In a recent post the solvatochromic behaviour of cobalt chloride in acetone was discussed. There were some ideas how the acetone might influence the ratio between the blue $\ce{[CoCl4]^2-}$ and the pink $\ce{[Co(H2O)6]^2+}$.
An interesting thought was the change in the dielectricity constant, which would influence the dissociation of $\ce{CoCl2}$ and therefore, if lowered, cause the $\ce{Co^2+}$ and $\ce{Cl^-}$ ions to form the blue $\ce{[CoCl4]^2-}$ complex again. Atleast this is how I understood the explanation.
Using this idea any solvent, that is miscible with water and does not form a differently colored complex with $\ce{CoCl2}$ should have a similar effect if the dielectricity constant is close or lower than the one for acetone. So I gave it a test and used the first solvents I could find in our lab, dmf, acetone, isopropanol and thf (in descending order of permittivity). The solutions were prepared by dissolving a spatula of $\ce{CoCl2 * 6 H2O}$ in a few ml of solvent so it was fully saturated. The liquid layer was transfered to a test tube and diluted with water until the color had changed to pink. Solvent was then added until a blue hue was visible again. To limit the amount of solvent used it was not added until a strong blue color appeared.
For the acetone you can see the best effect. It quickly turned blue again and upon adding more acetone even became quite intense. Dmf had a much lighter solution to begin with, the blue complex was formed quite quickly but it was not as intense as for acetone. Isopropanol gave only a poor result. A few drops of water were enough to completely discolor the solution (still more drops were added to have about equal volumes in all tubes). Even after adding a large amount of isopropanol no intense blue color showed. THF gave good results when the layers were not allowed to mix but once the solution was stirred it quickly discolored again.
As you can see all solvents gave quite intense blue colors with the pure $\ce{CoCl2}$ but even though THF has a much lower dielectricity constant it only gave poor results. I don't really know if this can be explained by the given theory but perhaps someone can find some clues through this experiment.