You are on page 1of 11

Golden Ratio Project Statistics 1510 (Day) Chase Comfort

Abstract The purpose of this experiment was to establish whether or not the human body can be shown to be based on a "Golden Ratio" of 1.618. In order to determine this, I took measurements of supposed "Golden Ratios" on the human body of different groups, divided into race and gender. By using TC-Stats, I was able to determine that none of the groups exhibited adherence to a "Golden Ratio", and that none of the groups were appreciably closer to the "Golden Ratio". Introduction "You are a beautiful, unique snowake, just like everyone else." -Unknown The idea of the "Golden Ratio" as it applies to humans is that there is a special ratio that is common throughout nature, and that this ratio is somehow responsible for our ability to evolve in a way that allowed us to survive. To arrive at the "Golden Ratio", one must rst understand the Fibonacci sequence. In the Fibonacci sequence, the proceeding two numbers on a number line are added to make a third number, starting with 0 and 1. What is interesting about this sequence of numbers is that, as you move through the sequence, the ratio of the adjacent numbers approaches a certain number. That ratio is the "Golden Ratio", approximately 1.618. The purpose of this project is to make an assertion as to whether or not the anatomy of different human groups are based on the Golden Ratio, and to back it up with evidence. Hypothesis: I hypothesize that the ratios that are given will be reasonably close to the Golden Ratio regardless of race or gender. Data Collection To collect data for this project, I rst decided on what groups I wanted to compare. To make sample collection easy, and to roughly represent the population of Taft, I decided to go with groups divided into gender (male/female) and race (white/ Hispanic). With these groups in mind, I decided on two examples of golden ratios given to us in the project video, length of wrist to elbow divided by length of nger to wrist and distance from navel to top of head divided by distance from shoulder line to top of head. To make these measurements, I used a simple measuring tape, and recorded the results in centimeters. I wanted to keep the groups as even as possible, so I decided to take measurements from 5 white males, 5 Hispanic males, 5 white females and 5 Hispanic females, for a total of 20 subjects. Once I had decided on the amount of people from each group I would need to measure, I took simple random samples of students in the Taft College library that met the criteria of the groups. I measured the lengths of interest on these students and recorded the results in TC-Stats. Once I had my raw data, I was able to use the column arithmetic function of TC-Stats to solve for the ratios I was interested in. Data Analysis

With this method of grouping, I was able to compare the ratios of 10 males to 10 females, and 10 whites to 10 Hispanics, and provide the summary statistics below: Summary Statistics of Ratios for Males:

Summary Statistics of Ratios for Females:

Legend: Ratio 1= Length of wrist to elbow divided by length of nger to wrist. Ratio 2= Length from navel to top of head divided by length from shoulder-line to top of head. While we could make comparisons based off of the summary statistics, it would help to have a clear visual representation of the data in order to understand what is happening.To visualize what the summary statistics are telling us, I utilized the box-andwhisker plot function of TC-Stats to make the following comparative charts. Male versus Female: (M=Male F=Female) Legend: Ratio 1= Length of wrist to elbow divided by length of nger to wrist.

Ratio 2= Length from navel to top of head divided by length from shoulder-line to top of head

This chart allows us to see what the data is telling us. For Ratio 1, Females had a minimum of 1.3714, a median average of 1.47105, a maximum of 1.6, and a standard deviation of approximately 0.0798. For the same ratio, Males had a minimum of 1.4211,

a median average of 1.5076, a maximum of 1.7222 and a standard deviation of

approximately 0.1047. This data shows that these groups were fairly similar with respect to the number line, even though the Males had a slightly higher variance. Also, both of the groups provided ratios close to the Golden Ratio. For Ratio 2, Females had a minimum value of 1.6667, a median average of 1.9483, a maximum at 2.1154 and a

Legend: Ratio 1= Length of wrist to elbow divided by length of nger to wrist. Ratio 2= Length from navel to top of head divided by length from shoulder-line to top of head White versus Hispanic Ratio1

For Ratio 1, Hispanics had a minimum of 1.375, a median average of 1.50675, a maximum of 1.667 and a standard deviation of approximately 0.1. In comparison, the Ratio 1 values for Whites had a minimum of 1.3714, a median average of 1.50675, a maximum of 1.7222 and a standard deviation of approximately 0.0981. The rst value that I wanted to double-check was the White maximum, as I wanted to know if a single anomalous value had "thrown off" my box plot. Upon reviewing the data, however, I saw that there was another value between the 3rd Quartile and the maximum. What is most striking about this data comparison is that the median values are the same, emphasizing the fact that these groups are very similar, although the White group had a slightly lower variance. White versus Hispanic Ratio 2

For Ratio 2, the Hispanic group had a minimum of 1.7857, a median average of 1.9909, a maximum of 2.1154, and a standard deviation of approximately 0.115. Ratio 2 for the White group had a minimum of 1.667, a median average of 1.8938, a maximum

of 2.1852, and a standard deviation of approximately 0.1529. Again, the Ratio 2 data had every value over the Golden Ratio, suggesting that I either introduced bias to my data or that this ratio is not in fact a Golden Ratio. As we can see above, the White group had a higher variance, but the groups are still similar relative to the number line. True Average

Using TC-Stats, we can construct a 95% condence interval for the true average of the ratios in each group. The method of constructing a condence interval depends on the normality of the data, and (Figure 1) shows that these groups all failed a normality test. Using the Sign Test function of TC-Stats, we are able to determine the following 97.9% condence intervals for the median:

Ratio 1 Males:( 1.4211,1.6) Ratio 1 Females:(1.3714,1.5143) Ratio 2 Males:(1.8065,2) Ratio 2 Females(1.6667,2.08) Ratio 1 Whites:(1.3714,1.5152) Ratio 1 Hispanics:(1.375,1.6) Ratio 2 Whites:(1.6667,2.08) Ratio 2 Hispanics:(1.7857,2) Comparison In order to to test my assertion that there was no statistically signicant difference in the expression of the Golden Ratio amongst the groups, I opted to use a 2 Sample TTest. In order to use the 2 Sample T-Test , we must satisfy the following assumptions: Each sample must be randomly and independently selected from the population and the populations must be distributed normally. Also, in order to decide whether or not we pool the variances, we must take an F-test. The assumption of populations being distributed normally was violated for all of the groups, as shown in the normal plots (Figure 1). Because the normality assumption was violated, we have to use a non-parametric test to compare the populations. The non-parametric test we used for these populations was the Wilcoxon Rank-Sum test to nd the following p-values: Male versus Female Ratio 1: P-value: .2047 Decision: Fail to reject.

Conclusion: There is not enough evidence to suggest that one group is different from the other. Male versus Female Ratio 2: P-value: .9131 Decision: Fail to reject. Conclusion: There is not enough evidence to suggest that one group is different from the other White versus Hispanic Ratio 1: P-value: .7424 Decision: Fail to reject. Conclusion: There is not enough evidence to suggest that one group is different from the other White versus Hispanic Ratio 2: P-value: .7709 Decision: Fail to reject. Conclusion: There is not enough evidence to suggest that one group is different from the other
Conclusion My initial hypothesis was not entirely correct. While the groups did show similarities to each other for every comparison of the ratios, only Ratio 1 had averages near the Golden Ratio. For Ratio 2, while the groups were similar to each other, they were not as close to the Golden Ratio. Also, none of the ratios that were compared on the basis of race or gender could be said to be dissimilar. Bibliography http://qt-server.taftcollege.edu/podcasts/stat1510/youtube/ Golden_Ratio_Human_Body.mp4 http://qt-server.taftcollege.edu/podcasts/stat1510/youtube/The_Golden_Mean.mp4 Appendix Raw Data Legend: M=Male F=Female W=White H=Hispanic
Sex M M M Race Shoulder-Top of Head W 29.5 W 32 W 27 Navel-Top of Head Ratio 1 55 1.4211 59 1.5 59 1.5152 Ratio 2 1.8644 1.8438 2.1852

M M M M M M M F F F F F F F F F F

W W H H H H H W W W W W H H H H H

31 28.5 27.5 29.5 28 24 29 28 24 27 29 25 28 23 25 26 28

56 54 54.5 55 52 50 58 53 50 45 55 52 50 46 45 55 56

1.4211 1.7222 1.5 1.6667 1.6 1.5556 1.4211 1.5429 1.4211 1.3714 1.5135 1.5143 1.5135 1.3889 1.375 1.4286 1.6

1.8065 1.8947 1.9818 1.8644 1.8571 2.0833 2 1.8929 2.0833 1.6667 1.8966 2.08 1.7857 2 1.8 2.1154 2

Figure (1): Male Ratio1:

Female Ratio 1: Male Ratio 2:

Female Ratio 2:

White Ratio 1:

Hispanic Ratio 1:

White Ratio 2:

Hispanic Ratio 2:

You might also like