Testing for Impact of Sunlight
Performing an experiment to test for a relationship between sunlight and the presence of purple goop would require a design that is similar to that which examined the influence of fish on the water. Two identical bowls of water with no fish would be needed with one placed in the sunlight for 24 hours and the other placed in the dark for the same amount of time to act as a control. The hypothesis for such a test is that the water in the sunlight will develop the purple slime over this period. Possible outcomes of this experiment include the prediction that only the water in the sunlight develops slime, that both bowls develop slime, that neither bowl will have slime, and even the possibility that only the bowl in the darkness ends up with slime.
An evaluation for the effect of corn flakes on the growth of purple stuff, like the previously discussed experiments, would include two bowls of water that are identical in all ways but one. In this case both bowls would be placed in the dark to discount the role of sunlight and only one would receive corn flakes at the beginning of the testing period. The bowls would also be free of fish to avoid a potentially confounding effect. Control for the test would be provided by the bowl of water with no cornflakes as long as it was exposed to the same environmental conditions and time effects as the bowl exposed to the corn flakes. The hypothesis would be that only the corn flakes bowl would develop purple slime over the 24 hour period. Like the previous tests there are four possible outcomes regarding the presence of slime: develops in only the corn flakes bowl which would favor the hypothesis, neither bowl develops slime, both bowls get slime, or the non-flakes bowl develops the goo.
It is entirely possible that more than one variable is responsible for the growth of the purple slime. There are several combinations of factors that could be the key to the slime including fish presence with sunlight, fish and corn flakes, sunlight and corn flakes, or even all three variables. Relationships where multiple factors are required to see the result are known as interactions and can be evaluated using methods analogous to those examining the factors as single contributors. First, the presence of an interaction can be recognized by the failure to develop purple goo in all of the previous tests. To check for the impact of fish and sunlight together we would use two identical bowls of water differing only in two ways with the non-control bowl being placed in sunlight and having a fish placed in it. Fish and corn flakes together in one bowl and placed in darkness is a sufficient condition to test for the second interaction combination, while the third possibility would use a bowl with corn flakes placed in the sunlight without a fish. The three-variable interaction requires both fish and cornflakes in a bowl that is also placed in the sun.
Unfortunately none of the discussed designs would be adequate to gain a scientifically acceptable insight of the relationship between purple slime development and the three factors. The main reason is that there are many other variables that could be having a significant impact on the situation. The water itself is a variable that could be controlled for by using another water source in a control bowl (a water-free empty bowl could technically suffice but would be cruel to the fish in interaction conditions). The bowl may also contain rocks or other environmental items that are part of the relationship and require a controlled test using identical bowls differing only in item presence. Many context variables can also be confounding including temperature and air quality, which are especially troublesome when different rooms are required to test for the impact of light. We can never say with certainty that the slime is a result of any test variables or their interactions because we cannot observe every instance. Instead we aim for a high scientific probability using controls and statistical methods.