used when you want to find associations between two (or more) variables (let´s say, phosphorus concentration and growth of algae, for example).

Do you want to know if there are differences between your treatments? If that is the case, you can use a t-student (if you want to compare only two groups, lets say treated vs control) or better a one way anova. But, as I said, it depends on what is your aim and on the set up of your experiment. If you can provide more information, maybe I can write a more specific example on what test to use and why!

alqga on Thu Aug 30 19:09:58 2012 said:

For example I treat cells on a time course on 48, 24, 12, 8, 6 and 4 with a chemical.
For example I test protein X with western blot and RT-PCR.
now I want to find relationship between timecourse and protein X.


-----
Babak

Ok!

Then, if what you want to see it`s if your protein´s increase (or decrase, whatever) is time-dependant, then I think you can use a correlation. But to see whether your treatment had an effect you should compare treated vs control with an anova. Did you only use one concentration of your chemical?The same concentration during the whole time course?

The same concentration during the whole time course

I think you need to plot the values on a graph. To actually perform statistics you need replicates of your data so that a mean and standard deviation/error can be generated. I don't think that an ANOVA will be suitable in this case, but you can do statistics on the difference of the lines between the controls and treatments.

You (both) should also look up the difference between parametric and non-parametric statistics!

Sorry, but I think that to do a correlation you don't neccesarily need replicates. If for example you want to know if there is a correlation between temperature in summer and development of algal blooms you don' t have replicates and still you will be able to do it. In any case, he didn't say nothing about replicates, so we don' t know about the set up. I agree he can plot the values on a graph, but additionally I think he can perform a correlation to see the strenght of the association. But to see if the treatment had an effect he does need to perform a test.

Either parametric or non-parametric, I think I know the difference. Parametric tests such anova when you have normal and homocedastic data, non parametric, such as anova on ranks (kruskal-wallis) when any of those test fail, am I wrong?

alqga on Fri Aug 31 11:36:47 2012 said:

Correct, however, to compare between a control line and a treatment line, you do need replicates, so that you can establish if the difference(s) seen are due to natural variation or are actually statistically valid difference.

Nearly but not quite, Parametric tests are indeed for distributed data (assuming the data fits a distribution is the primary criterion), but they also require large (n>30) sample sizes, and assume certain things about your data, meaning that they are not particularly robust tests, despite their power. Non-parametric are used where one or more of those conditions are not met, and can still be used on data that fits a distribution - they have less power, but are more robust.

It is incredibly common to see the wrong tests used in the literature, I'm not sure if this is because of lack of statistical knowledge (to be honest I know bugger all about stats myself) or if it is due to poor reviewing of the papers. I suspect, because parametric tests are much easier to calculate, this is the reason they are taught primarily in schools and universities, despite not being particularly useful in the real world.