Factor Analysis of Body Hair Distributions DONALD G . BARKER Texas A f M University, College Station,Texas 77840

KEY WORDS

Factor . Body . Hair.

ABSTRACT Using observations on the presence or absence of terminal hair in 13 locations on 700 white adult males, descriptive statistics were tabulated, phi coefficients of intercorrelation were computed, principal axis factor loadings were extracted, and varimax rotation of factors to simple structure was performed. The resulting rotated factors were interpreted as five independent components of body hair distribution: dorsal (less sacral), pubic extension, thoracic, sacral, and abdominal.

A series of researches by L. R. Setty has classified and described the pattern and distribution of hair on the scalp (Setty, '70a), ears (Setty, '69a), face (Setty, '71), back of neck (Setty, '72), front of neck (Setty, '66a), chest (Setty, '61), abdomen (Setty, '66b and 67), back (Setty, '62), penis (Setty, '69b), scrotum (Setty, '70b), arms (Setty, '64), legs (Setty, '68), and hands and feet (Setty, '66c). One of Setty's ('62) most extensive series of observations described the distribution of hair on the trunks of 700 white adult males. Setty generously made these data available for the further statistical analyses reported here. METHOD

As reported by Setty ('62), 700 white adult males were examined for the presence or absence of terminal hair in 13 locations on the back and front of the trunk. Each individual was coded 0 (hair absent) or 1 (hair present) for each of the following locations. 1. Acromial (top of shoulders). 2 . Infracervical (lower part of back o f neck). 3. Scapular (shoulder blade regions). 4. Lumbar (loin areas). 5. Sacral (between and just above buttocks). 6. Infrascapular (below scapular and above lumbar areas). 7. Circumareolar (around the nipples). 8. Pectoral (on the breasts). 9. Sternal (center or lower part of area occupied by breastbone). AM J

P H Y S A N T H R O P ,44

27-30

10. Infraclavicular (below the collar bones). 11. Sagittal (narrow upward extension of' pubic hair along the midline). 12. Acuminate (triangular upward extension of pubic hair). 13. Disperse (distributed over most o f lower abdomen in roughly quadrangular pattern). Because of overlapping of the body areas involved, individuals coded positive for pectoral hair were coded positive for circumareolar hair, and those coded positive for acuminate were coded positive for sagittal. Men classed as positive for the disperse category were coded as exhibiting hair in both the sagittal and acuminate positions. The 700 observations of 13 variables were intercorrelated, factor analyzed, and rotated to simple structure, using phi coefficients, a principal axis extraction routine, and a varimax rotation routine described by Veldman ('67). Fifty percent of the variance of the 13 variables was accounted for by two factors, 59% by three factors, 66% by four factors, 72% by five factors, 77% by six factors, 82% by seven factors, 86% by eight factors, and 100% by 13 factors. The 5-factor solution was selected for presentation here. R E S U L T S A N D DISCUSSION

The results of the analyses are shown in tables 1, 2, and 3. Table 1 lists the number and percent of the 700 subjects positive for presence of terminal hair in each of the 13 locations examined. The table shows body hair 27

28

DONALD G. BARKER

Number iind percent of 700 white cldzilt mules with hair on errch o f 1 3 locations of the trunk Location

1 . Acromial Infracervical Scapular Lumbar Sacral Infrascapular Circumareolar Pectoral Sternal Infraclavicular 1 1 . Sagittal 12. Acuminate 13. Disperse

2. 3. 4. 5. 6. 7. 8. 9. 10.

Number

185 275 86 202 319 232 665 ._ 568 570 474 677 668 329 ~

Percent

26.43 39.29 12.29 28.86 45.57 33.14 95.00 81.14 81.43 67.71 96.71 95.41 47.00

to be nearly universal in the sagittal (97%), acuminate (95%), and circumareolar (95%) regions and to be least frequent in the scapular (12% ), acromial (26% ), lumbar (29% ), and infrascapular ( 3 3 % ) areas. In general, body hair was noted most frequently on the ventral surface of the trunk (abdomen and chest) and least frequently on the dorsal surface (particularly the upper and middle back). Table 2 shows the intercorrelations among locations of body hair, that is, the extent to which the presence of hair in one location is associated with its presence in each of the other body areas. Since each variable was coded dichotomously, phi coefficients of correlation were computed and are shown. All correlations were positive, although several were not significantly different from zero correlation. In general, the presence of hair in an area was associated - to varying extents - with its presence in other areas. The principal axis factor analysis was intended to represent most of the variability of the 13 intercorrelated variables by means of a smaller number of orthogonal (uncorrelated) factors or components of body hair distribution. The resulting pattern of factor loadings was rotated so as to simplify the factor structure. The rotated factor loadings appear as table 3. The highest factor loadings on each factor are boxed to assist in interpreting the factor. Factor I. Dorsal (less sacral). Factor 1 has high loadings on all upper and middle

back locations: acromial, infracervical, scapular, infrascapular, and lumbar. It could be identified unqualifiedly as a component of dorsal or back hair, except that it is almost completely unrelated to sacral hair (loading of only 0.13). Factor 2. Pubic extension. Factor 2 is almost exclusively loaded on the variables of sagittal and acuminate abdominal hair patterns, both of which are characterized by a n upward extension of the pubic hair, the sagittal in a narrow vertical line and the acuminate in a narrowing upward pointing triangular configuration. This factor has little or no association with the disperse or quadrangular abdominal hair pattern (loading of only 0.12) even though the latter is adjacent to and includes the former. Factor 3 . Thoracic. Factor 3 has its highest loadings on the four chest locations: circumareolar, pectoral, sternal, and infraclavicular. Factor 4 . Sacral. Factor 4 is clearly identifiable as a sacral component of body hair. It is essentially independent of all locations except the sacrum (loading of 0.93), except for a modest loading on the adjacent lumbar region (0.26) and a negative loading for hair in the scapular region ( - 0.24). Factor 5 . A b d o m i n a l . Although labeled here as an abdominal component, Factor 5 has its single high loading on the disperse or quadrangular pattern and is essentially unrelated to the sagittal and acuminate patterns, which have high loadings on Factor 2 . Factor 5 seems based on a lateral expanse of abdominal hair, while Factor 2 seems related to medial hair on the abdomen. While all but the disperse abdominal location had low loadings on Factor 5, there was a tendency toward positive loadings for lateral regions (circumareolar, acromial, lumbar, and pectoral) and toward negative loadings for medial regions (sternal, infraclavicular, and infracervical). CONCLUSIONS

Most of the variation in body hair distribution can be accounted for by five independent factors or components: dorsal (less sacral), pubic extension, thoracic, sacral, and abdominal (less sagittal and acuminate).

29

FACTOR ANALYSIS OF BODY HAIR DISTRIBUTIONS TABLE 2

Intercorrelutions umong locutions of body hair f o r 700 wh ite udult mules Locations

1. Acromial 2. Infkacervical 3. Scapular 4. Lumbar 5. Sacral 6. Infkascapular 7. Circumareolar 8. Pectoral 9. Sternal 10. Infkaclavicular 11. Sagittal 12. Acuminate 13. Disperse

1

2

55 40 60 21 49 12 28 23 31 091 12 30

39 48 19 51 13 32 26 41 13 13 15

3

4

5

41 031 51 091 18 18 20 071 08 1 22

27 60 15 30 17 29 12 14 31

6

7

18 16 27 25 25 15 16 101

8

15 31 20 30 11 14 31

9

48 38 28 33 39 18

1

0

47 46 34 42 26

1

47 34 41 0 2 1

1

22 23 101

1

84 17

2

21

Note All decimals omitted Indicates phi coefficients not significantly different from zero beyong 0 05 level of confidence

1

TABLE 3

Rotated fuctor loudings of fuctor analysis of body hair distributions f o r 700 wh ite adult mules Rotated factors Variable

1

2

3

4

5

Communality

1. Acromial

74

- 01

16

17

15

62

2. Infracervical

72

01

28

08

- 11

62

3. Scapular

74

10

02

- 24

- 02

62

4. Lumbar

75

04

08

26

21

69

5. Sacral

13

09

16

01

92

06

12

04

15

67

22

- 03

33

63

19

12

25

69

28

10

- 24

67

6. Infrascapular

7. Circumareolar

- 04

8. Pectoral

20

9. Sternal

17

10. Infkaclavicular

32

1 F!

12

- 14

64

18

06

05

91

28

04

09

04

02

pq

91

17

08

09

02

11. Sagittal

05

12. Acuminate

05

13. Disperse

28

12

24

15

Proportion of variance

El

83

~

Note: All decimals omitted. Entries in boxes represent highest factor loadings for each factor.

ACKNOWLEDGMENT

LITERATURE CITED

The author thanks h.L. R. Setty, Department of h a t o m y , Howard University College of Medicine, Washington, D. C., for providing data and encouraging the study.

Setty, L. R. 1961 The distribution of chest hair in Caucasoid males. Am. J. Phys. Anthrop., 19:

28~288.

- 1962

Hair patterns on the back of white males. Am. J. Phvs. AnthmD.. 20: 365373. 1964" The distrib;tion of h i r - o f the

30

~

~

DONALD G . BARKER

upper limb in Caucasoid males. Am. J. Phys. Anthrop., 22: 143-148. 1966a The distribution of anterior cervical hair in white and negro males. Amer. J. Phys. Anthrop., 24: 321-324. 1-b Varieties of the acuminate abdominal hair pattern of white males. J. Nat. Med. ASSOC., 58: 191-193. 1966c A comparative study of the distribution of hair of the hand and the foot of white and negro males. Am. J. Phys. Anthrop., 25: 131-137. 1967 Varieties of the quadrangular abdominal hair pattern of white males. J. Nat. Med. Assoc.. 59: 4 5 4 7 . 1968 The distribution of hair of the lower limb in white and negro males. Am. J. Phys. Anthrop., 29: 51-55.

1969a Hair patterns of the pinna of white and negro males. Am. J. Phys. Anthrop., 31: 15S161. 1969b Penile hair patterns of whites and negroes. J. Nat. Med. Assoc., 61: 67-69. 1970a Hair patterns of the scalp of white and negro males. Am. J. Phys. Anthrop., 33: 49-56. 1970b Scrota1 hair patterns of whites and negroes. J. Nat. Med. Assoc., 62: 15f3-158. 1971 Hair patterns of the face of white and negro males. J. Nat. Med. Assoc., 63: 128131. 1972 The distribution of posterior cervical hair in white and negro males. J. Nat. Med. Assoc., 64: 239-241. Veldman, D. J. 1967 Fortran Programming for the Behavioral Sciences. Halt-Rinehart-Winston, New York.

Factor analysis of body hair distributions.

Using observations on the presence or absence of terminal hair in 13 locations on 700 white adult males, descriptive statistics were tabulated, phi co...
232KB Sizes 0 Downloads 0 Views