ABORIGINAL MARRIAGE AND SURVIVAL Author(s): Alan Gray Source: Journal of the Australian Population Association, Vol. 1, No. 1 (September 1984), pp. 18-30 Published by: Springer Stable URL: http://www.jstor.org/stable/41110670 . Accessed: 15/06/2014 16:56 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. . Springer is collaborating with JSTOR to digitize, preserve and extend access to Journal of the Australian Population Association. http://www.jstor.org This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PM All use subject to JSTOR Terms and Conditions
Transcript
ABORIGINAL MARRIAGE AND SURVIVALAuthor(s): Alan GraySource: Journal of the Australian Population Association, Vol. 1, No. 1 (September 1984), pp.18-30Published by: SpringerStable URL: http://www.jstor.org/stable/41110670 .
Accessed: 15/06/2014 16:56
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp
.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].
.
Springer is collaborating with JSTOR to digitize, preserve and extend access to Journal of the AustralianPopulation Association.
http://www.jstor.org
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
Recent studies of the demography of the Aboriginal population of Australia have demonstrated a recent steep decline in fertility (Gray, 1983), coincident rapid decline in infant mortality (Smith, 1980; Thomson, 1983), continuing very high levels of adult mortality (Smith and others, 1983; Gray, 1983: 95-128) and an overall slowing of Aboriginal population growth from the very high levels of the 1950s and 1960s (Gray and Smith, 1983). The changes have dramatically altered the age profile of the Aboriginal population already, although Smith and Gray (forthcoming) now argue that echo effects of high fertility of the 1960s may bolster short-run growth rates as the large birth cohorts of that period move through the peak child-bearing age groups.
It is important to emphasize that Aboriginal fertility decline is not merely a reflection of coincident decline in the fertility of Australian women in general. On the contrary, it appears that there are few similarities between the patterns of Aboriginal and non-Aboriginal fertility change.
For instance, all non-Aboriginal age-specific birth rates fell more or less together and by similar proportions in the 1970s, while for Aboriginal women there was little change in age-specific birth rates below age 25 but very steep decline above age 30. In 1966-71 fertility above age 30 accounted for 38 per cent of the Aboriginal total fertility rate of 5.9, while in 1976-81 it accounted for only 17 per cent of the total fertility rate of 3*3. The age pattern of Aboriginal fertility was very different to that of non-Aboriginal women and remains different after the decline. And the absolute average level of Aboriginal fertility was and is higher than that of nonAboriginal women*
Continuing high fertility of young Aboriginal women has resisted the influences which had such a dramatic effect on their fertility at higher ages. Attitudinal data from five Aboriginal communities in different parts of Australia (Gray, 1983: 217-39) show that Aboriginal women overwhelmingly disapprove of the use of family planning methods to delay the birth of a woman's first child, although they may approve of its use for spacing children or for stopping having children.
The reason that high levels of child-bearing at low ages have remained characteristic of Aboriginal fertility patterns is that child- bearing plays an essential part in Aboriginal family formation, rather than being just incidental or consequential to family formation (ibid.: 92).
(1) The views expressed in this paper are those of the author and not necessarily of the Australian Bureau of Statistics.
18
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
Analysis of the attitudes of Aboriginal women towards family size reveals that both reasons for wanting more children and reasons for not wanting more children have strong economic foundations (ibid.: 284-311) . The economic advantages of having children and having no more children, moreover, now actually link together to promote small family size. There are three aspects to this. The first is that although care in their old age remains a very important benefit that Aboriginal women hope to gain from having more children, the quality of this care can be expected to depend on the future economic status of their children, which is now much more dependent on education. The second is that the economic benefit of assistance with housework, in which Aboriginal women place some store, is eroded by longer education of children and this erosion is reinforced by lower demand for this kind of assistance in households with smaller numbers of children. The third is that recognition of the costs of caring for children, substantially affected by education costs, tends to be associated with low parity among those women who want no more children.
The role played by change in educational opportunities, costs and benefits in this analysis is a central one: because of changes in education, the micro-economic costs and benefits of children shifted in favour of small family size and provided motivation for the acceptance by Aboriginal women of family planning services, through concurrently-expanding health services in Aboriginal communities.
This synopsis has left important questions unanswered. What is the economic significance of a method of family formation in which child-bearing is an essential component of the process? What changes in the micro-economies of Aboriginal households and families are occurring now, and what is their economic significance? Even to address these questions requires a much greater understanding of Aboriginal marriage and family formation than has been attempted so far in demographic studies. This paper is an attempt to lay part of this foundation.
The paper considers data on Aboriginal "marital status" as measured by Australian population censuses, in conjunction with adult mortality estimates. The aim is to model the lives of Aboriginal marriages and the attrition factors contributing to their ending.
Modelling the Dynamics of Aboriginal Marriage
A static description of Aboriginal marriage can be had from population census data. Doubts about the meaning of these data have resulted in avoidance of extended analysis of marital status in recent studies of Aboriginal demography. One reason that has been given for this avoidance (Gray, 1983: 69) has been the supposition that ABS imputation of marital status in cases of non-response could have substantially distorted the pattern of non-response in the Aboriginal population, if there was a difference between Aboriginal and non- Aboriginal marriage customs. However, unpublished 1981 Census data show that while marital status was imputed for 11 per cent of Aboriginal males and 10 per cent of females, most of the non- response (78 per cent) was confined to the 0-14 age groups: for no age group
19
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
above age 20 of either sex, nor for any marital status other than "never married11, was the incidence of non-response greater than 5 per cent. A proportionate redistribution of imputed marital statuses corresponds very closely with the imputed distribution. There are therefore sound reasons for believing that the 1981 Census data are an accurate reflection of how Aboriginal people describe their marital statuses.
Legal solemnized marriage is a step undertaken by few of the Aboriginal couples in conjugal relationships anywhere in Australia (Barwick, 1963, 1974; Kitaoji, 1976; Gray, 1983; 62-7). Kitaoji points out that the forms of sexual union constitute a logically ordered sequence: casual liaison and pregnancy often precede the first nuptial union, which may or may not be permanent, and consensual union usually precedes legal union if the marriage is legalized at all. It can be shown (Gray, 1983: 67-72) that the census data refer basically to people who have been married either consensually or legally, with the provision that measurement has been inconsistent between censuses.
Table 1 shows the age distribution of Aboriginal marital status as measured by the 1981 Australia Census.
Several aspects of this distribution require comment. The first comment is about the apparent lateness of marriage. The "never married" category is the largest for Aboriginal males below age group 30-34, and for females below age group 25-29. The proportions of 25-29 year-olds never married, at 49 per cent for males and 36 per cent for females, are larger than in the Australian population as a whole (35 per cent for males and 19 per cent for females in 1981). And while in the total population the age-specific proportion described as "now married" reaches a high point, for each sex, of almost 83 per cent in the 40-44 age group in 1981, for Aboriginals the highest levels are 63 per cent for males aged 40-44 and 61 per cent for women aged 35-39.
Interpretation of apparent late marriage, and relatively low proportions described as married, must be qualified by the observation that for Aboriginals marriage is less an event that a state of existence reached at an unidentified point in the sequential process of family formation. The fact that many Aboriginal people describe themselves as "never married" does not necessarily mean that they may not already be in a permanent sexual union that they will later describe as marriage, and may in some cases reflect uncertainty about whether the census question applies to their marriages or only to church and registry marriages.
A second point about Table 1 concerns the discrepancy between the numbers of Aboriginal men and women described as currently married. What needs to be realized here is that a proportion of the marriages described in the table are between Aboriginals and non- Aboriginals, so that there is no theoretical reason that the figures for the two sexes should balance even approximately.
A third point concerns the substantial numbers described as separated or divorced. It is commonplace in Aboriginal communities to find that, after breakdown of consensual marriages, the former
20
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
partners will describe themselves as separated. Legal separation or divorce is rare (Gale, 1969; Barwick, 1974) precisely because formalization of marriage is itself uncommon and a step usually taken only after the partners have been long married already. The choice of the term "divorced" by a substantial number of Aboriginal respondents whose marriages have broken down may reflect an attempt to select the most appropriate of several inappropriate categories of marital status (non- Aboriginal style) to describe an Aboriginal marital status.
A further observation about Table 1 is that it contains strong evidence of a marriage squeeze for unmarried Aboriginal women above age 30. Age group by age group, there are excesses of Aboriginal women not currently married over men not currently married, and these excesses become more substantial with increasing age.
The table also contains information about the lack of sensitivity of Aboriginal fertility levels to exposure to risk of pregnancy measured in terms of marital status. There would appear to be little relationship between these variables and probably less now than in the past. Child-bearing by Aboriginal women begins early and now also ends early but, as has been seen, the state described by Aboriginal respondents as marriage begins relatively late. It is not difficult to show, using the marital status distribution in Table 1 and a suitable model life table, that at age 15 Aboriginal women can expect to spend only fourteen of the next thirty-five years of life in the married state. In the ten peak years of child-bearing between ages 15 and 25, they will on average spend only two years describing themselves as married. While it is possible to conclude that exposure to risk of pregnancy measured by marital status is an almost irrelevant concept in analysis of Aboriginal fertility, this conclusion must still be placed in the context of sequential family formation and is not a measure of dysfunction in Aboriginal family life.
By far the most outstanding feature of Table 1, however» is the large size of the category "widowed11 - especially at lower ages. For example, Aboriginal women account for only 1.2 per cent of the age groups 15-34 of Australian women in 1981, but Aboriginal widows in these age groups account for 5.1 per cent of total widows. At all ages below 60 and for both sexes, the widowed marital status category is relatively several times larger for Aboriginals than for non- Aboriginals. By age group 40-44, more than 10 per cent of Aboriginal women describe themselves as widowed (not ever widowed, but widowed and not remarried) and by age group 55-59 this proportion rises to one-third. In the latter age group, 10 per cent of Aboriginal men also describe themselves as widowed.
These high proportions of widowhood are a reflection of the very high adult mortality rates which afflict the Aboriginal population. It is obvious that very high adult mortality is a significant consideration in determining the likelihood of survival of Aboriginal marriages.
The dynamic forces which have shaped the marital status distribution of Table 1 can theoretically be disentangled from this distribution under certain assumptions. The theory is an analogue of
22
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
stable population analysis, in that if one assumes that the probabilities of change from one marital status to another have remained fixed over time, then the distribution of marital status will itself be an invariate of each age. To be really rigorous about this, account would need to be taken of the existence of a marriage squeeze and other influences on probabilities in a non-stable population, but essentially this kind of model can be proposed for any non-stable population structure, and the underlying probabilities can be estimated accordingly* Of course, a marriage squeeze is not of itself inconsistent with a stable state, although a temporary squeeze would be. The main considerations are to arrive at suitable specifications for a model of marital status change, and to be reasonably confident that the model has been historically close to constant .
The last part of this, the assumption of past fixed probabilities of change of marital status, is impossible to justify empirically for the Aboriginal population. It is also by no means as unreasonable an assumption as it would be for the non-Aboriginal Australian population. I have already referred to the strength of the Aboriginal institution of family formation and its resistance to external influences as important issues in explaining the pattern of Aboriginal fertility change. In a sense, moreover, the model to be discussed is not particularly sensitive to the stability assumptions, since estimates of change will be restricted to ages only five years apart.
These matters, and the details of specification and estimation for the selected model of marital status change, are discussed fully in a separate technical paper (Gray, in preparation). The main features of the model are as follows. First, it is assumed that between ages five years apart all Aboriginals of a given sex are subject to the same probabilities (where applicable) of marriage, separation, divorce and widowhood. The marriage probability is assumed to apply to all persons whose marital status is not "now married", the separation probability only to the status "now married", the divorce probability to those who are either "now married" or "separated11, and the widowhood probability to any person who is "now married", "separated" or "divorced". In addition, it is assumed that there is a probability of reclassification of marital status to "never married" from the other marital statuses, except "now married". The existence of this kind of reclassification is indicated by slight increases in the proportion described as never married between some of the successive higher age groups in Table 1, and although there are other plausible explanations for this, it is in any case an expected phenomenon: Aboriginal people do sometimes describe themselves as never married, meaning never legally married, when they are no longer in a marital relationship. Finally, the model also allows for some variation in the mortality probabilities for different marital statuses, within an overall fixed probability of death between successive ages of each sex. (The fixed mortality schedule is from Gray (1983, Appendix 3), with infant mortality rate parameter of 30 per thousand live births) .
Variation in mortality probabilities according to marital status is a much -ignored characteristic of populations. It is rarely given more than a passing reference (Spiegelman. 1955: 61; Cox, 1970: 139-41) along the lines that higher probabilities of death for never-married
23
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
people reflect congenital or chronic health problems. The alternative social explanation, that the state of marriage promotes Wellness, is mentioned in these passing references but deserves more thorough investigation. Recent studies (Vaupel and others, 1979; Mantón and others, 1981) go in quite another direction: they explain heterogeneity of mortality experience within populations as the result of a level of "frailty" with which each individual is born and which is carried through life. It is, however, not difficult to calculate from Australian* data that death rates are several times lower for married persons, at most ages, than for those who are divorced or widowed as well as for those who have never been married. Does frailty cause divorce? Does it cause one's husband or wife to die? While the relationship between marital status and mortality is worthy of much greater research attention, it was nevertheless found in testing different mortality assumptions for the model of Aboriginal marital status change that a different level of mortality for never-married Aboriginals was more consistent with the data than any other assumption. Again, details are given in the separate technical paper.
Estimates of the main marital status change probabilities of the model are shown in Table 2.
TABLE 2 Estimated probabilities of changes of aarltal status over 5-year period* Aboriginal sen and voaen
* Age Marriage Separation Divorce Widowhood Redassi f I cat Ion
These estimates, it must be emphasized, are artefacts of the model and the assumptions underlying it. Nevertheless, it was found that changes to the model's specifications always resulted in only minor variations to these basic parameters* The results in Table 2 can therefore be taken to be close to true values. In this way, they may be used with some confidence to describe the "lives" of Aboriginal marriages*
Survival of Aboriginal Marriages
The probabilities of marital status change given in Table 2 may be used in conjunction with mortality probabilities to calculate the changes that will occur to a cohort of Aboriginal men or women as they pass through life* An estimate of the effect of high adult mortality on the survival of partners to Aboriginal marriages will be seen to be one of the meaningful outcomes of this analysis* The results will also show that marital breakdown is easily the largest factor contributing to attrition of Aboriginal marriages during the first dozen years after the partners come to describe themselves as married .
A summary of life cycle events that will occur to cohorts of one thousand Aboriginal men or women is shown in Table 3*
TABLE 3 Life cycle «vents In cohorts of Aboriginal sen and fjoe^B mm- ejotfel ledi
Events Holes Females
Births (standardized size of cohort) 1000 1000
Deaths: Before age 19 46 46 While ne^or married, ages 194 284 206 While married 369 196 While no longer married 301 992
First marriages: Number 669 747
Ending In separation/divorce 227 349 Ending In death of self 314 132 Ending In death of spouse 129 270
All marriages: Number 771 1001
Ending In separation/divorce 248 413 Ending In death of self 369 196 Ending In death of spouse 193 392
The table shows that, given an infant mortality rate of 30 per thousand live births in the assumed model life table, about one third of Aboriginal males and one quarter of Aboriginal females die before settling into the married state. Most of these deaths occur at ages 15 or higher. Similar proportions (37 per cent of males and 20 per cent of females) die while married, leaving more than half of Aboriginal women to die while no longer married. Given low life expectancy, 56 years at birth for females by the model life table, this finding is a reflection of the observable reality of large numbers of women who are household heads rearing families in Aboriginal communities. It is also suggestive that this reality entails some risk of orphanhood for Aboriginal children. (A more detailed analysis would show that 25 of the cohort of 1000 would die while no longer married at ages below 40. It can be inferred that recent decreases in the age of termination of child-bearing are probably associated with significant reduction in the risks of orphanhood for Aboriginal children.)
It has been mentioned (in discussion of Table 1) that disparity between numbers of male and female marriages can be accounted for by intermarriage between Aboriginals and non -Aboriginals . The same kind of disparity can be seen in Table 3, not only in the total number of marriages into which cohorts of Aboriginal men and women enter, but in first marriages as well. It can also be calculated that males who ever marry have an average of 1.15 marriages each, while females who ever marry on average do so 1.34 times each. These ratios reflect the different rates at which men and women survive their first marriages: 47 per cent of men who marry for the first time die while still in that marriage, compared with only 18 per cent of women. It is already possible to see in this pattern that the marriage squeeze for Aboriginal women over age 39 is attributable to high mortality of adult Aboriginal men.
The line of analysis in Table 3 is carried a little further in Table 4, which shows the contribution of various sources of attrition of first Aboriginal marriages by duration. The results suggest that, at marriage, the expected duration of the partnership is about twenty years, slightly more for men and slightly less for women. Within the first 12.5 years of marriage, marital breakdown is easily the largest source of attrition. Within this period, 25 per cent of male marriages and 36 per cent of female marriages end in separation or divorce. Of marriages which have not broken down, the death of either partner accounts for considerable and increasing proportions of marriage attrition. By the male analysis (which produces somewhat lower proportions than the female analysis), 14 per cent of marriages which have not ended in breakdown have ended in death after 12.5 years, 31 per cent after 22.5 years, and 55 per cent after 32.5 years.
The summary of attrition and life cycle events given in Tables 3 and 4 provides insight into the extent of brute demographic pressures on Aboriginal marriages. High adult mortality reduces the security of Aboriginal marriage, underlies the high incidence of single-parent households in many Aboriginal communities and squeezes the marriage market for single Aboriginal women over the age of 30. It is therefore at the root of many of the social pressures on Aboriginal communities .
26
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
TABLE 4 Attrition of first ««rrlages; cohorts of Aboriginal «en and voaen
Time lapse Surviving Deaths of Deaths of Separations/ Expected from first marriages self spouse Divorces duration marriage (years)
Male first marriages fron birth cohort of 1000
0 years 669 -12 -3 -51 22.7 2.5 years 602 -22 -6 -72 22.6 7.5 years 501 -22 -7 -42 21.7
12.5 years 430 -25 -9 -21 19.8 17.5 years 375 -31 -11 -11 17.3 22.5 years 322 -37 -14 -6 14.8 27.5 years 264 -39 -16 -6 12.5 32.5 years 203 -37 -17 -4 10.5 37.5 years 145 -32 -16 -3 8.6 42.5 years 93 -24 -13 -3 7.2 47.5 years 53 -31 -17 -5 5.8
Female first marriages fro« birth cohort of 1000
0 years 747 -7 -13 -«5 18.2 2.5 years 641 -9 -25 -119 18.5 7.5 years 488 -8 -25 -68 18.6
12.5 years 386 -10 -26 -32 17.8 17.5 years 318 -14 -27 -14 16.0 22.5 years 264 -17 -30 -9 13.8 27.5 years 208 -18 -30 -5 11.9 32.5 years 155 -17 -29 -3 10.1 37.5 years 106 -13 -25 -3 8.5 42.5 years 66 -8 -17 -3 7.2 47.5 years 37 -11 -22 -4 5.9
Discussion
The analysis in this paper has relied heavily on the use of new and previously untried techniques to recreate the dynamics of marital status change from a static description. The results should therefore be regarded as tentative and indicative, and there are so many unknown variables estimated through the model that it would be difficult to derive any measure of accuracy for the estimates that have been presented« Nevertheless, the results accord so closely with the discernible reality in many Aboriginal communities that it is possible to be confident that the basic model and the broad pattern of the results are both sound.
27
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
It is also to be noted that the analysis has been undertaken using census data for Australia as a whole, with no attempt to disaggregate the base data into geographical divisions or areas. There are no guarantees that aggregate-level analysis will replicate the actual experience of Aboriginals in any area or location. While this is so, it should be remembered that the high adult mortality pattern which creates the distinctive pressures on survival of Aboriginal marriages appears to show little geographical variation (Gray, 1983: 114-5).
In the first section of this paper, two questions were posed, questions which can now be addressed with somewhat less temerity than at the outset. The first question concerned the economic significance of a method of family formation in which child-bearing is an essential component of the process. An hypothesis that the question should be inverted should now be considered. Possibly it is the child-bearing that has the economic significance, and that family formation is incidental to an economic necessity for Aboriginal women to bear children. It cannot be doubted that children are regarded as an economic asset by Aboriginal women (ibid.: 286-91), not because of immediate benefits but because of the necessity "to have someone to help you when you're old". In five Aboriginal communities in different parts of Australia, 74 per cent of Aboriginal women who wanted more children said that this reason was very important to them in their wish for more children, and in a factor analysis of reasons for wanting more children economic benefits figured in large terms. In the context of marital insecurity and a prospective single state for much of their later adult lives, the strength of this reason for child- bearing is entirely explicable«
The second question concerned change in the micro-economies of Aboriginal households and families. The effect of interaction between change in educational opportunity and fertility has been mentioned. What is also apparent from the analysis in this paper is that the micro-economies of Aboriginal households must also be considered in terms of the pressures on marital security from high adult mortality*
28
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
BARWICK, D.E. (1963) A Little More than Kin : regional affiliation and group identity among Aboriginal migrants in Melbourne. PhD thesis, Australian National University.
BARWICK, D.E. (1974) "The Aboriginal family in south-eastern Australia." In J. KRUPINSKI and A. STOLLER (eds.), The Family in Australia: social, demographic and psychological aspects: 153-67. Pergamon, Rushcutters Bay.
GALE, F. (1969) "A changing Aboriginal population." In F. GALE and G.H. LAWTON (eds), Settlement and Encounter: geographical studies presented to Sir G r en fell Price: 65-88. Oxford University Press, Melbourne.
GRAY, A. (1983) Australian Aboriginal Fertility in Decline. PhD thesis, Australian National University.
GRAY, A. (in preparation) "Dynamics from structure of marital status: method and application to Aboriginal Australians."
GRAY, A. and L.R. SMITH (1983) "The size of the Aboriginal population." Australian Aboriginal Studies 1983, 1: 2-9.
KITAOJI, Y. (1976) Family and Social Structure among Aborigines in Northern New South Wales. PhD thesis, Australian National University .
MANTÓN, K.G., E. STALLARD and J.W. VAUPEL (1981) "Methods for comparing the mortality experience of heterogeneous populations." Demography , 18 (3): 389-410.
SMITH, L.R. (1980) Aboriginal Vital Statistics: an analysis of trends . Aboriginal Health Bulletin No. 1, Commonwealth Department of Health. Australian Government Publishing Service, Canberra.
SMITH, L.R. and A. GRAY (forthcoming) "The Australian Aborigines1 demographic transition." Paper prepared for Conference of the International Union for the Scientific Study of Population, Florence, June 1985.
SMITH, L., N. THOMSON and A. GRAY (1983) Aboriginal Mortality in New South Wales Country Regions 1980-81. New South Wales Department of Health, Sydney.
SPIEGELMAN, M. (1955) Introduction to Demography. The Society of Actuaries, Chicago.
29
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
VAUPEL, J.W., K.G. MANTÓN and E. STALLARD (1979) "The impact of heterogeneity in individual frailty on the dynamics of mortality.11 Demography , 16 (3): 439-54.
30
This content downloaded from 185.44.77.128 on Sun, 15 Jun 2014 16:56:46 PMAll use subject to JSTOR Terms and Conditions
