This paper examines the reliability and stability of developmental idealism (DI) measures in Nepal. DI is a set of cultural schemas that contains beliefs and values favoring modern societies and families over traditional ones and that views modern families as causes and effects of modern societies. It also views the world as dynamic, with change from traditionality toward modernity. Earlier studies have shown that DI has been disseminated widely internationally, but provide little evidence concerning whether individual views of DI can be reliably measured or the extent to which such views are stable across time. We estimate the reliability and stability of DI measures using panel data collected in Nepal. Our results indicate substantial reliability, equal or nearly equal to the reliability of standard value and belief items measured in general American surveys. There is also considerable stability of DI views across our study interval from 2008 to 2011.

## INTRODUCTION

This paper examines the measurement reliability and stability of individual views concerning developmental idealism (DI), which is a set of cultural schemas that contains beliefs and values concerning modern societies and families. DI builds on the ideas of modernization and development, values modern societies and families over traditional ones, views modern families as causes of modern societies, and posits modern societies as causes of modern families. DI also views the world as dynamic, with change moving from traditionality to modernity.

DI is part of an increasingly influential world culture that has been spreading around the world (Krücken and Drori 2009; Meyer et al. 1997). This world culture emphasizes the importance of individualism, freedom, equality, development, science-based education, human rights, and family forms seen as modern. World culture has helped foster many international changes, including increases in school enrollment and the standardization of educational content (Baker and Letendre 2005; Chabbott 2003), the spread of laws supporting human rights (Cole 2005; Meyer, Bromley, and Ramirez 2010), the establishment and spread of family planning programs (Barrett and Frank 1999; Thornton 2001, 2005), the spread of programs for gender equality and the elimination of female circumcision (Berkovitch 1999; Boyle 2002), and changing laws about sexual behavior (Frank, Camp, and Boutcher 2010). The primary focus of this research literature has been on the spread of world culture in laws, school programs and textbooks, governmental policies and programs, and the programs of nongovernmental organizations rather than on the beliefs, values, and behavior of individuals.

Nevertheless, research has shown that the ideas of modernization have been disseminated in many places around the world (Ahearn 2001; Dahl and Rabo 1992; Deeb 2006; Ferguson 1999; Osella and Osella 2006; Pigg 1992). There is also a growing body of survey research showing that DI has been disseminated widely, with many ordinary people around the world understanding the basic ideas of development and developmental hierarchies (Binstock and Thornton 2007; Dorius 2016; Melegh et al. 2012, 2016; Thornton, Dorius, and Swindle 2015; Thornton et al. 2012a). Many people also believe that societal modernity is both a cause and an effect of family and demographic attributes (Abassi-Shavazi, Nodoushan, and Thornton 2012; Binstock and Thornton 2007; Mitchell 2009; Thornton et al. 2012b, 2014).

The DI literature has argued that the international spread of DI beliefs and values has had important influences on family and demographic behavior. Particularly relevant here are the arguments of Allendorf and Pandian (2016) that DI may have influenced marital change in India. Thornton and Xie (Thornton 2005; Thornton and Xie 2016) have made similar arguments for the influence of DI on Chinese family and demographic change—as have Yount and Rashad (2008) and Kavas and Thornton (2013) for several Middle Eastern countries. Similarly, Thornton and Philipov (2009) have suggested that the large declines in marriage and childbearing in Central and Eastern Europe after the collapse of state socialism may have been at least partially a result of the spread of DI.

There is also limited micro-level evidence that DI beliefs and values influence the behavior of individuals. This has been shown by Allendorf and Thornton (2015) using panel data from Nepal. DI measured in a baseline survey had significant influences on self-choice and inter-caste marriage occurring after the baseline.

However, further research on the influence of DI on individual behavior has been limited by the lack of information about the reliability and stability of individual DI measures. Fortunately, there is a small literature suggesting that answers to at least certain DI questions are of high quality. For example, survey respondents in Nepal answer DI questions straightforwardly and distinguish between questions that are worded in opposite directions (Thornton, Ghimire, and Mitchell 2012). Latent class analysis also shows that Nepalis can be divided reliably into groups with different levels of endorsement of DI (Mitchell 2009). Furthermore, data from Argentina, China, and Egypt suggest that when the dimensions of DI being measured are very similar, the measurement reliability is high (Thornton et al. 2016). There is also evidence from Taiwanese college students that individual responses concerning developmental hierarchies are measured with substantial reliability (Thornton and Yang 2016).

Measurement reliability is important because investigations of the causes and consequences of DI require that we have reliable measures of DI. By reliable measures, we mean measures that are closely related to what individuals actually believe and value—that is, the measures match in a reasonable way the thoughts of the individuals. Without such reliable DI measures, it is difficult to estimate the relationship between DI and other aspects of life.

There is also the issue of the stability of DI values and beliefs. Here we recognize two different dimensions of stability. One, which we label absolute stability, concerns the absolute change in an individual's values or beliefs—becoming more or less positive toward DI over some period of time. Depending on the nature of individual changes, these changes could combine in such a way that the endorsement of DI in the population increased, decreased, or stayed the same over time. The second dimension of stability, which we label relative stability, is the extent to which the positions of individuals in the population distribution of DI beliefs or values remain constant over time. If the positions of individuals stay precisely fixed relative to other individuals over time, this dimension of stability would be perfect. This could occur even though there was significant individual or aggregate change over time. It is this second dimension of stability—relative stability—that is the focus of this paper, and we investigate it empirically over a three-year period in Nepal.

The relative stability of DI beliefs and values is important because it is unlikely that DI beliefs and values would be related to other dimensions of life if the distribution of individual beliefs and values were constantly being shuffled. That is, in order for DI beliefs and values to be importantly related to other life dimensions that are not constantly changing, they must have at least some relative stability.

We know of only one previous study about the relative stability of individual DI beliefs, and that study concerns the stability of beliefs about developmental hierarchies among Taiwanese college students. Empirical evidence from a panel study of these students indicates that the beliefs these students had in the nature of the international developmental order were indeed quite stable over the course of their college careers (Thornton and Yang 2016). We know of no research estimating the relative stability of DI values and beliefs concerning the interrelations of development and family and demographic matters.

For this paper, we have multiple goals concerning the reliability and relative stability of DI survey measures. First, we investigate the reliability of several batteries of DI beliefs and values collected in a panel study in Nepal. We do so using data and methods that permit us to estimate the measurement reliability of individual survey items in ways that separate measurement reliability from conceptualization issues. Second, we compare the measurement reliability for DI values with that for DI beliefs. Third, we compare the reliability of DI measures in Nepal with the reliability of survey measures on other topics in the United States. Finally, we provide estimates of the relative stability of individual beliefs and values across time. Our data and methods allow us to estimate relative stability while both controlling for measurement reliability and avoiding issues of measurement conceptualization. We estimate both the reliability and the relative stability of several measures of DI using panel data collected in Nepal from 2008 to 2011.

We now discuss DI and why its measurement is important. We then discuss our study site in Nepal. Then, we present our data and analytical methods, results, and conclusions.

## DEVELOPMENTAL IDEALISM

DI emerged from the modernization/development model that has been powerful among social scientists and public policy makers for centuries (Nisbet 1969). Although the model of modernization has been heavily criticized in recent years as a framework for scholarship and policy, its influence continues to be far-reaching (Nisbet 1969). For centuries, the development model has placed societies along a developmental staircase, with Northwest Europe and its overseas populations identified as developed, and other countries as less developed. The model tells adherents that their world is dynamic, with change moving from traditionality toward modernity. The good life is specified to be located in Northwest Europe and its diasporas, and more advanced societies are specified as models for less developed ones.

DI is a cultural schema that draws from the developmental model of beliefs and values just described to guide the lives of individuals and communities. It tells people that the societal attributes it defines as modern are good and to be sought after. DI also places high value on modern families, which it defines to include the following attributes: autonomy of children, romantic love, self-choice marriages, marriages contracted at mature ages, individualism, nuclear families, gender equality, contraceptive use, and low fertility. In recent decades, the rise of divorce, sex outside marriage, and non-marital cohabitation and childbearing in Europe and North America have caused these behaviors to be associated with modernity, but they are often perceived negatively by elites and the general public. Traditional families are assumed to reflect the other end of the continuum from modern families. DI also includes causal statements about modern families fostering societal development and societal development bringing modern families. Also, equality and freedom are identified as fundamental human rights.

The spread of DI as a cultural schema is particularly important because ideational factors are increasingly seen as having substantial consequences for family and demographic behavior (Cherlin 1992; Lesthaeghe 1983; Pearce 2002; van de Kaa 1987; Yount and Rashad 2008). Cultural models provide frameworks for understanding the world and how it works, how people should live, and what things are good and how to achieve them (Fricke 1997; Geertz 1973; Johnson-Hanks et al. 2011; Thornton et al. 2001). Contradictory schemas often exist simultaneously across and within societies, and within individuals. Of particular importance is that for centuries the societies of the world have had their own beliefs and values about the world, appropriate goals, and appropriate means to achieve those goals. DI explicitly contradicts many aspects of these historical family systems, with clashes of culture frequently resulting. Although resistance and adaptation are often the results, the introduction of DI often brings important changes in many aspects of family life.

A growing body of research has documented the widespread dissemination of DI, but we have little information about the factors influencing its adoption by individuals. We also know little about the consequences of individuals' holding DI beliefs and values for their demographic, familial, religious, and political behavior. Addressing these important research agendas requires knowledge of the reliability and stability of the survey items measuring DI. Our goal is to acquire such knowledge for one population in the country of Nepal.

As suggested above, DI covers many aspects of personal, religious, political, and family life. Some aspects have to do with beliefs, and others have to do with values. We conceptualize beliefs as statements of “what is,” “what was,” or “what will be.” Beliefs may or may not be true, or they may be contingent—true in some circumstances and not others. For example, an individual may believe that having fewer children increases a family's income, which may be true under some circumstances but not others. However, individuals' beliefs are real in their consequences for that individual (Thomas et al. 1929).

Values, by contrast, are representations of what is good or valuable—“standards” that govern behavioral choices. Values are “abstract ideals, positive or negative, not tied to any specific attitude object or situation, representing a person's beliefs about ideal modes of conduct and ideal terminal goals” (Rokeach 1970:124). Values are important because many central human institutions such as marriage, childbearing, and childrearing have strong evaluative components (Rokeach 1970, 1973).

It is useful to note that while we have separated the concepts of values and beliefs, the boundary between them can be fuzzy, and they can be interrelated to each other in several ways. For example, a belief that something leads to a valued outcome may result in that thing also being valued. Also, people can hold conflicting values (and beliefs) at the same time. Such contradictory values and beliefs may have important effects on the reliability and relative stability of DI measures ascertained at different times and circumstances.

Although we cannot evaluate every belief and value concerning DI in a single paper, we examine the reliability and relative stability of several. These include the following three belief dimensions: that family changes are moving toward modernity; that modern society brings modern families; and that modern families help produce a modern society. We also examine values concerning family attributes defined as modern compared to family attributes defined as traditional. We do this in two ways: by asking whether individuals value a modern family trait over a traditional one, and by asking whether individuals think a family change in the direction of DI would be good or bad.

By examining these five dimensions of DI together, we can compare their reliability and stability. Our expectation is that values are measured more reliably than beliefs. We base this expectation on Alwin's (2007) extensive research in the United States, where he found that survey measures of values were, on average, more reliable than survey measures of beliefs. Our research will permit us to see whether this conclusion generally holds for the very different setting of Nepal.

We also expect that values involving more complex cognition will be measured less reliably than values involving less complex cognition. More specifically, we expect that evaluations using straightforward comparisons between two family elements will be measured more reliably than evaluations involving future family trends. The former involves no time reference beyond the present, while the latter concerns a hypothetical future.

Finally, our research permits comparisons of our measures of reliability with those obtained by Alwin (2007) using data from the United States. Although we use different questions than Alwin, which prevents strict comparisons, we can investigate whether such measures in Nepal are in the same general range as those in the US.

## DATA AND MEASURES

### Setting

Our study setting is the western Chitwan Valley, which lies in the south-central part of Nepal. Nepal was kept in relative isolation from the rest of the world until the mid-1950s, when it opened up considerably. However, beginning in the mid-1970s, Nepal has experienced dramatic social and economic change. Nepal started receiving large amounts of foreign aid from various countries, multinational and bilateral organizations, and international nongovernmental organizations. This flow of international aid was primarily targeted at improving the living conditions of the rural poor through improvements in educational systems, health services, and rural infrastructure. The result has been a dramatic spread of formal education, wage work, government services, transportation and communication infrastructure, and the mass media. Even though Nepal is still predominantly a rural country, urbanization has gained momentum in the last couple of decades.

Despite these changes, Nepal is still one of the lowest-income countries in the world—only USD 490 per person per year in 2010 (Asian Development Bank 2012). Almost a third of the population lives below the absolute poverty line. The economy has been characterized by dependence on agriculture, with over 85% of the population involved in agriculture in 2010.

Before the 1950s, the western Chitwan Valley, our study setting, was primarily covered with dense forests and was infamous for malaria. With U.S. assistance, the Nepalese government initiated a reclamation program in the valley in the 1950s by clearing the forests. Since then, the area has witnessed a rapid inflow of migrants. People were attracted by the free distribution of land for agricultural purposes at the beginning of settlement, and by the subsequent growth of amenities and services. Currently, the valley is inhabited mostly by in-migrants, especially from the Nepali hills, adjacent districts in Nepal, and from northern India.

The valley was connected to the rest of Nepal by all-weather roads in the late 1970s and experienced changes similar to those in other parts of the country (Axinn and Yabiku 2001; Shrestha 1989). Moreover, Chitwan's central location and relatively extensive transportation network have been catalytic in transforming Chitwan into a hub for business and tourism. This has resulted in a rapid proliferation of schools, health services, mass media, transportation and communication infrastructure, and government wage labor opportunities (Shivakoti et al. 1999). This spread of community-level changes has produced a reorganization of individuals' daily social lives within the lifetimes of Chitwan's current residents, with increasing participation in education and employment, exposure to mass media, and international communication and travel (Axinn and Yabiku 2001; Beutel and Axinn 2002; Ghimire and Axinn 2006), more specifically exposure to and interaction with the outside world and new ideas.

### Sample Design

The data for our analysis come from a study of 151 neighborhoods scattered throughout the western Chitwan Valley. For this study, a neighborhood was initially defined as a geographic cluster of 5 to 15 households. These neighborhoods were chosen as an equal-probability, systematic sample of neighborhoods in the western Chitwan valley, and the characteristics of this sample closely resemble the characteristics of the Chitwan Valley population (Barber et al. 1997).

The main universe of the study included all residents of the sampled neighborhoods aged 15–59 at the time of a household census in 2008. In addition to this main sample, we interviewed parents of unmarried people aged 15–34 and spouses of people aged 15–34, as long as the parents and spouses were living in Nepal. Because our sample comes from one valley, our results cannot be extrapolated to all of Nepal. The distribution of respondents on several key demographic characteristics is shown in Table 1.

TABLE 1.
Respondent characteristics (N = 5,190)
 Gender Female 58% Ethnicity Brahmin/Chhetri 45% Dalit 11% Newar 6% Hill Indigenous 17% Terai Indigenous 21% Education None 24% 1–5 years 9% 6–10 years 22% 11–14 years 27% 15+ years 18% Marital status Ever married 76% Number of children 0 29% 1–2 30% 3–4 26% 5+ 15% Age in 2008 15–24 30% 25–34 25% 35–44 20% 45+ 25%
 Gender Female 58% Ethnicity Brahmin/Chhetri 45% Dalit 11% Newar 6% Hill Indigenous 17% Terai Indigenous 21% Education None 24% 1–5 years 9% 6–10 years 22% 11–14 years 27% 15+ years 18% Marital status Ever married 76% Number of children 0 29% 1–2 30% 3–4 26% 5+ 15% Age in 2008 15–24 30% 25–34 25% 35–44 20% 45+ 25%

### Data

Our panel study was conducted from 2008 to 2011. The baseline interviews (July–December 2008) included 5,190 individuals, with a response rate of 97.1%. Of the people interviewed in the baseline, 92% were also interviewed in the subsequent interviews analyzed here. This data collection includes socioeconomic and ideational measures. Each of the ideational measures used in this paper was initially gathered in the Wave 1 baseline and in two subsequent interviews, with the time between the Wave 1 and 2 measures and between the Wave 2 and 3 measures being approximately one year.

When we started this research, there were no measures of DI, and we needed to create new measures. Several approaches were used to accomplish this. We used the ethnographic fieldwork of others (Ahearn 2001; Bista 1991; Fricke 1997; Guneratne 1998, 2001; Pigg 1992) and conducted our own ethnographic fieldwork, informal discussions, focus groups, and in-depth interviews to understand Nepali thinking on the meaning of development and family matters. We followed these investigations by designing and pretesting survey questions about DI.

We use data from several sets of questions, with each set asking about several family elements. Many of the family elements included have long been part of DI's traditional–modern continuum: arranged marriage; age at marriage; intergenerational co-residence; respect for authority; fertility; and putting family needs above one's own needs. We also asked about same-caste marriage and widow remarriage, which are part of historical Nepali culture and often placed on a traditional–modern continuum. Our final family elements are premarital sex, nonmarital cohabitation, marital dissolution, and never marrying, which have become common in Europe and North America in recent decades and are frequently seen as aspects of modernity. English translations of the Nepali questions are provided in  Appendix A.

Our first set of questions asked whether development would decrease or increase the prevalence of the family elements mentioned above, and the second set asked whether changes in family elements would make Nepal richer or poorer. The third question set asked about expectations concerning future changes in various family elements, and whether respondents thought that changes in family elements in the DI direction would be good or bad, or didn't matter. For our fourth set, we asked people to compare family elements at different ends of a continuum and say which one they preferred.

With the exception of the questions about future family changes, all of the answers were coded into dichotomies: either accepting or rejecting the DI response. The few people who said that they did not know or provided an answer that could not be coded as endorsing or rejecting the DI response were coded as not giving the DI response. For the questions about evaluations of future trends, respondents were invited to say whether that change would be good or bad, or didn't matter. To show univariate distributions, we dichotomized as DI responses those answers saying that a modern family change was good, while all other answers were coded as not DI. To estimate correlations, reliability, and stability, we used the ordered trichotomy of support for DI change ranging from bad, to indifferent, to good.

### Methods for Estimating Reliability and Stability

Most estimates of the reliability of survey measures are derived using multiple survey items that are assumed to measure the same underlying construct, with the unfortunate consequence that estimated reliability is biased downwards by violations of that assumption. To avoid this confounding of reliability and conceptualization problems, we estimated reliability in a model using data for a series of one-item variables that are each measured three times in a panel study (Alwin 1989, 2007).

Our procedure is summarized in Figure 1, where we specify that an underlying variable with a true score of η is measured by a single indicator y for the same individual at three points in time. The observed indicator y in each year is a product of the underlying variable η in the year that the indicator is measured. As shown in Equations (1), (2), and (3), the observed indicator y is linked to η, the true score, through λ, the regression coefficient, also referred to as the factor loading. The observed indicator is also affected by the unobserved error of measurement ε, which we assume to be random.

FIGURE 1.
Single-item reliability and stability model
FIGURE 1.
Single-item reliability and stability model

Reliability is the extent to which the variance of the observed indicator (y) is “true” rather than “error” variance. It is measured as the ratio of the true score's (η's) variance to the indicator's (y's) variance. The square root of this reliability is equal to λ, the factor loading. When reliability is 1, the variance in the indicator reflects only the variance of the underlying construct; there is no error variance. When reliability is 0, none of the indicator's variance reflects the underlying construct; all of it is error variance.

$y1=λ1η1+ε1$
1

$y2 =λ2η2+ε2$
2

$y3 =λ3η3+ε3$
3

Equations (4) and (5) link together the three underlying variables (also see Figure 1). These equations indicate a process where each underlying construct is assumed to be caused by the immediately preceding underlying construct and a substantive error term, ζ (Alwin 1989, 2007; Jöreskog 1970). This means that the time 1 underlying construct is assumed to affect the time 3 construct only because it influences the time 2 construct. The β coefficients in the equations are standardized regression coefficients that indicate the temporal stability of the underlying variables, where the underlying constructs have been purged of measurement error. As discussed earlier, this is the relative stability of individuals within the distribution of scores of the entire population.

$η2=β21η1+ζ2$
4

$η3=β32η2+ζ3$
5

These standardized regression coefficients (β) indicate the standard-deviation difference produced in the dependent variable by a standard-deviation difference in the predictor variable. Coefficients less than 1 indicate a regression to the mean, as individuals both high and low in the distribution tend to move toward the mean, whereas coefficients greater than 1 indicate movements away from the mean, as individuals high in the distribution tend to move even higher and individuals low in the distribution tend to move even lower.

We estimated these models by first calculating the appropriate tetrachoric and polychoric correlations among Wave 1, Wave 2, and Wave 3 variables for a particular family construct using Mplus (Muthén 1981; Muthén and Muthén 1998). For the variables measured as dichotomies, we based our estimates on tetrachoric procedures. For variables measured as ordinal trichotomies, we based our estimates on polychoric procedures.

With the correlations estimated, we estimated λ, β21, and β32 using the procedures described by Heise (1969) and Alwin (1989, 2007). This procedure required making simplifying assumptions to estimate Equations (1), (2), and (3), with an essential assumption being that measurement reliability is equal across time.

## RESULTS

### Aggregate Distributions

We begin our discussion of results with the univariate distributions of each of the variables analyzed, with the numbers in Table 2 indicating the percentage of people endorsing a particular aspect of DI. The first three sets of columns represent the level to which the respondent's beliefs reflect what is found in the DI model—that development causes family change, that family change causes development, and that DI is expected to increase in the future. The fourth and fifth sets of columns represent respondent values or preferences relative to DI, both in the future and in general. We have bolded the table entries where more than 50% endorsed the DI belief or value. We also show (in the next-to-last row) the average number of DI responses given by respondents for each set of questions. These averages are calculated for people with good data in all waves of the interviews, whereas the individual percentages are based on people having good data for just that wave.

TABLE 2.
Percentage of respondents giving developmental idealism (DI) answer, by DI dimension, family element, and wave
BeliefsValues
(1) Dev. → Family(2) Family → Dev.(3) Expect future DI(4) Prefer future DI(5) Prefer DI
W1W2W3W1W2W3W1W2W3W1W2W3W1W2W3
Non-marital cohabitation 74.8 75.1 76.7 16.3 22.2 21.4 82.5 84.0 85.8 15.9 18.2 16.4 12.6 13.5 12.5
Age of first marriage for a womana 45.7 42.2 38.3 54.8 51.6 46.7 75.2 75.4 76.5 66.8 65.7 66.3
Divorce 48.1 52.2 56.7  7.3 11.6 10.8 62.8 66.7 70.0 12.4 12.5 11.9 57.3 61.8 65.7
Married children living with parents 65.3 70.2 71.3 37.6 37.0 33.5 70.5 73.2 76.0 17.2 16.2 13.5 12.4 14.3 13.2
Respect for authority       39.9 42.8 46.1 16.5 13.5 12.6
Individual needs before family 77.3 76.1 77.1 30.6 34.9 32.4 76.3 78.7 81.1 39.4 40.1 34.8 13.0 19.4 19.3
Premarital sex 80.6 79.0 82.3  6.4 10.4 10.0 88.0 88.3 89.5  6.4  6.5  5.9  1.7  1.8  2.3
Marriages arranged by parents 86.3 86.3 84.6 50.7 48.9 42.6 87.7 86.4 87.5 37.9 32.8 29.6 70.8 74.5 74.1
Inter-caste marriage 87.6 87.5 88.7 40.9 46.3 43.6 92.4 91.9 92.6 35.4 36.5 32.6 13.3 15.6 15.6
Women never marryingb 39.5 38.4 41.1 41.6 42.2 40.7 42.8 42.2 45.6 26.4 27.2 28.1  8.5  7.5  7.3
Widow remarriage 80.1 80.7 82.8 47.0 48.2 48.5 79.4 83.6 86.1 59.1 62.5 64.1 77.8 80.5 83.8
Time from marriage to first Birth 72.4 68.3 68.9 75.8 73.0 70.6 71.9 72.0 75.1 71.1 71.2 72.1
Mean no. of developmental responses  7.58  7.56  7.68 d  4.09  4.26c  4.00d  8.70  8.86c  9.13d  4.04  4.02  3.87 d  2.67  2.89c  2.94d
N 4,628 4,631 4,631 4,628 4,631 4,631 4,652 4,656 4,655 4,652 4,656 4,655 4,652 4,656 4,656
BeliefsValues
(1) Dev. → Family(2) Family → Dev.(3) Expect future DI(4) Prefer future DI(5) Prefer DI
W1W2W3W1W2W3W1W2W3W1W2W3W1W2W3
Non-marital cohabitation 74.8 75.1 76.7 16.3 22.2 21.4 82.5 84.0 85.8 15.9 18.2 16.4 12.6 13.5 12.5
Age of first marriage for a womana 45.7 42.2 38.3 54.8 51.6 46.7 75.2 75.4 76.5 66.8 65.7 66.3
Divorce 48.1 52.2 56.7  7.3 11.6 10.8 62.8 66.7 70.0 12.4 12.5 11.9 57.3 61.8 65.7
Married children living with parents 65.3 70.2 71.3 37.6 37.0 33.5 70.5 73.2 76.0 17.2 16.2 13.5 12.4 14.3 13.2
Respect for authority       39.9 42.8 46.1 16.5 13.5 12.6
Individual needs before family 77.3 76.1 77.1 30.6 34.9 32.4 76.3 78.7 81.1 39.4 40.1 34.8 13.0 19.4 19.3
Premarital sex 80.6 79.0 82.3  6.4 10.4 10.0 88.0 88.3 89.5  6.4  6.5  5.9  1.7  1.8  2.3
Marriages arranged by parents 86.3 86.3 84.6 50.7 48.9 42.6 87.7 86.4 87.5 37.9 32.8 29.6 70.8 74.5 74.1
Inter-caste marriage 87.6 87.5 88.7 40.9 46.3 43.6 92.4 91.9 92.6 35.4 36.5 32.6 13.3 15.6 15.6
Women never marryingb 39.5 38.4 41.1 41.6 42.2 40.7 42.8 42.2 45.6 26.4 27.2 28.1  8.5  7.5  7.3
Widow remarriage 80.1 80.7 82.8 47.0 48.2 48.5 79.4 83.6 86.1 59.1 62.5 64.1 77.8 80.5 83.8
Time from marriage to first Birth 72.4 68.3 68.9 75.8 73.0 70.6 71.9 72.0 75.1 71.1 71.2 72.1
Mean no. of developmental responses  7.58  7.56  7.68 d  4.09  4.26c  4.00d  8.70  8.86c  9.13d  4.04  4.02  3.87 d  2.67  2.89c  2.94d
N 4,628 4,631 4,631 4,628 4,631 4,631 4,652 4,656 4,655 4,652 4,656 4,655 4,652 4,656 4,656
a

For dimensions 1 and 2 the question read “people” instead of “a woman.”

b

c

Mean difference in number of developmental responses between Waves 1 and 2 significant at p = .05 or better.

d

Mean difference in number of developmental responses between Waves 2 and 3 significant at p = .05 or better.

The Wave 1 data show that the endorsement of DI depends greatly on the family element being studied and on the dimension of DI being asked about. In general, the support for DI was higher for the two dimensions focused on development causing family change and for expectations for future change than for the other three dimensions (columns 1 and 3). This can be seen most clearly in the average number of DI answers given, where the means are above seven (out of 11 or 12 items) for both development causing family change and for expectations for future change, and the means for each of the other three dimensions are below five (out of 9–12 items).

More specifically, for the dimension of future expectations, more than 50% expected increases in DI in the next 20 years in all but two cases. In 9 of the 12 future questions, 70% or more expected changes in the direction of DI, and in several cases more than 80% did. The two family elements where less than 50% expected family change in the DI direction were respect for authority and never marrying.

Similar patterns existed for the dimension of development causing family change. In all but three questions more than 50% said that development would bring family change in the direction of DI. And, for 7 of the 11 family elements, more than 70% said that development would bring family change in the DI direction.

As already noted, there was less overall support for the other three dimensions of DI—saying that family change would bring development (column 2), evaluating DI change as positive (column 4), and valuing family elements in the direction of DI (column 5). For all three of these dimensions, the modal support was less than 50%. A majority said that family change in the direction of DI would bring change only in the cases of age at first marriage increasing, more self-choice in spouse selection (but just barely), and time from marriage to first birth (column 2). Endorsement of trends in the direction of DI exceeded 50% only in the cases of age at first marriage, widow remarriage, and the time from marriage to the first child (column 4). Fewer questions were asked about general values concerning DI, but the picture here is one of relatively little endorsement of DI (column 5). In only three questions out of nine—divorce, arranged marriage, and widow remarriage—was the DI option chosen over the opposite option. In each of the other six family items the percentage endorsing the option was less than 20%.

Comparing the data from Wave 1, Wave 2, and Wave 3, we see that there is considerable aggregate stability in the percentage endorsing a particular DI item across these waves. For three of the five DI dimensions—development causes family change, family change causes development, and evaluations of future family changes (columns 1, 2, and 4), the distributions are remarkably similar across the three waves. There are modest ups and downs for individual items and for the means, but these are not in a consistent direction.

However, for the questions about expectations for future trends (column 3) and for values for DI (column 5), there seems to be a modest upward trend in the percentage endorsing the DI position. This holds true for 9 of the 12 items measuring expectations for future trends and for 6 of the 9 items measuring general values for DI. It also holds true for the averages, with the increases in average endorsement between Waves 1 and 2 and between Waves 2 and 3 being statistically significant for these two dimensions of DI.

### Individual Measurement Reliability

We now shift our attention to individual measurement reliability in Table 3, where we present the reliability for each measured variable. As illustrated in Figure 1, our concern here is the extent to which the “true” scores (η) of people's DI beliefs and values (what individuals actually believe and value) are reflected in our measures (y) of these DI beliefs and values. We focus on the coefficient λ, which reflects the extent of the overlap between the true score and the observed score. The reliability reported in Table 3 equals the square of the standardized effect (λ) of each of the underlying η variables on its corresponding observed y variable.

TABLE 3.
Reliability of individual measures on latent variables, by developmental idealism dimension and family element
BeliefsValues
(1) Dev → Fam(2) Fam → Dev(3) Expect Future DI(4) Prefer Future DI(5) Prefer DI
Non-marital cohabitation 0.43 0.44 0.50 0.54 0.39
Age of first marriage for a womana 0.61 0.28 0.44 0.47
Divorce 0.50 0.33 0.60 0.54 0.76
Married children living with parents 0.35 0.32 0.48 0.43 0.66
Respect for authority   0.60 0.53
Individual needs before family 0.30 0.44 0.31 0.51 0.54
Premarital sex 0.68 0.49 0.74 0.56 0.55
Marriages arranged by parents 0.52 0.38 0.55 0.44 0.70
Inter-caste marriage 0.51 0.47 0.54 0.57 0.78
Women never marryingb 0.48 0.48 0.35 0.44 0.68
Widow remarriage 0.48 0.35 0.56 0.45 0.77
Time from marriage to first birth 0.30 0.41 0.40 0.46
Average reliability 0.47 0.40 0.51 0.50 0.65
Average reliabilityc 0.47 0.40 0.53 0.51 0.64
BeliefsValues
(1) Dev → Fam(2) Fam → Dev(3) Expect Future DI(4) Prefer Future DI(5) Prefer DI
Non-marital cohabitation 0.43 0.44 0.50 0.54 0.39
Age of first marriage for a womana 0.61 0.28 0.44 0.47
Divorce 0.50 0.33 0.60 0.54 0.76
Married children living with parents 0.35 0.32 0.48 0.43 0.66
Respect for authority   0.60 0.53
Individual needs before family 0.30 0.44 0.31 0.51 0.54
Premarital sex 0.68 0.49 0.74 0.56 0.55
Marriages arranged by parents 0.52 0.38 0.55 0.44 0.70
Inter-caste marriage 0.51 0.47 0.54 0.57 0.78
Women never marryingb 0.48 0.48 0.35 0.44 0.68
Widow remarriage 0.48 0.35 0.56 0.45 0.77
Time from marriage to first birth 0.30 0.41 0.40 0.46
Average reliability 0.47 0.40 0.51 0.50 0.65
Average reliabilityc 0.47 0.40 0.53 0.51 0.64
a

For dimensions 1 and 2 the question read “people” instead of “a woman.”

b

c

Each average calculated using only those family elements measured across all dimensions of developmental idealism.

Table 3 indicates that the average reliability is generally similar for three dimensions of DI—development causing family change, expectations for the future, and evaluations of the future (columns 1, 3, and 4). Looking first at the development-causing-family-change questions (column 1), we see that reliability ranges from 0.30 to 0.68, with an average of 0.47. For the future-expectation questions (column 3), it ranges from a low of 0.31 to a high of 0.74, with an average of 0.51. The reliability of the evaluation of future DI trends (column 4) ranges from 0.43 to 0.57, with an average of 0.50, similar to the average reliability for the future-expectations questions.

The general values for DI (column 5) tend to be more reliable than the three sets of variables just mentioned, and those for family change causing development (column 2), generally less reliable. The reliability of the general DI values ranges from 0.39 to 0.78. The mean reliability of general values is 0.65, compared to the 0.47, 0.51, and 0.50 averages for the three sets of variables just mentioned.

As already mentioned, the dimension of family change causing development (column 2) is generally less reliable than the other four dimensions of DI. Here reliability ranges from 0.28 to 0.49, with the average being 0.40.

We are not surprised by the general-value dimension being more reliable, on average, than the other four dimensions. As we noted earlier, Alwin (2007) found that in the United States measures of values were more reliable than measures of beliefs. On the basis of Alwin's U.S.-based work, we expect our general-value variables to be more reliable than the three dimensions of DI that we categorize as belief variables: development causing family change, family change causing development, and expectations about future family change.

As expected, the questions measuring the general-family-value dimension (column 5) are generally more reliable than questions measuring evaluations of future family change (column 4). We expect that this is at least partially due to the lower cognitive difficulty of evaluating value questions without a time reference compared to value questions about a hypothetical future.

Also note that there is an overall, but not consistent, tendency for the development-causing-family-change variables (column 1) to be more reliable than the family-change-causing-development variables (column 2). We conjecture that this difference may be due to the belief that development causes family change having higher salience than the belief that family change causes development. Individual Nepalis may also have stronger beliefs in development causing family change than in family change causing development, which is consistent with the relatively weak endorsement of beliefs about family change causing development.

We now turn to the question, how do these measures of reliability compare to studies of other populations? Here we compare our findings for average reliability to the averages calculated by Alwin (2007) using the same method for many questionnaire items used in American surveys. Alwin estimated the average reliability of American value questions to be 0.65 and the average reliability of American belief questions to be 0.58.

Remarkably, our average reliability of 0.65 for the Nepali general-value items is identical to Alwin's average American value reliability of 0.65. Thus, our results from a Nepali population on this set of value variables are very similar to results in the United States for other variables. This average reliability of 0.65 for both the American value questions and the Nepali questions about general values is higher than the reliability of 0.50 for the Nepali value questions asking people to evaluate future family changes. As discussed above, this is likely the result of the greater cognitive complexity of this set of Nepali questions.

As we noted earlier, we classify our questions about development causing family change, family change causing development, and expectations of the future as beliefs. Our average reliability of 0.47 for the development-causing-family-change question and 0.51 for expectations-of-future-family-change questions are only somewhat lower than the average reliability of belief measures in the United States (0.58). Our average reliability for the questions about family change causing development is, as we discussed earlier, somewhat lower, at 0.40. Again, we emphasize that our questions about beliefs are complex: they ask respondents about causation and to peer 20 years into the future and report their expectations about that future compared to the present. We expect that this kind of belief is more complex and difficult to report reliably than most of the belief questions that were asked in the United States. The reliability of belief questions in Nepal may also be reduced by the extensive period of civil conflict and uncertainty in the country. The data are certainly consistent with these expectations, although the differences in reliability are not very large in the case of questions about development causing family change and about expectations of the future.

Earlier we mentioned that the square root of a measurement reliability is equal to the coefficient λ, which is the standardized loading, or effect of the true score on the observed variable. The square roots of the average reliability scores for the five sets of items are 0.69, 0.63, 0.71, 0.71, and 0.81, respectively. These standardized effects of the true scores on the observed scores are quite high.

### Individual Stability across Time

We now turn to our estimates of individual-level relative stability across time (Table 4). As we discussed earlier and illustrated in Figure 1, these are standardized regression coefficients of a Wave 2 variable on the same Wave 1 variable and standardized regression coefficients of a Wave 3 variable on the same Wave 2 variable. Thus, they represent the extent to which individual positions in the overall distribution remain constant across waves of observation.

TABLE 4.
Stability coefficients from Wave 1 to Wave 2 and from Wave 2 to Wave 3, by dimension of developmental idealism (DI) and family element
BeliefsValues
(1) Dev. → Family(2) Family → Dev.(3) Expect future DI(4) Prefer future DI(5) Prefer DI
W1–W2W2–W3W1–W2W2–W3W1–W2W2–W3W1–W2W2–W3W1–W2W2–W3
Non-marital cohabitation 0.72 1.01 0.75 0.86 0.66 0.94 0.72 0.92 0.76 0.84
Age of first marriage for a womana 0.79 0.92 0.77 1.04 0.80 0.92 0.62 0.78
Divorce 0.65 0.90 0.99 1.04 0.81 0.90 0.72 0.90 0.81 0.99
Married children living with parents 0.75 0.96 0.88 0.97 0.82 0.89 0.74 0.79 0.81 0.82
Respect for authority     0.76 0.92 0.55 0.63
Individual needs before family 0.78 1.10 0.76 0.94 0.75 0.92 0.86 0.95 0.83 0.92
Premarital sex 0.74 0.90 0.88 0.85 0.78 0.85 0.83 0.95 1.04 1.12
Marriages arranged by parents 0.85 0.98 0.76 0.95 0.96 1.0 0.75 0.90 0.86 0.89
Inter-caste marriage 0.84 0.96 0.76 0.95 0.87 1.0 0.80 0.92 0.88 0.93
Women never marryingb 0.69 0.89 0.76 0.92 1.0 1.15 0.78 0.97 0.81 0.90
Widow remarriage 0.70 1.02 0.88 1.05 0.64 0.83 0.86 0.94 0.86 0.92
Time from marriage to first birth 0.77 0.84 0.76 0.88 0.70 0.96 0.74 0.88
Average stability 0.75 0.95 0.81 0.95 0.80 0.94 0.75 0.88 0.85 0.92
Average stabilityc 0.75 0.98 0.83 0.95 0.79 0.92 0.78 0.91 0.86 0.93
BeliefsValues
(1) Dev. → Family(2) Family → Dev.(3) Expect future DI(4) Prefer future DI(5) Prefer DI
W1–W2W2–W3W1–W2W2–W3W1–W2W2–W3W1–W2W2–W3W1–W2W2–W3
Non-marital cohabitation 0.72 1.01 0.75 0.86 0.66 0.94 0.72 0.92 0.76 0.84
Age of first marriage for a womana 0.79 0.92 0.77 1.04 0.80 0.92 0.62 0.78
Divorce 0.65 0.90 0.99 1.04 0.81 0.90 0.72 0.90 0.81 0.99
Married children living with parents 0.75 0.96 0.88 0.97 0.82 0.89 0.74 0.79 0.81 0.82
Respect for authority     0.76 0.92 0.55 0.63
Individual needs before family 0.78 1.10 0.76 0.94 0.75 0.92 0.86 0.95 0.83 0.92
Premarital sex 0.74 0.90 0.88 0.85 0.78 0.85 0.83 0.95 1.04 1.12
Marriages arranged by parents 0.85 0.98 0.76 0.95 0.96 1.0 0.75 0.90 0.86 0.89
Inter-caste marriage 0.84 0.96 0.76 0.95 0.87 1.0 0.80 0.92 0.88 0.93
Women never marryingb 0.69 0.89 0.76 0.92 1.0 1.15 0.78 0.97 0.81 0.90
Widow remarriage 0.70 1.02 0.88 1.05 0.64 0.83 0.86 0.94 0.86 0.92
Time from marriage to first birth 0.77 0.84 0.76 0.88 0.70 0.96 0.74 0.88
Average stability 0.75 0.95 0.81 0.95 0.80 0.94 0.75 0.88 0.85 0.92
Average stabilityc 0.75 0.98 0.83 0.95 0.79 0.92 0.78 0.91 0.86 0.93
a

For dimensions 1 and 2 the question read “people” instead of “a woman.”

b

c

Each average calculated using only those family elements measured across all dimensions of developmental idealism.

Table 4, which reports these measures of relative stability, demonstrates considerable stability of true scores between Waves 1 and 2 for each of the variables. Stability between Waves 1 and 2 is quite high, ranging from 0.55 to over 1. Furthermore, the average stability does not vary greatly over the five dimensions—from 0.75 for development causing family change to 0.85 for values about DI. Such stability is substantial, although we note that the intervals between waves were relatively short.

Also, for almost all of the variables there is more stability between Wave 2 and Wave 3 than between Wave 1 and Wave 2. There is only one exception to this observation in the entire table. Furthermore, the average stability between Waves 2 and 3 is noticeably higher than the average stability between Waves 1 and 2. The smallest increase in average stability was from 0.85 to 0.92, for the general-DI-values dimension, and the greatest increase in average stability was from 0.75 to 0.95, for the development-causing-family-change dimension.

## CONCLUSION

We began this paper with the observation that DI is an important set of ideas with implications for a broad range of behaviors. We also noted that empirical investigations of the factors influencing DI and the consequences of these values and beliefs require questionnaire items that can be reliably measured.

Our research is reassuring about the quality of our measures of DI, as it has shown that for five dimensions of DI, the measurement reliability in Nepal is quite favorable. Our average estimated reliability ranged from 0.40 to 0.65, with the factor loadings ranging from 0.63 to 0.81.

Furthermore, the average reliability of our straightforward value questions concerning different family elements is very similar to the average reliability of value questions measured in American surveys. However, as expected, our more cognitively complex value questions about future change were measured with somewhat less reliability than value questions focused on the present. Furthermore, the average reliability of two of our dimensions measuring DI beliefs was only a bit lower than the average reliability of beliefs found in American surveys. These results are reassuring for the research community's ability to study reliably the causes and consequences of DI beliefs and values.

Our research also demonstrates that rather than the various dimensions of DI being ephemeral, they are quite stable over time. Of course, in this research our time interval between waves is relatively short. But this significant level of relative stability suggests that DI beliefs and values may also be important features of people's mental maps of the world, with potentially strong implications for decision-making and behavior.

We end our paper with the caveat and challenge that our research has only been conducted in one region of Nepal and may not be generalizable to the entire country. Nepal is also only one country in a very diverse world. This suggests the need for the reliability and stability of DI measures to be examined in other parts of the world.

## REFERENCES

REFERENCES
2012
.
“Family Life and Developmental Idealism in Yazd, Iran.”
Demographic Research
26
:
207
38
.
Ahearn, Laura M.
2001
.
Invitations to Love: Literacy, Love Letters, and Social Change in Nepal
.
Ann Arbor
:
University of Michigan Press
.
Allendorf, Keera, and Roshan K. Pandian.
2016
.
“The Decline of Arranged Marriage? Marital Change and Continuity in India.”
Population and Development Review
42
(
3
):
435
64
.
Allendorf, Keera, and Arland Thornton.
2015
.
“Caste and Choice: The Influence of Developmental Idealism on Marriage Behavior.”
American Journal of Sociology
121
(
1
):
243
87
.
Alwin, Duane F.
1989
.
“Problems in the Estimation and Interpretation of the Reliability of Survey Data.”
Quality and Quantity
23
:
277
331
.
Alwin, Duane F.
2007
.
The Margins of Error: A Study of Reliability in Survey Measurement
.
Hoboken, NJ
:
Wiley-Blackwell
.
Asian Development Bank.
2012
.
Basic Statistics 2012
Axinn, William G., and Scott T. Yabiku.
2001
.
“Social Change, the Social Organization of Families, and Fertility Limitation.”
American Journal of Sociology
106
(
5
):
1219
61
.
Baker, David, and Gerald LeTendre, eds.
2005
.
National Differences, Global Similarities: World Culture and the Future of Schooling
.
Stanford, CA
:
Stanford University Press
.
Barrett, Deborah, and David John Frank.
1999
. “Population Control for National Development: From World Discourse to National Policies.” In
Constructing World Culture: International Nongovernmental Organizations since 1875
, edited by J. Boli and G. M. Thomas.
Stanford, CA
:
Stanford University Press
.
Barber, Jennifer S., Ganesh P. Shivaki, William G. Axinn, and Kishor Gajurel.
1997
.
“Sampling Strategies for Rural Settings: A Detailed Example from the Chitwan Valley Family Study.”
Nepal Population Journal
6
(
5
):
193
203
.
Berkovitch, Nitza.
1999
. “The Emergence and Transformation of the International Women's Movement.” Pp. 100–26 in
Constructing World Culture: International Nongovernmental Organizations since 1875
, edited by J. Boli and G. M. Thomas.
Stanford, CA
:
Stanford University Press
.
Beutel, Ann M., and William G. Axinn.
2002
.
“Gender, Social Change, and Educational Attainment.”
Economic Development and Cultural Change
51
(
1
):
109
34
.
Binstock, Georgina, and Arland Thornton.
2007
.
“Knowledge and Use of Developmental Thinking about Societies and Families among Teenagers in Argentina.”
Demografia
50
(
5
):
75
104
.
1991
.
Fatalism and Development: Nepal's Struggle for Modernization
.
Calcutta
:
Orient Longman
.
Boyle, Elizabeth Heger, ed.
2002
.
Female Genital Cutting: Cultural Conflict in the Global Community
.
Baltimore, MD
:
Johns Hopkins University Press
.
Chabbott, Colette, ed.
2003
.
Constructing Education for Development: International Organizations and Education for All
.
New York
:
Routledge Falmer
.
Cherlin, Andrew.
1992
.
Marriage, Divorce, Remarriage
.
Cambridge, MA
:
Harvard University Press
.
2005
.
“Sovereignty Relinquished? Explaining Commitment to the International Human Rights Covenants, 1966-1999.”
American Sociological Review
70
(
3
):
472
95
.
Dahl, Gudrun, and Annika Rabo.
1992
.
Kam-Ap or Take-Off: Local Notions of Development
.
Stockholm
:
Stockholm Studies in Social Anthropology
.
Deeb, Lara, ed.
2006
.
An Enchanted Modern: Gender and Public Piety in Shi'i Lebanon
.
Princeton, NJ
:
Princeton University Press
.
Dorius, Shawn F.
2016
.
“Chinese and World Cultural Models of Developmental Hierarchy.”
Chinese Journal of Sociology
2
(
4
):
577
608
.
Ferguson, James.
1999
.
Expectations of Modernity: Myths and Meanings of Urban Life on the Zambian Copper Belt
.
Berkeley
:
University of California Press
.
Frank, David J., Bayliss J. Camp, and Steven A. Boutcher.
2010
.
“Worldwide Trends in the Criminal Regulation of Sex, 1945 to 2005.”
American Sociological Review
75
(
6
):
867
93
.
Fricke, Tom E.
1997
.
“The Uses of Culture in Demographic Research: A Continuing Place for Community Studies.”
Population and Development Review
23
(
4
):
825
32
.
Geertz, Clifford.
1973
.
The Interpretation of Cultures: Selected Essays
.
New York
:
Basic Books
.
Ghimire, Dirgha, and William G. Axinn.
2006
.
“Family Change in Nepal: Evidence from Western Chitwan.”
Contributions to Nepalese Studies
33
(
2
):
177
201
.
Guneratne, Arjun.
1998
.
“Modernization, the State, and the Construction of Tharu Identity in Nepal.”
Journal of Asian Studies
57
(
93
):
749
73
.
Guneratne, Arjun.
2001
.
“Shaping the Tourist's Gaze: Representing Ethnic Differences in a Nepali Village.”
Journal of the Royal Anthropological Institute
7
(
3
):
527
43
.
Heise, D. R.
1969
.
“Separating Reliability and Stability in Test-Retest Correlations.”
American Sociological Review
34
:
93
101
.
Johnson-Hanks, Jennifer A., Christine A. Bachrach, S. Phillip Morgan, and Hans-Peter Kohler.
2011
.
Understanding Family Change and Variation: Toward a Theory of Conjunctural Action
.
London
:
Springer
.
Jöreskog, Karl G.
1970
.
“Estimation and Testing of Simplex Models.”
British Journal of Mathematical and Statistical Psychology
23
(
2
):
121
45
.
Kavas, Serap, and Arland Thornton.
2013
.
“Adjustment and Hybridity in Turkish Family Change: Perspectives from Developmental Idealism.”
Journal of Family History
38
(
2
):
223
41
.
Krücken, Georg, and Gili S. Drori, eds.
2009
.
World Society: The Writings of John W. Meyer
.
Oxford
:
Oxford University Press
.
Lesthaeghe, Ron.
1983
.
“A Century of Demographic and Cultural Change in Western Europe: An Exploration of Underlying Dimensions.”
Population and Development Review
9
(
3
):
411
35
.
Melegh, Attila, Tamás Kiss, Sabina Csánóová, Linda Young-DeMarco, and Arland Thornton.
2016
.
“The Perception of Global Hierarchies: South-Eastern European Patterns in Comparative Perspectives.”
Chinese Journal of Sociology
2
(
4
):
497
523
.
Melegh, Attila, Arland Thornton, Dimiter Philipov, and Linda Young-DeMarco.
2013
.
“Perceptions of Societal Developmental Hierarchies in Europe and Beyond: A Bulgarian Perspective.”
European Sociological Review
29
(
3
):
603
15
.
Meyer, John W., John Boli, George M. Thomas, and Francisco O. Ramirez.
1997
.
“World Society and the Nation-State.”
American Journal of Sociology
103
(
1
):
144
81
.
Meyer, John W., Patricia Bromley, and Francisco O. Ramirez.
2010
.
“Human Rights in Social Science Textbooks: Cross-National Analyses, 1970-2008.”
Sociology of Education
83
(
2
):
111
34
.
Mitchell, Colter M.
2009
. “Three Essays on Worldviews, Autonomy and the Family in Nepal.”
Dissertation, Sociology, University of Michigan
.
Muthén, B. O.
1981
. “Factor Analysis of Dichotomous Variables: American Attitudes toward Abortion.” In
Factor Analysis and Measurement in Sociological Research: A Multi-Dimensional Perspective
, edited by D. J. Jackson and E. F. Borgatta.
London
:
Sage
.
Muthén, L. K., and B. O. Muthén.
1998
.
Mplus Users Guide
. 6th ed.
Los Angeles, CA
:
Muthén & Muthén
.
Nisbet, Robert A.
1969
.
Social Change and History: Aspects of the Western Theory of Development
.
Oxford
:
Oxford University Press
.
Osella, Filippo, and Caroline Osella.
2006
.
“Once Upon a Time in the West: Stories of Migration and Modernity from Kerala, South India.”
Journal of the Royal Anthropological Institute
12
(
3
):
569
88
.
Pearce, Lisa D.
2002
.
“The Influence of Early Life Course Religious Exposure on Young Adults' Dispositions toward Childbearing.”
Journal for the Scientific Study of Religion
41
(
2
):
325
40
.
Pigg, Stacy Leigh.
1992
.
“Inventing Social Categories through Place: Social Representations and Development in Nepal.”
Comparative Studies in Society and History
34
(
3
):
491
513
.
Rokeach, Milton.
1970
.
Beliefs, Attitudes, and Values
.
San Francisco, CA
:
Jossey-Bass
.
Rokeach, Milton.
1973
.
The Nature of Human Values
.
New York
:
Free press
.
Shivakoti, G., W. G. Axinn, P. Bhandari, and N. Chhetri.
1999
.
“The Impact of Community Context on Land Use in an Agricultural Society.”
Population and Environment
29
(
3
):
191
213
.
Shrestha, Nanda R.
1989
.
“Frontier Settlement and Landlessness among Hill Migrants in Nepal Tarai.”
Annals of the Association of American Geographers
79
(
3
):
370
89
.
Thomas, D. S., et al
1929
.
Some New Techniques for Studying Social Behavior
.
New York
:
Columbia University Press
.
Thornton, Arland.
2001
.
Demography
38
(
4
):
449
65
.
Thornton, Arland.
2005
.
Reading History Sideways: The Fallacy and Enduring Impact of the Developmental Paradigm on Family Life
.
Chicago, IL
:
University of Chicago Press
.
Thornton, Arland, William G. Axinn, Thomas Fricke, and Duane F. Alwin.
2001
. “Values and Beliefs in the Lives of Children and Families.” Pp. 215–43 in
The Well-Being of Children and Families: Research and Data Needs
, edited by A. Thornton.
Ann Arbor
:
University of Michigan Press
.
Thornton, Arland, Shawn F. Dorius, and Jeffrey Swindle.
2015
.
“Developmental Idealism: The Cultural Foundations of World Development Programs.”
Sociology of Development
1
(
2
):
277
320
.
Thornton, Arland, Dirgha J. Ghimire, and Colter Mitchell.
2012
.
“The Measurement and Prevalence of an Ideational Model of Family and Economic Development in Nepal.”
Population Studies
66
(
3
):
329
45
.
Thornton, Arland, and Dimiter Philipov.
2009
.
“Sweeping Changes in Marriage, Cohabitation and Childbearing in Central and Eastern Europe: New Insights from the Developmental Idealism Framework.”
European Journal of Population
25
:
123
56
.
Thornton, Arland, Rachael S. Pierotti, Linda Young-DeMarco, and Susan Watkins.
2014
.
“Developmental Idealism and Cultural Models of the Family in Malawi.”
Population Research and Policy Review
33
(
5
):
693
716
.
Thornton, Arland, and Li-shou Yang.
2016
.
“Perceptions of Developmental Hierarchies in Taiwan: Conceptual, Substantive, and Methodological Insights.”
Chinese Journal of Sociology
2
(
4
):
547
76
.
Thornton, Arland, and Yu Xie.
2016
.
“Developmental Idealism in China.”
Chinese Journal of Sociology
2
(
4
):
483
96
.
Thornton, Arland, et al
2012a
.
“Knowledge and Beliefs about National Development and Developmental Hierarchies: The Viewpoints of Ordinary People in Thirteen Countries.”
Social Science Research
41
(
5
):
1053
68
.
Thornton, Arland, et al
2012b
.
“International Fertility Change: New Data and Insights from the Developmental Idealism Framework.”
Demography
49
:
677
98
.
Thornton, Arland, et al
2016
.
“Evaluating the Measurement Reliabilities and Dimensionality of Developmental Idealism Measures.”
Chinese Journal of Sociology
2
(
4
):
609
35
.
van de Kaa, Dirk J.
1987
.
“Europe's Second Demographic Transition.”
Population Bulletin
42
(
1
):
1
59
.
Yount, Kathryn M., and Hoda Rashad, eds.
2008
.
Family in the Middle East: Ideational Change in Egypt, Iran, and Tunisia
.
Oxford
:
Routledge
.

## NOTE

NOTE
Funding and acknowledgements: We appreciate the research support provided by a grant (R01 HD054847) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD). Our research was also supported by NICHD grants (R24HD041028 and P2CHD041028) to the Population Studies Center of the University of Michigan.