8
Ready, Willing, and Able: A Conceptualization of Transitions to New Behavioral Forms

RON LESTHAEGHE AND CAMILLE VANDERHOEFT

INTRODUCTION

In this paper we shall try to present a simple mathematical model for describing the adaptation to new forms of behavior and for studying the subsequent generalization of these forms among populations. Such transitions obviously involve processes of innovation and diffusion. In this conceptualization we shall use basic concepts that correspond to three preconditions for the adaptation to a new mode of behavior. These three preconditions are readiness, willingness and ability. This formulation is taken directly from Coale (1973), who grouped the preconditions for a fertility transition under these headings. To the best of our knowledge, this simple conceptualization has not received any further attention in the 25 years following its introduction.

The notion of readiness refers to the fact that the new forms of behavior must be advantageous to the actor; that is, their utility must be evident and outweigh their disutility. As such, the condition of readiness refers to the microeconomic cost-benefit calculus that actors utilize in decision processes.

The notion of willingness refers to considerations of legitimacy and normative (e.g., ethical, religious) acceptability of the new pattern of ac-

Ron Lesthaeghe and Camilla Vanderhoeft are professors in the department of social research at Vrije University, Brussels. The authors extend a special note of gratitude to Jan Mariën at Vrije University for further mathematical research on the location of the minimum of three beta-distributions, and Martin Vaessen at Macro International for making available several tabulations from the raw data.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives 8 Ready, Willing, and Able: A Conceptualization of Transitions to New Behavioral Forms RON LESTHAEGHE AND CAMILLE VANDERHOEFT INTRODUCTION In this paper we shall try to present a simple mathematical model for describing the adaptation to new forms of behavior and for studying the subsequent generalization of these forms among populations. Such transitions obviously involve processes of innovation and diffusion. In this conceptualization we shall use basic concepts that correspond to three preconditions for the adaptation to a new mode of behavior. These three preconditions are readiness, willingness and ability. This formulation is taken directly from Coale (1973), who grouped the preconditions for a fertility transition under these headings. To the best of our knowledge, this simple conceptualization has not received any further attention in the 25 years following its introduction. The notion of readiness refers to the fact that the new forms of behavior must be advantageous to the actor; that is, their utility must be evident and outweigh their disutility. As such, the condition of readiness refers to the microeconomic cost-benefit calculus that actors utilize in decision processes. The notion of willingness refers to considerations of legitimacy and normative (e.g., ethical, religious) acceptability of the new pattern of ac- Ron Lesthaeghe and Camilla Vanderhoeft are professors in the department of social research at Vrije University, Brussels. The authors extend a special note of gratitude to Jan Mariën at Vrije University for further mathematical research on the location of the minimum of three beta-distributions, and Martin Vaessen at Macro International for making available several tabulations from the raw data.

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives tion. Such an evaluation occurs against the backdrop of internalized normative structures existing in societies at any point in time. The basic question addressed by willingness is to what extent new forms of behavior run counter to established traditional beliefs and codes of conduct, and to what extent there is a willingness to overcome moral objections and fears. The adoption of new forms of behavior may also depend on the availability of new techniques. The notion of ability then refers to the accessibility of these innovations. Also, this access may have a cost that reduces ability, even if it is merely psychological. Obviously this third precondition disappears when the issue of accessibility to new facilitating factors does not arise. The conceptual model built around “ready, willing and able” (R,W,A for short) may have many applications in a variety of fields. In general, the R and W conditions arise in all matters that have both an economic and a moral dimension. The use of the R,W,A preconditions also has the advantage of creating links between the various social science disciplines, and particularly between economics concentrating on the R condition, and the other social sciences that pay more attention to normative and cultural aspects, that is, to the W condition. The present conceptualization is therefore also meant as an overarching framework for the integration of hitherto segregated “narratives” existing in the various social science disciplines (compare van de Kaa, 1996; Burch, 1996; Lesthaeghe, 1997). Finally, the model will also attempt to build bridges to the literature dealing with processes of diffusion or contagion and with social learning (self-initiated) and social influence (other initiated) (compare Montgomery and Casterline, 1996). The structure of the paper is as follows. First we shall revisit the R,W,A preconditions and their use in various narratives of the fertility transition. After all, this was the empirical field where this general formulation was initiated. Then we shall present transitions as a function of changing distributions of R, W, and A. Here we shall adopt three beta-distributions and define the outcome variable S as the minimum of the R, W, and A scores. If success (S) with respect to the adoption of a new form of behavior is dependent on meeting the three preconditions jointly, that is, S=R⋂W⋂A and if R, W, and A are distributions on a zero to unity scale, then for an individual i: Si=Min(Ri, Wi, Ai) which means that the weakest link (the smallest of the three scores) will determine the outcome.

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives In the next section, we shall link the shape of the beta-distributions to the Montgomery-Casterline formulation of social learning and social influence, thereby introducing outside influences in the decision process and degrees of heterogeneity within a population with respect to all three preconditions. In the last section, we return to a demographic application by relating the R,W,A concepts to actual data taken from the Demographic and Health Service (DHS) surveys in African countries. The purpose here is to establish where the bottleneck conditions are located. R,W,A AND FERTILITY TRANSITION NARRATIVES As indicated in the introduction, the RWA preconditions were introduced in 1973 by Coale in an article that attempted to summarize the findings of the Princeton European Fertility Transition Project (EFT). Coale clearly meant that the onset and the speed of European fertility transitions was contingent on the joint meeting of the three preconditions, that is, S=R⋂W⋂A. But just like in the “nature-nurture” debate in psychology, the findings of the EFT project were quickly converted by others into a “culture versus economics” debate despite the fact that R⋂W specifies a “culture and economics” model. This misinterpretation continues today. In this paper we consider the “economics versus culture” formulation as a dead-end street (see Lesthaeghe, 1997), and we shall not devote any more time to it. Rather, we shall give a short overview of the “sub-narratives” attached to R, W, and A. First, the R precondition has been extensively discussed and conceptually modeled in the economic literature dealing with demographic outcome variables. All schools of thought in the microeconomics of the family give a great weight to the classic cost-benefit calculus. The starting point is simple: the essence of the model is the presumption that families would balance utilities against disutilities ascribed to the nth child to determine whether a family wanted this child (Liebenstein, 1957). The neoclassic formulation that followed introduced the assumptions of fixed preferences, maximizing behavior and equilibrium solutions. In 1960, Becker introduced the concept of a household production function. The demand for children depends on the utility (economic, social, and psychological) of offspring to the parents and on the costs of children (i.e., costs of parental time, labor, and external inputs). Caldwell’s “wealth flow reversal” (1982) equally states that a fertility decline starts when the “wealth flow” over a lifetime from children to parents changes into a “wealth flow” in the opposite direction. So far, and this holds for Easterlin’s, Caldwell’s, and the early neoclassical versions, the parental decisions are based solely on the parental

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives interests. The much older theory of “social capillarity,” formulated by Dumont in 1880, introduced the welfare of the children themselves and altruistic behavior of parents in favor of the children’s future well-being. In Dumont’s conceptualization all individuals aspire to upward social mobility, but when the parents cannot achieve this for themselves, they project this ambition onto their children and invest in children’s health and especially education. This is an early formulation of Becker’s “dynastic multi-generational model” that introduces a preference shift in favor of “higher quality” children. From this a “quantity-quality swap” is derived. In Dumont’s version, industrialization, urbanization, and economic growth opened up new opportunities for the incoming generations and higher real wages allowed parents to invest more in the education of a smaller set of children, thereby maximizing the social mobility chances of their offspring. It is clear that in this version bequests and investments are added to parental time, labor, and external inputs. In Easterlin’s version extra attention is being paid to several other crucial factors. First, a corrective response can also be generated by an increase in the supply of children. Such an increase can stem from a variety of factors, such as declining infant and childhood mortality (increasing the supply of surviving children), reduced birth spacing (decreasing length of breastfeeding and postpartum abstinence), and increased fecundity. Even with a constant demand for children, an increase in the supply would produce excess fertility and generate a corrective response in the other direction. Furthermore, Easterlin and Crimmins (1985) pay considerable attention to factors associated with the costs of fertility regulation, which, in our framework, fall under the ability precondition. He also emphasizes that the key variables are reflecting the subjective perceptions and not the objective costs and benefits. The advantage of economic theories dealing with the R condition has been their conceptual richness and the predilection for formal specifications. The disadvantages are related to the facts that (i) many concepts (e.g., child utility, child quality) are multidimensional and therefore difficult to measure, (ii) the nature of motivations is very difficult to extract from respondents, and (iii) the calculation of a balance between costs and benefits is not easy for actors, let alone observers (compare Burch, 1997; Robinson, 1997). The outcome is that we have a set of theories that explain conceptually why fertility control may be advantageous, but that we are still far removed from reliable and valid measurements of the key ingredients. Incidentally, the studies that tried to measure the key concepts pertaining to child utility in a direct way, rather than through rough proxies, were fielded by social psychologists rather than by economists. The “Value of Children Project” by Fawcett, Arnold, and Bulatao (Fawcett, 1972; Fawcett and Arnold, 1995; Bulatao, 1979) is a prominent example of such attempts.

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives The W condition, by contrast, has received far less attention than the R condition. The main reason for this is that willingness is taken to follow immediately in the wake of readiness. In other words, there is no moral dilemma or “cultural lag.” This may be true in problems of firms adopting a new technology, but not in the field of fertility transitions. Much of the discussion of the W condition in narratives of fertility transition stems from the Princeton EFT project and is therefore linked to the concept of secularization, meaning the reduced credibility given to religious prescriptions. Also, the measurement of secularization in European historical settings was facilitated by the fact that secularism was often an overt element of the political-ideological dimension of social organization. This permitted operationalizations through voting behavior or through adherence to religious practices (e.g., Lesthaeghe and Wilson, 1986; Livi-Bacci, 1977; Le Bras and Todd, 1981). But the fact that the degree of secularization was readily measurable only in Western Europe does not mean that the W condition is irrelevant elsewhere. Clearly, the W condition refers to a much broader set of issues than Western-style secularization in relation to Christianity. Secular political mobilization (e.g., Nag, 1989) and growing female empowerment in developing countries (e.g., Mason, 1985), all in relation to fertility control and health, show the relevance of the W condition. First, the W condition deals with the legitimacy of interfering with nature or with a “natural order” as a cultural construction. Second, it deals with the belief in the power that individuals have to alter this natural order, and hence W depends, inter alia, on dimensions such as fatalism. Third, the W condition depends on the degree of internalization of traditional beliefs and codes of conduct. And fourth, W depends on the seventy of sanctions (even imaginary ones such as those stemming from avenging spirits) attached to transgressions of normative prescriptions. Much of this is not only dependent on individual psychological dispositions, but equally on institutional agency and on what Delumeau describes as the “politics of culpabilization” (1983). Occasionally sociological studies conducted in other than Western countries have attempted to operationalize such dimensions of “control over nature” or of “fatalism versus self-directed destiny” (e.g., Inkeless and Smith, 1974, is a classic in this field), but these batteries of questions have never found their way into the large-scale demographic surveys (e.g., World Fertility Survey, DHS). Generally speaking, the broader context of the W condition has remained inadequately documented in the areas of fertility or health transitions. The A condition has again received ample attention, predominantly in the family planning literature. In fact, the precursor of the World Fertility Survey (WFS) has been the series of knowledge, attitude, and practice (KAP) surveys dealing with the assessment of knowledge, atti-

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives tudes, and practice of contraception in developing countries. These studies were predominantly designed to show that there was a knowledge gap, that is, it was essentially the lack of knowledge about contraception and the lack of accessibility to reliable contraception that formed the bottleneck. Others argued vividly that a lack of motivation constituted the weakest link. Stronger still, if there was no “reversal of the wealth flow,” family planning efforts would run against the interests of large segments of populations of developing nations. In short, we had a clear debate about the relative locations of the W and A distributions. Also, national politics in many countries got involved in both local and worldwide debates on the feasibility of promoting ability, and the United Nations (World Population Conferences, United Nations Fund for Population Activities) assumed a leading role in promoting the legitimacy (W) and the accessibility (A) of family planning. More recently, academic interest in the issue of ability has taken the forms of studies in diffusion mechanisms and models (e.g., Rosero-Bixby and Casterline, 1993; Montgomery and Casterline, 1996). Several of these ideas will be used in this paper as well. To sum up, the R and A conditions for fertility transitions are covered extensively by the literature, but the W precondition in Coale’s formulation has been given much less attention. The various dimensions involved in cultural change in developing countries need to be given greater priority. R,W,A DISTRIBUTIONS AND THE WEAKEST LINK MODEL In the following section, we assume that the degree of fertility control (S) is an outcome variable with a continuous intensity ranging from 0 to 1. This outcome variable is, as in Coale’s original verbal formulation, dependent on three preconditions, R, W, and A, as shown in the Boolean expression: S=R⋂W⋂A that is, all three conditions must be met jointly. However, for S to be a continuous variable, we must also assume that R, W, and A are continuous and comprised between 0 and 1. In this new formulation, a score of 0 for R would mean that limiting fertility would have 0 advantages and only entail disadvantages. A score of 0.5 would typify the situation where advantages and disadvantages are in perfect balance, and obviously a score of unity would mean there are only advantages in adopting the new strategy. Similarly, a score of 0 on W means that fertility control is ethically or religiously totally unacceptable, a score of 0.5 identifies the point of indecision, and a score of unity implies that there are no moral or

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives cultural obstructions to adopting the new form of behavior. Finally, a score of 0 on A means that the individual has no means whatsoever to control fertility, a score of 0.5 implies that there would only be ineffective traditional methods, and a score of unity corresponds to complete ability to regulate fertility. An index of contraceptive use efficiency would be equally appropriate. In this model one could convert these scores into a dichotomy (controller/no controller) if the score on the outcome variable is larger than a given cutting point, say 0.5. For each individual in a population, a score is available on all three preconditions (Ri, Wi, and Ai). In the weakest link model, the outcome score for that individual, i.e., Si, is the smallest value of the three, Ri, Wi, or Ai. Hence: Si=Min (Ri, Wi, Ai). This means, for instance, that precondition A would be the bottleneck if Ai is the lowest score: the individual could be highly ready and willing, but has few means of controlling fertility (e.g., only abstinence). This principle is readily generalizable to entire populations. In this instance we deal with three distributions for R, W, and A, respectively, and the weakest link rule gives the distributions of the outcome variable S as S=Min(R, W, A). These distributions need a particular shape. Here we have opted for a beta-distribution, because this distribution is contained between 0 and 1 and because it has the feature of bell-shaped distributions if its mean is 0.5 and if the variance is small. If the mean is lower than 0.5, the distribution is positively skewed, and if it is larger than 0.5, the distribution is negatively skewed. In Figure 8–1, we have produced three such beta-distributions, respectively with means=.1667(var=.0106), .5(var=.0357), and .7778(var=.0173). The distribution to the left in Figure 8–1 would show the population distribution at the onset of an R, W, or A transition. The vast majority would see little economic advantage in controlling fertility, or would largely be unwilling or unable to do so. However, there would already be an upper tail of “innovators” for whom R, W, or A would come closer to the 0.5 mark or even surpass it. Halfway during the transition of the three preconditions, the distribution would assume a classic bell shape and half the population would be located beyond the 0.5 cutting point. Finally, near the end of the transition, only the lower tail of the skewed distribution would drag behind the majority of the population. Such a general movement of the distribution from left to right in Figure 8–1 seems an attractive representation of a general transition because it does accommodate the features of “early initiators” and “late joiners.” As indicated, our problem consists of finding the distribution of the minimum of Ri, Wi, and Ai. Assuming stochastic independence between the

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives FIGURE 8–1 Shift over time of the beta distribution of the intensity of either R, W, or A from low (less than 0.5) to high (greater than 0.5). random variables R, W, and A (subscripts are dropped to simplify the notations), the distribution of S=Min(R,W,A) can easily be obtained from the following probabilistic statement (which holds for any s between 0 and 1): which, in terms of the cumulative distribution functions of R, W, and A, also can be written as: Differentiating with respect to s gives the following expression for the probability density function (pdf) of S: Using the interpretation of a random variable’s density in s as the probability that the random variable takes the value s, this formula becomes

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives intuitively appealing and clear: the probability that the minimum S assumes the value s is the probability that one of the three underlying variables assumes that value s, while the other two have at least that value s. Moreover, if, for fixed s, both 1–FW(s) and 1–FA(s), for example, are large (i.e., close to 1), then fs(s) is close to fR(s). Thus, if two of the underlying random variables (e.g., W and A) are heavily right skewed, then the distribution of S is close to that of the third random variable (e.g., R). We used the above formula to calculate and draw the pdf of S in Figures 8–2 to 8–4, which will be discussed hereafter. Notice that although R, W, and A are assumed to be beta-distributed, S generally will not be beta-distributed. An explicit formula for the pdf of S, however, is not our concern here, and would not even be useful for our purposes, as it involves incomplete beta functions (which are to be evaluated by numerical integration). This can also be understood intuitively. In Figure 8–2 we have reproduced the same three beta-distributions as those of Figure 8–1. Assume that the left-hand distribution now represents the individuals’ scores for one of the preconditions, say W, and that the other two are representing R and A. From the “weakest link” rule S=Min (R, W, A), it follows that the FIGURE 8–2 Location of W (left), R (middle), and A (right) at one point in time (example) and location of the distribution of the minimum (Ri, Wi, Ai) (=dotted line).

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives outcome for S would closely resemble the distribution of the weakest link, that is, of W. In fact, an overwhelming majority of individuals have scores Wi that would be the lowest of the three, and only for a few persons, mostly located at the upper tail of W, one would find scores of Ri and Ai that could be smaller than their Wi. Hence, the distribution of S must always be slightly to the left of the distribution of the weakest link condition (here W). Hence, the upper tail of W will be pulled in, S would have a slightly higher peak than W, and consequently the mean of S must be smaller than the mean of W. Similarly, the variance of S also will be reduced compared to the variance of W. As expected, the calculation of the S-distribution (see dotted line on Figure 8–2) shows exactly these features. Two other examples will bring this out in a more striking way. In Figure 8–3 we have plotted (full lines) three beta-distributions with the same mean (=0.25) but a different variance (respectively .0208, .0144, and .0110). The distribution of their minimum (dotted line) has a mean of 0.14 only, and also a variance that is much smaller, that is, .0052. This example further illustrates that each of the three beta-distributions for R, W, and A FIGURE 8–3 Location of W (left), R (middle), and A (right) at one point in time (second example) and location of the distribution of the minimum (Ri, Wi, Ai) (= dotted line).

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives may have an upper tail (innovators) larger than the “indecision” -cutting point of 0.5, but that such an upper tail for the distribution of the minima would be virtually nonexistant. In the third example (Figure 8–4) we present a situation in which the distributions again have different means (0.4, 0.5, and 0.7) and different variances (.04, .0278, and .0191). Suppose we are dealing with a situation in which the vast majority of the population is already quite ready to control fertility (right-hand distribution), that willingness is following in the wake of readiness (middle distribution), but that availability and accessibility to efficient contraception would be lagging (left-hand distribution). In this instance, the distribution of the minima of scores (dotted line) would typically be situated further to the left than the distribution of the weakest link and a much smaller proportion would have Si scores greater than 0.5 than in any of the other three distributions. In this example the mean of the S distribution is only 0.33 and the variance is again smaller than that of the weakest link distribution (.022 compared to .04). (In the next section we shall see that Figure 8–4 very closely resembles the situation found in Niger.) This section has illustrated the rules of the game. We shall now take up the issue of diffusion and shifting distributions. FIGURE 8–4 Location of W (left), R (middle), and A (right) at one point in time (third example) and location of the distribution of the minimum (Ri, Wi, Ai) (=dotted line).

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives R,W,A AS SEEN THROUGH THE AFRICAN DHS SURVEYS A conceptual model should also derive some credibility from an application. In this section we shall try to locate the proportions of women of reproductive age in African countries in eight categories, ranging from obviously ready, willing and able (RWA) to none of these three (rwa). In this application the conditions are seen as discrete, that is, satisfied or not, and this will be denoted by upper case or lower case letters. The following eight categories can obviously be established: RWA RWa RwA Rwa rWA rWa rwA rwa We shall apply this classification to all women who are currently married, fecund, and exposed to risk of becoming pregnant (i.e., excluding those who are amenorrheic, or already pregnant). Among such women, those who are current users of contraception plainly fall in category 1, RWA. The others are nonusers and must be distributed over the remaining seven slots. Those among them who are nonusers in order to conceive (“want another child soon”) are obviously members of categories 5 through 8, and have the attribute r, meaning not ready to delay the next pregnancy. Those nonusers who want to delay the next birth or to avoid it altogether are ready to control, but do not do so, either because they are not willing and/or not able. They must therefore belong to categories 2, 3, or 4. The three-way classification can now be abbreviated as follows: *RWA: current users *r..: nonusers who want their next pregnancy soon *R-RWA: all other nonusers who want to delay the next pregnancy (2+years) or avoid it altogether (i.e., RWa+RwA+Rwa). Such a three-way classification can be obtained from the African DHS surveys for all currently married, fecund, and exposed women; the results are presented in Table 8–1. Before going into the details of this table, we shall first establish a link with the theoretical distributions presented in Figure 8–4. Suppose that the beta-distributions of Figure 8–4 would represent, from left to right, the distributions of willingness (mean=0.4), readiness

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives TABLE 8–1 Distribution of Currently Married, Fecund, and Exposed Women According to Their Planning Status of the Next Birth; African DHS surveys Nonusers (Proportion) DHS Country and Date N of Women Pregnancy Wanted r.. Next Pregnancy to be Delayed (2+years) or No More Wanted R-RWA Users (proportion) RWA CAR 1994–1995 2,306 .62 .22 .16 Niger 1992 1,840 .52 .34 .14 Mali 1995–1996 4,160 .46 .42 .12 Uganda 1995 2,382 .46 .28 .26 Benin 1996 2,041 .44 .29 .27 Nigeria 1990 2,478 .39 .45 .16 Senegal 1992–1993 1,722 .39 .42 .19 N.Sudan 1989–90 2,187 .39 .40 .21 Tanzania 1991–1992 2,543 .35 .40 .25 Cameroon 1991 1,337 .35 .31 .34 Zambia 1992 2,006 .34 .32 .34 Burkina F. 1993 2,338 .33 .49 .18 Zimbabwe 1994 2,331 .28 .13 .59 Madagascar 1992 1,727 .25 .39 .36 Malawi 1992 1,471 .24 .45 .31 Namibia 1992 1,308 .24 .26 .50 Rwanda 1992 1,627 .21 .30 .49 Ghana 1993 1,502 .16 .41 .43 Morocco 1992 3,129 .14 .20 .66 Kenya 1993 2,657 .12 .31 .57 Egypt 1992 6,370 .11 .21 .68 NOTE: Exposed=not amenorrheic or pregnant; also women reporting not having sex, infrequent sex, menopausal/hysterectomy, subfecund and infecund or in postpartum abstinence are eliminated from N. SOURCES: Adapted from (before 1994): computed from Westoff & Bankole (1995) table 4.1 p. 5; (from 1994–1996) computed from special output prepared by Macro International, personal communication, Dr. M.Vaessen. (mean=0.5), and ability (mean=0.7). Assume, furthermore, that we use a cutting point value of 0.5 for dichotomizing these distributions. The proportions in each of the eight discrete categories are then: RWA: 0.142 rWA: 0.142 RWa: 0.014 rWA: 0.014 RwA: 0.313 rWA: 0.313 Rwa: 0.031 rWA: 0.031 R..: 0.500 r..: 0.500

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives Given that the mean for readiness in this example has been set at 0.5, the population would obviously be split equally over the R.. and r.. slots. Furthermore, since willingness is defined as the weakest link, each of these halves must contain much smaller proportions satisfying W than A. Using the three-way classification adopted in Table 8–1 for real population, the above example corresponding to Figure 8–4 would yield the following outcomes: RWA: 0.142 r..: 0.500 R-RWA: 0.358 This can be compared to the values observed for Niger in 1992 (see Table 8–1): RWA: 0.140 r..: 0.520 R-RWA: 0.340 Hence, Figure 8–4 can be taken as a fairly close representation of the Niger situation. Roughly half the population of married, fecund, and exposed women would not be ready to postpone or avoid the next pregnancy (r=0.520) at any rate, and of the other half, more than two-thirds (R-RWA=0.340) would either be unwilling, unable, or both. The bottleneck condition is, furthermore, especially a lack of willingness (left-hand distribution on Figure 8–4), and hence we would expect that ethical or religious objections, health fears and beliefs, or social pressure from others would be the key factors in pulling the S curve for Niger to the left, thereby preventing a contraceptive breakthrough. In their study of “unmet need,” Westoff and Bankole (1995, nr. 4.1:5) present a table that allows us to establish this first three-way division for many other African countries. Those classified as RwA, Rwa, or Rwa in this paper differ from the Westoff-Bankole women with unmet need in a number of ways. First, our denominator only contains exposed women, whereas theirs also includes currently pregnant or amenorrheic women. Second, our numerator only contains the nonusers with a desire to postpone or avoid the next pregnancy, whereas theirs also uses the nonexposed women who report a mistimed or unwanted previous birth. The classification we adopt has the advantage of concentrating exclusively on the next birth (which we need conceptually to assess R or r), but it has the disadvantage of excluding substantial numbers of women who are pregnant or amenorrheic. Information on their future intentions rather than past experience would have helped. The proportions that we derive for Rwa+RWa+RwA are often larger than the figures derived by Westoff and Bankole for unmet need, not only because of the smaller denominators used in our computation, but also because we suspect that the num-

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives ber of mistimed or unwanted last or current pregnancies is likely to be underreported in African populations. In other words, we suspect that Westoff-Bankole unmet need is underestimated (which, in fact, makes their argument for countries with large unmet need even more powerful). The other distinction is that Westoff and Bankole imply, by virtue of the label “unmet need” (we assume: need for family planning), that the bottleneck condition is nonability (a). In our conceptualization, the bottleneck can equally be nonwillingness (w) or nonwillingness and nonability jointly (wa). Finally, a short note on the calculations is required. The results in Table 8–1 stem from the Westoff-Bankole table for all DHS surveys prior to 1994. The percentages were recalculated by eliminating the infecund women and the pregnant or amenorrheic women from the Ns used in the Westoff and Bankole table. For DHS surveys with dates 1994 or later, the results were obtained from special tabulations provided by Macro International starting from the raw data tapes. In these tables, although produced for fecund, married, and exposed women, a number of respondents still give reasons for not using contraception pertaining to not being married, having no or infrequent sex, being infecund or subfecund, or having reached menopause. These women were also eliminated from the analysis. We can now turn to Table 8–1. The outcomes for Morocco and Egypt were added to Table 8–1 for comparison. In our logic we start with a first dichotomy pertaining to readiness, that is, to r.. or R.. Two countries have more than half the population of married, fecund, and exposed women who are not ready to postpone or avoid the next pregnancy (r..): Niger and the CAR (Central African Republic). Another three have proportions for r.. in excess of 40 percent: Mali, Uganda, and Benin (see column 2). However, Uganda and Benin must have distributions of W and A that have shifted further to the right than in the other three countries, because their values of R,W,A are already larger than 0.25. The next group of countries has values for r.. of between 30 and 39 percent, indicating that a larger part of the R distribution has moved to the right. This group contains Senegal, Nigeria, Burkina Faso, Northern Sudan, Cameroon, Tanzania, and Zambia. But, in addition, Cameroon and Zambia have significantly higher proportions in R,W,A, meaning that they must have more favorable locations of the W and A distributions as well. In the third group of countries, the subpopulation with r.. is already smaller than 30 percent; some, such as Ghana and Kenya, have proportions lower than 20 percent, which is already typical for Northern Africa. Yet, in this group, the A or W distributions seem to act as a stronger brake in Malawi or Madagascar, since proportions in R,W,A are still below 40

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives percent. To a lesser extent, this also holds for Ghana, especially when compared to Kenya, Rwanda, Namibia, and Zimbabwe with proportions in R,W,A close to or in excess of 50 percent. The analysis conducted so far illustrates that the planning status of the next birth already sheds some light on the approximate locations of the R, W, and A distributions. The three-way classification can, however, be refined a bit further for women falling in the R-RWA category (column 3 in Table 8–1) because more information is available that helps to clarify the respective roles of W and A. The DHS surveys of the late 1980s probed reasons for not using contraception among married, fecund, and exposed women who also stated that they would “be unhappy to have the next pregnancy soon” or for whom “such a pregnancy would cause problems.” The results are also published in the DHS country reports for these years (chapter 4). Among the answers, some categories are indicative of infecundity or subfecundity or nonexposure, and we have eliminated such respondents from our analysis. The recalculated percentages are reproduced in Table 8–2. The DHS country reports for the 1990s either do not have such tables or do not publish them for married, fecund, and exposed women. However, Macro International could produce tabulations at our request for five surveys between 1994 and 1996 that satisfy our needs. Again, women who want to postpone the next pregnancy (2+years) but were not using contraception for reasons of infecundity or nonexposure were eliminated. These results for the later five surveys should be comparable to those published for the late 1980s, and they are reproduced in Table 8–3. In both tables we have regrouped the response categories in two large classes. First, the reasons for not using contraception, despite a manifest need for postponing or altogether avoiding a next pregnancy pertaining to a lack of knowledge about methods of contraception, a lack of knowledge about Family Planning (FP) services, difficulty of access to FP, or costs, are grouped in the category “nonability” (i.e., condition a). Reasons related to personal opposition to FP, to opposition from others, to religious objections, to fatalistic attitudes, or to fears for health are regrouped in the category “nonwillingness” (i.e., condition w). Only one response item could be specified, so that no information is available for the proportion satisfying both conditions, that is, aw. Finally, in some countries the frequencies for “other reasons” without further specification and/or the nonresponse are fairly high—sometimes in excess of 30 percent—so that extra caution is needed in interpreting the outcomes. The first question that can be addressed with this additional information is whether, for those in R-RWA, the dominant bottleneck is either a or w. A ratio a/w is therefore calculated in Tables 8–2 and 8–3. For the late 1980s, the a/w ratio is larger than unity in all but three countries. The

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives TABLE 8–2 Breakdown of Reasons for Not Using Contraception Among Fecund and Exposed Women Who Want to Delay or Avoid the Next Pregnancy (Condition R), but Who Are Also Nonusers (Conditions a, w, or aw); Various DHS Sub-Saharan Countries in the Late 1980s   Mali 1987 Senegal 1986 Togo 1988 Liberia 1986 Ghana 1988 Burundi 1987 Uganda 1988–89 Kenya 1989 Zimbabwe 1988 Botswana 1988 A.Bottleneck=nonability (a) N=835 264 610 331 786 486 1388 1818 400 697 —Lack of information 48.3 30.3 38.9 11.8 32.1 39.7 37.6 25.8 8.0 6.5 —Access difficult 2.3 1.1 2.6 12.7 2.5 3.3 9.9 13.9 23.2 0.1 —Too expensive Na Na 4.4 15.1 2.8 2.3 1.8 2.2 4.5 12.9 Total 50.6 31.4 45.9 39.6 37.4 45.3 49.3 41.9 35.7 19.5 B.Bottleneck=nonwillingness (w) —Religion opposed 10.1 20.4 5.4 2.7 4.4 1.0 22.0 5.7 5.8 1.4 —Others opposed, social control 2.0 Na Na Na 0.9 0.2 0.6 0.9 1.2 3.9 —Husband opposed 12.7 8.0 Na 9.4 5.1 4.3 4.3 11.2 11.3 8.0 —Opposition to family planning Na Na 14.3 6.3 4.8 4.8 5.5 4.2 6.2 18.2 —Fatalistic Na Na Na Na 0.6 3.3 0.9 1.4 1.8 0.9 —Inconvenient, bad for health 4.6 8.0 15.7 20.8 15.0 4.3 10.5 11.4 20.3 20.2 Total 29.4 36.4 35.4 39.2 30.8 17.9 43.8 34.8 46.6 52.6 Ratio a/w 1.72 .86 1.30 1.01 1.21 2.53 1.13 1.20 .77 .37 C.Bottleneck-not specified —Other reason 15.0 21.6 17.9 21.1 17.6 28.3 5.8 16.0 13.8 3.7 —Don’t know 4.8 10.6 Na Na 13.4 8.2 Na 6.2 3.5 23.0 —No answer 0.2 Na 0.8 Na 0.8 0.2 0.9 1.0 0.5 1.1 Total 20.0 32.2 18.7 21.1 31.8 36.7 6.7 23.2 17.8 27.8 NOTE: Excluded from the calculations are: breastfeeding or amenorrheic women, women with “infrequent sex” and, for Togo, Senegal, and Mali, also women who want a birth soon (in the other countries, such women were already eliminated from the published analysis). Na=response category not used in published table. SOURCE: Adapted from DHS individual country report.

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives TABLE 8–3 Breakdown of Reasons for Not Currently Using Contraception Among Fecund and Exposed Married Women Who Have Indicated That They Want to Postpone (2+years) or Avoid the Next Pregnancy (R-RWA); Selected DHS Countries in the 1990s   Mali 1995–96 Benin 1996 Centr. Afr. Rep. 1994–95 Uganda 1995 Zimbabwe 1994 A.Nonability (a) —Lack of information 42.5 47.4 36.1 41.2 3.7 —Access difficult 0.6 0.4 0.3 2.2 3.9 —Too expensive 1.0 3.0 0.4 4.0 1.3 Total a 44.0 50.8 36.8 47.4 8.9 B.Nonwillingness (w) —Religion opposed 2.6 3.3 5.4 2.9 10.6 —Husband opposed 4.8 4.8 6.4 13.9 12.4 —Others opposed 0.2 0.2 0.2 0.2 0.7 —Opposition to FP 10.8 18.1 10.6 4.2 14.3 —Health fears 19.2 14.5 8.1 19.9 45.5 —Inconvenient to use 2.6 0.8 Na 1.4 1.6 Total w 40.2 41.7 30.7 42.5 85.1 C.Not specified —Other reasons 2.2 4.5 6.0 8.2 3.3 —Don’t know 13.5 3.0 1.4 1.8 2.6 —No answer Na Na 25.2 Na Na Total unspecified 15.7 7.5 32.6 10.0 5.9 a/w ratio 1.10 1.22 1.20 1.12 0.10 NOTE: Na=response category not used in published table. SOURCE: Adapted from DHS data files; personal communication, Dr. M.Vaessen, Macro International. first of these is Senegal, but unfortunately this is a case with more than 30 percent of unspecified or missing answers. The other two are Zimbabwe (a/w=.77) and Botswana (.37), which are countries with high proportions in RWA and low proportions in R-RWA. By 1994 the a/w ratio for Zimbabwe (see Table 8–2) further declines to only 0.10, and in Mali, the ratio diminishes from 1.72 in 1987 to 1.10 in 1995–1996. This suggests that a/w ratios decline when proportions of users (RWA) increase. In such circumstances, the bottleneck condition at the onset would be primarily the A distribution, which is logical for most of Sub-Saharan Africa given the lower knowledge levels and the much weaker FP organization during the 1980s. But when overall need for contraception increases over time, that is, when the R distribution shifts to the right, the W distribution rather than the A distribution increasingly becomes the weakest link. Hence, one can expect for the future that the reasons for not using contraception among those with a spacing or stopping need will increasingly be

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives associated with nonwillingness rather than nonability, as was already the case in Botswana and Zimbabwe in the 1980s. This does not imply that the W distribution remains static—in fact, it too shifts to the right—but that in the course of the transition the distributions for R and A are moving faster. At this later stage, despite greater willingness than before, willingness becomes the bottleneck condition. Finally, Tables 8–2 and 8–3 also lend more support to the hypothesis that reasons for nonwillingness may be increasingly associated with health fears (bad for health, side effects, inconvenient to use) rather than with social opposition to fertility control in general. The items concerning health fears already had the highest frequencies in the 1980s in Togo, Ghana, Liberia, Kenya, Zimbabwe, and Botswana (Table 8–2) and in the 1990s in Mali, Uganda, and Zimbabwe (Table 8–3). Moreover, these items were more frequently cited in the 1990s than in the late 1980s in the three countries for which we have two observations: an increase in Mali from 4.6 to 21.8 percent, in Uganda from 10.5 to 21.3 percent, and in Zimbabwe from 20.2 to 47.1 percent. Admittedly, the evidence from Tables 8–2 and 8–3 is not yet conclusive and needs to be checked out for more countries with at least two observations tabulated for the R-RWA subpopulations. But it does at least advance two new hypotheses: The take-over hypothesis: As the three distributions for R, W, and A shift to the right, the A distribution is likely to move faster than the W distribution, leading to a situation in which increasing willingness still becomes the bottleneck condition. The shifting objections hypothesis: As the W distribution moves to the right, nonwillingness becomes increasingly associated with beliefs about the health impact of contraception and less with general ethical, religious, or social opposition. Nevertheless, country-specific features associated with differences in culture, social organization, and FP program implementation are likely to exert their influence as well. CONCLUSIONS The reintroduction of the triple concepts of readiness R, willingness W, and ability A in social demography has a set of advantages: It allows us to integrate economic and noneconomic paradigms of transitions to new forms of behavior, a crucial requirement for the study of fertility transitions in particular.

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives It avoids dead-end streets such as the “economics versus culture” debate. It sharpens awareness of the fact that transitions can take many forms depending on the shapes of the R, W, and A distributions and the speed at which they move. The model presented here hinges on the weakest link principle, that is, it is the minimum of either R, W, or A that determines the final speed of the adoption of fertility regulation (either for spacing or stopping). Such a bottleneck model elucidates the role of leads and lags and recognizes that, during the course of a transition, different factors may be responsible for slower change or for barrier effects in diffusion. With respect to the latter effects, models should not only be constructed with respect to diffusion of contraceptive knowledge and availability (i.e., ability), but equally pay attention to the diffusion of readiness or perceptions of economic advantage and of willingness or perceptions of cultural, social, and psychological obstacles. The R, W, A model further allows for the detection of bottleneck conditions. The application to the data from African DHS surveys illustrates that a simple three-way classification of the fecund and exposed population according to the planning status of the next pregnancy can already shed light on the approximate locations of the R, W, and A distributions in each of the countries concerned. As such, the application is a variant of the “unmet need” concept, but it fully recognizes that such unmet needs also can be associated with a lack of willingness, and not solely with a lack of ability. The heterogeneity of the sub-Saharan populations with respect to the planning status of the next birth (i.e., the distribution over the categories r.., R-RWA, and RWA) testifies to this effect. This heterogeneity indicates that factors associated with low readiness and ability tend to be responsible for the bottleneck at the onset, but that the willingness condition is likely to become the weakest link at a later stage. In other words, as the distributions of R, W, and A move to the right, the shift in the W distribution may be slower than that of the other two. In such circumstances policies become necessary that confront cultural, social, and psychological barriers to the use of contraception, in addition to policies that further facilitate access to FP. Finally, a closer inspection of the reasons given for not using contraception among fecund and exposed women who manifestly want to delay or avoid their next pregnancy (i.e., those in the R-RWA category) reveals that a shift may be occurring in the nature of nonwillingness. More specifically, as the distribution of W also shifts to the right, the remaining obstacles seem to be increasingly associated with health-related fears rather than with more general ethical, religious, or social objections. This equally implies that

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives public messages related to FP should be increasingly attentive to such fears, particularly in countries that have to pull in the tail of “late joiners.” Further research starting from Coale’s three preconditions can easily be imagined. First, locations and shapes of beta-distributions for the R, W, and A conditions can easily be constructed, and the location of S determined. As was done for the case of Niger, the actual proportions in the categories RWA, r.., and R-RWA can be obtained from the DHS data, and these can be compared to a set of model situations to infer the approximate locations of R, W, and A distributions. Second, the DHS data on reasons for not using contraception among the R-RWA subpopulation of fecund and exposed women should be produced systematically and in a comparable fashion. To estimate Rwa, the questionnaire should also allow for the specification of multiple reasons rather than just one. The breakdown of the fecund and exposed female population in the categories r.., RWA, RwA, RWa, and Rwa would further facilitate the estimation of the location of the R, W, and A distribution in each country and their subgroups, thereby shedding more light on the prevailing weakest link at various points in time. REFERENCES Bulatao, R. 1979 On the Nature of the Transition in the Value of Children. EWPI Paper 60-A. Honolulu: East-West Center. Burch, T.K. 1996 Icons, straw men and precision: Reflections on demographic theories and fertility decline. The Sociological Quarterly 37(1):59–81. 1997 Fertility Decline: Toward a Synthetic Model. Unpublished paper for the International Conference on Computer Simulation and the Social Sciences, September, Cortona, Italy. Caldwell, J.C. 1982 Theory of Fertility Decline. New York: Academic Press. Coale, A.J. 1973 The demographic transition reconsidered. In IUSSP—Proceedings of the International Population Conference. Liège, Belgium: Eds. Ordina. Delumeau, J. 1983 Le Péché et la Peur—La Culpabilisation en Occident. Paris: Fayard. Dumont, A. 1880 Dépopulation et Civilisation—Etude Démographique. Paris: Eds. Lecrosnier et Babé. Easterlin, R., and E.Crimmins 1985 The Fertility Revolution: A Supply-Demand Analysis. Chicago: University of Chicago Press. Fawcett, J., ed. 1972 The Satisfactions and Costs of Children: Theories, Concepts and Methods. Honolulu: East-West Population Institute.

OCR for page 240
Diffusion Processes and Fertility Transition: Selected Perspectives Fawcett, J., and F.Arnold 1975 The value of children: Theory and method. Representative Research in Psychology 4(1):23–26. Inkeless, A., and D.H.Smith 1974 Becoming Modern: Individual Change in Six Developing Countries. Cambridge, MA: Harvard University Press. Kohler, H.-P. 1997 Fertility and Social Interaction—An Economic Approach. Unpublished Ph.D. thesis, University of California, Berkeley. Lave, C, and J.G.March 1975 An Introduction to Models in the Social Sciences. New York: Harper & Row. Le Bras, H., and E.Todd 1981 L’Invention de la France. Paris: Hachette. Lesthaeghe, R. 1997 Imre Lakatos’ views on theory development: Applications to the field of fertility theories. In IPD-Working Papers 97–1. Brussels: Vrije Universiteit. Lesthaeghe, R., and C.Wilson 1986 Modes of production, secularization and the pace of the fertility decline in Western Europe, 1879–1930. Pp. 261–292 in The Decline of Fertility in Europe, A.J. Coale and S.C.Watkins, eds. Princeton: Princeton University Press. Liebenstein, H. 1957 Economic Backwardness and Economic Growth. New York: Wiley & Sons. Livi-Bacci, M. 1977 A History of Italian Fertility During the Last Two Centuries. Princeton: Princeton University Press. Mason, K.O. 1985 The Status of Women. New York: Rockefeller Foundation. Montgomery, M., and J.Casterline 1996 Social learning, social influence, and new models of fertility. Population and Development Review 22(supplement): 151–175. Nag, M. 1989 Political awareness as a factor in accessibility of health services. Economic and Political Weekly (Bombay) 24(8):417–426. Robinson, W.C. 1997 The economic theory of fertility over three decades. Population Studies 51:63–74. Rosero-Bixby, L., and J.Casterline 1993 Modeling diffusion effects in fertility transition. Population Studies 47:147–167. van de Kaa, D.J. 1996 Anchored narratives: The story and findings of half a century of research into the determinants of fertility. Population Studies 50:389–432. Westoff, C., and A.Bankole 1995 Unmet need: 1990–1994. In Demographic & Health Surveys, Comparative Studies no. 16. Calverton, MD: Macro International Inc.