Dishwasher Soap Experiment Part 1

The TDE is in the initial stages of the next experiment: Dishwasher soap!

Actually, this experiment has been casually happening for quite some time now. However, I will now formalize it and do an official experiment to confirm the results. Today I'll take you through my initial planning for this experiment.

Step 1: Identify the phenomenon.
This is just noticing things around you and asking questions. It can come from personal experiences or from what you hear from others. In the dishwasher soap case, I read somewhere a while back that often people think that their dishwasher is underperforming, but simply changing to a better soap can help. This article stated that dishwasher soap has a much larger effect on results than the dishwasher itself. So, I want to test that idea.

Step 2: Define the problem/question. (Hypothesis)
In grade school I was taught  that the first thing you do in an experiment is create your hypothesis and that this is your "guess" of what will happen. We were told that you have to state which conclusion you expect to happen. Wikipedia states that a Hypothesis is a "proposed explanation for a phenomenon" and that "For a hypothesis to be a scientific hypothesis, the scientific method requires that one can test it."

I think many school teachers get it a bit wrong. I remember being told that my hypothesis has to state what I think the outcome will be, very specifically. For my dishwasher soap example, the teacher may have given an example of "Dishwasher soap A will give the best results." To me, this is just setting yourself up for confirmation bias issues in your experiment. Everyone likes to be right. A better hypothesis is "Dishwasher soap type has an effect on cleanliness of dishes after washing in an automatic dishwasher." I'm pretty sure my teachers in elementary and secondary school would have rejected this hypothesis, but I also think my science peers would agree with me on this being the better hypothesis. Ultimately I do want to know which soap is best, but my initial assumption to test is that the soap type has a significant effect on dish cleanliness. If results of this initial experiment show that Dishwasher Soap A appears to be best, then that could be a hypothesis of a future more detailed experiment.

In "real" science you don't really ever write a hypothesis anyway. You write a grant proposal instead.

Step 3: Investigate the previous research
Look for other articles and papers that have been previously written on the topic or related topics. Maybe someone has already answered your question. If not, previous research will help guide you. In my case, I looked up dishwasher soap results on Consumer Reports Online. I'd love to post a screen shot of the ratings, but I'm assuming that's frowned upon. This previous research helped me narrow my list of soaps to test and gave me ideas of what to collect as results.

Step 4: Decide what/how to measure
My hypothesis states that soap type will have an effect on cleanliness of my dishes. How am I going to measure cleanliness of dishes? There are many ways I could define cleanliness of dishes. Using my personal life experiences and the previous research I have decided to define cleanliness based on 2 common post-wash issues that I find to be annoying and make me feel that my dishes didn't get clean. I will measure:
1) Number of bits of food remaining (or re-stuck) on dishes after washing
2) Shininess of silverware after washing

How I will measure these:
1) Count bits of food found when unloading the dishwasher.
2) This is trickier. I wish I had a gloss meter! Instead I will attempt to accurately rate 2 factors of shininess: 2a) Overall shininess of each piece as either shiny (3), dull (1), or in-between (2) and report this as an average numerical rating (1-3 scale) for each load and 2b) % coverage of water spots or other dullness for each piece and report this as an overall average % (0-100 scale) for the load.

Step 5: Variables
Define dependent variables (dish cleanliness based on food bits and shininess in this case) and independent variables. There are many other possible independent variables:
-Water temperature
-Water pressure
-Wash cycle type
-Amount of soap used
-Pre-washing conditions, if any
-Food type on dirty dishes
-Length of time food has been on dirty dishes

I'm sure there are more as well as additional dependent variables that I did not mention.

Step 6: Understand your limitations, and do what you can to overcome them.
Thankfully, many of the possible variables are controlled by the dishwasher itself. Some of them, like amount of soap, are rather easy to control. Others such as food type are harder to control. However, one of the reasons I want to look specifically at silverware is that my experience tells me that peanut butter often ends up caked onto silverware if not properly removed before the wash. We eat a lot of peanut butter and ideally I'd be able to put silverware in the dishwasher without pre-washing. So, I will be sure to include a few pieces of peanut butter covered silverware in every load, without pre-washing it.

Another limitation in my case is number of test runs. I have small amounts of multiple types of soap right now. I will not be spending the money to buy new containers of soap just for this experiment. So, in some cases I may only be able to do one load per soap type. I will do as many iterations of each soap type as possible, targeting a minimum of n=3 and an ideal of n=5.

Next up, the actual DOE plan. But, I think I'll leave that for next time. There will be spreadsheets and screenshots and controls oh my! This will also give me time to brainstorm additional variables to consider. Feel free to join my brainstorm; leave your suggestions in the comments!