Obesity & Aggregate Fertility: An Extension of the Low Fertility - - PDF document

Obesity & Aggregate Fertility: An Extension of the Low Fertility Trap Hypothesis* Layton M. Field, PhD – Mount St. Mary’s University Lindsey B. Field, M.S., RDN, LD – Mount St. Mary’s University *Please contact the corresponding author (Layton Field; lfield@msmary.edu) for the most recent version of this paper.

Background Virtually every corner of the globe is experiencing some degree of fertility decline. According to the Central Intelligence Agency’s World Factbook 118 countries currently have an estimated total fertility rate below 2.1 (U.S. Central Intelligence Agency 2016). Popular discourse often views these declines as successful completions of the demographic transition in which lagging fertility rates finally decline to replacement levels. Undoubtedly, the early work of Notestein (1945, Notestein 1953) and Davis (1945, Davis 1963) in crafting this eloquent theory helps make sense of fertility decline at various points in modern history. Yes, demographic transition theory works well for some regions of the world. However, we also know that fertility decline in some countries does not quite fit with the demographic transition paradigm (Connelly 2008, Hodgson 1988, Szreter 1993). More importantly, several countries have witnessed continued fertility decline once the country’s respective fertility reached replacement levels. This has become known as the second demographic transition (Sobotka 2008). The second demographic transition is often characterized by decreases in desired family size, the postponement of union formation, and the delay in childbearing, all of which negatively influences overall fertility levels. Ultimately, these sustained fertility declines in much of Western Europe and Southeast Asia have necessitated a new approach to understanding fertility decline as fertility decline has given rise to substantial population aging in those regions. Some have suggested ways in which the graying of populations will impact future social, economic, and international arrangements (Last 2013, Wattenberg 2005). As such, population aging may require more countries to invest in pronatalist policies to positively impact fertility, though current pronatalist policies appear to have a very limited effect on fertility levels (Boling 2008, Lutz and Skirbekk 2005, McDonald 2006).

Thus, assisting the efficacy of pronatalist policies may require authorities to rethink many

• f the current constraints that negatively impact fertility. Recently, the Low Fertility Trap

Hypothesis (LFTH) provides an excellent theoretical framework for understanding the conditions which suppress fertility. Essentially, LFTH posits that countries that fall below a total fertility rate of 1.5 are likely to become trapped “if a trap is defined as an unpleasant situation (governments would rather see higher fertility) into which one enters unintentionally and which it is very difficult to get out of” (Lutz, Skirbekk and Testa 2006:173). Furthermore, LFTH states there are three concurrent mechanisms operating to suppress fertility levels including demographic, social, and economic factors. Demographically, as fertility declines and populations age future mothers decline as a share in the total population. For example, we know that the share of women in childbearing ages is similar between the United States and Europe, 40 percent to 38 percent respectively (Johnson, Field and Poston 2015). However, the stock of future mothers is quite different with Europe’s future mothers representing only 15 percent of the population compared to the United States at 20 percent. In other words, a smaller number of future mothers in Europe will have to have more children in order to offset the effect of population aging in Europe. Socially, LFTH suggests that norms and expectations surrounding family formation and ideal family size will continue to negatively influence fertility (Lutz, Skirbekk and Testa 2006). For instance, as couples opt to have fewer children, the next generation will internalize that norm and continue to have fewer children. This is also apparent in the demographic literature as some evidence suggests continual declines in desired fertility levels (Goldstein, Lutz and Testa 2003). Finally, from an economic standpoint couples balance the cost of childbearing against their desired standard of living and future earning potential (Lutz, Skirbekk and Testa 2006). Not

surprisingly, as couples become accustomed to a more expensive lifestyle, many will likely opt to have fewer children based on the rationale that their earning potential is incapable of providing for multiple children while maintaining their standard of living. As noted previously, LFTH provides a relatively simple explanation for why fertility continues to decline and may in fact become trapped in low fertility countries. Of course, a single theory will not likely accurately explain every scenario. In fact, we suggest that a fourth mechanism may also pressure fertility in a downward direction, namely, biology. Clearly, there are heritable characteristics that span the range of genetic variance which impact both fecundity and fertility (Kohler et al. 2006). However, we suggest there is growing evidence that increases in excess body fat, leading to obesity, also contribute to declines in fertility. Prior demographic research has substantiated a connection between obesity and declines in fertility although the direct biological connection was less clear and much of the relationship was attributed to union formation patterns that vary by obesity status (Jokela, Elovainio and Kivimäki 2008). Today, however, the maternal nutrition literature suggests a more direct connection between obesity and fecundity. Carrying around excess body fat has been directly linked to decreased fecundity and fertility among women (Stang and Huffman 2016). Women who are overweight or obese have increasing amounts of central adiposity, fat that accumulates around the abdominal area. Adipose tissue has been shown to reduce fertility among women through it role in the metabolism of sex hormones and associated enzymes and can delay fecundability compared to normal weight women (BMI of 18.5-24.9 kg/m2) (Diamanti‐Kandarakis and Bergiele 2001, Law, Maclehose and Longnecker 2007, Stang and Huffman 2016, Wise et al. 2009). For example, Robker and colleagues (2009) studied the pre-ovulatory follicular environment among women from different

BMI groups: moderate (BMI = 18-24.9 kg/m2); overweight (BMI= 25.0-29.9); and obese n=32 (BMI= ≥30 kg/m2). The women were patients at an infertility clinic undergoing in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI). Women in the obese group had significantly fewer oocyte (immature egg cell) extraction from the ovary compared to the

• verweight group (P=0.016) and obese women produced a lower number of embryos per female

compared to the other two groups (P=0.44). The researchers found a positive correlation between increases in BMI and elevated levels of insulin (P=0.0008), glucose (P=0.00), lactate (P=0.01), and triglycerides (P=0.0003), and C-reactive protein (P<0.0003), metabolites collected from the follicular fluid. C-reactive protein is an inflammatory marker that is usually elevated in obese

• individuals. These significant alterations in the follicular microenvironment can influence the

functionality of the ovary and therefore the quality of the oocycte. Subsequently, egg fertilization and implantation into the uterine wall may be compromised. Certainly, the link between fecundity and obesity will become more pressing as the number of individuals categorized as obese in the United States, and abroad, continues to

• increase. Currently, 38.3% of U.S. women in their peak fertility years are considered obese and

another third are overweight decreasing their fecundability ratio (Wise et al. 2009). Another 17.1% of adolescent girls/women 12-19 years of age, many of whom are at or nearing reproductive age, are also obese (Ogden CL et al. 2015). Similarly, other countries including Australia (28.4%), Mexico (26.8%), the United Kingdom (26%), Germany (23.3%) and New Zealand (32.0%) report a high prevalence of obesity that could be associated with low levels of fertility (Australian Bureau of Statistics 2012, Health and Social Care Information Centre 2016, New Zealand Government 2016).

Analysis Thus, the objective of this paper is to examine these parallel trends of increasing obesity and declining fertility. As stated previously, the biological link between obesity and fecundity is now better defined in the nutritional literature. However, to date no study examines to what extent the biological characteristic of weight impacts aggregate fertility. As such, we intend to analyze data at the county level for all counties in the United States. We draw on data from three primary sources. Our key independent variable comes from the United States Center for Disease Control and Prevention (CDC), which measures the estimated percentage of women ages 18 and older within each county that are classified as obese,

• r having a BMI of 30 or greater. Second, we collected fertility estimates for each county based
• n the five year estimates contained in the 2013 American Community Survey conducted by the

United States Census Bureau. We are using fertility data from 2013 averaged over five years to match the data we have available on obesity, which includes data from 2009 to 2013. Finally, we are using data from the “Atlas of Rural and Small-Town America” developed by the United States Department of Agriculture (USDA) to control for other known correlates of fertility decline including income, education, percent minority, and employment at the county level. The dependent variable for this study is the fertility rate for each county as calculated by the five year estimates contained in the 2013 American Community Study. The fertility rate represents the number of women between the ages of 15 and 50 that have given birth over the past 12 months (fertility). Our key independent variable examines the estimated percent of females over the age of 18 who have a BMI of 30 or greater (obesity). We also include several important controls including the total number of women between the ages of 15 and 50 for each

county (females), percent of the population that is non-Hispanic white (white), percent of the population that has less than a high school diploma (education), median household income (income), and the status of the county as a metropolitan or rural county (metro). The descriptive statistics for these variables are displayed in Table 1. Table 1 displays the descriptive statistics for all counties. Table 1. Descriptive Statistics of All U.S. Counties, Circa 2013 (N=3,117)

Mean

• Std. Dev.

Min. Max

Fertility (births per 1,000 women) 56.95 20.89 252 Obesity (% Females with BMI 30+) 29.82 4.97 10.98 53.64 Female Population Ages 15-50 23,807.79 81,574.73 10 2,624,407 Non-Hispanic White (%) 78.40 19.82 2.86 99.16 Less than High School (%) 15.02 6.76 1.27 53.28 Median HH Income ($) 47,097.73 12,066.63 21,658 125,635 Metropolitan County .3702

• 1

On average, counties across the United States report that about 57 women per 1,000 women report having a birth in the past twelve months. Interestingly, on average 29.82 percent

• f women in the United States are classified as obese. Some counties reported obesity prevalence

rates as low as 10.98 percent and others as high as 53.64 percent. Similarly, the percent of the population at the county level that is non-Hispanic white averages 78.4 percent but ranges substantially from a low of approximately 3 percent to a high of around 99 percent. Education and median household income are also varied across the United States with the average county reporting that 15 percent of the population has less than a high school diploma with a median household income of approximately 47,000 dollars. However, these averages hide substantial

• ranges. For example, in some counties as little as 1 percent of the population has less than a high

school diploma while in other counties as much as 53 percent of the population has failed to achieve that educational benchmark. Median household income also ranges from a low of

$21,658 to a high of $125,635. Finally, approximately 37 percent of counties in the United States are considered metropolitan counties. However, since the county serves as the unit of analysis we have decided to reclassify the

• besity measure into five different categories assessed at the county level and run the descriptive

statistics on each group. In other words, we have broken counties into one of five groups including counties in which less than 15 percent of women are considered obese (low), counties where between 15 percent and 24.9 percent of women are currently obese (below average), counties where between 25 percent and 34.9 percent of women are classified as obese (average), counties with women between 35 percent and 44.9 percent of women are obese (above average), and finally counties where more than 45 percent of women are obese (high). In Table 2 we explore the descriptive statistics based on the obesity classification.

Table 2. U.S. County Descriptive Statistics by Obesity Category

Mean

• Std. Dev.

Min. Max Lowest Obesity (N=13) Fertility (births per 1,000 women) 50.08 10.61 29 65 Obesity (% Females with BMI 30+) 13.10 1.25 10.98 14.9 Female Population Ages 15-50 24067.15 27614.93 2186 79034 Non-Hispanic White (%) 79.85 12.99 43.90 93.33 Less than High School (%) 6.18 2.75 2.58 12.41 Median HH Income ($) 71079.31 18968.44 44508 107250 Metropolitan County .3069

• 1

Below Average Obesity (N=409) Fertility (births per 1,000 women) 52.17 21.57 252 Obesity (% Females with BMI 30+) 22.10 2.37 15.1 24.98 Female Population Ages 15-50 71594.63 178138.3 15 2624407 Non-Hispanic White (%) 76.66 19.57 12.79 98.42 Less than High School (%) 10.47 5.19 1.87 45.02 Median HH Income ($) 57389.10 16305.23 27627 122641 Metropolitan County .5330

• 1

Average (N=2,328) Fertility (births per 1,000 women) 57.58 20.90 187 Obesity (% Females with BMI 30+) 29.19 2.58 25 34.98 Female Population Ages 15-50 17356.68 52109.75 10 1357453 Non-Hispanic White (%) 81.89 17.21 2.86 99.16 Less than High School (%) 14.91 6.44 1.27 53.28 Median HH Income ($) 46851.58 9944.52 23047 125635 Metropolitan County .3565

• 1

Above Average (N=344) Fertility (births per 1,000 women) 58.85 19.27 144 Obesity (% Females with BMI 30+) 37.76 2.39 35 44.52 Female Population Ages 15-50 11908.33 32924.04 180 439055 Non-Hispanic White (%) 60.23 22.11 9.50 98.68 Less than High School (%) 20.87 5.52 7.91 40.50 Median HH Income ($) 36930.80 6816.47 21658 75682 Metropolitan County .2907

• 1

Highest Obesity (N=23) Fertility (births per 1,000 women) 53.35 24.46 9 113 Obesity (% Females with BMI 30+) 47.75 2.21 45.1 53.64 Female Population Ages 15-50 4825.04 4471.15 1746 22431 Non-Hispanic White (%) 25.20 7.13 13.67 41.25 Less than High School (%) 25.65 4.81 13.67 34.48 Median HH Income ($) 27958.52 2663.92 22640 33644 Metropolitan County .0870

• 1

The range of these descriptive statistics is substantial when we examine all counties in the United States. However, the patterns identified in the descriptive statistics are even starker when recalculated for each of the defined obesity categories. For example, non-Hispanic White on average make up 79 percent of the population among counties with the lowest level of obesity prevalence compared to just 25 percent among counties with the highest prevalence. Similarly, counties with higher levels of obesity tend to have lower levels of education and much lower median household incomes. For example, the median household income in counties at the lowest end of the obesity spectrum averaged $71,000 compared to approximately $28,000 among counties in which obesity is the most prevalent. We also noticed that many of the high obesity counties tended to reside in rural areas of the country. Therefore, we decided to test for spatial autocorrelation prior to proceeding to our regression analysis, which resulted in a Moran’s I of .67 suggesting that a typical ordinary least squares regression model would be inappropriate (Anselin 2002). As such, we used the software package STATA13 to complete all of the data preparation and descriptive statistics but executed

• ur spatial lag models using the package GeoDa. Generally, spatial lag models are very similar

to spatial error models and return results that are often statistically identical (Ward and Gleditsch 2008). Nonetheless, we selected the spatial lag model based on the selection process as defined by Anselin (2005) and because we are comfortable with the logical assumption that contiguous counties have a direct influence on the obesity prevalence and fertility rates of neighboring

• counties. Thus, we have calculated utilized spatial lag models with first level queen contiguous
• counties. The results from the spatial lag models are reported in Table 3 below.

Table 3. Regression Coefficients for Spatial Lag Models (N=3,117) Model 1 Model 2 Model 3 b Std. Error b Std. Error b Std. Error

Obesity (%)

.22** .07 .14 .09

• 1.07**

.23

Female Population Ages 15-50 (100,000’s)

• .87**

.49

• .69

.48

Non-Hispanic White (%)

• .1**

.02

• .13**

.02

Less than High School (%)

.17* .08 .21** .07

Median HH Income ($10,000)

1.82** .43

• 5.32**

1.34

Metropolitan County

• 6.42**

.85

• 6.97**

.85

Lagged Fertility

.24** .03 .20** .02 .18** .03

Constant

36.72** 2.58 41.09** 5.2 74.97** .02

Income/Obesity Interaction

.28** .05

R- Squared

.041 .067 .080

Log Likelihood

• 13818.5
• 13770.1
• 13754.5

Rho

.242 .199 .183 *p<.05 **p<.01 Our primary hypothesis suggests that higher levels of obesity will negatively affect fertility levels. The results in Model 1 fail to support that conclusion. The coefficient associated with obesity is statistically significant but moves in the opposite direction of what we expected such that as obesity increases, the number of women reporting a birth in the past 12 months also

• increases. Furthermore, the addition of the remaining control variables in Model 2 also returned

unexpected results. The coefficient associated with obesity in Model 2 fails to achieve statistical significance suggesting that obesity isn’t a key predictor of aggregate fertility once we take into account various predictors of fertility. However, we know from our initial descriptive statistics that household income varies tremendously across the United States and was particularly varied among the different categories

• f obesity reported in Table 2. Therefore, we decided to test for an interaction effect between
• besity prevalence and median household income. We know that low income households are at

higher risk of obesity (Ogden et al. 2010). Similarly, we typically assume a negative correlation between income and fertility (McCall and Percheski 2010). In our case, household income appears to be positively correlated with fertility in Model 2. Consequently, we believe there to be an interaction effect driving both of these atypical results. Model 3 includes the interaction effect, which is statistically significant (b=.28, p<.001). Moreover, the coefficients associated with obesity and household income both change to the hypothesized direction. Interactions are often best understood by calculating expected values using selected

• figures. The resulting equation for Model 3 could be written as:

𝑍 ̂

𝐺𝑓𝑠𝑢𝑗𝑚𝑗𝑢𝑧 = 74.97 − 1.07𝑃𝑐𝑓𝑡𝑗𝑢𝑧 − . 13𝑋ℎ𝑗𝑢𝑓 +. 21𝐹𝑒𝑣𝑑𝑏𝑢𝑗𝑝𝑜 − 5.32𝐽𝑜𝑑𝑝𝑛𝑓 − 6.97𝐷𝑝𝑣𝑜𝑢𝑧 +. 28𝐽𝑜𝑢𝑓𝑠𝑏𝑑𝑢𝑗𝑝𝑜

Assuming we input the mean for each variable in the equation, as found in Table 1, we would

• btain an expected fertility rate of approximately 43.2 births per 1,000 women. Yet, if we impute

the mean value of obesity and household income for counties with the lowest obesity prevalence (as defined in Table 2) the expected fertility rate falls to 35.22. Moreover, if we once again hold the variables of race, education, and metropolitan status constant but impute the mean values for

• besity and household income for the highest obesity counties the expected fertility falls to

32.41. The same pattern emerges if we run the calculations for non-metro counties. Essentially, this interaction between income and obesity suggests that as income increases, at the county level, the slope of the effect of obesity on fertility weakens. Thus, counties with more financial means have the resources to perhaps negate some of the negative effects of obesity, or mitigate the rise of obesity all together.

Discussion The objective of this paper is to examine the effect of obesity on fertility rates at the aggregate level. More specifically, we hoped to extend the Low Fertility Trap Hypothesis to include a physiological characteristic, namely obesity. Finally, we intended to compare the characteristics of counties with a high prevalence of obesity to those with lower rates. We have found limited support for our original hypothesis suggesting that rising obesity levels will likely have a negative impact on already declining fertility rates. However, the effect is reduced when we account for the interaction between obesity and income. Thus, counties with higher median household incomes will likely face fewer issues than more impoverished regions of the country. We think our results represent a natural extension of the Low Fertility Trap Hypothesis. We expect fertility rates will continue to fall across the globe, particularly in response to the demographic, social, and economic factors negatively correlated to fertility as defined by LFTH. Moreover, such large shifts in biological conditions, like rising obesity rates, will likely also negatively impact fertility at an aggregate level. Our results certainly recognize the importance of the main factors of LFTH but suggest that the addition of a biological component makes logical sense. Additionally, obesity rates are concentrated in counties with lower than average median household incomes, smaller proportions of non-Hispanic Whites, lower levels of education, and predominately rural parts of the country. The obesity rates alone in these counties climb as high as nearly 54 percent. The impending health care costs associated with these communities is going to increase exponentially as these counties are plagued with the large number of negative health conditions associated with obesity increase. Moreover, these communities will likely have the least financial capital to contend with these trends. In future research we hope to examine

these high obesity counties more closely, particularly as obesity relates to fertility. For example, we would like to asses fertility intentions, parity, and other fertility related behaviors in hopes of shedding more light on the relationship between obesity and fertility. Similarly, this analysis has focused exclusively on obesity and fertility of women. We would like to extend the project to assess the impact of fluctuations in male obesity at the county level to determine if male obesity may also serve as an influential factor. In sum, countries around the globe are experiencing continued declines in fertility rates. The Low Fertility Trap Hypothesis provides one elegant explanation as to why many more countries can expect fertility decline in the near future. However, we feel that a biological trapping mechanism also contributes to the social, economic, and demographic mechanism currently identified in LFTH. We believe our results provide limited support that this biological link between obesity and fertility does directly impact aggregate fertility rates. This is more concerning given current trends in obesity rates around the globe. We certainly are not arguing that obesity is the most important factor in fertility decline. However, this link may be important to the many countries that continue to introduce pronatalist policies in an attempt to offset declining fertility rates.

References: Anselin, Luc. 2002. "Under the Hood: Issues in the Specification and Interpretation of Spatial Regression Models." Agricultural Economics 27(3):247-67. Anselin, Luc. 2005. Exploring Spaital Data with Geoda: A Workbook. Santa Barbara, CA: Center for Spatially Integrated Social Science. Australian Bureau of Statistics. 2012. "Australian Health Survey: First Results, 2011-12." Canberra: Australian Bureau of Statistics 4364.0.55.001. Boling, Patricia. 2008. "Demography, Culture, and Policy: Understanding Japan's Low Fertility." Population and Development Review 34(2):307-26. Connelly, Matthew James. 2008. Fatal Misconception: The Struggle to Control World

• Population. Cambridge, Mass.: Belknap Press of Harvard University Press.

Davis, Kingsley. 1945. "The World Demographic Transition." Annals of the American Academy

• f Political and Social Science 237:1-11. doi: 10.2307/1025490.

Davis, Kingsley. 1963. "The Theory of Change and Response in Modern Demographic History." Population index 29(4):345-66. Diamanti‐Kandarakis, E and A Bergiele. 2001. "The Influence of Obesity on Hyperandrogenism and Infertility in the Female." Obesity reviews 2(4):231-38. Goldstein, Joshua, Wolfgang Lutz and Maria Rita Testa. 2003. "The Emergence of Sub- Replacement Family Size Ideals in Europe." Population Research and Policy Review 22(5/6):479-96. doi: 10.2307/40230837. Health and Social Care Information Centre. 2016. "Statistics on Obesity, Physical Activity and Diet." England.

Hodgson, Dennis. 1988. "Orthodoxy and Revisionism in American Demography." Population and Development Review 14(4):541-69. doi: 10.2307/1973624. Johnson, Kenneth M., Layton M. Field and Dudley L. Poston. 2015. "More Deaths Than Births: Subnational Natural Decrease in Europe and the United States." Population and Development Review 41(4):651-80. doi: 10.1111/j.1728-4457.2015.00089.x. Jokela, Markus, Marko Elovainio and Mika Kivimäki. 2008. "Lower Fertility Associated with Obesity and Underweight: The Us National Longitudinal Survey of Youth." The American Journal of Clinical Nutrition 88(4):886-93. Kohler, Hans-Peter, Joseph Lee Rodgers, Warren B. Miller, Axel Skytthe and Kaare Christensen.

• 2006. "Bio-Social Determinants of Fertility." International Journal Of Andrology

29(1):46-53. doi: 10.1111/j.1365-2605.2005.00606.x. Last, Jonathan V. 2013. What to Expect When No One's Expecting: America's Coming Demographic Disaster. New York: Encounter Books. Law, DC Gesink, Richard F Maclehose and Matthew P Longnecker. 2007. "Obesity and Time to Pregnancy." Human Reproduction 22(2):414-20. Lutz, Wolfgang, Vegard Skirbekk and Maria Rita Testa. 2006. "The Low-Fertility Trap Hypothesis: Forces That May Lead to Further Postponement and Fewer Births in Europe." Vienna Yearbook of Population Research 4:167-92. doi: 10.2307/23025482. Lutz, Wolfgang and Vegard Skirbekk. 2005. "Policies Addressing the Tempo Effect in Low- Fertility Countries." Population and Development Review 31(4):699-720. doi: 10.2307/3401522.

McCall, Leslie and Christine Percheski. 2010. "Income Inequality: New Trends and Research Directions." Annual Review of Sociology 36(1):329-47. doi: doi:10.1146/annurev.soc.012809.102541. McDonald, Peter. 2006. "Low Fertility and the State: The Efficacy of Policy." Population and Development Review 32(3):485-510. doi: 10.2307/20058901. New Zealand Government. 2016, "Tier 1 Statistics 2015/16: New Zealand Health Survey". (http://stats.govt.nz/browse_for_stats/snapshots-of-nz/nz-social- indicators/Home/Health/obesity.aspx). Notestein, Frank W., ed. 1945. Population - the Long View, Edited by T. W. Schultz. Chicago: University of Chicago Press. Notestein, Frank W. 1953. Economic Problems of Population Change: Oxford University Press. Ogden CL, Carroll MD, Fryar CD and Flegal KM. 2015. Prevalence of Obesity among Adults and Youth: United States, 2011–2014. Nchs Data Brief, No 219Congress. Ogden, Cynthia L., Molly M. Lamb, Margaret D. Carroll and Katherine M. Flegal. 2010. "Obesity and Socioeconomic Status in Children and Adolescents: United States, 2005– 2008." in NCHS data brief no 51. Hyattsville, MD: National Center for Health Statistics. Robker, R. L., L. K. Akison, B. D. Bennett, P. N. Thrupp, L. R. Chura, D. L. Russell, M. Lane and R. J. Norman. 2009. "Obese Women Exhibit Differences in Ovarian Metabolites, Hormones, and Gene Expression Compared with Moderate-Weight Women." The Journal of clinical endocrinology and metabolism 94(5):1533-40. Sobotka, Tomas. 2008. "Overview Chapter 6: The Diverse Faces of the Second Demographic Transition in Europe." Demographic Research 19:171-224. doi: 10.4054/DemRes.2008.19.8.

Stang, Jamie and Laurel G. Huffman. 2016. "Position of the Academy of Nutrition and Dietetics: Obesity, Reproduction, and Pregnancy outcomes." Journal of the Academy of Nutrition and Dietetics 116(4):677-91. doi: http://dx.doi.org/10.1016/j.jand.2016.01.008. Szreter, Simon 1993. "The Idea of Demographic Transition and the Study of Fertility Change: A Critical Intellectual History." Population and Development Review 19(4):659-701. doi: 10.2307/2938410. U.S. Central Intelligence Agency. 2016, "Total Fertility Rate" The World Factbook. Retrieved November 21, 2016, (https://www.cia.gov/library/publications/the-world- factbook/rankorder/2127rank.html#us). Ward, Michael Don and Kristian Skrede Gleditsch. 2008. Spatial Regression Models. Thousand Oaks: Sage Publications. Wattenberg, Ben J. 2005. Fewer: How the New Demography of Depopulation Will Shape Our

• Future. Chicago: Ivan R. Dee.

Wise, Lauren A., Kenneth J. Rothman, Ellen M. Mikkelsen, Henrik Toft Sørensen, Anders Riis and Elizabeth E. Hatch. 2009. "An Internet-Based Prospective Study of Body Size and Time-to-Pregnancy." Human Reproduction. doi: 10.1093/humrep/dep360.