Now, onto my first experiment!
The Shampoo Experiment Part 1: Design of Experiment
A little while ago I was sitting on my couch feeling like I really needed a shower. Yep, feelin dirty. But I had already taken a shower that day. And no, I had not done any tap dancing after my shower. Shocking, but true. I found myself once again noticing that if I took a shower without washing my hair I felt almost as "dirty" as if I had taken no shower at all. The interesting part about this is that I often skip shampoo* when I shower. On those days I don't feel dirty. In fact, it seems no different than the days that I use shampoo.
*Did you know that it's better if you don't use shampoo every day? This is especially true for curly hair that gets extra frizzy when it gets dried out. Natural oils are good for your hair. Some people never use shampoo at all. These people call their movement "no poo." They really should have ran that past a few more people before deciding on a name. Its real. Google it.
So, I found myself realizing/confirming that if I don't do anything to my hair I feel dirty, and if I get it wet and use conditioner I feel clean. I wanted to know why. Or, if not exactly why*, then exactly what was the main causal factor.
*Note: We often get think we've discovered "why" something happens but we really haven't truly figured out the why. There's this surprisingly confusing fuzzy area between narrowing down to the main factor and truly understanding the full mechanism of why. More on this later.
Clearly the only thing to do in this situation is to design an experiment to test all possible factors!
I started by listing out the possible factors (variables):
- Water
- Conditioner
- Scalp Scrub (when I use conditioner I work it through my hair and therefore indirectly "scrub" my scalp in the process.)
My research on the No Poo movement (seriously, Google it) led me to add dry shampoo as another aspect of my experiment. And, after the first couple of days of my experiment I realized I had another factor, my leave in conditioner.
The final list of possible factors:
- Water
- Conditioner
- Scalp Scrub
- Leave In Conditioner
(Plus dry shampoo as an extra test to see if it could make me feel less dirty on dry hair days.)
Hypothesis!
To be honest, I never wrote down a clear hypothesis statement. Going into this my guess (hypothesis) was that it was either the use of water or the scalp scrub (or most likely the combination of the two) that would be the main cause of my dirty feeling.
So, if I were to write it now for what I was thinking then, it would be something like this:
Hypothesis: Getting my hair wet and massaging/scrubbing the scalp causes me to feel clean.
Controls!
Here are things I did to try to have scientific controls in my experiment:
1) Take a shower at approximately the same time every day. This didn't always happen, but I did my best to stick to it and feel it was acceptably controlled.
2) Try to keep level of exercise/sweat consistent each day. This was harder, but mostly successful. There were some tap dance days and lots of packing and moving boxes as we were in the process of moving. So I tended to get a little warm/sweaty every day but not a lot. The day that I tap danced ALL day and got really sweaty was not included in my data.
3) Leave my hair down after showers as much as possible so that it dries in approximately the same length of time each day. (It was long at the time and I often pulled it up to keep it out of my way.)
4) Collect data at approximately the same time every day.
5) Brush my hair the same amount each day (to avoid extra "scalp scrub" factors.)
6) Wear white cotton undies every day (to prevent any outside contributions to feminine itch or discomfort that could have skewed my dirtiness ratings.)*
7) Include data from a blind experimenter. (More below.)
8) Take a full shower, including shampoo, every other day to "reset" for the next set of experimental conditions.
*In the name of science and thorough explanations of science, I am sharing everything. And ladies, there is no shame in admitting that part of what makes us feel we need a shower is feminine discomfort.
Concerns!
There were more of these than I'd like there to be. As explained above, I tried to include as many controls as possible but this is still a VERY subjective experiment. Here are things I had concerns about in terms of eliminating bias.
1) Not double blind and impossible to be double blind.
1a) I was acting as both the lead experimenter and the only test subject. This is not good science.
1b) The testing involved physically doing things to the test subject (me) such that it is impossible for the test subject to be blind. This is the worst case scientifically as there is no way to eliminate bias from the test subject.
1c) Given 1a and 1b, there was no way for the lead experimenter to be blind. Logical deduction: The test subject (me) and the experimenter (also me) were the same person, and the test subject could not be blind, therefore it follows that the experimenter could not be blind.
2) Data collected is subjective. I self-rated on a 1-5 scale of dirtiness (where 1 is the same as a full shower and 5 is the same as no shower.) Add to that the fact that there was no way to eliminate bias via blind trials and we have a recipe for really bad science.
To help with these concerns I implemented a second set of data after the first 2 days of my experiment. (I later re-ran the first 2 trial sets and ignored the initial dirtiness results in order to have complete data on everything.) In addition to my self rating of "dirtiness" I added a second experimenter who rated odor of the top of my head and odor of my armpit each day. This experimenter was blind. However, smell rating (also a 1-5 scale, where 1 is no bad smell and 5 is really bad smell) is still subjective and we are still far from an ideal experiment.
n!
(not factorial, I'm just excited about science)
In a scientific experiment, "n" refers to the number of times you repeat the same experimental conditions. The lowest this number should ever be is 3.* In my experiment, n=1. The reason for this is time. I could only test a new set of variables every other day. My secondary experimenter wasn't thrilled with the idea of smelling me every day for any length of time, and especially not interested in doing this for months. Even with n=1, my experiment took 30 days. That was long enough. Please note that n=1 is really REALLY bad science.
*n=3 is acceptable for middle school science projects. It is also commonly used in industry for failure prediction and analysis (as opposed to research.) This is because someone already did the research to develop a test method that you are following. Using n=3 should be able to confirm if your product meets your test method or not. n=3 also works for confirmation of contamination as many people previously did a whole lot of research to develop equipment to identify material composition accurately enough that n=3 is OK for CSI. For research, ideally you should target n=5 for a respectable experiment. If you're looking to publish in a scientific journal then n should equal A LOT.
So, we've established my variables, my hypothesis, my controls, and my concerns. We've also established that this was a really poor experiment in terms of bias control.
DOE!
Now, the meat of it. The DOE!
I set up my DOE such that I changed only 1 variable at a time. There were a few situations that were impossible however. Mostly, the use of conditioner (regular or leave in) on dry hair. I tested these indirectly by doing a set of trials: wet hair/conditioner/leave in, wet hair/conditioner/no leave in, wet hair/no conditioner/leave in, and wet hair/no conditioner/no leave in. When everything was done, I ended up with the following table of trials:
Definitions:
Control 1 = Full shower with Shampoo
Control 2 = No shower at all
Baseline/Conditioner = Everything but shampoo (which I anecdotally noticed seems the same as Control 1.)
Baseline 2 = Body shower without getting hair wet or doing anything else to hair (which I anecdotally noticed seems the same as Control 2.)
Additional Info:
1) The red highlighted days are when I had my period*
2) The yellow highlighted days I was a bit sick.
3) Days 1 and 2 were ignored as they did not include smell data. Day 1 reoccured numerous times (every other day) and Day 2 reoccured on Day 19.
4) No ratings were done on Fridays. This is because my secondary experimenter and I often do not sleep in the same place on Thurs nights and therefore there was no ability to take smell data on Friday mornings. (This is why a second "control 1" is sometimes seen on Thurs)
5) I did the same trial on Day 11 and Day 13. Not on purpose. Got the same results though so thats promising for repetition.
6) If you're looking carefully you'll see some additional repetition. For example, Control 2 is shown on Day 17 and Day 19. We missed a few days. Or, it was so late in the day by the time we remembered that I decided to skip that day and redo it.
7) Day 30 was an extra test that I added based on my data up until then. We'll talk more about that later!
*Did you know that until very recently NO clinical research experiments were conducted on women, especially women of childbearing age, because everyone thought that our menstrual cycles would mess everything up? Or, they were concerned that the women in the experiments might be pregnant and did not want to endanger the unborn. Or who really knows why. This means that nearly ALL medical knowledge has not been adequately confirmed on women. We are now learning that women's experience of heart disease is very different than men's and that women react differently to many medications (both effectiveness and side effects.) Its pretty scary stuff, especially when you read that women were not really added to clinical research trials until the 1990s (!!!) and that even today they are still not well represented in clinical trials, especially phase 1 trials. Here is a short and readable summary. Here is the 1993 FDA publication that started to create true change. Here is is a more readable summary of that 1993 publication on the FDA website. And here is a 2008 paper discussing how women were still under-represented.
If you find this interesting and you are an ADULT I highly recommend this book. Not safe for prudes!*
*Did I just add a footnote to my footnote? Yep. Prudes are cool! You be prudish if you are. Be you. But, if you are prudish, don't read that book. That is all.
A final note. I shall do my best to back up claims with either data or reputable publications. You should too. Don't believe everything you see on Facebook. :)
Stay tuned! Next week I'll briefly summarize my procedures, and present my results and conclusions! There are graphs! Yay! Graphs!
Also to be covered next week:
- That fuzzy area between the main factor and the mechanism of why
- Day 30 extra trial. Fun!
Peace and rhythm to you all,
The Tap Dancing Engineer
No comments:
Post a Comment