Was the Jelly Belly experiment scientific?

The Jelly Belly Experiment was an experiment to test the level at which sight affects taste. If you have a really bad memory, or weren't there, this is how the experiment happened. There were two testers who were blindfolded; two were not. There were 13 different flavours of jelly bean. Using a list of possible flavours, the testers were fed beans and guessed their flavours. The data was assembled in a table, and compared with the correct answers. Here are the results:
The blindfolded testers missed 11 out of 13 and 8 out of 12 (one bean was missing apparently) respectively. The sight-enabled testers missed 3 and 2 respectively.

Taste is a sense that is a lot weaker than our sight, and it's used far less. So it is logical that seeing what you are about to eat is going to alter your experience of the taste, and therefore how it actually tastes. For instance, If you see a yellow Jelly Bean, it probably won't be cotton-candy flavoured, and even if it tastes cotton-candy flavoured, you won't put that down on paper because, well, how could it be? In this experiment, the added advantage of sight meant that the testers got most of the flavours right. But the blindfolded tasters weren't biased by sight, and judged on taste alone. Also they didn't know all of the flavours possible (at least at first) so they might be thinking, for example, "blueberry" only to find it's not a choice. It was a choice, but it's just an example. Anyway, the blindfolded testers scored far worse. This shows that sight does affect our reaction to taste. But just how seriously should we take that result?

Was the experiment "scientific"?

Conditions
The testers did not have a drink of water in between each bean, which meant that they had remnants of previous beans in their teeth. This does not help get accurate results. A scientific investigation should not allow for such error when it can be avoided.
Quantity
There were only two testers per group. A good scientific investigation needs a sample size of larger than two.
Controllability
The start of the experiment was chaos. During the process of figuring out which numbers the beans went on, some of the testers discovered accidentally the flavour of one or two beans. A good scientific experiment should not be so largely improvised, as this one was. The beans should have been ready to be tested from the beginning. At the very least a plan should have been laid out. This was not the case.
Measurability? Repeatability?
A good experiment is repeatable, which means that a repeat performance will achieve similar results. In this case, the testers are likely to get better with each test, as they become more familiar with the flavours. If the test was repeated with different testers though, and different flavours, then it probably would get similar results each time, but of course there is the variability of different people's ability to taste and previous familiarity with the Jelly Bean flavours.
A good experiment has measurable results. This is where our experiment is the strongest. Our results were certainly measurable, and a comparison of those measurements was all we needed to test our hypothesis.

In conclusion, the process at the beginning of our experiment would make real scientists laugh, and the lack of quantitative evidence makes the reliability of the data lower than it could be. However, the results do match our hypothesis and what is simply logical. I will say that the experiment was quasi-scientific.