The apparel industry is replete with assumptions regarding the body-garment relationship. Traditional anthropometry focuses on linear body measurements, which are inadequate to describe and classify the human body-form for apparel pattern development. To enable the development of a body-form based block system, this case study explored the body-garment relationship for a sheath dress to determine if apparel block shapes could be categorized based on distinct body-form variations. A modified version of Gazzuolo’s (1985) body-garment relationship theory guided the development and analysis of the study. Pattern blocks were fit to 39 female subjects, with 16 dimensions extracted from specific pattern components and graphed to reveal between one and five groups per dimension. Visual analysis of the sample’s body scans revealed 27 body-form variations with 99 categorical descriptions. Categorical descriptions were compared to the dimensional values resulting in ten suggestions for a body-form based block system, and seventeen assumptions that require further analysis. In conclusion, this case study discovered multiple body-form variations across a single size, but block shapes could not be identified due to the wide variation in the sample. Future studies should assess a statistically significant sample of individuals with in-depth analysis of a single body region to determine if there are generalizable body-form variations across the population.
This case study explores the relationship between the human body and the clothing that covers it by empirically testing the common apparel assumption: If ten women of the same size wear the same dress, it will fit them all differently. Anyone who shares their clothing with a sibling/friend of the same size can state this fact, but their stories constitute disparate anecdotal evidence affected by individual fit preference. Empirical, objective assessment of the body–garment relationship for women who share the same size has not been conducted. The body–garment relationship covers interactions between objective and subjective measures of fit, as well as the design features of a garment. Each area requires separate research prior to assessing the associations between them. This study focused on objective measures of fit for American women aged 18 to 54.
Apparel is traditionally a trial-and-error industry, basing decisions on assumptions or ideals rather than on empirical data anchored in content analysis. In addition, sizing and fit are considered competitive advantages and treated as trade secrets in the industry, requiring every manufacturer to define their own body type and sizing system based on their own individual experiences and beliefs. The US government has attempted to alleviate this by offering standardized sizing systems, but research has shown that these systems fail to fit the US female population well (ex. Salusso-Deonier et al. 1985; Goldsberry et al. 1996; Ashdown 1998; Alexander et al. 2005).
Traditional anthropometry provides researchers with the ability to describe the body via linear measurements. However, linear measurements are inadequate to describe and classify human body-forms in a way that is useful to pattern-making practice. To improve patternmaking practice, research into body-form variations and how they affect pattern blocks is necessary. This study is grounded in Gazzuolo’s (1985) body–garment relationship theory which provides an avenue for empirically testing specific relationships between body-form variations and pattern block components (i.e. bust prominence vs. bust dart depth). Such tests help determine which relationships are crucial to the final pattern block shape, and thereby improve the patternmaking process by focusing the technical designer on the most impactful dimensions. This research provides a starting point for examining which body-form variations affect pattern block shapes, as well as how each body-form variation affects each pattern dimension. Knowledge of the body–garment relationship engenders the development of a body-form based block system. Such a system would benefit the apparel industry through faster and more accurate pattern block generation for specific target consumers. Faster turn-around times in the design stage lead to faster overall time-to-market and allow companies to more quickly capitalize on market trends, with fewer markdowns due to poor fit.
The purpose of this case study was to determine if apparel block shapes could be categorized based on distinct body-form variations, with the goal of empirically establishing that similar body measurements do not produce similar body forms. The research questions are:
What are the body form variations across a single size?
What do these findings suggest for the development of a body-form based block system?
To assess the relationship between the body and the garment, it was essential to review literature pertaining to both subjects. Literature concerning the body focused on systems for classifying bodies, while literature concerning the garment focused on the underlying assumptions of patternmaking and grading.
Body-form classification systems can be split into two main categories: sizing systems and form assessment. Sizing systems divide a given population into groups based on body measurements so that the majority of the population is represented in the system using the least number of sizes possible (Petrova 2007). The best sizing systems are based on anthropometric data taken from a large, representative population. Only six anthropometric sizing surveys have been conducted in the US in the past 75 years: The O’Brien and Shelton survey (1941), ANSUR (1988), NCTRF (1990), the Reich and Goldsberry survey (1993), CAESAR (1998), and SizeUSA (2002). These surveys partially influenced the following US government standards for women’s apparel: CS215-58, PS 42–70, ASTM D5585, D5586, D6829, D6960, D7197, and D7878. Most research on sizing systems focuses on illustrating how poorly the government standards fit the US population, which is generally accomplished by testing linear measurements from the standard against the linear measurements from a population to discover statistically significant differences (ex. Patterson and Warden 1983; Simmons et al. 2004; Salusso et al., 2006; Alexander et al. 2012). The focus of these studies was on the linear measurements, not on the body-form or how the body-form could impact pattern-shape, suggesting a gap in the literature related to body-form.
The second most common body-from classification system focuses on body-form assessment. Body-form assessment (aka ‘figure evaluation’) scales classify human bodies into specific categories, such as: sizes, numbers, heights, volumes, letters, and shapes. Figure evaluation relies on comparisons between an observed form and a standard form and can be broken down into four categories: Proportions, Posture, Whole Body, and Body Components.
Proportions are the relationships between different body component lengths (Palmer and Alto 2005). These relationships are the fundamental building blocks for garment patterning systems. Understanding the locations of major body components assists with the accurate placement of pattern features (i.e. seamlines and darts) and helps determine grading rules. Patterning texts agree that the standard figure is evenly divided lengthwise at the hips, and the knees are halfway between the hip and the floor (Latzke and Quinlan 1940; Liechty et al. 1986; Maehren and Meyers 2005), but differ on figure height, elbow placement, and waist level. There are three accepted deviations from the standard figure: short-waisted, long-waisted, and asymmetrical. Short- and long-waisted are calculated by ratios of length measurements between the underarm and hips (Maehren and Meyers 2005). Asymmetrical proportions refer to a difference between the right and left half of the body (Minott 1974, 1978) or between the front and back of the body (Liechty et al. 1986).
Poor posture alters the body configuration, causing key body components, such as the shoulders, breasts, and buttocks, to move out of alignment (Rasband and Liechty 2006). The five most common incorrect posture variations include: overly erect posture, slumped posture, swayed back (Liechty et al. 1986), tilted hip-forward posture, and tilted hip-backward posture (Minott 1974).
Whole body assessment in pattern-making and fitting texts typically flatten the human body to assess for shape instead of form (ex. Maehren and Meyers 2005; Rasband and Liechty 2006). This practice ignores the height, weight, volume, angle, and arc variations intrinsic to the human body and limits the applicability of body-form classification to pattern-making practice. Common body shapes include: average/hourglass, triangle, inverted triangle, rectangular, tubular, oval/rounded, elliptical, and diamond (Latzke and Quinlan 1940; Maehren and Meyers 2005; Rasband and Liechty 2006).
Research on whole-body classification generally echoes popular literature. In their analysis of women from the SizeUSA database, Simmons et al. (2004) discovered nine statistically significantly different whole-body shapes (hourglass, bottom hourglass, top hourglass, spoon, triangle, inverted triangle, rectangle, and diamond) and termed their system the Female Figure Identification Technique (FFIT) for Apparel. The researchers used the FFIT for Apparel system to test the body shapes present in ASTM D5585-95, discovering that only the spoon category was represented. In a study of 222 scans from the TC2 and NC State body scan databases, the spoon category was the third largest category with 17.1% of the population; the bottom hourglass category was largest at 40% ( Istook et al. 2004). In a following study, 6310 American women were analyzed, with the spoon category being the second largest category, and the rectangle category being the largest (Lee et al. 2007). In either case, the body shape of the government standard does not meet the needs of the general American female population.
One study that does not reduce the human body to a two-dimensional shape for whole-body classification was conducted by Olds et al. (2013). Twenty-nine dimensions were extracted from 301 Australian adult body scans and clustered into groups described by the ecto-, endo-, and mesomorph classification system. This approach focused on overall volume, revealing markedly different forms (i.e. oval vs. top hourglass) when comparing the group’s average and most extreme subjects. Simmons et al. (2004) ran into this problem when developing the FFIT for Apparel classification system and disregarded K-means cluster analysis as a viable option for sorting body shapes. These studies indicate that body form can vary across similar volumes, indicating that it may also vary within a single size.
These studies suggest that: (a) population lengths and widths, though not necessarily circumferences, vary more widely than assumed in government sizing standards (Salusso-Deonier et al. 1985), and (b) linear measurements from voluntary standards are inappropriate for fitting the general US population (Simmons et al. 2004; Alexander et al. 2012). These findings indicate that even with similar circumference measurements, subjects may still vary in body-form, as linear measurements do not indicate the depth or volume of body features. In addition, findings from the Olds et al. (2013) study indicate that overall body volume alone does not adequately describe body-form variations.
Historical analysis of patternmaking provides clues for why patterns poorly fit their intended populations in the current market. Before the industrial revolution made ready-made apparel available cheaply, all clothing was custom-made. Dressmakers and tailors analyzed their clients’ body-form and movements to produce garments that fit them perfectly (Kidwell and Christman 1974). The shift from custom to ready-made required a re-imagining of the pattern-drafting process. Tailors invented two drafting systems: direct and proportional. Direct systems were abbreviated versions of custom-made, while proportional systems relied on the principle that the human body is proportional, and that a single measurement could predict the rest (Kidwell and Christman 1974; Aldrich 2007). Proportional drafting led to proportional sizing, and in 1881, Charles Hecklinger combined the ‘body’ (a muslin fit to a specific client; origin of basic blocks) with proportional drafting, developing the first systematic adaptation for pattern blocks, which became the basis for applying size charts to patterns (Kidwell and Christman 1974; Aldrich 2007). These changes to the patternmaking and grading systems essentially eliminated the complexity of the body form from the patternmaking process. The complexity of the body form must be considered during the patternmaking process if a garment is to fit its intended population, hence a need for studies such as this one that empirically assess the relationships between the body and the garment.
Research on pattern shape focuses either on grading or pattern shape changes driven by the body, but does not provide analysis of specific relationships between body-form variations and pattern block dimensions. Grading is the act of increasing and decreasing a pattern by set increments to create a range of sizes. Schofield and LaBat (2005a, b) analyzed 40 US sizing charts from 1873 to 2000 to determine the underlying principles of industry grading practice and then testing these grading assumptions against grade rules developed via regression analysis of the 1988 ANSUR survey. They concluded that none of the grading assumptions were supported by empirical data. Bye, LaBat, McKinney, and Kim (2008) compared traditional grading practices to optimum ones by evaluating sheath dresses graded using traditional grade rules against those custom-fit to subjects. Analysis of the number of adjustments, overall differences between sizes, specific comparisons between key pattern segments, and visual assessment of fit led to the conclusion that traditional grade rules do not provide good fit across a size range. Taken together, these studies suggest that traditional grading practices do not take the body-form into consideration across a size range.
Schofield et al. (2006) explored satisfaction with pant seat shape (flat vs. full) for women aged 55 and older discovering through expert analysis the flat-seat pants fit the majority of the 176 subjects best. Song and Ashdown (2012) tested the final fit of a pair of custom-fit pants when the original pattern was drafted from pant pattern blocks using three lower-body hip variations (curvy, hip tilt, and straight), as well as a standard industry pattern, concluding that the basic blocks created using the hip variations generated better fitting customized pants. Sohn and Bye (2012) investigated changes in sheath dress patterns throughout three pregnancies; concluding that (a) grading for maternity sizing should not be proportional because humans do not grow proportionally, and (b) that different bodies change differently, and that these changes do affect pattern blocks. All of this research on pattern shape changes suggests that the body-form should be a key consideration during patternmaking and that specific body-form variations do affect patterns and grade rules.
A self-sorting method was used to find an appropriate sample of female subjects who share a single size. Unlike the FFIT for Apparel (Simmons et al. 2004) system, the self-sorting method allowed for the deferment of body-form classification until after pattern-block assessment, a crucial way for this study to retain its validity. By sorting into sizes using the most basic key measurements necessary for fitting clothing to the torso (bust, waist, and hips), more detailed body-form variations could be assessed after the garment was fitted to the body, but would ensure a similar basic body type.
The 18 to 54 age range was chosen to follow the American Society for Testing and Materials (ASTM) sizing standard age groupings, which split adult women into two groups: 18–54 and 55 + . To focus on body-form variations within a single size, height was bounded to ensure extremes in height did not skew the results. The lower limit for the height range was set via analysis of the ASTM adult female sizing standards, and was the shortest height found (via ASTM D7878-13e1: Misses Petite, 00P-20P). The upper limit was set in consultation with two experts as no acceptable upper limit was presented by ASTM standards. The three measurements (bust, waist, hip girths) are important measurements for body-form classification (Istook et al. 2004; Alexander et al. 2012) and factor into fit issues with clothing (Alexander et al. 2005).
The goal of the self-sorting method was to find a set of women who were similar enough in body measurements that they could be considered the same size. Subjects within ± 1″ of each other were considered the same size. The ± 1″ tolerance was deemed reasonable, given that a 1″ grade is accepted as an industry standard, a total possible range of 2″ keeps subjects within a single size, and the same tolerance has been utilized in similar research on sizing. For example, Alexander et al. (2012) used the ± 1″ tolerance when testing how well measurements from plus-sized SizeUSA subjects fared against ASTM D6960-04.
The 1036 subjects were sorted via a matrix method such that each subject was tested against every other subject. Four matrices were created in Excel, one each for bust, waist, and hip girths (Fig. 1), and one summing these three girths (Fig. 2). Matrix cells were color coded for ease of visual confirmation while searching the matrix. The equation used to determine if subjects were within ± 1″ of each other is:
‘SUBR’ corresponds to the subject number in the subject row and ‘SUBC’ corresponds to the subject number in the subject column. If the row and column subject labels match, this indicates they are the same subject, and the equation produced a response of ‘0′. When the row and column subject labels did not match and had girths (‘SRGIR’ and ‘SCGIR’) within ± 1″ of each other the equation produced a response of ‘1′, indicating a match; if they did not have girths within ± 1″ of each other the equation produced a response of ‘0′, indicating they did not match.
The cumulative matrix amalgamated the results of the bust, waist, and hip girth matrices to determine the final number of matches between subjects (Fig. 2). The equation for the cumulative matrix is:
‘C#’ refers to the specific cell within a matrix shared by two subjects. A ‘1′ indicated the subject pair had one girth measurement within the ± 1″ tolerance. A ‘2′ indicated the subject pair had two girth measurements within the ± 1″ tolerance. A ‘3′ indicated the subject pair had all three girths within the ± 1″ tolerance and were considered a ‘match’. The ‘Matches’ column, on the left-most side of Fig. 2, counts the number of times subjects in the ‘Subject #’ column matched all three girths with subjects in the ‘Subject #’ row. Groups were formed based on the number of matches between ‘column’ subjects and ‘row’ subjects. One hundred and one subjects did not match anyone else, the average group consisted of 13.8 members, there were 19 groups with 40 + matches, and the largest group had 47 members.
The nineteen groups with 40 + subjects were chosen for further analysis. The ‘column’ subject was the one subject that matched everyone else in the group and was designated the ‘fit model’ for the group. A single group was chosen after analyzing how well the fit model’s age, height, weight, bust, waist, and hip girths matched her sample groups’ average (Table 1).
A modified version of Gazzuolo’s (1985) Body–Garment Relationship (BGR) framework guided this research (Fig. 3). The original BGR is composed of four major components:
The analytical component abstracts the garment, determines the operational definitions of garment orientation, and identifies essential dimensions.
The dimensional component uses the operational definitions from the analytical component to generate, collect, analyze, and sort data from the pattern blocks.
The visual component analyzes critical values (lengths, widths, angles, and radii) of one subject’s body to another’s to understand the proportionate and spatial relationships between body sites and to uncover the extent of physical prominences.
The physiological component focuses on in-depth analysis of the potential reasons why the body formed as it did, including heredity, nutrition, and the environment.
The analytical component set the foundation and bounded the research; providing a thorough description of the chosen garment (a sleeveless sheath dress, Fig. 4) and the fitting rules used to ensure consistent fit across the range of custom-fitted dresses. A sleeveless sheath dress was selected as it covers the body components most often associated with body-form: bust, waist, stomach, abdomen, high hip, hips, and thighs (Simmons et al. 2004; Lamport 2008, 2010). Basic blocks were chosen as they are the closest approximation of the body that is possible for a garment (Fig. 5).
Garment abstraction is the specification of all the components of pattern-shape variance [level of abstraction (complexity), grain orientation, means of suspension, reduction/enlargement, division, and correspondence], such that all the elements of the body-form are considered and applied to the garment (Gazzuolo 1985). This is a correspondence-level garment (highest level of complexity for garment abstraction). The front and back of the dress are differentiated and the seamlines and darts are located relative to the body-form (Fig. 5). The grain falls vertically along the center front and back of each piece. Dress suspension occurred at the shoulders and the location of greatest lower-body prominence. Dimensional reduction/enlargement, used for increasing/decreasing a pattern-block component’s value, could occur at seamlines, hem, and darts. Contour reduction/enlargement, used to align the garment to the body’s natural contours, could occur at the neckline, armhole, and skirt side seams. Each of the six block pieces had vertical divisions that occurred at the side-seam and center back, and horizontal divisions that occurred at the waist and shoulders.
Correspondence specified the anatomical locations of the major pattern points, which occurred at all block borders and the points of greatest prominence. The correspondence points and seams are: high-point shoulder, shoulder point, shoulder seam, center back neck point, center front neck point, neckline, shoulder blade apex, underarm point, armhole, bust apex, center back waist point, center front waist point, side waist points, waist seam, greatest lower-body front prominence, buttocks prominence, greatest lower-body side prominence, knees, side seam, center back seam, and center front line. Analysis of correspondence led directly to the development of the eight fitting rules (Table 2). Visual analysis of fit was employed for fit evaluations and the principles of reduction and enlargement were used to achieve it.
Basic blocks created by the University of Minnesota were the basis for this pattern. The most representative size of blocks was chosen based on comparisons between the fit model’s and block’s bust, waist, and hip girth. The front shoulder dart was moved into the side seam, at bust level, allowing for more accurate triangulation of the bust prominence on the bodice block. Reducing the number of waist darts in the skirt from two per side to one per side made it easier to track changes in the skirt darts. The fit rules were then applied to the basic blocks, resulting in a custom set of blocks for the fit model. These blocks became the starting blocks for fitting the sample and were intended to reduce the amount of alterations and end with better fitting final garments (Song and Ashdown 2012).
The goal of the dimensional component was to describe the major block-shape variations in this sample. The authors have worked extensively with both physical and virtual fitting and could accurately analyze the final fit of the garments without outside assistance. To generate data, the right-hand side of the blocks were altered until the fitting rules were met for each subject. Optitex’s CAD system automatically mirrors changes made on the “working half” of the garment to the “mirrored half” of the garment, so that both halves of the garment are identical. Data collection consisted of gathering length and width dimensional values from the right-hand side of the blocks that directly corresponded to specific body-form variations, which are color-coded in Fig. 5. Dimensions needed to be directly comparable to the body-form so that they physiological component could be smoothly carried out (Table 3).
Dimension values were entered into a spreadsheet, sorted from smallest to largest, and graphed. Each subject received an identifier (a1–a44) to protect their identities. Descriptive frequencies were calculated for each dimension. The dot graphs were set so that the minimum and maximum y-axis values equated to the smallest and largest standard deviations necessary to show all data points for each dimension. The graphs visually represented the range of the measurements within a single dimension and allowed for group identification.
Content analysis of the body through in-depth inspection of the body scans resulted in categorical descriptions of multiple body-form variations. Coding terminology used in this study was subjective and relates only to the sample analyzed; it was not meant to be generalizable. In this study, the term ‘average’ indicated that a variation did not belong in either the upper or lower categories of the body-form variable. The term ‘obscured’ indicated that the body-form variation could not be assessed and does not count as a body-form variation categorical descriptor. Likewise, ‘combo’ designations in the GLBFP region do not count as a categorical descriptor because they account for subjects with equally prominent stomachs and abdomens (stomach was determined to be higher on the body than the abdomen).
The analytical component defined seven key regions (neck, shoulder, shoulder blades, bust, GLBFP, buttocks, and GLBSP) for analysis, ensuring accurate and focused content analysis of the body scans. Body-form variation categorical data was organized in an Excel spreadsheet. Each variation had at least two categories, labeled by specific body part (i.e. shoulder or bust) and measurement entity (i.e. length or fullness). Tallies of how many subjects fell into each category allowed for comparisons within individual body-form variations.
In the modified BGR, the physiological component compares the block-shape variances to the body-form variances. Twenty-seven assumptions were developed based upon consideration of how the body could affect the pattern blocks at specific locations (Table 4), guided by the garment abstraction analysis conducted during the analytical component. A strict one-to-one comparison was used to bound the research. By thinking of the pattern blocks as points connected by lines, the garment was more easily abstracted and each point and line were considered separately. Block points can move either horizontally or vertically, changing the length, steepness, and/or curvature of the connected lines.
Dimensional values were plotted on dot graphs to determine groupings; some groups were more obvious than others. Dot graphs were chosen to allow the researcher to see how the sample dimensions ranged naturally (Fig. 6). The dot graphs were visually analyzed to discover where the groups split naturally, relying on long spaces between dots and locations where the dots levelled off to distinguish groups. Every attempt was made to eliminate subjectivity in group formation, though alternative methods for group formation should be assessed to ensure empirical objectivity.
While the range of some dimensions is quite small, the average measurements for those dimensions are also quite small, thus small differences have a big impact on the number of groups within a dimension. For example, even though the range for averaged shoulder drop is 0.70″, the dot graph suggests that for this sample, there are five distinct groups based on spaces between clusters of dots.
For the dot graphs, the y-axis values indicate the range of measurements for each dimension while the x-axis values indicate the individual subject number. Groupings were color coded, with a red dot denoting the fit model. The fit model was included in the graphs, but not in the calculations of descriptive frequencies or in the groups. Descriptive frequencies as well as the number of groups identified from the graphs are presented in Table 5.
Twenty-seven body-form variations with ninety-nine variation categories were discovered during visual content analysis of the body scans. The inductive coding was developed by the researchers based on the scans in the sample and did not rely on other body classification methods since those were developed through deductive coding. This method is aimed at describing a specific population thoroughly, not on generalizing the findings of a specific population to the general population. Variation of body-form components was evident, even in such a small sample and had to be documented with as many categories per body-form variation as necessary. The number of groups per body-form variation ranged from two to seven.
Analysis of the neck included neck thickness, the neck-to-shoulder transition, collarbone visibility, and neck tilt. Neck thickness produced three groups: thin (13), average (12), and thick (14). The neck-to-shoulder transition produced two groups: sharp (18) and smooth (11). Collarbone visibility ranged from flat (2), nearly flat (15), visible (16), and prominent (6). Neck tilt ranged from straight (8), slightly forward (8), forward (18), and far forward (5) (Fig. 7).
Analysis of the shoulder included shoulder length description, shoulder point sharpness, shoulder point alignment, and shoulder slope description. Shoulder length description produced three groups: short (11), average (10), and long (8). Shoulder point sharpness produced two groups: sharp (16) and soft (23) (Fig. 8). Shoulder point alignment was assessed by the placement of sagittal planes at both shoulder points and analysis of the relation of the planes to the bust, high-hip, and thigh, with alignment either inside, aligned, or outside of each body component. Shoulder point alignment produced seven groups: inside bust, high-hip, and thigh (1), aligned with bust, outside high-hip, inside thigh (1), aligned with bust and high-hip, inside thigh (1); aligned with bust, inside high-hip and thigh (10), outside bust and high-hip, inside thigh (8), outside bust, aligned with high-hip, inside thigh (2), outside bust, inside high-hip and thigh (16). Shoulder slope description ranged from flat (3), slightly sloped (4), sloped (20), more sloped (6), and steep (6).
Analysis of the shoulder blades included prominence point alignment, blade prominence, blade description, and blade width. Prominence point alignment was assessed by marking the prominence points with a transverse plane and seeing where on the body it matched. Prominence point alignment occurred at the armpit (7), at the arm join (30), and above the arm join (1); one subject’s alignment was obscured by their sports bra. Blade prominence ranged from flat (10), almost flat (4), visible (20), and prominent (5) (Fig. 9). Shoulder blade width was assessed by marking the blade prominence points with sagittal planes and determining the distance between them in relation to the entire back. Blade width produced three categories: narrow (15), average (7), and wide (14); three subject’s widths were obscured by their sports bras.
Analysis of the bust included descriptions of bust fullness and bust point width, as well as determination of ribcage containment. Bust fullness ranged from very small, small, average, full, and very full (Fig. 10). Ribcage containment was determined by assessing if the breasts were wider than the torso at the bustline, and produced two categories: contained (26), not contained (13). Bust point width was assessed by marking the bust points with sagittal planes and determining the distance between them in relation to the entire front. Bust point width produced three categories: narrow (9), average (13), and wide (17).
Analysis of the GLBFP included waist indentation, GLBFP identification, GLBFP description, GLBFP alignment and GLBFP extension. Waist indentation ranged from none (4), barely (5), slight (22), and indented (8). To identify the GLBFP a frontal plane was positioned against the body at the abdomen; if the stomach aligned with this plane, the abdomen and stomach were considered equally prominent, but if the stomach extended past the plane, the stomach was deemed the greater prominence. The GLBFP was the abdomen for 35 subjects and the stomach for two, and two subjects had equal stomach and abdomen prominences (Fig. 11). GLBFP descriptions included: flat (9), oval (7), softly pointed (3), softly rounded (4), and rounded (15), and one person had a combination of rounded and softly pointed. GLBFP alignment produced five categories: below waist (1), above high-hip (3), at high-hip (21), slightly below high-hip (4), and below high-hip (8); two subjects had combinations: at waist and at high hip (1), and at waist and below high-hip (1). The GLBFP extension was determined by identifying if the plane that was used to identify the GLBFP passed through the bust, indicating that the GLBFP was more prominent than the bust. The GLBFP extension produced three categories: yes extended (22), aligned (4), and not extended (13).
Analysis of the buttocks included descriptions of the prominence, length, fullest part, and alignment. Buttocks prominence produced two categories: flat (15) and prominent (24). Buttocks length was determined by how much space the buttocks took up between the crotch and the waist; there were two categories: short (10) and long (29). The fullest part of the buttocks was discovered by placing transverse planes at the top, bottom, and fullest part of the buttocks, resulting in three categories: low (12), middle (23), and high (4) (Fig. 12). The transverse plane located at the fullest part of the buttocks bisected the entire body, allowing assessment of alignment; this produced four categories: at the hip (16), slightly below hip (12), below hip (10), and far below hip (1).
Analysis of the GLBSP included location and alignment identification, and description of the prominence. GLBSP location produced two categories: high-hip (4) and thigh (35) (Fig. 13). GLBSP alignment produced five categories: below crotch (3), at crotch (20), above crotch (12), below abdomen (1), and at abdomen (3). GLBSP description produced five categories: flat (5), softly pointed (10), pointed (2), softly rounded (8), and rounded (14).
Each body-form variation corresponded to a pattern dimension. In Excel, pattern dimension values were sorted from smallest to largest, simultaneously sorting the body-form variations. Tallies of each category within each body-form variation were calculated to see how many of each category fell within each group. For pattern dimensions with only one group, the group was split at the mean and the upper half was compared to the lower half.
Neck circumference was compared to neck thickness, the neck-to-shoulder transition, and collarbone visibility. There were 19 subjects below the mean and twenty subjects above the mean for neck circumference. Thin necks were the majority below the mean (56.2%), while thick necks were the majority above (50%). A sharp neck-to-shoulder transition was the majority above and below the mean at 68.4% and 75% respectively. The ‘nearly flat’ collarbone category was the majority below the mean (52.6%), while the visible collarbone category was the majority above (45%).
Front neck drop was compared to neck tilt. Five groups were identified for front neck drop during the dimensional component. Straight neck tilt was the majority for group 1 (50%), forward neck tilt was the majority for groups 2 (66.7%), 3 (71.4%), and 4 (54.5%). Group 4 also had a large number of subjects with slightly forward neck tilt (36.4%) evenly spaced throughout. Far forward neck tilt was the majority for group 5 (71.4%).
Shoulder seam was compared to shoulder length, shoulder point sharpness, and shoulder point alignment. Five groups were identified for shoulder seam during the dimensional component. For the shoulder length description, group 1 consisted entirely of short shoulders, while groups 4 and 5 consisted entirely of long shoulders. Groups 2 and 3 included all three categories, with the short and average categories tied for the majority in group 2 (42.9%) and the long category the majority for group 3 (52.2%). For shoulder point sharpness, groups 1 (74%) and 2 (71%) had a majority of sharp shoulder points, while groups 3 (73.9%) and 4 (66.7%) had a majority of soft shoulder points. Group 5 was evenly split between sharp and soft shoulders. For shoulder point alignment, groups 1 (75%) and 3 (39.1%) had a majority of the ‘outside bust, inside high-hip and thigh’ category. The ‘aligned with bust, inside high-hip and thigh’ category was the majority for group 2 (42.9%). Group 4 was evenly split between three categories and group 5 was evenly split between two categories.
Averaged shoulder drop was compared to shoulder slope description, shoulder point sharpness, shoulder length description, and the neck-to-shoulder transition. Five groups were identified for averaged shoulder drop during the dimensional component. For shoulder slope description, the flat category was the majority for group 1 (66.7%), the sloped category appeared in groups 1, 2, 3, and 4 and was the majority for groups 2 (40%) and 3 (64%), and the steep category comprised the majority of group 4 (60%) and the entirety of group 5. For shoulder point sharpness, both sharp and soft categories appeared in groups 1, 2, 3, and 4. Sharp shoulder points comprised the entirety of group 1 and the majority of group 2 (60%), while soft shoulder points were the majority for groups 3 (68%) and 4 (60%), and comprised the entirety of group 5. For shoulder length description, short shoulders comprised the entirety of group 1 and were the majority of group 2 (80%). Long shoulders were the majority for groups 3 (48%) and 4 (80%), and comprised the entirety of group 5. Average shoulders appeared in groups 3 (36%) and 4 (20%) as the second largest contingent. For the neck-to-shoulder transition, the sharp transition comprised the entirety of group 1, and was the majority for groups 2 (60%), 3 (72%), and 4 (80%), while the smooth transition comprised the entirety of group 5.
The bodice back waist dart depth was compared to the shoulder blade prominence point alignment. Five groups were identified for the bodice back waist dart depth during the dimensional component. The ‘at arm join’ category occurred in every group, was the majority for groups 2 (81.3%), 3 (83.3%), and 4 (83.3%), and tied for majority with groups 1 (33.3%) and 5 (50%). The armpit category was spread through the groups and the ‘above arm join’ category only occurred in group 2.
The bodice back waist dart width was compared to shoulder blade prominence and shoulder blade description. Four groups were identified for the bodice back waist dart width during the dimensional component. For shoulder blade prominence, the flat category occurred in every group and was the majority for group 1 (66.7%). The visible category was the majority for groups 2 (66.7%), 3 (50%), and 4 (56.3%). For the shoulder blade description, both the flat and rounded categories appeared in every group. The flat category was the majority for group 1 (66.7%), the softly pointed category was the majority for group 2 (50%), and the rounded category was the majority for groups 3 and 4 (50% each).
The between back waist darts distance was compared to shoulder blade width. Three groups for the between back waist darts distance were identified during the dimensional component. For shoulder blade width, the narrow category was the majority for group 1 (57.1%), the wide category was the majority for group 2 (62.5%) and comprised the entirety of group 3. The average category only appeared in group 1 (33.3%).
Bust dart depth was compared to bust fullness and ribcage containment. Four groups were identified for bust dart depth during the dimensional component. For bust fullness, the full category appeared in all four groups and comprised the entirety of group 2 and the majority for groups 3 (54.5%) and 4 (50%). The small category was the majority for group 1 (61.1%). For ribcage containment, the majority of groups 1 (77.8%) and 3 (72.7%) had their breasts contained within their torso, while the majority of group 4 (66.7%) did not have their breasts contained within their torso.
The between front waist darts distance was compared to bust point width. Four groups for the between front waist darts distance were identified during the dimensional component. For bust point width, the narrow category was the majority for group 1 (75%) and tied for majority with the average category for group 2 (37.5%). The average category was the majority for group 3 (66.7%), and the wide category was the majority for group 4 (70.6%).
The waist circumference was compared to waist indentation. Four groups for waist circumference were identified during the dimensional component. For waist indentation, the slight category was the majority for group 1 (58.3%) and comprised the entirety of group 4. The indented category comprised the entirety of group 2, and the ‘none’ category comprised the entirety of group 3.
The front waist width was compared to the GLBFP location, GLBFP description, and the GLBFP extension. Six groups for front waist width were identified during the dimensional component. For the GLBFP location, the abdomen comprised the entirety of groups 1, 2, and 5, and was the majority for group 3 (96.3%). The abdomen and ‘both’ categories were tied for majority in group 4 (50%), and all three categories tied in group 6. For the GLBFP description, the rounded category was the majority for groups 2 (50%), 3 (37%), and 6 (66.7%) and tied for majority in group 4 (50%). The softly rounded category comprised the entirety of group 1 and tied for majority with the oval category in group 5 (50%). For the GLBFP extension, subjects with GLBFP’s that extended past the bust comprised the entirety of groups 1 and 2, and were the majority for group 3 (51.9%). Subjects with no GLBFP extension were the majority of group 6 (66.7%).
The skirt front waist dart depth was compared to the GLBFP alignment. Five groups for skirt front waist dart depth were identified during the dimensional component. For the GLBFP alignment, the high-hip category tied for majority with the above high-hip category in group 1 (40%), tied for majority with the below high-hip category in group 2 (37.5%) and was the majority for groups 3 (69.2%) and 4 (63.5%). The below waist and below high-hip categories tied for majority in group 5 (50%).
The skirt front waist dart width was compared to the GLBFP description. Three groups for the skirt front waist dart width were identified during the dimensional component. For the GLBFP description, the softly rounded category comprised the entirety of group 1. The rounded category was the majority for groups 2 (37.1%) and 3 (66.7%), while the flat category was the second largest contingent in group 2 (25.7%).
The skirt back waist dart depth was compared to buttocks length, buttocks fullest part, and buttocks alignment. Three groups for skirt back waist dart depth were identified during the dimensional component. For buttocks length, the long category comprised the entirety of group 1 and was the majority for group 3 (75.7%). The short category comprised the entirety of group 2. For buttocks fullest part, the low category comprised the entirety of group 1 the high category comprised the entirety of group 2, and the middle category was the majority for group 3 (62.2%). For buttocks alignment, the ‘slightly below true hip’ category comprised the entirety of group 1, the hip category comprised the entirety of group 2 and was the majority for group 3 (40.5%).
The skirt back waist dart width was compared to buttocks prominence description. Four groups for skirt back waist dart width were identified during the dimensional component. For the buttocks prominence description, the flat and prominent categories tied for majority in group 1 (50%). The flat category was the majority for group 2 (75%) and the prominent category was the majority for groups 3 (80%) and 4 (56.3%).
Skirt curve length was compared to GLBSP location, GLBSP alignment, and GLBSP description. Two groups for skirt curve length were identified during the dimensional component. For the GLBSP location, the high-hip category comprised the entirety of group 1, while the thigh category comprised the entirety of group 2. For the GLBSP alignment, the abdomen category was the majority for group 1 (75%), while the crotch category was the majority for group 2 (57.1%). For the GLBSP description, the rounded category comprised the entirety of group 1, and tied with the softly pointed category for the majority of group 2 (28.6%).
Results from the visual component provide the answer to the first research question: What are the body-form variations across a single size? All seven torso regions had multiple body-form variables, and each body-form variable had at least two categorical descriptors.
While many of the categorical descriptions of the body-form variations can be found in popular sewing literature (indicating many of these variations are known), this study provides a method for systematic analysis of a group of individuals missing from the literature. While Simmons et al. (2004) and Connell et al. (2006) provide methods for systematic body-form analysis, this study provides a method for deeper analysis of the body and relates body-form variations to specific pattern block components. Body-form analysis in the apparel industry is only useful in the context of pattern block generation or alteration.
Interestingly, the sample differed in many ways from the fit model’s categorical descriptions of the body-form variations, with sixteen matches and eleven non-matches. As seen in Table 6, the neck and shoulder regions have the most non-matches with different categories for 75% of each region. This means that garments that fit the upper torso of the fit model well, fit the sample’s upper torso poorly. Differences in neck thickness, neck tilt, shoulder length, and shoulder point alignment affected total garment balance and caused the lower section of the garment to appear poorly fitted on most of the sample. The remaining five body regions matched well, which makes sense as the sample was sorted by bust, waist, and hip girths, and these measurements directly impacted the shoulder blade, bust, GLBFP, buttocks, and GLBSP regions of the pattern blocks. This suggests that the addition of the neck circumference or shoulder length linear measurements to fit model designation for a target market may improve garment fit and speed up the garment sampling process.
The assumptions from the physiological component provide answers to the second research question: What do these findings suggest for the development of a body-form based block system? Assumptions were split almost into thirds: ten were upheld, eight were partially upheld, and nine were not upheld. The upheld assumptions provide specific suggestions for how specific body-form variations affect specific pattern components. The remaining seventeen assumptions require further analysis before suggestions can be created. While not all assumptions provided concrete suggestions for the creation of a body-form based block system, a promising start has been made.
Assumptions 4, 5, 6, 8, 9, 10, 15, 16, 17, and 26 are upheld, meaning that there are clear groupings between pattern dimension values and body-form variation categories. These assumptions deal with the neck, shoulder, shoulder blade, bust, and GLBSP regions of the pattern blocks. Assumption 4 suggests that pattern blocks may need at least four distinct neck drop lengths while Assumption 5 suggests that pattern blocks may need three distinct shoulder seam lengths. Assumptions 6, 8, 9, and 10 suggest that when someone has soft shoulder points, they need a pattern with longer shoulder seams, and when someone has sharp shoulder points, they need a pattern with shorter shoulder seams. Assumption 15 suggests that pattern blocks need three alternative dart placement locations depending on shoulder blade width. Assumption 16 suggests that at least five bust dart depths are necessary for the wide range of bust fullness while Assumption 17 suggests that busts contained within the ribcage require patterns with smaller bust dart depths and busts not contained within the ribcage require patterns with larger bust dart depths. Assumption 26 suggests two different locations for the hip curve point, as well as different angles of curvature (one centered at the abdomen with a deeper curve; one centered at the crotch with a shallow curve).
Assumptions 1, 11, 14, 18, 20, 24, and 27 are partially upheld, meaning there are possible connections between body-form variations and pattern dimension values, but require more research to confirm. In general, there are large numbers of categories within each body-form variation for these assumptions, accounting for the difficulty in determining clear groupings. For assumption 1, thin and thick necks are clearly split, but average necks span the entire measurement range. For assumption 11, the smooth transitions cluster loosely around the largest shoulder drop values, but the sharp neck-to-shoulder transitions span the entire measurement range. For assumption 14, flat shoulder blades as well as softly pointed shoulder blades group together, but softly rounded and pointed categories have too few subjects for clear groupings, and the rounded category spans the entire measurement range. For assumption 18, the narrow category groups around the smallest measurements, but the average and wide categories span the entire range, and the four largest measurements belong to subjects with average bust point widths. For assumption 20, the abdomen category comprises 90% of the sample, meaning that the stomach and ‘both’ categories are too small to group; additionally, the data suggests that the GLBFP location has a stronger influence on the front waist width than the GLBFP description. For assumption 21, the data suggests that the oval, softly pointed, and rounded GLBFP categories extend past the bust, while the flat category does not, but the groupings are loose. For assumption 24, the two buttock length categories are evenly spaced throughout the measurement range forming no groups, though they do correspond strongly with the location of the fullest part of the buttocks, with the long + low and long + middle categories showing the most interaction. For assumption 27, the data suggest that some prominences may affect skirt curve length, especially at the high-hip, but there are too few subjects to draw a firm conclusion.
Assumptions 2, 3, 7, 12, 13, 19, 22, 23, and 25 are not upheld, meaning that groupings contradict expectations or that groupings were nonexistent. Five of these dimensions focus on darts, indicating that the prevailing understanding of dart function may be incomplete. Assumption 2 contradicts expectations as smooth neck-to-shoulder transitions appear to produce smaller neck circumferences while sharp neck-to-shoulder transitions appear to produce larger neck circumferences. Assumption 3 may contradict expectations as more prominent collarbones produce smaller neck circumferences. Assumption 7 suggests that the shoulder seam length cannot be categorized using shoulder point alignment as this variation compares multiple body-form components that do not interact with the shoulder. Assumption 12 may contradict expectations as shoulder blade prominences closer to the waist produce larger bodice back waist dart depths while shoulder blade prominences farther from the waist produce shorter bodice back waist dart depths. Assumption 13 suggests that there is no relationship between shoulder blade prominence and the width of the bodice back waist dart, which contradicts accepted pattern-drafting practice (i.e. the fuller and more prominent, the deeper and wider the corresponding dart). Assumption 19 suggests there is no relationship between waist indentation and waist circumference within a size. Assumption 22 suggests that GLBFP alignment alone cannot account for the skirt front waist dart depth while Assumption 23 found that there either may be too many categorical descriptors for the GLBFP description, or the skirt front waist dart width is not affected by the GLBFP. Assumption 25 suggests that greater buttocks prominence does not lead to larger skirt back waist dart widths. Assumptions 13, 23, and 25 suggest that instead of the dart acting as a way to accommodate a prominence, it acts solely as a reduction device, indicating that as long as overall circumference is reduced, darts do not need to point to a prominence.
This study concludes that for this sample there are multiple body-form variations across a single size and that the findings from comparing body-form variations to pattern dimensions can provide important suggestions for the development of a body-form based block system. Unfortunately, complete pattern block shapes could not be found from this sample, as there was too much variation in body-form. A body-form based pattern block system will be complex and require a new mode of thinking about pattern-drafting. Patterns should be thought of as puzzles, with the body-form variation-based pattern components making up the puzzle pieces. Such a system necessitates a large library of pattern components, but once compiled, they can be combined in infinite ways. This would allow not only traditional ready-to-wear manufacturers to create better fitting block patterns for their target markets, but also designers who wish to specialize in customization to more easily create block patterns for individual customers.
Due to the subjective nature of visual analysis, the results from studies such as this one cannot be generalized. New objective methods for describing body-form variation must be developed. Potentially useful body-form description may come from Gazzuolo’s (1985) original visual analysis strategy of comparing the linear measurements from a subject’s body against the linear measurements from their pattern blocks, but depth and volume calculations will be necessary to fully describe the body as purely linear measurements do not adequately describe body-form variation, as shown in this study. Due to the scope of the current study, statistical analysis could not be performed on the dimensional values, nor on the physiological comparisons between dimensional and categorical variables. Statistical analysis should be performed to validate the conclusions made herein. Next steps include testing each region of the body separately, determining exact dimensions for each body-form variation category in multiple sizes, and discovering common combinations of body-form variations in the population.
Availability of data and materials
Part of the data that support the findings of this study are available from SAE International (CAESAR—https://store.sae.org/caesar/#3dna). The remaining data analyzed during this study are not publicly available due to privacy restrictions set by the University of Minnesota’s IRB.
Aldrich, W. (2007). History of sizing systems and ready-to-wear garments. In S. Ashdown (Ed.), Sizing in clothing: Developing effective sizing systems for ready-to-wear clothing (pp. 1–56). Cambridge: Woodhead Publishing Limited.
Alexander, M., Pisut, G. R., & Ivanescu, A. (2012). Investigating women’s plus-size body measurements and hip shape variation based on SizeUSA data. International Journal of Fashion Design, Technology and Education,5(1), 3–12.
Ashdown, S. (1998). An investigation of the structure of sizing systems: A comparison of three multidimensional optimized sizing systems generated from anthropometric data with the ASTM standard D5585–94. International Journal of Clothing Science and Technology,10(5), 324–341.
Connell, L. J., Ulrich, P. V., Brannon, E. L., Alexander, M., & Presley, A. B. (2006). Body shape assessment scale: Instrument development for analyzing female figures. Clothing and Textiles Research Journal,24(4), 80–95.
Petrova, A. (2007). Creating sizing systems. In S. Ashdown (Ed.), Sizing in clothing: Developing effective sizing systems for ready-to-wear clothing (pp. 57–87). Cambridge: Woodhead Publishing Limited.
Salusso-Deonier, C. J., DeLong, M. R., Martin, F. B., & Krohn, K. R. (1985). A multivariate method of classifying body form variation for sizing women’s apparel. Clothing and Textiles Research Journal,4(1), 147–160.
Schofield, N. A., & Labat, K. L. (2005). Exploring the relationships of grading, sizing, and anthropometric data: Table 1 Comparison of sizing charts from 1873 to 2000. Clothing and Textiles Research Journal,23(1), 14.
Schofield, N. A., Ashdown, S. P., Hethorn, J., LaBat, K., & Salusso, C. J. (2006). Improvin pant fit for women aged 55 and older through an exploration of two pant shapes. Clothing and Textiles Research Journal,24(2), 147–160.
Simmons, K. P., Istook, C. L., & Devarajan, P. (2004). Female figure identification technique (FFIT) for apparel, Part I: Describing female shapes. Journal of Textile and Apparel, Technology and Management,4(1), 1–16.
RC analyzed and interpreted the data and wrote the manuscript for her master’s thesis. EB acted as graduate advisor, provided guidance, and suggested key revisions. Both authors read and approved the final manuscript.
RC is an Assistant Professor of Fashion Design and the Fashion Design Program Director at Brenau University and holds a MS and a Ph.D. in Apparel Studies from the University of Minnesota. Her research interests focus on fit, sizing, and apparel industry design practice. RC is the corresponding author for this paper.
EB is the Department Head of the University of Minnesota’s Design, Housing, and Apparel program. Her research specialties and expertise focus on fit, sizing, functional apparel, and the relationship between design, manufacturing, and sustainability. She has received the ITAA Lectra Innovation award for faculty research twice, is co-editor of a focused issue of the CTRJ (Advancing Design Scholarship in Textiles and Apparel), and editor of a focused issue of the Fashion and Textiles journal (New Approaches to Apparel Making in the twenty-first century).
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.