correlational research - article[1]

Upload: anny2171

Post on 08-Apr-2018

221 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/7/2019 Correlational Research - Article[1]

    1/5

    nutrition survey

    TheAmerican Journal of C linica l Nutrition30: MAY 1977, pp.721-725. Printed in U .S.A . 721

    Level o f educa tion , level o f in com e, andleve l o f fa tness in adu lts12

    Stanley M . Stephen M . Patric ia E . and Ian T. T. H igg ins4

    ABSTRACT Socioeconomic sta tus is system atically re lated to the level of fatness , and

    therefore the incidence of obesity , in a total community survey of nearly 5,000 adults . Among

    males w ith more than 12 years of schooling, the average thickness of 4 fa tfo lds is 10% greater,

    amounting to about 2 kg of to tal fat, than those w ith 8 years or less of education. Inema l e s ,

    however, the opposite trend is observed, those in the higher educational group averaging 20%

    thinner fatfo lds , or about 5.5 kg total fa t, than fem ales in the lower educational group. These

    findings confirm the need for standards of obesity that take socioeconomictatus into ac-

    count. Am. J. C lin . N utr.30: 72 1-725, 1977.

    Although socioeconom ic sta tus is known

    to affect the level of fatness and therefore

    the incidence of obesity, neither the nature

    nor the direction of the rela tionship is c lear.

    In some studies, the poor are leaner and the

    more affluent are more frequently obese

    ( 1). In most reports , obesity is more com -mon in those of lower socioeconom ic sta tus,

    and less frequent in the highest socioeco-

    nom ic class (2-8). In our own analysis of

    fa tfold data from the Ten State Nutrition

    Survey, fa tness increases w ith income level

    in the adult male, yet decreases w ith in-

    creasing income in the adult female (9-1 1).

    Some of these discrepancies appear to

    arise from differing socioeconom ic indices,

    different measures of occupation and educa-

    tion, and different definitions of obesity .

    In other cases, there are obvious sampling

    problems, w ith different population seg-m ents involved. Most previous investiga-

    tions have made use of selected groups , and

    not natura l or demographic populations,

    nor have they included the complete socio-

    econom ic range.The present study of adult fa tness , there-

    fore, is based upon a single community, w ith

    representa tion from the entire socioeco-

    nom ic range. Both sexes are represented in

    nearly equal numbers, and the wom en are

    for the m ost part the w ives of the males.W ith a large sample size (N5 ,000) de-

    rived from a tota l popula tion sample sev-

    era l socioeconom ic indices and four differ-

    ent fa tfolds, w e now have definitive infor-m ation on the relationship between various

    socioeconom ic m easures and the level of

    fa tness in an agricultura l-industria l Amen-

    can com munity.

    M ethods and m ater ia ls

    This study is based upon four fatfold m easurem ents

    and three socioeconom ic measures on 4,936 adult par-

    ticipants in the Tecumseh (M ichigan) Project o fhe

    University of M ich igan School of Public Health12).

    A ll measurem ents were taken during exam ination

    round 2, between 1962 and 1965.

    The three socioeconom ic measures includedI ) edu-

    cational level of head of household ,2) educational

    level of the subject and3) total household income at

    the time of the examination. Since educational level

    I This study was supported by Grant HD 09538

    (Socioeconomic and Genetic D eterm inants of Obesity)

    from the N ational Institute of Health .

    2 Address reprin t requests to : D r. Stanley M . Garn,

    Cent e r f or Human Growth, 1111 E. Catherine S t.,

    Ann Arbor , Mich igan , 48109.

    Center for H uman Growth, University of M ichi-

    gan and the Nutrition Unit, School of Public H ealth ,

    Ann A rbor, M ichigan. D epartm ent of Epidem iol-

    ogy, U nivers ity of M ichigan. School of Public Health ,

    Ann Arbor, M ichigan.

  • 8/7/2019 Correlational Research - Article[1]

    2/5

    Age group N

    Subscapular

    8 Y r or Over 12

    less y r

    Triceps

    8 Y r or Over 12

    less y r

    Abdomina l

    8 Y r or Over 12less y r

    8 Y r orless

    Iliac

    Over 12yr

    Ma l e s

    25-35 178 14.8 13.8 13.5 16.8 29.7 30.2 24.9 25.23 5 - 4 5 2 7 4 1 5 . 5 1 6 . 8 1 4 . 1 1 6 . 4 2 9 . 5 3 0 . 0 2 1 . 2 2 3 . 5

    4 5 - 5 5 1 7 2 1 7 . 2 1 7 . 9 1 3 . 9 1 5 . 4 2 9 . 6 3 1 . 32 3 . 0 2 5 . 35 5 - 6 5 1 3 0 1 7 . 3 2 2 . 4 1 5 . 1 1 5 . 12 9 . 6 3 0 . 1 2 2 . 4 2 2 . 865-75 91 13.5 17.2 12.4 14.7 24.7 29.8 16.8 19.7

    M ean dif ference (mm ) +2.0 +1.9 +1.7 +1.8Percent dif ference (%) +1 3 + 1 4 +6 +8

    Females

    25-35 174 17.4 13.0 22.4 23.1 34.5 24.6 25.3 16.2

    35-45 254 18.9 14.8 25.4 24.4 35.5 26.9 24.6 16.545-55 176 24.0 18.0 29.6 25.0 44.9 33.3 34.6 23.355-65 129 22.4 19.2 27.8 27.2 40.9 35.0 31.8 25.265-75 96 24.2 20.8 30.0 25.0 44.9 38.2 35.1 30.8

    M ean difference (m m ) -4.2 -2.1 -8.5 --7.9Percent dif f erence (% ) -2() -8 -2 1 -26

    72 2 GARN ET A L .

    (, Com puter calculated using random ly generated fractional integers to allow decim al values of m edians .N

    for the tw o educational lev els show n.

    ranged from no schooling through postgraduate univ er-

    sity level, three educational groupings w ere used.

    T hese groupings com prised1 ) 8 y ears or less (16.5 % ofhouseholds), 2) 9 to 12 years (60.5% of households)and 3) more than 12 years (23% of households). In

    com parable f ashion, the incom e groupings w ere) be -low $5,000 per y ear,2) $5,000 to 10,000 and3)

    $10,000 and above. A pprox im ately 60% of the house-

    holds w ere in the second ($5,000 to 10,000) incom e

    c a t e g o r y .

    Of the four fatfolds , tw o are w ell k now n-the tn-ceps fatfold and the subscapular fatfold. T he less-f am il-

    iar fatf olds included the abdom inal fatfold, 2 cm to the

    right of tile um bilicus, and the iliac fatfold, at the

    m idax illary line just above the cres t of the ilium . A ll

    w ere tak en w ith the L ange constant pressure calipers,

    as described by the Comm ittee on N utritional A nthro-

    pom etry , N .R .C . (13), and the encoded data w ere

    carefully edited before initial data analy sis. Fatfold-fatfold correlations av eraged 0.79 on a decade-by -dec-

    ade basis , w ith correlations as high as 0.91 for som e

    fatfold-com binations , and fatf old-w eight corre lationsav eraged 0.71 for m ales and 0.75 for fem ales.

    Com parisons w ere decade-specif ic , to correct forage trends. In order to elim inate changing educational

    lev els , the y oungest age category w as 25 to 35The

    basic question w as the relationship betw een socioeco-

    nom ic lev el and level of f atness in adult m ales and

    fem ales in a single comm unity .

    Findings

    A s show n in Table 1(partially abridgedto sav e space) the level of education and thelevel of fatness are system atically related inboth sex es, but in diam etrically opposite di-rections for adult m ales and adult fem ales.

    TA BLE IEffect of educational lev el on fatfold thickness (mm )

    For the 2,310 adult m ales, the level of

    fatness increases w ith educational lev el such

    that those w ith m ore than 1 2 years of educa-

    tion are fatter than those w ith 8 years of

    education or less. T ak ing all four fatfoldsinto account (triceps, subscapular, abdom i-nal, and iliac) and all three decade groups

    (4th through 8th), this trend is highly signif i-cant by the k -square test. T he2 value is1 1 .26, w ith corrections for continuity .

    Expressed in m illim eters of the double

    fatfold, T ecum seh m ales w ith college and/orprofessional education are approx im ately

    10% fatter than those w ith less than high-school education . Converted into estim atedw eight of fat (FW ) and percent f at (%F),

    using the m ethod developed by Oarn and

    Harper (14) and elaborated by B roek (15),males w ith a college education have approx-

    imately 2 kg greater fat w eight than those

    w ith 8 y ears of education or less and a com -

    parably greater percentage of fat.For 2,626 adult fem ales (sim ilarly age-

    grouped) fatnessdecreases w ith increasing

    educational level. This is show n for all fourfatfolds in Table 1and for f ive decadegroupings. A gain the trend is highly signif i-cant, holding for 19 out of 20 paired com -

    panisons, w ith a stochasticx2 value of 14.46.W om en w ith college and/or professiunal ed-ucation are far leaner than those lack ing the

    high school years .

  • 8/7/2019 Correlational Research - Article[1]

    3/5

    AGE

    LEV ELS OF EDUCA T ION , INCOM E. A ND FA TNESS IN A DULTS72 3

    a Com puter calculated using random ly generated fractional integers to allow decim al v alues of m edians.N

    for the tw o incom e levels show n.

    Expressed as an av erage for all four fat-

    folds, w om en in the third (highest) educa-tional category average 6 mm leaner than

    those in the low est category (8 years andless). T his is a 20% difference in the aver-

    aged thickness of outer fat. T he FW is.5

    kg less in the m ost educated grouping, and

    % F decreases from 32% to 26% w ith in-

    creasing educational level.M uch the sam e trend appears w hen

    household incom e is em ploy ed as the socio-

    econom ic indicator, using all four fatfoldsand all f ive decade groupings (T able2). A s

    show n, the higher incom e m ales are fatter

    for the triceps, subscapular, abdom inal, and

    iliac fatfolds. Y et the higher incom e fem alesare leaner (at every site m easured) than

    those in the low er incom e category ; this

    reiteration of fatness trends, detailed in T a-

    ble 2, raises the tw o questions as to w hetherthe respondents educational lev el is the m a-

    jon param eter (on that of the spouse) andwhether educational lev el attained or house-

    hold incom e are m ore closely related to the

    fatness trends.

    Com paring fatness lev el both by educa-tional lev el of the respondent and educa-

    tional bevel of the head of the household, as

    in Figure 1 the f irst question is imm ediatelyansw ered. T he trends are very m uch thesam e, by respondents educational level and

    by education level of the husband, but there

    is a som ew hat greater fatness spread w henthe w om ans ow n educational lev el is used(cf. solid lines i n Fi g. 1 ) .

    The interaction betw een education and

    incom e is best dem onstrated by com paringfatness levels w ith each of the socioeco-

    nom ic v ariables held constant, in turn. This

    design includes all three lev els of incom e

    and all three lev els of education first holding

    EE

    -J

    U-

    (.3

    -I

    FIG . 1 Relationship betw een educational lev eland iliac fatfold thick ness in w om en betw een the 4th

    and 8th decade. A s show n, w om en w ith 8 years of

    education or less are sy stem atically f atter than w om enw ith 12 years of education or m ore. A s show n, this

    inv erse relationship betw een education and f atness in

    the fem ale is only partially attenuated w hen the hus-

    bands education is used instead.

    TABLE 2Effect of incom e level on fatfold thickness (mm )

    A ge group N

    Subscapular

    in- H igh in-com e com e

    Triceps

    Low in- High in-com e com e

    Abdomina l

    Low in - H igh in -com e com e

    Iliac

    L ow in- High in-com e com e

    Male s

    2 5 - 3 5

    35-45

    4 5 - 5 555 -65

    6 5 - 7 5

    1 3 6

    18 9

    1 491 23

    8 2

    11 . 0 1 4 . 2

    17.2 16.9

    14 . 4 1 8 . 31 5 . 3 2 3 . 4

    1 4 . 0 1 6 . 7

    1 5 . 2 1 6 . 2

    13.6 15.8

    14 . 3 1 6 . 01 3 . 6 1 7 . 3

    1 2 . 6 1 3 . 7

    2 4 . 7 2 7 . 1

    31.6 29.5

    29 . 6 3 0 . 6

    2 6 . 8 3 3 . 1

    2 5 . 0 2 0 . 0

    1 9 . 7 2 3 . 3

    22.4 24.9

    21 . 5 2 6 . 1

    2 1. 7 25 .5

    16 . 9 1 6 . 7

    Mean di ff er ence (mm)

    P ercen t difference (qf,)

    + 3 . 5+24

    + 1 .9

    +14

    +0.6

    +2

    +2.9

    +14

    Fema l e s

    25-35

    35-4545-55

    5 5 - 6 5

    65-75

    15 6

    22 216 9

    132

    10 1

    15.1 13.5

    18.8 16.825.0 17.0

    2 1 . 7 1 7 . 7

    23.0 16.3

    22.4 22.9

    23.5 23.230.0 24.8

    27 . 3 2 8 . 4

    27.6 25.1

    30.0 26.5

    34.8 27.443.4 30.7

    44 . 8 3 7 . 0

    41.0 29.7

    20.9 15.9

    24.8 16.932.0 24.1

    32 . 6 2 5 . 7

    35.0 26.2

    Meandifference(mm)

    Per cent d if fe r ence (% )-4.5

    -2 2

    -1.3

    -5

    -8.5

    -2 2

    -7.3

    -2 5

  • 8/7/2019 Correlational Research - Article[1]

    4/5

    724 GARN ET AL.

    L6 FEMALES

    FIG. 2. Effects of education on fatness(above)and of income on fatness(below) in adult males and

    females, pooling all ages. Restricting this comparisonto the iliac fatfold, it is possible to show for three levelsof education and three levels of income, that both

    educational level and income are related to the level offatness but itt cons stently oppos te directions inma l e s

    and females. (These comparisons involve the $5,000 to$10,000 income grouping and the 9 to 12 year cduca-

    flon grouping and so differ from Tables I and 2 and

    Fig. 1.)

    income level constant and then holding edu-cational level constant. The results graphedin Figure 2 for the iliac fatfold show a de-

    crease in fatness with increasing educationallevel in the female amounting to some 33%and a slight increase in fatness with increas-ing educational level for males. Holding ed-

    ucational level constant, income alone has aconsiderable effect on fatness in both sexes.The females show a9% decrease in iliacfatfolds with i n c r e a s i ng i n come while malesregister a 20% increase over the same three

    income levels. lhere is, therefore, a systematic relation-

    ship between socioeconomic status and fat-ness in a total population sample, but inopposite directions for the two sexes as ti-dily summarized in Figure 3. Of the twosocioeconomic variables considered (educa-

    tion and income) both relate to fatness.W ithin families, however, income appearsto be the more important of the two conre-lated variables. For comparable educationallevel, more money makes for leaner females

    and - within the income range of the sam-

    pie-leaner males.

    Discussion

    This study confirms the highly systematic

    relationship between socioeconomic statusand the level of fatness in adults as well asthe fact that the relationships are in oppositedirections for the two sexes. Among malesstudied in a total community sample highereducation or higher income is associatedwith greater fatness. Among females in thevery same community those with higher ed-

    ucational attainment or higher family in-come are dramatically leaner at four differ-

    ent measuring sites.

    MALES FEMALES

    FIG. 3.Systematic but contrasting relationship be-tween educational level and fatness in males and fe-

    males respectively. U sing pooled-age data, 4,936 par-

    ticipants. and three educational groupings, it is clearl h a t t h e ma l e increases approximately 15% in iliacf a t f o l d t h i c k n e s s wh i l e t h efemale decreases nearly30% in iliac fatness in the progression from less thanhigh school education through to college and beyond.

    (This comparison includes all 4,956 participants and

    thus differs from the education-restricted comparisonshown in the lower part of Fig. 2.)

  • 8/7/2019 Correlational Research - Article[1]

    5/5

    LEVELS OF EDUCATION, INCOM E, AND FATNESS IN ADULTS725

    These findings, based on nearly 5,000

    adults, are consistent with what we havepreviously reported from the lower-incomeTen State Nutrition Survey (9-1 1), but theyinclude additional socioeconomic measures,additional fatfolds, and attention to the in-teraction between education and income.These findings help to explain the obsenva-tion by Stunkard et al. (Ref. 7, p.7 9 ) thatupper class women were systematically thin-ner whereas among boys as among men,there were no such differences.

    Both of the trends here reported are pat-ently inconsistent with the genetic hypothe-sis, and they are inconsistent with the stratifi-cation hypothesis (i.e., that the fat are con-centrated in the lower income group of re-cent immigrants). Nor can the relationshipsbetween socioeconomic status and adult fat-ness be reconciled with the early-inductionhypothesis. One would have to assume thatthe poor overfeed their daughters and un-denfeed their sons in early infancy, or thatthe rich underfeed their infant girls yet over-feed their baby boys.

    W ith data such as these drawn from asingle community (and with the women beingthe wives of the men), there is further rea-son to rethink the conceptual basis of presentfatness norms. I f fatness is related to socio-economic status, as it certainly proves to be,then whose measurements should constitutethe norm? Should the fatness norm be de-nived from the poor (with fatter women), orthose more affluent (but with fatter men)?Should the proper definition of obesity begroup related , or purely statistical and based

    on a National Probability Sample as we havepreviously discussed (16).

    W hat we have here are major and defini-tive indications that the level of fatness (andtherefore the incidence of obesity) differsaccording to educational level and level ofincome in an American community, and indiffering directions in the two sexes. These

    findings do not fit available animal modelsof fatness and obesity and, indeed, theypoint out the limitations of available animalmodels for obesity research. They point tothe need to investigate human obesity incultural context, in the context of the com-munity, in the family, and in behavioral con-text. The fact that women of lower educa-

    tional attainment, but married into higher

    income families, are considerably leanerthan their low income peers indicates theextent to which social factors are far fromtrivial in the development of obesity. fl

    The authors wish to thank Kenneth E. Guire for hisaid in data reduction and statistical operations. The

    manuscript was completed by Linda Kelly.

    R eferences

    1 . H A M M O N D , W. H.M easurement and interpreta-tion of subcutaneous fat with norms for children

    and young adult males. Bnit. J. Prey. Soc. M ed. 9:

    2 0 1 , 1 9 5 5 .

    2. B O N N E T , F., A N D H . LOZET. Le contexte medico-social de lob#{ 233} sit#{ 233}hez l enfant.Acta Paediat.

    B e l g . 2 2 : 2 1 1 , 1 9 6 8 .

    3. GO LDB LA i -r , P. B., M . F.M ooRE A N D A. SniN-K A R D . Social factors in obesity.J. Am. M ed. A s-s o c . 1 9 2 : 9 7 , 1 9 6 5 .

    4. M ooRE, M . E., A .T U N K A R D A N DL. S R O L E . Obe-

    sity, social class and mental illness. J. Am. M ed.A ssoc. 181: 138, 1962.

    5 . S ILV E R S TON E , J . T. Obesity and social class. Psy-chother. Psychosom. 18: 226, 1970.

    6. S ILV E R S TON E , J. T ., R. P.GORDON ANDA. SWN-K A R D . Social factors in obesity in London. Practi-tioner. 202: 682, 1969.

    7. STUNKARD, A., E.D AQU IL I , S. Fox A N D R. D . L .FILI0N . Influence of social class on obesity and

    thinness in children. J. Am. M ed. Assoc. 221:

    5 7 9 , 1 9 7 2 .

    8. W H IT E LAW , A. G. L. Theassociation of socialclass and sibling number with skinfold thickness inLondon schoolboys. Human Biol. 43: 414, 1971.

    9. G A R N , S. M ., A N D D. C. C L A R K . Economics andfatness. Ecol. Food Nutr. 3: 19, 1974.

    10. GARN, S. M .,A N D D. C. C L A R K . The effect of

    socioeconomic status on growth, size, fatness andproportions. In: Anthropology of the United

    States, edited by W . C. Sturtevant, W ashington,

    D. C .: Anthropological Society of W ashington,1977.

    1 1 . G A R N , S. M ., A N D D. C.C L A R K . Trends in fatnessand the origins of obesity. Pediatrics 57: 443,1976.

    12. NA P I E R , J. A ., B. C.JO H N SO N A N D F. H. Ep-STEIN. The Tecumseh community health study. In:Case Book of Community Study, edited by I. I .

    K essler and M . L. Levin. Baltimore: Johns Hop-kins Press, pp. 25-44, 1970.

    13 . Committee on Nutritional Anthropology. Recom-mcndations concerning body measurements for

    the characterization of nutritional status. Human

    Biol. 28: 115, 1956.14. G A R N , S. M ., A N D R. V . H A R P E R . F at accum ula-

    tion and weight gain in the adult male. Human

    Biol. 27: 39, 1955.

    1 5 . B R O K , J . Quantitative description of body com-position: Physical anthropologys fourth dimen-

    sion. Current Anthrop. 4: 3, 1963.16. GARN , S. M .The measurement of obesity. Ecol.

    Food Nutr. 1: 333, 1972.