In the last post we discussed my variables, controls, hypothesis, and DOE/methods so we're ready for:
PROCEDURE!
We covered much of this in the previous sections, so I'm just going to do a quick summary. For each set of test variables (line) in the DOE, the procedure was the following:
1) On day before testing each set of conditions, run "Control 1"
- Shower between 11am and 12pm
- Get hair wet
- Use all products (shampoo, conditioner, leave-in conditioner)
- Leave hair down to dry as quickly as possible
2) The following morning, around 7am, perform the odor test
- Smell top of head and give rating
- Smell right armpit and give rating
3) Just prior to next shower, rate dirtiness and record all data (if not yet recorded for odor test.)
4) On test day, shower between 11am and 12pm. Use test conditions shown in DOE for that day.
5) The following morning, around 7am, perform odor test
- Smell top of head and give rating
- Smell right armpit and give rating
6) Just prior to next shower (between 11am and 12pm) rate dirtiness and record data.
As you can see, my procedure is quite simple. Shower every day around the same time, rate odor and dirtiness around the same time each day, follow the test conditions in the DOE.
RESULTS!
Now, what you've all been waiting for... the results!
First, lets look at armpit smell:
Scale:
1 = No unusual/bad smell
5 = Worst smell imaginable
Average armpit smell with shampoo = 1.86
Average armpit smell without shampoo = 1.09
Next, Head Smell:
Scale:
1 = No unusual/bad smell
5 = Worst smell imaginable
Average head smell with shampoo = 1.29
Average head smell without shampoo = 2.25
And, with a little more test conditions detail:
Finally, dirtiness rating:
Scale:
1 = Same as full shower (control 1)
5 = Same as no shower (control 2)
Average dirtiness rating with shampoo = 1.38
Average dirtiness rating without shampoo = 2.55
And, with more detail:
And there you have it! The results!
.
.
.
What? Disappointed? I have reported my results. I think what you were really looking for is:
CONCLUSIONS!
So, what does it all mean? There are some interesting things to note in my data.
First, lets get armpit smell out of the way. I was looking for a way to check my level of "overall dirtiness" other than my self rating of how I felt. I decided that armpit smell might be a way to do this. However, you can see by looking at the data that there is NO correlation between armpit smell and the use of shampoo or how clean I felt. Just to make that perfectly clear, here are some scatter plots:
Same as above. No correlation between armpit smell and use of shampoo. Or really, some suggestion of an inverse relationship where using shampoo makes my armpits smell worse.
Compared to Dirtiness rating:
And, compared to head smell:
I did not take the time to add R squared values to these scatter plots. but I think its pretty clear that there is no correlation worth discussing here. We can eliminate armpit smell from the rest of our discussion.
And, onto some interesting data in my dirtiness ratings:
Check out this graph again:
There are a few key things to note here:
1) Look at the first (blue) bar. That is the bar for the average of "Control 1." If you recall my scales above, a rating of "1" is defined as "the same as Control 1." Therefore, all of the dirtiness ratings for "Control 1" days SHOULD be a rating of "1." But, the average is greater than one. There were 3 days in my experiment where I rated a "Control 1" day a rating of "2."
My conclusion here is that because I was actively considering how "dirty" I felt, I was more likely to feel "dirty" and even days where I took a full shower I sometimes felt a little less than clean. This is one of the indications that this subjective rating is not the best metric to use.
2) Look at the last (also blue) bar. This is no shower. Again, by my own definition of my rating scale this should be a 5 but it is a 4. I didn't feel as "dirty" as I have in the past.
3) Look at the second to last (dark red) bar. This is my "baseline 2" and the condition that started the whole experiment. I had anecdotally noticed that on days where I used the conditions in "baseline 2" I felt "as dirty" as if I took no shower at all. Here, my own self-rating of dirtiness does not support this initial evidence.
For notes 2 and 3 above here is my conclusion:
As the experiment progressed, I got more "used" to non-shampoo days and days where I did not even get my hair wet. There were a few times that I ran a dry-hair day that we did not rate and it had to be repeated. By the time I got to the dry-hair days that we did rate, I was more used to skipping my hair entirely and rated them lower. If a replication of this experiment is done, I suggest that these conditions be tested both early and late in the experiment and compare them.
4) Use of dry shampoo had little to no effect on how "dirty" I felt. The experimenter noticed that scalp/hair appeared a little less dirty visually, but this evidence was not documented properly.
There are still some general trends that we can pull out of this data. However, based on above, these should be taken lightly as the results are likely biased by many uncontrollable factors that have previously been discussed.
Here we see the trend more clearly:
What we see here is:
1) Shampoo does have some effect (the jump from the blue bar to the red bar is elimination of shampoo)
2) Leave-In conditioner has no effect (the lack of jump between the red and orange/yellow bar shows the elimination of leave-in conditioner has no effect.)
3) Use of water has the largest effect (the jump between the orange/yellow bar and the green bar is the elimination of water.)
So, based on my self ratings, the best way to feel "clean" each day is to get my hair wet.
Finally, Head Smell Conclusions. Lets look at that bar graph again:
1) Look at the first blue bar compared to the next three (red, orange/yellow, and green) bars. Shampoo does not have a significant effect on head smell.
2) Look at the first 4 bars clumped together and compare that to the remaining 6 bars clumped. The first 4 include leave-in conditioner and the remaining 6 do not.
3) Scalp scrub seems to have a negative effect. Compare orange/yellow bar to green bar, eliminating scalp scrub MAYBE reduces smell slightly on wet hair. It seems to have a larger effect on dry hair. Compare the pink bar to the dark red bar. Eliminating scalp scrub (dark red) increases odor rating. More data is needed to make a solid conclusion on scalp scrub.
4) Again we see little effect from dry shampoo (compare pink bar and light green bar next to it.)
And again, we can pull some overall trends from this data. Here is a summary:
1) Shampoo has basically no effect on head smell. (Compare blue and red bars.)
2) Leave in conditioner has a significant effect (compare red bar to orange/yellow bar.)
3) An interesting drop in smell ratings when hair did not get wet (compare orange/yellow bar to the green bar.) My conclusion here is that getting my hair wet actually washes away the previous day's leave in conditioner. On days when I did not get my hair wet, the smell of the the previous day's leave in conditioner was still present enough to give a smell rating in between a day when leave-in conditioner was used (red) and a day when I wet my hair but did not use leave in conditioner (orange/yellow.)
4) The data confirms that skipping a shower completely does give a higher smell rating than any of the test conditions.
And, one final conclusion is that the experimenter has no idea what color that one bar is. Is it orange? Is it yellow? Perhaps marigold. Leave a comment and tell me what color you think it is.
So, there you have it. A quick summary of what we have learned:
1) This experiment was full of bias and much more repetition is needed to draw solid scientific conclusions from it.
2) The initial conclusions (that should be validated by other experiments) are the following:
- The biggest factor on "dirtiness" is whether or not hair got wet.
- The biggest factor on "head smell" is use of Leave-in conditioner.
- Armpit smell had no correlation to this experiment and future validation of this work does not need to include this data.
NEXT WEEK!
Next week I will milk a 3rd post out of this one experiment by covering the following:
- Why I still don't know WHY and the Day 30 Extra Trail.*
- Do women's menstrual cycles actually mess up the data?
*Yes, I promised these topics would occur "next week" last week which means that they should be included in this post. However, this post is already quite long so I have decided to do a Part 3 to the Shampoo Experiment posts. Sorry. I'm in charge and I can change things. At least I delivered on the graph promise.
No comments:
Post a Comment