ANNALS OF HUMAN BIOLOGY,
1992,
VOL.
19, NO. 6,
589-593
Stature, upward social mobility and the nature of statural differences between social classes T. BIELICKI a n d H . WALISZKO Institute of Anthropology, Polish Academy of Sciences
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
Received 8 April 1991; revised 3 February 1992
Summary. Stature and educational level achieved were studied in 10 groups of 19-year-old Polish men born in 1967 and examined in 1986. Each group consisted of subjects equated for (1) parental education and occupation, (2) urban-rural residence and (3) number of children in family. It was found that within each group subjects who were secondary school students or graduates were on average taller than their age-mates who by the time of examination had never moved beyond the level of elementary or basic trade school. This result is consistent with the long-debated hypothesis that in industrial societies upward social mobility tends to be selective with regard to body height. Theoretically, such social selection could be expected to inflate the magnitude of social-class differences in stature by adding to them a genetic component. However, a mechanism can be envisaged by which preferential recruitment of taller individuals to upper social strata might indeed be at work and yet produce no differential distribution of genotypes along the social scale.
1. Introduction A number of studies have suggested that in modern industrial societies taller individuals are for some reason more likely to advance, or less likely to regress, in socioeconomic status during their lifetime than are short individuals--advancement or regress being measured with regard to some initial condition, e.g. parental status (cf. Bielicki and Charzewski 1983 for references). The subject is interesting in its own right, but it also has important implications for a much broader area of human biology, namely for the study of social-class gradients in physical growth. Such gradients are as a rule interpreted as purely phenotypical manifestations of inequalities in living standards between the better-off and less well-off social strata. Indeed, it is this very interpretation that forms the foundation of 'epidemiological auxology': the use of growth data as sensitive instruments for measuring the extent of socioeconomic inequalities in a society (Tanner 1981, 1989, Bielicki 1986). It has, however, also been recognized that if a differential social mobility of the kind described above were at work, then the commonly observed social-class differences in stature might contain not only an environmental but also a genetic component, even in an ethically homogeneous society; these differences, therefore, might present an exaggerated, over-dramatized picture of the impact of economic conditions on children's growth (cf. Tanner 1989, Bielicki 1986). In this report new evidence is provided for the existence o f an association between tallness and upward social mobility; and the possible nature and consequences of that association are discussed.
2. Materials and methods Fot the purpose of this analysis we used part of the national random sample of 19-year-old Polish military conscripts born in 1967 and examined in 1986 (Bielicki and Waliszko 1991, Bielicki, Malina and Waliszko 1992). Out of that material totalling c a . 26,000 individuals, 10 different (non-overlapping) groups of subjects were chosen,
0301-4460/92 $3.00@ 1992Taylor & Francis Ltd.
T. Bielicki and H. Waliszko
590
Table 1. Stature (means and SD's) of 19-year old Polish conscripts from 10 socially homogeneous groups, by education of conscript.
Social group 1. Urban, parents secondary school, two children
Difference A-B
177.31 6-28 (456)
176.05 6.15 (202)
+ 1.26"
177.17 6-04 (125)
175.06 5.66 (97)
+2.14"
3. Urban, parents primary or trade school father skilled worker grandfather worker, two children
176.61 5-81 (311)
175.30 6.37 (658)
+ 1.31"*
4. Urban, parents primary or trade school father skilled worker grandfather worker, three children
176-24 6-23 (136)
173.84 6.24 (395)
+2.40**
5. Urban, parents primary or trade school father skilled worker, grandfather peasant, two children
176.58 6-62 (242)
175.30 6.13 (446)
6. Urban, parents primary or trade school father skilled worker, grandfather peasant, three children
176.21 6.06 (146)
175.41 6.36 (309)
7. Rural, parents primary or trade school, father skilled worker, grandfather peasant, two children
176.50 6.14 (98)
174.94 6"02 (331)
8. Rural, parents primary or trade school, father skilled worker, grandfather peasant, three children
175.49 6"55 (114)
173.87 5.86 (469)
+ 1.62"*
9. Rural, father primary school, father and grandfather peasant, two children
174.53 5"37 (102)
174-20 5.90 (450)
+ 0.33
10. Rural, father primary school, father and grandfather peasant, three children
175.23 5-92 (138)
173.67 5"96 (822)
2. Urban, parents secondary school, three children
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
Education of conscript Level A Level B
+ 1-28"
+ 0.80 + 1.56"
+ 1.56'*
Urban residence: cities, population 100,000-500,000; rural residence: villages. Number of children includes conscript. Sample size in parentheses. See also explanations in text.
each group h o m o g e n e o u s with regard to four criteria: (1) parental education; (2) parental (plus, for some groups, paternal g r a n d f a t h e r ' s ) occupation; (3) u r b a n - r u r a l residence a n d (4) n u m b e r of siblings. The 10 groups are defined i n table 1. They represent the m a x i m u m degree o f h o m o g e n e i t y of social b a c k g r o u n d attainable, in this material, for groups of a size j u d g e d sufficient for statistical analysis. W i t h i n each group only those subjects were t a k e n into a c c o u n t who represented either o f the two following levels of educational status: (A) conscripts who at the time o f e x a m i n a t i o n (age 18.6-19-5 years) were secondary school students or graduates (including those who have already gone to college); a n d (B) conscripts who have t e r m i n a t e d their education at the level o f p r i m a r y school or have gone to basic trade schools, i.e. who by the time of e x a m i n a t i o n have never enrolled in secondary schools. ('Basic trade school' in P o l a n d is 2 or 3 years o f a b o v e - p r i m a r y schooling in rather specialized vocations, such as 'electrician', ' a u t o m e c h a n i c ' , ' b u i l d i n g machines o p e r a t o r ' , etc.; it gives n o right to enter college, it is n o r m a l l y finished at the age of 16-17 years a n d it
Stature and upward social mobility
591
almost invariably constitutes for the individual the highest-achieved level of formal education). Subsequently, within each of the 10 socially homogeneous groups subjects of level A were compared with those of level B with regard to mean stature. The results are given in table 1.
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
3.
Results
The pattern seen in table 1 is unmistakable. Within each one o f the 10 groups considered the 19-year-olds who have gone to secondary school are on average taller than their age-mates who come from a very similar social background but who have never moved beyond the level of the basic school. Thus among 10 independent comparisons all are concordant and none discordant with the hypothesis of differential mobility, a highly significant (/7 < 0.01 on the sign test) departure from the null expectation. Moreover, in as many as eight out of the 10 groups the statural superiority of subjects o f level A over those o f level B is statistically significant. It should be added that mean ages of subjects A and B at the time of examination (as calculated over all groups) were practically identical: 18.90 and 18.96 years, respectively. 4.
Discussion and conclusions
The above results appear to be in perfect agreement with the supposition that an association exists between upward social mobility and tallness. How can such an association come about? Theoretically, three different explanations seem possible.
4.1. The hypothesis of genetic "linkage" Tallness is genetically linked with some attributes of the intellect or personality which facilitate educational achievement or enhance motivation for such achievement. The above is practically equivalent to postulating that the multiple genes affecting stature (or at least some of them) have pleiotropic effects on certain socially important psychological characteristics of the individual. This supposition, although theoretically admissible, would constitute a typical example of an ad hoc explanation; it is not supported by any evidence known to us and seems unlikely.
4.2. The "trump card" hypothesis Tallness is a phenotypic trait which in many social situations is an asset--rather than a handicap or a socially neutral character. For example, tallness in males tends to be perceived as one of the elements of physical attractiveness and hence may help to open the gates to social acceptance. Also (or alternatively) it may tend to enhance the individual's self-confidence and self-assertiveness. For either o f these two reasons tallness may act as an aid in climbing the social scale. Generally speaking, this supposition seems plausible. However, it is not readily apparent how it could be used as an explanation of the specific results reported in the present study. Decisions on which type o f above-primary school to attend are taken at an age (normally around 12-14 years) when the boy's physical growth is, in the great majority of cases, far from completed. And while it is conceivable that such a decision might, to some extent, be influenced by the degree of maturational advancement of the subject at that age, it would be harder to understand why it should be related to his stature at 19 years, especially in view of the well-known fact that adult stature does not depend upon the tempo of maturation during adolescence (e.g. Marshall and Tanner 1986).
592
T. Bielicki and H. Waliszko
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
4.3.
The "third factor" hypothesis Within each group of families homogeneous with regard to the social criteria here taken into account there remains much unrecorded variation in economic standards a n d / o r the quality of parental care; i.e. each group contains some families which are 'better' and some which are 'poorer' in these respects. The 'better' ones have, in each group, contributed sons who are on average taller (because of better nutrition a n d health care) and who also enjoyed better educational opportunities (because of greater parental encouragement and assistance). Thus, the association between body height and educational achievement, observed in young adults, is spurious: it has been brought about by an underlying 'third factor' of a purely environmental nature. As noted by Tanner (personal communication), one specific variant of this environmental interpretation may be singled out as potentially applicable to the results of the present study. It can be described in the form of the following sequence of causes and effects: lower living standards in the family---->slower maturation of son---~the delayed maturer does less well in primary school than do his maturationally more advanced age-mates--->his poorer scholastic performance prevents or discourages him from going on to secondary school; and, being a slow developer, he has not yet completed his growth in height at the age of 19 years and therefore tends to be shorter (cf. the classic dispute between Boas and Porter, as outlined in Tanner 1966). Of the three above-described possible interpretations the 'third factor' model seems to us the most likely one. It is--one should n o t e - - n o t discordant with the results of earlier studies based on analyses of pairs of same-sexed adult siblings. In those studies it was shown that men judged as 'social climbers' were on average taller than their brothers who were 'non-climbers' (Schumacher and Knussman 1979); and that in a large sample of pairs of adult brothers the proportion of pairs in which the taller sib was the better educated of the two was significantly higher than the proportion of pairs in which the opposite was the case (Bielicki and Charzewski 1983). Admittedly, between-sib comparisons may be regarded as a more rigorous test of the hypothesis of differential social mobility than the approach used in this study, where comparisons are made between groups of individuals who, although coming from the same social background, have not shared the same family environment. Yet we are of the opinion (contra Schumacher and Knussman 1979) that even sib-sib comparisons do not provide conclusive evidence that the observed statistical association between tallness and upward social mobility is due to anything else but a common environmental factor affecting, simultaneously, both the physical growth of the child and his capacity for future educational or occupational advancement (see discussion in Bielicki and Charzewski 1983, p. 407). What are the implications of each of the three models here hypothesized for the problem of social-class differences in stature? It is evident that the situation envisaged by the hypothesis of genetic 'linkage' would have genetic consequences: with such selection at work statural differences between upper and lower social strata would not be purely phenotypic in nature, even in an ethnically homogeneous society. The same is true for the 'trump card' model. Admittedly, in this case it is the phenotypically tall who would be favoured, irrespective of his genes. But amongst tall individuals carriers of genes predisposing to tallness occur more frequently than they do among the shorter ones. Consequently, it is genotypes, not only phenotypes, that would become unequally distributed along the social scale. In contrast, the mechanism suggested by the 'third factor' model is 'blind to
Stature and upward social mobility
593
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
genes': it would have no effect whatsoever on the distribution o f genotypes in the population. Yet the fact remains that even in this case the process of social mobility envisaged is differential: in each generation the better educated strata are continually reinforced 'from below' by young upwardly mobile individuals who, as adults, will tend to become phenotypically taller than their non-mobile age-mates coming from the same social class. Recruitment o f this kind, although random with regard to genes for body height, is nevertheless selective with regard to adult height: it is the ones w h o end up as tall adults that are more likely to move up on the social scale! This fact should not be lost to sight when interpreting the significance o f statural differences between social classes.
References BIELICKI,T., 1986, Physical growth as a measure of the economic well-being of populations. In Human Growth, edited by F. Falkner and J. M. Tanner, voI. 3 (New York: P l e n u m Press), 2nd edition, pp. 283-305.
BIELICKI, Z., and CHARZEWSKI,J., 1983, Body height and upward social mobility. Annals o f Human Biology, 10, 403-408. BIELICKI, T., and WALISZKO, A . , 1991, Urbanization-dependent gradients in stature a m o n g Polish conscripts in 1976 and 1986. American Journal o f Human Biology, 3, 419-424. BIELICKI, Z., MALINA,R. M., and WALISZKO,A., 1992, Monitoring the dynamics of social stratification: statural variation a m o n g Polish conscripts in 1976 and 1986. American Journal o f Human Biology (In press). MARSHALL, W. A., and TANNER, J. M., 1986, Puberty. In: Human Growth, edited by F. Falkner and J. M. Tanner, vol. 2 (New York: Plenum Press). SCHUMACHER, A., and KNUSSMAN, R., 1979. Are differences in stature between social classes a modification or an assessment effect? Journal o f Evolution, 8, 809-812. TANNER,J. M., 1966, Galtonian eugenics and the study of growth. The relation of body size, intelligence test score, and social circumstances in children and adults. Eugenics Review, 58, 122-135. TANNER, J. M., 1981, A History o f the Study o f Human Growth (Cambridge: Cambridge University Press). TANNER, J. M., 1989, Foetus into Man (London: Open Books), 2nd edition. Address for correspondence: Institute o f Anthropology, Polish Academy o f Sciences, Kuznicza 35, 50-951 Wroclaw, Poland. Z u s a m m e n f a s s u n g . FOr 10 Gruppen 19j~thriger polnischer M~nner, geboren 1967 u n d untersucht 1986, wurden die KOrperhOhe und der erreichte Ausbildungsgrad untersucht. Die Gruppen wurden nach den Kriterien 1. Ausbildung und Beruf der Eltern, 2. Wohnort: Stadt-Land und 3. Zahl der Kinder in der Familie, h o m o g e n zusammengesetzt. Innerhalb jeder Gruppe wiesen Individuen mit einer hOheren Schulbildung eine durchschnittlich grOBere KOrperh6he a u f als Individuen mit geringerer Schulbildung. Diese Beobachtung 19.Bt sich mit der lang debattierten Hypothese in Einklang bringen, nach der in industriellen Gesellschaften eine aufwarts gerichtet soziale Mobilit~tt im Hinblick a u f die KOrperhOhe selektiv zu sein scheint. Daraus lieBe sich theoretisch die Erwartung ableiten, dab eine solche Selektion die Unterschiede in der K6rperh6he zwischen den sozialen Strata vergrOBert, indem sie ihnen eine genetische Komponente hinzufiagt. Es kOnnte allerdings auch ein M e c h a n i s m u s wirksam sein, der eine Rekrutierung von Individuen mit groBerer KOrperhOhe in den oberen sozialen Strata zwar favorisiert, ohne dab daraus jedoch eine unterschiedliche Verteilung von Genotypen a u f der sozialen Achse resultiert. R6sum6. La stature et le niveau scolaire atteint sont 6tudi6s parmi 10 groupes de jeunes h o m m e s polonais ag6s de 10 ans, n6s en 1967 et examin6s en 1986. C h a q u e groupe est compos6 par des sujets nivel6s pour 1. l'6ducation parentale et la profession, 2. la r6sidence urbaine-rurale et 3. le n o m b r e d'enfants dans la famille. On trouve dans chaque groupe que les individus qui ont efectu6 des 6tudes secondaires ou qui sont dipt6m~s, sont en m o y e n n e plus grands que leurs pairs qui au m o m e n t off ils ont 6t6 examin6s, n'avaient jamais d6pass6 le niveau scolaire 616mentaire ou le niveau de base d ' u n e 6cole de commerce. Ce r6sultat est c o n f o r m e / t l'hypoth6se longuement d~battue, que la mobilit6 sociale vers le haut tendrait ~t &re s61ective pour ce qui concerne la stature, dans une soci6t6 industrialis6e. En th6orie, une telle s61ection sociale pourrait augmenter la magnitude des diff6rences de stature entre classes sociales, en leur apportant un c o m p o s a n t gdndtique. Cependant, on peut envisager l'existence d ' u n m6canisme par lequel u n recrutement pr6f6rentiel d'individus plus grands par les classes sociales sup6rieures pourrait ~tre en oeuvre et cependant ne pas produire de distribution diff6rentielle des g6notypes selon l'6chelle sociale.
1992,
VOL.
19, NO. 6,
589-593
Stature, upward social mobility and the nature of statural differences between social classes T. BIELICKI a n d H . WALISZKO Institute of Anthropology, Polish Academy of Sciences
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
Received 8 April 1991; revised 3 February 1992
Summary. Stature and educational level achieved were studied in 10 groups of 19-year-old Polish men born in 1967 and examined in 1986. Each group consisted of subjects equated for (1) parental education and occupation, (2) urban-rural residence and (3) number of children in family. It was found that within each group subjects who were secondary school students or graduates were on average taller than their age-mates who by the time of examination had never moved beyond the level of elementary or basic trade school. This result is consistent with the long-debated hypothesis that in industrial societies upward social mobility tends to be selective with regard to body height. Theoretically, such social selection could be expected to inflate the magnitude of social-class differences in stature by adding to them a genetic component. However, a mechanism can be envisaged by which preferential recruitment of taller individuals to upper social strata might indeed be at work and yet produce no differential distribution of genotypes along the social scale.
1. Introduction A number of studies have suggested that in modern industrial societies taller individuals are for some reason more likely to advance, or less likely to regress, in socioeconomic status during their lifetime than are short individuals--advancement or regress being measured with regard to some initial condition, e.g. parental status (cf. Bielicki and Charzewski 1983 for references). The subject is interesting in its own right, but it also has important implications for a much broader area of human biology, namely for the study of social-class gradients in physical growth. Such gradients are as a rule interpreted as purely phenotypical manifestations of inequalities in living standards between the better-off and less well-off social strata. Indeed, it is this very interpretation that forms the foundation of 'epidemiological auxology': the use of growth data as sensitive instruments for measuring the extent of socioeconomic inequalities in a society (Tanner 1981, 1989, Bielicki 1986). It has, however, also been recognized that if a differential social mobility of the kind described above were at work, then the commonly observed social-class differences in stature might contain not only an environmental but also a genetic component, even in an ethically homogeneous society; these differences, therefore, might present an exaggerated, over-dramatized picture of the impact of economic conditions on children's growth (cf. Tanner 1989, Bielicki 1986). In this report new evidence is provided for the existence o f an association between tallness and upward social mobility; and the possible nature and consequences of that association are discussed.
2. Materials and methods Fot the purpose of this analysis we used part of the national random sample of 19-year-old Polish military conscripts born in 1967 and examined in 1986 (Bielicki and Waliszko 1991, Bielicki, Malina and Waliszko 1992). Out of that material totalling c a . 26,000 individuals, 10 different (non-overlapping) groups of subjects were chosen,
0301-4460/92 $3.00@ 1992Taylor & Francis Ltd.
T. Bielicki and H. Waliszko
590
Table 1. Stature (means and SD's) of 19-year old Polish conscripts from 10 socially homogeneous groups, by education of conscript.
Social group 1. Urban, parents secondary school, two children
Difference A-B
177.31 6-28 (456)
176.05 6.15 (202)
+ 1.26"
177.17 6-04 (125)
175.06 5.66 (97)
+2.14"
3. Urban, parents primary or trade school father skilled worker grandfather worker, two children
176.61 5-81 (311)
175.30 6.37 (658)
+ 1.31"*
4. Urban, parents primary or trade school father skilled worker grandfather worker, three children
176-24 6-23 (136)
173.84 6.24 (395)
+2.40**
5. Urban, parents primary or trade school father skilled worker, grandfather peasant, two children
176.58 6-62 (242)
175.30 6.13 (446)
6. Urban, parents primary or trade school father skilled worker, grandfather peasant, three children
176.21 6.06 (146)
175.41 6.36 (309)
7. Rural, parents primary or trade school, father skilled worker, grandfather peasant, two children
176.50 6.14 (98)
174.94 6"02 (331)
8. Rural, parents primary or trade school, father skilled worker, grandfather peasant, three children
175.49 6"55 (114)
173.87 5.86 (469)
+ 1.62"*
9. Rural, father primary school, father and grandfather peasant, two children
174.53 5"37 (102)
174-20 5.90 (450)
+ 0.33
10. Rural, father primary school, father and grandfather peasant, three children
175.23 5-92 (138)
173.67 5"96 (822)
2. Urban, parents secondary school, three children
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
Education of conscript Level A Level B
+ 1-28"
+ 0.80 + 1.56"
+ 1.56'*
Urban residence: cities, population 100,000-500,000; rural residence: villages. Number of children includes conscript. Sample size in parentheses. See also explanations in text.
each group h o m o g e n e o u s with regard to four criteria: (1) parental education; (2) parental (plus, for some groups, paternal g r a n d f a t h e r ' s ) occupation; (3) u r b a n - r u r a l residence a n d (4) n u m b e r of siblings. The 10 groups are defined i n table 1. They represent the m a x i m u m degree o f h o m o g e n e i t y of social b a c k g r o u n d attainable, in this material, for groups of a size j u d g e d sufficient for statistical analysis. W i t h i n each group only those subjects were t a k e n into a c c o u n t who represented either o f the two following levels of educational status: (A) conscripts who at the time o f e x a m i n a t i o n (age 18.6-19-5 years) were secondary school students or graduates (including those who have already gone to college); a n d (B) conscripts who have t e r m i n a t e d their education at the level o f p r i m a r y school or have gone to basic trade schools, i.e. who by the time of e x a m i n a t i o n have never enrolled in secondary schools. ('Basic trade school' in P o l a n d is 2 or 3 years o f a b o v e - p r i m a r y schooling in rather specialized vocations, such as 'electrician', ' a u t o m e c h a n i c ' , ' b u i l d i n g machines o p e r a t o r ' , etc.; it gives n o right to enter college, it is n o r m a l l y finished at the age of 16-17 years a n d it
Stature and upward social mobility
591
almost invariably constitutes for the individual the highest-achieved level of formal education). Subsequently, within each of the 10 socially homogeneous groups subjects of level A were compared with those of level B with regard to mean stature. The results are given in table 1.
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
3.
Results
The pattern seen in table 1 is unmistakable. Within each one o f the 10 groups considered the 19-year-olds who have gone to secondary school are on average taller than their age-mates who come from a very similar social background but who have never moved beyond the level of the basic school. Thus among 10 independent comparisons all are concordant and none discordant with the hypothesis of differential mobility, a highly significant (/7 < 0.01 on the sign test) departure from the null expectation. Moreover, in as many as eight out of the 10 groups the statural superiority of subjects o f level A over those o f level B is statistically significant. It should be added that mean ages of subjects A and B at the time of examination (as calculated over all groups) were practically identical: 18.90 and 18.96 years, respectively. 4.
Discussion and conclusions
The above results appear to be in perfect agreement with the supposition that an association exists between upward social mobility and tallness. How can such an association come about? Theoretically, three different explanations seem possible.
4.1. The hypothesis of genetic "linkage" Tallness is genetically linked with some attributes of the intellect or personality which facilitate educational achievement or enhance motivation for such achievement. The above is practically equivalent to postulating that the multiple genes affecting stature (or at least some of them) have pleiotropic effects on certain socially important psychological characteristics of the individual. This supposition, although theoretically admissible, would constitute a typical example of an ad hoc explanation; it is not supported by any evidence known to us and seems unlikely.
4.2. The "trump card" hypothesis Tallness is a phenotypic trait which in many social situations is an asset--rather than a handicap or a socially neutral character. For example, tallness in males tends to be perceived as one of the elements of physical attractiveness and hence may help to open the gates to social acceptance. Also (or alternatively) it may tend to enhance the individual's self-confidence and self-assertiveness. For either o f these two reasons tallness may act as an aid in climbing the social scale. Generally speaking, this supposition seems plausible. However, it is not readily apparent how it could be used as an explanation of the specific results reported in the present study. Decisions on which type o f above-primary school to attend are taken at an age (normally around 12-14 years) when the boy's physical growth is, in the great majority of cases, far from completed. And while it is conceivable that such a decision might, to some extent, be influenced by the degree of maturational advancement of the subject at that age, it would be harder to understand why it should be related to his stature at 19 years, especially in view of the well-known fact that adult stature does not depend upon the tempo of maturation during adolescence (e.g. Marshall and Tanner 1986).
592
T. Bielicki and H. Waliszko
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
4.3.
The "third factor" hypothesis Within each group of families homogeneous with regard to the social criteria here taken into account there remains much unrecorded variation in economic standards a n d / o r the quality of parental care; i.e. each group contains some families which are 'better' and some which are 'poorer' in these respects. The 'better' ones have, in each group, contributed sons who are on average taller (because of better nutrition a n d health care) and who also enjoyed better educational opportunities (because of greater parental encouragement and assistance). Thus, the association between body height and educational achievement, observed in young adults, is spurious: it has been brought about by an underlying 'third factor' of a purely environmental nature. As noted by Tanner (personal communication), one specific variant of this environmental interpretation may be singled out as potentially applicable to the results of the present study. It can be described in the form of the following sequence of causes and effects: lower living standards in the family---->slower maturation of son---~the delayed maturer does less well in primary school than do his maturationally more advanced age-mates--->his poorer scholastic performance prevents or discourages him from going on to secondary school; and, being a slow developer, he has not yet completed his growth in height at the age of 19 years and therefore tends to be shorter (cf. the classic dispute between Boas and Porter, as outlined in Tanner 1966). Of the three above-described possible interpretations the 'third factor' model seems to us the most likely one. It is--one should n o t e - - n o t discordant with the results of earlier studies based on analyses of pairs of same-sexed adult siblings. In those studies it was shown that men judged as 'social climbers' were on average taller than their brothers who were 'non-climbers' (Schumacher and Knussman 1979); and that in a large sample of pairs of adult brothers the proportion of pairs in which the taller sib was the better educated of the two was significantly higher than the proportion of pairs in which the opposite was the case (Bielicki and Charzewski 1983). Admittedly, between-sib comparisons may be regarded as a more rigorous test of the hypothesis of differential social mobility than the approach used in this study, where comparisons are made between groups of individuals who, although coming from the same social background, have not shared the same family environment. Yet we are of the opinion (contra Schumacher and Knussman 1979) that even sib-sib comparisons do not provide conclusive evidence that the observed statistical association between tallness and upward social mobility is due to anything else but a common environmental factor affecting, simultaneously, both the physical growth of the child and his capacity for future educational or occupational advancement (see discussion in Bielicki and Charzewski 1983, p. 407). What are the implications of each of the three models here hypothesized for the problem of social-class differences in stature? It is evident that the situation envisaged by the hypothesis of genetic 'linkage' would have genetic consequences: with such selection at work statural differences between upper and lower social strata would not be purely phenotypic in nature, even in an ethnically homogeneous society. The same is true for the 'trump card' model. Admittedly, in this case it is the phenotypically tall who would be favoured, irrespective of his genes. But amongst tall individuals carriers of genes predisposing to tallness occur more frequently than they do among the shorter ones. Consequently, it is genotypes, not only phenotypes, that would become unequally distributed along the social scale. In contrast, the mechanism suggested by the 'third factor' model is 'blind to
Stature and upward social mobility
593
Ann Hum Biol Downloaded from informahealthcare.com by Chulalongkorn University on 12/27/14 For personal use only.
genes': it would have no effect whatsoever on the distribution o f genotypes in the population. Yet the fact remains that even in this case the process of social mobility envisaged is differential: in each generation the better educated strata are continually reinforced 'from below' by young upwardly mobile individuals who, as adults, will tend to become phenotypically taller than their non-mobile age-mates coming from the same social class. Recruitment o f this kind, although random with regard to genes for body height, is nevertheless selective with regard to adult height: it is the ones w h o end up as tall adults that are more likely to move up on the social scale! This fact should not be lost to sight when interpreting the significance o f statural differences between social classes.
References BIELICKI,T., 1986, Physical growth as a measure of the economic well-being of populations. In Human Growth, edited by F. Falkner and J. M. Tanner, voI. 3 (New York: P l e n u m Press), 2nd edition, pp. 283-305.
BIELICKI, Z., and CHARZEWSKI,J., 1983, Body height and upward social mobility. Annals o f Human Biology, 10, 403-408. BIELICKI, T., and WALISZKO, A . , 1991, Urbanization-dependent gradients in stature a m o n g Polish conscripts in 1976 and 1986. American Journal o f Human Biology, 3, 419-424. BIELICKI, Z., MALINA,R. M., and WALISZKO,A., 1992, Monitoring the dynamics of social stratification: statural variation a m o n g Polish conscripts in 1976 and 1986. American Journal o f Human Biology (In press). MARSHALL, W. A., and TANNER, J. M., 1986, Puberty. In: Human Growth, edited by F. Falkner and J. M. Tanner, vol. 2 (New York: Plenum Press). SCHUMACHER, A., and KNUSSMAN, R., 1979. Are differences in stature between social classes a modification or an assessment effect? Journal o f Evolution, 8, 809-812. TANNER,J. M., 1966, Galtonian eugenics and the study of growth. The relation of body size, intelligence test score, and social circumstances in children and adults. Eugenics Review, 58, 122-135. TANNER, J. M., 1981, A History o f the Study o f Human Growth (Cambridge: Cambridge University Press). TANNER, J. M., 1989, Foetus into Man (London: Open Books), 2nd edition. Address for correspondence: Institute o f Anthropology, Polish Academy o f Sciences, Kuznicza 35, 50-951 Wroclaw, Poland. Z u s a m m e n f a s s u n g . FOr 10 Gruppen 19j~thriger polnischer M~nner, geboren 1967 u n d untersucht 1986, wurden die KOrperhOhe und der erreichte Ausbildungsgrad untersucht. Die Gruppen wurden nach den Kriterien 1. Ausbildung und Beruf der Eltern, 2. Wohnort: Stadt-Land und 3. Zahl der Kinder in der Familie, h o m o g e n zusammengesetzt. Innerhalb jeder Gruppe wiesen Individuen mit einer hOheren Schulbildung eine durchschnittlich grOBere KOrperh6he a u f als Individuen mit geringerer Schulbildung. Diese Beobachtung 19.Bt sich mit der lang debattierten Hypothese in Einklang bringen, nach der in industriellen Gesellschaften eine aufwarts gerichtet soziale Mobilit~tt im Hinblick a u f die KOrperhOhe selektiv zu sein scheint. Daraus lieBe sich theoretisch die Erwartung ableiten, dab eine solche Selektion die Unterschiede in der K6rperh6he zwischen den sozialen Strata vergrOBert, indem sie ihnen eine genetische Komponente hinzufiagt. Es kOnnte allerdings auch ein M e c h a n i s m u s wirksam sein, der eine Rekrutierung von Individuen mit groBerer KOrperhOhe in den oberen sozialen Strata zwar favorisiert, ohne dab daraus jedoch eine unterschiedliche Verteilung von Genotypen a u f der sozialen Achse resultiert. R6sum6. La stature et le niveau scolaire atteint sont 6tudi6s parmi 10 groupes de jeunes h o m m e s polonais ag6s de 10 ans, n6s en 1967 et examin6s en 1986. C h a q u e groupe est compos6 par des sujets nivel6s pour 1. l'6ducation parentale et la profession, 2. la r6sidence urbaine-rurale et 3. le n o m b r e d'enfants dans la famille. On trouve dans chaque groupe que les individus qui ont efectu6 des 6tudes secondaires ou qui sont dipt6m~s, sont en m o y e n n e plus grands que leurs pairs qui au m o m e n t off ils ont 6t6 examin6s, n'avaient jamais d6pass6 le niveau scolaire 616mentaire ou le niveau de base d ' u n e 6cole de commerce. Ce r6sultat est c o n f o r m e / t l'hypoth6se longuement d~battue, que la mobilit6 sociale vers le haut tendrait ~t &re s61ective pour ce qui concerne la stature, dans une soci6t6 industrialis6e. En th6orie, une telle s61ection sociale pourrait augmenter la magnitude des diff6rences de stature entre classes sociales, en leur apportant un c o m p o s a n t gdndtique. Cependant, on peut envisager l'existence d ' u n m6canisme par lequel u n recrutement pr6f6rentiel d'individus plus grands par les classes sociales sup6rieures pourrait ~tre en oeuvre et cependant ne pas produire de distribution diff6rentielle des g6notypes selon l'6chelle sociale.
