Yes, it affects the growth. Some  researchers  use black plastic sheets for covering tanks for preventing such factors. It also depend upon the fish species (some fishes stressed by their reflection as another fish)and its behavior (aggressive or social) as well crowding ,tank size do also affect. Light intensity on each tank is also  important . I suggest to avoid paint you  can use wooden blocks.

If your tanks transperent, you must cover each one with black or blue plastic sheets... No need paint (but if you wish do it ;)

Light is another issue, if they dont have equal light on top (some near windows, some far etc) you can distrubute trial's replicates equally accourding to day light... 

I guess outer row arent effected but you can put at the bottom (fill the gap) by a strafoam 8better with table colour ;) 

Good luck Nick

painting is good, but I would stay away from wrapping in plastic (though we've done that for years!).  your animal care committee might object to the plastic wrapping because it is difficult to keep clean.  We've taken to using corrugated plastic sheets (~ 25 $ Can for a 4 x 8 foot sheet at your local home depot or similar outlet *no endorsement implied*.  you can cut it to fit, slide it between tanks and the animal care folks like the fact that it can be easily removed and cleaned.

Hey I think you saw this post already in a other question but I add a few publications.

In this study's they use big black curtains to be able to observe the fish.

Article Prey or predator: 0+ perch ( Perca fluviatilis ) in the tra...

Article Boldness in two perch populations - long-term differences an...

Article How To Avoid Seven Deadly Sins in the Study of Behavior

Article Balancing the response to predation – The effects of shoal s...

If your tanks transperent, you must cover each one with black or blue plastic sheets..

If the species of fish that you are working with are domesticated and do not react on any or every human activity (feeding, watching, cleaning the screens or walls of aquarium.. etc).  We generally cover our aquaria with plastic, washable grids or you  can custom make with any window screen.  Another tip would be to minimize human interference while feeding.  I understand the agony of researchers feeding wild fish in captivity, be persistent and you will succeed. 

You're not overthinking this. You should take every precaution possible to see to it that each and every tank represents an identical environment for the fish.

However, even if you do this, there will always be subtle differences among the tanks, e.g. in water flow, level of vibration, shade, light...

Thus, in all probability, even if you placed similar groups of fish in all twelve tanks, and treated then all in an identical manner, you'd still find one tank with the highest growth rate and one with the lowest growth rate, and if you're unlucky, the difference would be statistically significant. This is the dreaded "tank-effect"!

The only way to avoid this is to have multiple tanks for each treatment. You have 12 tanks. Thus, you should not design any experiment with more than 6 treatments, ideally four treatments, with triplicate tanks for each treatment. Then you make sure to randomize the treatment groups among the tanks. This is also essential if you want to be able to publish your data in a good scientific journal.

Good luck with your experiments.

The colour of the tank bottom was very important in our experience with carp. It was very important, that the bottom had a dark (black or dark brown) colour. Another important factor in some fish species (e.g. African catfish) is the number of individuals per tank. You should observe the behaviour of the fish when kept in different numbers. Aggressive behaviour should be avoided in a growth experiment.  

All of the environmental conditionals should be same in all treatment groups during the Study. Light intensity, shape of tank, water current and the other parameters like isolation, environmental stress etc. If you can do this with the tanks you have you do not need to paint. But, if no, you should paint or isolate them any methods like covering etc. 

http://cmsadmin.atp.co.il/Content_siamb/editor/65.2013.798.Can.pdf

All my colleagues answers are correct and their complete each other. 

Hi Nick Sakich

Before starting  any study, the conditions  should be it same in all treatments, except the variable who want to study it  impact, if factors were affect partiality on some experimental treatments, the resulting effect may not come back for a variable to be studied and therefore obtain inaccurate results.

It can disposal of this problem by increase the number of duplicates per treatment and it distribution have randomly