Indian Body Dimensions

Dr Deepti Gupta,
Associate Professor,
Department of Textile Technology,
Indian Institute of Technology Delhi,
Hauz Khas, New Delhi-110016,
Ph: 2659 1417, e mail: deepti@textile.iitd.ernet.in

March 2008

---

Abstract

An anthropometric survey has been conducted for young Indian women and  men in the age group of 18-24 years. A protocol for conducting anthropometric studies has been developed and the problems with existing body size systems have been analysed. Results have been analysed to gain understanding about the  body   proportions  and dimensions existing in the population.  Univariate and multivariate analysis of the anthropometric data has been carried out. Using an indigenously developed algorithm and a unique approach, accurate body sizes for top and lower body garments have been proposed. For the first time it has been possible to map the Indian population for the purpose of evolving a sizing system for garments.  Wearer trials conducted with proposed sizes showed that fit obtained was excellent.

Introduction

The need for having a systematic and scientific system for measuring and classifying human bodies in India for the purpose of developing a sizing system for garments need not be emphasized any longer. As we move into times characterized by unprecedented retail growth, this need is becoming more and more acute. As the masses start frequenting the “Malls” for their clothing requirements, they are becoming aware of the “fit” or rather the “lack of it” that the various brands offer. Fit has been cited as one of the major criteria that determine the purchase decisions for clothing- and thus could make or mar a brand’s popularity. Given the novelty of “ Mall culture” in India, the consumers are still on a steep learning curve and it is a matter of time before they become sensitive to the issue of  fit and become frustrated by the lack of it.

Anthropometric surveys have been conducted frequently in India for the last 30-40 years. Some have been conducted by Home Science colleges and the results have been used for pattern development exercises for various age groups. Many others have been conducted to assess  the nutritional status of  population groups but unlike US studies, very little information about the findings of such studies is available in the public domain. Besides, these surveys were mostly restricted to small, local population sets. A comprehensive all India survey  was undertaken by Dr KS Singh for the Anthropological Survey of India, from 1985 onwards,  to study the effect of several  parameters on growth patterns of the various ethnic groups in India. However, the measurements taken for anthropological studies are quite different from those required for the purpose of developing size charts for garment manufacture and hence cannot be used for the purpose. Some  small studies have also been conducted by ergonomists to collect data for the purpose of  design of  work places.

More recently, some major garment manufacturers retailing in India, commissioned anthropometric studies before launching their respective brands in the domestic market. These surveys were often conducted by marketing organizations which arguably had no  expertise in the area of anthropometry and thus the data obtained was often found to be unreliable. Besides, the companies have held on to the results very closely perhaps as a trade secret, and again no information is available about the results or the analysis conducted. Even when the surveys were conducted as mentioned above, the analysis was restricted to calculation of simple arithmetic means, percentages and percentiles. For the purpose of mass manufacture such results cannot be used. In summary,  it may be said  that  there exists no record of a systematic, all India anthropometric survey ever conducted. And so, today there exist as many sizes and sizing systems in the Indian clothing market as there are brands! The sizes are different, the nomenclature is different and the size intervals are also different in each case. There is no way a customer can buy her clothing without having to try on innumerable clothes in each brand. Then again, there is no consistency of sizes even within the same brand, for example, top and lower body garments. That is to say, that a woman maybe a small (S) in shirts but a large (L) in trousers. All this creates a lot of confusion in the mind of the customer mainly because the exact body measurements for which the garment is intended are not known to the customer and often there is no relation between the garment size and the body measurements.

The reason why we in India, do not have even a primitive garment sizing system in place is, one- that the traditional Indian garments comprised primarily of draped wear and whatever little stitching was required, could be catered to by the local tailor; secondly, we had no organized retailing set ups and thirdly, we were not a fashion conscious or a fit conscious population. Thus the need was never felt for a ready made garment sizing system. As the clothing preferences now shift from Indian to western wear, the need for a sizing system is being felt acutely. But this does not mean that the so called developed nations are not facing problems in this area. The number of research studies conducted and the amount of literature generated in the area of sizing and fit in the last 5 years throughout the world is phenomenal. In fact, it would not be an exaggeration to state that any prestigious textile/ apparel department or university in Europe or America is currently working in the area of garment sizing and fit!!

The complexity of garment sizing

The problem of garment sizing is extremely complex. Being interdisciplinary in nature, it makes the entire exercise of developing a sizing system a highly capital and time intensive one. Size UK, conducted in 2000 cost £1.2 billion and Size Sweden took 5 years to complete. The extent of time and money required has been major deterrents in taking up the issue of data collection and data analysis for garment sizing. Further, development of  sizing systems has  required inputs from a) Anthropologists (planning and conducting an anthropometric  survey) b) Statisticians and mathematicians(population sampling plan, statistical and data analysis) c) Pattern makers (choice of critical measurements) d) Garment technologists and designers (production related issues- size plan and fit). But by far, the weakest link has been the process of data analysis. That is, how to process the huge amount of data generated, so as to be able to arrive at a suitable set of sizes, which can provide the framework for mass manufacture of garments. Simple statistics is incapable of processing this data due to the inherently non linear nature of the problem.

Most sizing systems used today are faulty as they are based on a simple system of averages or percentiles - thereby assuming that human bodies follow a mathematical precision and increments in their shapes and sizes. In other words, each body is supposed to fit one of the sizes which some statistician arrived at by taking the average of a given population- of which one may or may not have been a part. Or worse still, some companies use professional models to test fit their garments before they are put into production!! So unless one has the so called “average figure” or the figure of a “professional model”, one cannot expect to find a garment that fits.

Analysis of the Indian population data in the current study shows that ~ 20% of the population matches the dimensions of an “average figure”! So what happens to the rest 80%? Given this understanding, it might be better to follow a system where the bodies would not be required to fit the sizes, rather the sizes are devised to fit the bodies - as they are in reality. In other words, to devise a sizing system which covers ~80% of the population? 10-20% of the population which is not covered would be comprised of the outliers or those which are significantly different in size or proportion from standard dimensions. These will have to be catered to as special categories ( eg. plus size category).

Garment sizing is also about catering to different body proportions, such as waist to hip ratio (WHR), Bust to hip ratio and so on. This means that no single measurement can be taken as the basis of garment sizing (most current systems do this), instead, some key measurements have to be identified for each garment type and the population has to be classified on the basis of these multiple dimensions. Also, the number of size categories should be kept to a minimum to keep the inventories manageable. Once the size groups have been identified, it should be possible to tell as to how many people are covered under each one of the sizes to facilitate the stock production and inventories. All these further add to the complexity of the problem.

Work carried out at IIT Delhi

Textile Technology Department at IIT Delhi has been involved in research in the area of garment sizing for the past 6 years. Some of the achievements and learning from these studies includes an anthropometric survey conducted over the last 2 years resulting in recent, authentic data for over 1000 women and 500 men in the age group of 18-24 years. A similar survey in 5 regional zones for men and women in the age category of 24-50 years is underway.

Surveys were conducted manually in various educational institutions. Only those who agreed voluntarily were measured. Due to social inhibitions, in general it was found to be very difficult to get subjects, especially girls to agree to participate in such surveys. Such sentiments have been expressed by other researchers too who have undertaken anthropometric surveys at various points of time. Once the venue was decided, specially designed posters were put up to create awareness amongst the subjects about the purpose and objectives of the survey. Male and female members of the staff were rigorously trained in the process of conducting accurate, consistent and reproducible measurements as prescribed in the ASTM Methods D 6240.

Measurements were, wherever possible, conducted in an enclosed area or room to ensure the privacy of subjects. All measurements were taken on level, non carpeted floor for maximum accuracy.  A calibrated tape measure, stadiometer and a digital scale were used to conduct the survey. For major girth measurements, a length of ribbon was tied to locate the body landmarks correctly.  The subjects were asked to stand erect, but in a natural pose. All subjects were measured in well fitting minimum clothes and were bare foot. The measurements were taken in proper order and in the same sequence for all subjects. The 29 measurements taken on young men are shown in Appendix and the description is given in Table 1. A total of 35 measurements were taken for each female subject.

Table 1 Body measurements taken on young men

S.No	Measurement
	Weight (kg)		Height
	Cervical height		Centre front waist length
	Highest shoulder to chest		Front chest width
	Front Waist		Head Circumference
	Mid neck girth		Chest girth
	Waist girth		Centre back waist length
	Armhole depth		Across back shoulder width
	Across back width		Back Waist
	Waist Height		Crotch height
	Crotch length		Crotch depth
	Hip girth		Thigh girth
	Knee girth		Ankle girth
	Arm length		Upper arm length
	Upper arm girth		Elbow girth
	Wrist girth

The data was thoroughly checked for any errors or inconsistencies and cleaned.

Data analysis

Detail statistical analysis of the raw data was carried out to gain a general understanding about the shapes and sizes characterizing the young adult population of Indian men and women. Important insights were gained. Based on such analysis, the Indian male population was divided into 15 categories comprising of 3 categories on the basis of height and 5 categories on the basis of body shape or Drop value. These categories provide a critical framework for developing a garment sizing system for this group of people. The categories and the distribution of the population under the same are shown in Table 2. These categories are the same as those used by some European countries to describe their population but the drop intervals for Indian population were found to be very different from them. Again, the drop intervals were not calculated by just dividing the complete range with the number of categories required. Instead, a complex analysis of the population data was used to arrive at these categories which truly reflect the distribution in the population.

The range of drop values obtained for this data set confirm closely to the values reported for Sweden, South Africa and Germany. But the range is quite different from the populations of France, Finland and Denmark. (John M. Winks, The Textile Institute,). It is important to point out that the data is essentially for very young adults who are more health conscious and hence significantly slimmer as compared to middle aged Indian men. Large abdominal sizes are quite common in higher age groups and this trend is likely to show up during the data analysis of the population in the 26-45 year age category. In general, it can be concluded that Indian men are shorter and have less defined body shapes as compared to their European counterparts.

Table 2 – Percent distribution of young Indian men (18-24 years) under various height/shape categories.

S.No.	Category	Short	Medium	Tall	Total %
1	Corpulent	0.4	0.2	0	0.6
2	Stout	0.8	0.6	0.4	1.8
3	Portly	5	14	5	24
4	Regular	8	37	8	53.6
5	Athletic	2	15	3	20
	Total %	16.2	66.8	17.4	100*

* The figures may not add up to 100 as they have been rounded off.  

Multiple correlation analysis of the anthropometric data showed that neck girth, waist girth, hip girth and chest girth are correlated well to the weight, while cervical height and waist height are correlated well to the height. However, height and weight within themselves have very poor correlation. So in general, for good fitting of garments, at least one lengthwise measure and one girth measure would be required. Measures to be selected would depend on the garment type and the critical measurements for the same.

Most current garment sizing systems are also based on the above mentioned premise. In most cases any one measurement is taken as the basis of sizing such as chest or hip girth and all other measurements are calculated as a function of the same. Such systems assume that if mathematical correlation is high, then the derived measurements would also be accurate. This assumption may very well be true in mathematical applications but in garment manufacture this will invariably lead to poor fit of garments for most of the intended wearers. An example cited below illustrates this point.

Fig.1 Chest vs. waist measurements for young Indian men

Fig. 2 Waist girth of individuals having 90-96 cm of chest girth

It can be seen from Figure 2 that for people having a chest girth of say 91 cm, the waist girth varies from 71 cm to 86 cm, if the outlier at 91 cm is disregarded. This is a range of 15 cm. Similarly, for chest girth of 93 cm, the waist varies from 72 cm to 86 cm – a range of 14cm. It is this spread of data which renders simple mathematics redundant as far as the issue of garment sizing is concerned. That is to say, that despite a high correlation, in the practical sense it is not possible to fit all people having a chest girth of 93 cm with any one waist size. This is because bodies have different shapes and it is these shapes which are covered under the shape analysis, Table 2. But supposing the manufacture did want to offer one size - then what is that   combination of chest and waist (shape) and how many people would be covered under that combination or size? This analysis would have to be performed for each combination of measurements (sizes). A thorough understanding of data spread, coupled with complex multi variable analysis is required to identify and quantify the prominent shapes and sizes existing in any given population.

Algorithm for developing garment sizes
Based on the understanding of these complexities, a special algorithm has been developed to locate the most prominent sizes (defined as a combination of two or more body dimensions) existing in any given population. The system is highly flexible in that it is possible to run it on any anthropometric data set, it is possible to define the accuracy of fit ( by increasing or decreasing the tolerance) and it is possible to define the number of sizes. As the demand on any one of the above parameters is made more stringent, the accuracy of another is compromised, so an optimal combination has to be identified every time.

1	Body size charts for top body garments of 18-24 year old men
2	Body size charts for lower body garments of 18-24 year old men
3	Body size charts for top body garments of 18-24year old women
4	Body size charts for lower body garments of 18-24 year old women

Fit Trials

Fit trials were conducted with ITC Wills, which is the industry partner to the project. The size charts obtained were compared with the size charts of ITC limited. It was found that the predicted sizes were very close to what ITC was using. The predicted sizes however covered a much wider range of measurements, incorporation which could increase the population cover to nearly 80% . Fit models were identified and test garments were stitched for them to optimise the pattern. Fit trials yielded satisfactory fit for ITC Wills’s brand “Miss Players’  tops.

Appendix

Figure 1
a) Circumference measurements,
b) Front measurements
c) Back Measurements,
d) Measurements in Sitting Posture

 

a)

 

b)	c)	d)

         

 

Library Index | Home