Which analysis

#1
Hello All,

I am currently analysing my experimental data on postharvest storage of pears.
Lets say that I had 2 fcators (light treatment and cold storage duration). First, light treatment with 3 level of lighting intensity (0, 30 and 50 micromole per second). I grew my pears in this 3 different light treatments. After harvest I stored pears from this 3 different treatment in 4 different cold storage intensities (0 day, 5 days. 10, and 15 days). After each cold storage. The pears are subsequently stored at 20 degress for 15 days. During this storage at 20 degress, I evaluated the quality of the pears on day 0, 3, 5,10 and 15 days. Since I only used non-destructive analysis, I could follow the same pears from each cold storage and light treatment combination during quality evaluation.
If I would like to see if there any differences between light treatment and cold storage intensity and their interaction in any particular day ( 0, 3, 5,10 and 15 days) during storage at 20 degrees, which statistical analysis should I used? I was thinking to apply mixed-anova but it seems not suitable in my case.
Any suggestion/explanation will be much appreciated!

Best,
Abu
 
#3
Hi Greta!

Thanks for your reply.
I think mixed anova is not relevant because I have more than 2 independent variables [light treatment, cold storage intensity, shelf life duration (days)]. Or perhaps the problem is that I didn't know how to do it with my statistical software (SPSS)?

I randomised the pears harvested from the same light treatment and then assigned them to different cold storage intentsity.
 
#4
I think mixed anova is not relevant because I have more than 2 independent variables [light treatment, cold storage intensity, shelf life duration (days)].
Of course you can have several fixed effects and several random effects in one model .


I randomised the pears harvested from the same light treatment and then assigned them to different cold storage intentsity.

I grew my pears in this 3 different light treatments.

If you talk about pears, the kind of fruit that grows on trees, then I assume that you could only randomize the whole tree to a specific light treatment. Is that correct? On the other hand you could randomize each pear (from within a tree) to different storage intensities.
 
#5
Hi Greta!
Thanks for your answer.
Exactly, I randomised the whole pear tree.
Another question, can we follow up significant interaction with a posthoc analysis using LSD test? And then, when we make graph of each storage intensities which plot dependent variable as function of time can we only report one general LSD value? I prefer to do that than to give letters or stars to show significant different in each time point in the graph.

Thanks in advance!
 
#6
can we follow up significant interaction with a posthoc analysis using LSD test?
To do an LSD is just to do a t-test, as far as I remember it.

You can do an overall test if there is an interaction. But interactions are very difficult to interpret. So you would need to test the sum of main effects plus interaction.

(So if you have 3 levels of A and 3 levels of B then you have 9 (=3*3) effects in total with main effects of A plus main effects of B plus interaction effects.)

when we make graph of each storage intensities which plot dependent variable as function of time can we only report one general LSD value?
To make a diagram is the more modern way.