Parental Status Connection and Social Network Variety in Adulthood

Abstract

This study enriches the literature on social networks and social capital by investigating how parental status potentially impacts social network diversity in adulthood. Using the 2018 iteration of the General Social Survey (GSS, n = 2348), a high quality nationally representative survey of the United States, we utilize latent class and regression analyses, finding that parental status, especially medium and cross-status occupational connections contribute to social capital in the form of network diversity. Yet, personal socio-economic factors, notably income and race, largely offset parental effects. This underscores the complexity of network composition, emphasizing the influential role of individual resources, attributes, and mobility in shaping social networks and forming bridging social capital.

1. Introduction

In recent decades, there has been a growing interest in ideas of social capital and the social capital accrued via social networks. This has been the case with scholars and the wider public, both of whom have demonstrated increased interest in these topics. However, this surge in interest has also led to a diversity of opinions and competing conceptualizations regarding its definition and the methods for assessing it. Social capital can be operationalized by looking at how tightly knit social networks are, the diversity of connections within those networks, and gauging the overall social involvement, even though these often do not show strong correlations to each other [1]. The various approaches broached for measuring and definitionally understanding social capital also have significant implications for our comprehension of its effects on social connections, communities, civil society, and social hierarchies.

The research of this paper primarily focuses on the prominent but still often less-utilized way of understanding social capital referred to as social network variety. Network variety has been argued to be facilitated from ‘weaker’ [2] social connections or ‘structural holes’ [3] in social networks, whereby individuals can form less redundant connections and thus increase network diversity [4,5]. These connections can be formed as a product of cultural variety by which shared social connection points can be made [6,7]. Social network variety can in turn contribute to ‘bridging’ forms of social capital, defined by Putnam [8] as cultivating connections between individuals from different backgrounds, which ultimately fosters inclusivity, trust, cooperation, and resource exchange [9,10,11].

Just as with other forms of capital, it has been asserted that there is an intergenerational transmission of social capital (e.g., [12,13]), parental and childhood influence upon social capital formation (e.g., [14,15,16,17]), and compounding effects from other forms of parental capital on the ability to efficaciously mobilize social capital resources [18]. A prominent theorist in the conceptualization of different forms of capital, Pierre Bourdieu [19] argues for the intergenerational transmission of social capital. Drawing upon Bourdieu’s prominent, if incomplete, social capital theory, this study examines how the social networks and capital inherited by children from their parents may shape their future opportunities and social connections. Bourdieu’s [19] concept of social capital as resources embedded within networks elucidates the potential advantage or disadvantage transferred intergenerationally, and can be used to take up an empirical approach of investigating occupational network connections.

However, when it comes to understanding if specifically bridging forms of social capital bonds between parents, such as those theorized by Putnam [8], effect the bridging social capital of their children (and especially in adulthood), we do not yet if there is such a connection and there is an absence of scholarly studies. To the best of our knowledge, there has not been an investigation into how the mix of parental status, such as between high and low status occupations or education levels for different parents, might influence the social network variety of their children. In this study we consider both cross status occupational prestige connections and educational status connections for these purposes. The research of this paper therefore seeks to investigate the relationship, if any, between cross-status (or intra-status) parental connection and the social network variety of their (in this case adult) children.

2. Social Capital, Network Diversity, and Intergenerational Transmission

2.1. Social Capital and Social Network Variety

The core of any social capital concept relies upon the premise that there is a powerful potentiality inherent to social relations [20]. Likewise, in order for social capital to be of use, there must be an understanding that there are social rewards to be gained from social relationships and investments therein [21]. Foundational research (e.g., [6,22,23,24,25]) in this area found that these benefits include more prestigious occupations for those with more extensive, prestigious, and varied social networks, enabling them to access sought after employment opportunities, smoother promotional processes, and other opportunities for quicker advancement in their careers. Conversely, when there is a shortage of these advantageous connections, it can result in restricted entry to different institutions and segments of society where these advantages accrue, thereby causing an increase in social disparities [26].

Since these benefits compound upon each other and perpetuate themselves, in these ways social capital also reflects Bourdieu’s (e.g., [19]) work, which underscored the multifaceted nature of social capital as both a marker of social class position and a driver of social inequalities, including intergenerationally [5,19]. Indeed, other research has indicated that a lack of social capital resources is a prominent reason for the existence and reproduction of social class-based, gendered, and racialized inequities in contemporary societies (e.g., [25,27,28,29]). For instance, Lutter [28] finds that women have a higher risk for failure along their career path than men when both are in similarly less open and diverse networks, highlighting potentially discriminating elements of ostensibly similar forms of social capital between societal groups. Bourdieu’s [19] perspective on social capital therefore highlights its multifaceted characteristics, although rather focusing on its discriminating rather than its positive or democratizing potentials, as argued prominently by Putnam [8,30]. For his part, Bourdieu [19] explains some dynamics and manifestations of his theory of social capital and network resources this way:

Because the social capital accruing from a relationship is that much greater to the extent that the person who is the object of it is richly endowed with capital (mainly social, but also cultural and even economic capital), the possessors of an inherited social capital, symbolized by a great name, are able to transform all circumstantial relationships into lasting connections. They are sought after for their social capital and, because they are well known, are worthy of being known (‘I know [them] well’); they do not need to ‘make the acquaintance’ of all their ‘acquaintances’, they are known to more people than they know, and their work of sociability, when it is exerted, is highly productive.

(pp. 250–251)

While Coleman [31,32] and Lin [10,33,34] approach the concept from slightly different macro-theoretical viewpoints from Bourdieu, they share similarities in their focus on social capital as valuable resources that are derived from social network connections. According to Coleman, these network-derived resources go beyond the sheer quantity of capital possessed by individuals within one’s social circle. Rather, they involved the added value that results from interpersonal connections and interactions. Lin [10,34], however, underscores the inherent embeddedness of social capital within social networks and defines it as resources that can be leveraged through membership and connections within these networks. Lin [33] also argues that access to and mobilization of social resources are key elements to status attainment.

There is thus much potentiality for both symbolic and tangible gains to be made from such capital networks. For instance, connections with close family or family friends can have a propelling effect on career trajectories [35]. Likewise, social capital is arguably more likely to be generated and reproduced via networks of ostensible acquaintances, or what sociologist Mark Granovetter (e.g., [2]) termed the ‘strength of weak ties’. Savage at al. [5] likewise argue that those in prestigious occupations may not as directly benefit from close connections to power as people presume, but rather their beneficial status arises from being linked by varying degrees of separation to a wide connection of acquaintances, often as a result of advantaged backgrounds.

The advantages gleaned by this wider network of weaker connections work primarily to the degree that having a lower threshold of connection is able to produce a more diverse social network. Granovetter [2] argues that the diversity which is largely facilitated by weak ties produces beneficial social capital since those who are less close to each other are also less likely to share traits, experiences, forms of knowledge, and social connections that are redundant to their own. The merits of weak social network ties for social capital formation is also supported by Burt [3], where he argues for ‘structural holes’ in networks, which are areas within less dense (i.e., less constricting) social networks that facilitate the ability to make less redundant connections [4].

Relatedly, Erickson [6] contends that having a diverse social network in the workplace can be highly advantageous. Much of this advantage stems from a diversity of cultural tastes and behaviors by which people can relate and communicate with people of differing interests and backgrounds. In this analysis, Erickson draws on Peterson’s [36] theory of the cultural ‘omnivore’, who are those individuals that engage with a diverse range of cultural tastes and behaviors, irrespective of the relative social status of those forms of culture. Both Peterson and a number of subsequent studies (e.g., [37,38,39,40]) consistently have linked this behavior and this cultural variety with elevated social position.

According to Erickson [6], it is this eclectic cultural engagement that contributes substantially to wider and more varied social networks. This is due to the fact that this type of cultural diversity not only expands one’s cultural exposure but also facilitates more effective participation in social networks, ultimately providing beneficial social capital to individuals via network variety effects. It can also work in the opposite direction, from social variety to cultural variety. This is to say, as Erickson [1996] (277) argued, ‘(social) network variety is important…because contact with different types of people includes contact with different types of culture’. Cultural variety is thus both product and cause of social network variety, and vice-versa.

One of the primary drivers of social network variety’s social capital dividends is variety’s ability to facilitate types of ‘bridging’ social capital. As defined by Putnam [8], bridging social capital is the cultivation of relationships and connections between individuals or groups from different social backgrounds, be them socio-economic, cultural, or demographic. Therefore, almost by definition these are diverse connections. However, ultimately the more of these cross-status connections are within a network the more bridging social capital is conferred. Networks made up of these types of connections have also been referred to ‘heterophilous networks’ (as opposed to ‘homophilous networks’) in social network theory, whereby such termed social networks include socially diverse individuals and their connections [11]. Such connections can help to facilitate resource and knowledge exchange between different segments of society [9], increase community and civic engagement [8], and foster inclusivity, interpersonal trust, and cooperation, even as these diverse connections are often hard to build due to existing prejudices, discrimination, and the large gaps of social inequalities that these connections seek to bridge [10].

2.2. Intergenerational Transmission of Social Capital

When it comes to the intergenerational influence of social network variety and cross-status parental bonds, there is little known. Nunn et al. [41] extend Bourdieu’s concept of social capital by exploring compound inequality in intergenerational contexts. Their findings highlight how parental status and strategies contribute to the reproduction of inequality across generations, aligning with Bourdieu’s theory [19] on the transmission of social capital [41]. Among studies that are adjacent to such investigations, many of them focus on the utility of parental social capital within the educational process, finding utility of parental social capital within the educational process of their children (e.g., [16,17]). Others find intergenerational transmission, both via educational systems and networks [12] and immigrant or other ethnic networks (e.g., [13]). The inheritance of social capital generally has been found to positively contribute to increased social mobility [42]. Even if such social capital is not inherited, the incremental accumulation of social capital is an important driver of social mobility [43]. However, since the upwardly mobile tend to have larger social networks without ties to the most high-status persons, and thus limited bridging forms of social capital with high status individuals, this has been suggested by others as evidence of inter-generational processes of the social capital reproduction [44].

Among other potentially related research, Ravanera and Rajulton [45] find that parents, and especially mothers, from nuclear families that are intact tend to have greater social networks and social capital in the form of interpersonal trust of others and social trust of institutions. Although not dealing with how these forms of social capital may be reflected in interpersonal and societal trust, it is possible that such forms of trust may be reflected in children as they age. Gemar [46] found strong intergenerational transmission of religious service attendance, which may be a specific site where social connections are reproduced, although the intergenerational transmission was differentially influenced by each parent. In another study, Song [47] finds a motherhood penalty, fatherhood premium, single-parenthood penalty, married parenthood premium, and single-motherhood penalty in the quality of social capital that are possessed by these groups. However, there was not a difference in accessing more diverse social networks amongst these different groups [47]. The fact that prior research has not found a difference in the structure of parental relationship is important to note for this paper, where we do not investigate these various possible iterations of parental connectivity. We rather suggest that it is possible that no matter the specific contours of the parental bonds, those with cross-status parents may benefit from their symbolic and tangible social networks as a result.

Finally, while again not dealing with intergenerational reflection or transmission of social capital, Andersson [18] finds that parental economic capital impacts on the ability of their children to effectively mobilize their own social capital networks in leaving the family home and finding a first house. Therefore, while some of these studies are tangential to topics and processes investigated here, none investigate the type of connection probed in this paper. There are to our knowledge no existing studies that focus on the type of parental influence upon social networks that this paper seeks to investigate. It is for instance possible that parents of differing social status convey upon children the ability or willingness to seek out or navigate inter-status (bridging) relationships in both a more purposeful and natural way. This would in turn result in more diverse social networks and greater bridging forms of social capital, in theory potentially throughout the life course. To this end we ask the following guiding research questions in this paper:

• RQ1: Are parental cross status ties predictive of social network variety?
• RQ2: If yes, is this relationship mitigated by socio-economic or demographic effects?

3. Materials and Methods

3.1. Data

The data for the research of this paper comes from the General Social Survey (GSS), which is a prominent nationally representative survey of current public attitudes, demographics, and behaviours within the United States. Carried out by the NORC (National Opinion Research Center) at the University of Chicago, the GSS has been a consistently running survey program since 1972, with 34 different survey rounds between 1972 and 2022. However, many specify questions and question types vary from survey round to survey round with changes in American society and topics of prevailing public and scholarly interest. This paper utilizes the 2018 iteration of the GSS survey (n = 2348) since it is the most recent round of the survey program that includes all of the necessary questions for this paper. Specifically, the 2018 survey is the only round to include questions that enable the calculation of social network variety on the basis of occupational status.

3.2. Measuring Social Capital and Parental Status

3.2.1. Social Network Variety (Occupational Connections)

The way that we measure social capital in this paper is via the breadth and variety of social networks. We do so due to various arguments that social capital can be understood as the resources that are available through one’s social network (e.g., [31]). This is also efficacious since social capital can be understood and operationalized as cross-type connectedness [1]. We operationalize cross-status relationships via a measure of the variety in the occupations of one’s network, an approach also taken by others in social capital research (e.g., [5,7]). One potential outcome includes observing variety across occupations of differing social status within social networks, an observation by which we could reasonably assert types of ‘bridging’ social capital, or connections between people of different social positions [48]. Alternatively, we may observe little variety in the social status of the occupations present in people’s social networks, in this way reasonably suggesting more network closure and more exclusionary forms of social capital (e.g., [19]).

As Bourdieu argued, credentialization, and by extension symbolic and actual membership in certain groups, is a key component to social network formation and membership. Since these processes in turn influence the amount of resources that can be effectively accessed via such a network, the variety of occupational resources which one might have access to presents strong potentiality for social capital. Indeed, variety is a critical component of both social and cultural capital formation and utilization [6,7], which is also a dynamic relating to the concept regarding ‘weak ties’ [2]. These ‘weaker’ ties oftentimes provide more advantages within one’s personal social network, and while the strength of one’s connections cannot always be effectively measured, we observe variety with as many different occupational connections as the data can provide, and give importance to those which perhaps bridge status boundaries.

The survey question that we used for this network variety measure of social capital is ‘Do you know a woman or man who is a(n) X’. This is the only survey year that the GSS has included a question of this sort. The occupations that the GSS includes are bus driver, cleaner, police officer, hair-person, executive, personnel manager, lawyer, car mechanic, nurse, and school teacher. This measure also helps to get to bridging forms of social capital since we are explicitly measuring cross status connections within social networks. However, this question was only asked of the respondent (it was not asked about the respondent’s parents).

3.2.2. Parental Status Connection Variables

To get at calculating parental status we did two things. First, we took both of the occupational status scores that were listed for each of the respondent’s mother and father, and compressed the numerical ranks into three approximately equal categorical variables, which roughly corresponded to low, medium, and high-status occupations. We then compared these parental occupational status categories against each other to form six different categories of parental status combination. These six are low-low status, low-medium status, medium-medium, low-high, medium-high, and high-high. The size of these six categories were approximately 16%, 31%, 23%, 7%, 17%, and 6%, respectively. There are therefore three inter-status categories and three categories representing intra-status parental connection.

The second way that we measured parental status is through parental education. Parental education levels were again compressed into three categories, one representing a high school degree or less as the parent’s highest level of education, one category for an associate’s degree, and one for a university degree or higher. Since respondents are adults, the average age of their parents is above the average age for the population. We therefore argue that while others have argued that postgraduate education is in many respects the new undergraduate education for social status [49], we argue that undergraduate education is sufficient for the higher status category of education for this paper since the surge in undergraduate enrolment generally post-dates the generation of respondents’ parents. This approach again produced six categories: high school-high school (HS-HS), HS-Associates, Associates-Associates, HS-Undergrad (UG), Associates-UG, and UG-UG. The size of these six categories were approximately 64%, 6%, 1%, 13%, 3%, and 13%, respectively. We thus have three inter-status categories for education and three intra-status categories.

3.3. Methods of Analysis

For the analyses in this paper, we primarily utilize two methods of statistical analysis. The first is Latent Class Analysis (LCA), which is a method of statistical analysis used for identifying subgroupings within datasets that share superficially hidden, or ‘latent’, characteristics and patterns [50]. We use LCA specifically for its efficacy in identifying groups within our data regarding social networks of occupational connections. LCA is able to produce both how many distinct groups most coherently exist within the data and the typographical profiles of these groupings. This method of analysis therefore allows us to assess how many different groups exist for occupational social network variety for the GSS survey and the characteristic nature of these groups.

After using LCA to define social network groups within the data, we then perform multinomial logistic regression analyses with these social network groups as dependent variables. In these regression models, independent variables include, separately and controlling for each other, the two measures of parental status that we use, which are occupational status and education. We also subsequently include socio-economic and demographic variables within an additional regression model to see if the relationships between social network variety and parental status are consistent when controlling for other factors that might themselves mediate the accumulation or possession of social capital through network variety, and thus might mitigate the predictive power of parental status. For this analysis we use family income, highest personal level of education, race, age, and current sex as these controlling variables.

4. Results

4.1. Latent Class Analyses for Social Network Variety Measure

We first outline the results of the latent class analysis for the social network variety of the GSS data. For choosing which latent class model best fits the data, we rely on using the Bayes Information Criterion (BIC) and the Consistent Akaike Information Criterion (CAIC). These statistics convey best fit when they reach their optimal, that is lowest, point for a model solution. In particular, the Bayes Information Criterion has been argued to be the most statistically rigorous measure and thus an oft-preferred statistic for such models [51]. The LCA model fit statistics can be viewed in

Table A1. Model fit statistics for latent class analysis of social network variety.

		LL	BIC(LL)	CAIC(LL)	L2	p-Value	Class. Err.
Model1	1-Cluster	−7253.8173	14,578.3371	14,588.3371	2559.0181	1.60 × 10−134	0
Model2	2-Cluster	−6610.0737	13,368.6225	13,389.6225	1271.5308	1.30 × 10−8	0.0706
Model3	3-Cluster	−6523.4451	13,273.1382	13,305.1382	1098.2738	0.0096	0.1555
Model4	4-Cluster	−6456.5515	13,217.1237	13,260.1237	964.4866	0.63	0.1947
Model5	5-Cluster	−6435.761	13,253.3155	13,307.3155	922.9056	0.85	0.2068
Model6	6-Cluster	−6421.657	13,302.8801	13,367.8801	894.6974	0.93	0.2074

(see Appendix A).

As introduced earlier, there are a total of ten occupations included in the GSS data, and we use all ten as active variables in the LCA to identity social capital groups representing social network variety. With the model of best fit suggesting that there are four groups which best fit the data,

Table 1. Latent class profile of social capital network variety ^a.

	Cluster1	Cluster2	Cluster3	Cluster4	Relative Frequency
Cluster Size	0.3159	0.3044	0.2122	0.1675
Bus Driver
Yes	0.3487	0.6731	0.1590	0.1842	0.3796
No	0.6513	0.3269	0.8410	0.8158	0.6204
Cleaner
Yes	0.4553	0.8188	0.4702	0.2224	0.5301
No	0.5447	0.1812	0.5298	0.7776	0.4699
Police Officer
Yes	0.5840	0.9518	0.6184	0.1655	0.6331
No	0.4160	0.0482	0.3816	0.8345	0.3669
Hair person
Yes	0.7863	0.9832	0.7037	0.1943	0.7295
No	0.2137	0.0168	0.2963	0.8057	0.2705
Executive
Yes	0.2073	0.6279	0.6801	0.1060	0.4187
No	0.7927	0.3721	0.3199	0.8940	0.5813
Personnel Manager
Yes	0.3957	0.8733	0.6534	0.1525	0.5550
No	0.6043	0.1267	0.3466	0.8475	0.4450
Lawyer
Yes	0.2715	0.8668	0.9798	0.1451	0.5818
No	0.7285	0.1332	0.0202	0.8549	0.4182
Car Mechanic
Yes	0.8446	0.9474	0.5406	0.2867	0.7179
No	0.1554	0.0526	0.4594	0.7133	0.2821
Nurse
Yes	0.8763	0.9960	0.9196	0.3974	0.8417
No	0.1237	0.0040	0.0804	0.6026	0.1583
School teacher
Yes	0.8019	0.9856	0.9397	0.3210	0.8065
No	0.1981	0.0144	0.0603	0.6790	0.1935

^a Figures in bold face type are representative of those probabilities for inclusion in the latent class profile that exceed the overall relative frequency for the sample. Underlined figures are those which are near the overall relative frequencies for the sample.

presents the substantive LCA profile of these groups. The size of the four groups represents approximately 32%, 30%, 21%, and 17% of the survey sample, respectively. To interpret these groups, we rely on comparisons between the probabilities that members of a specific group know someone from a specific occupation and the relative frequency of knowing someone from that occupation, with special attention paid to when these probabilities exceed the relative frequencies for the overall sample [52].

The first group of the four is characterized by probabilities of having car mechanics, hair people, and nurses in their social networks that are above the relative frequencies of knowing those occupations in the sample. This group also has probabilities of knowing bus drivers, cleaners, police officers, and school teachers which are near the relative frequencies for the same and higher than other groups, while probabilities of knowing personnel managers, executives, and lawyers are well below the relative frequencies for the sample. For these reasons this first group consists of those who have a ‘primarily blue collar’ social network. While each occupation will have higher and lower ranks within them, this group primarily consists of those with occupational networks traditionally considered ‘blue collar’, with contacts of generally lower occupational prestige.

The third latent class group shows a relationship that mirrors this first group in ways in which I characterize then as having ‘primarily white collar’ social networks, including occupations such as knowing an executive, personnel manager, and lawyer. While again containing varying ranks inherent to each occupation, this group primarily contacts those with elevated occupational prestige.

Using the same statistical rationale, the second and fourth latent class groups are characterized by knowing many and few of these occupations, respectively, and for these reasons those in these groups are labelled as having ‘varied’ and ‘narrow’ social networks. Therefore, those in the second group have many contacts across the occupational prestige spectrum, while the fourth group has few of these contacts. Therefore, in order, we have labelled these four groups the ‘primarily blue collar’, ‘varied’, ‘primarily white collar’, and ‘narrow’, respectively.

4.2. Regression Analysis

The first stage of our regression analysis assesses the relationship between each of our two parental status variables and our measure of social network variety (see

Table 2. Log odds ratios from regression analysis of social network groups and parental occupational and educational status. ^a

	Narrow		Primarily		Primarily
			White-Collar		Blue-Collar
	Gross Effects	Net Effects	Gross Effects	Net Effects	Gross Effects	Net Effects
Parental Occupational Status
High-High	−0.138	0.053	0.567	−0.488	−0.701	−0.178
Medium-High	−1.462 **	−1.237 *	0.265	−0.393	−0.971 **	−0.675
Low-High	−1.061	−0.714	−0.118	−0.452	−1.011 *	-0.876
Medium-Medium	−1.693 **	−1.452 **	0.138	−0.080	−0.640	−0.339
Low-Medium	−0.874 *	−0.702	−0.425	−0.524	−0.677*	−0.566
Low-Low	---		---		---
Parental Educational Status b
UG Degree-UG Degree	−0.366	0.212	1.400 ***	1.619	−1.010 **	−0.597
Associate-UG Degree	−2.062	−1.399	0.255	0.007	−0.815	−0.447
High School-UG Degree	−0.999 **	−0.705	0.359	0.422	−0.111	0.186
Associate-Associate	−2.585	−1.942	−1.717	−1.326	−1.185	−0.718
Associate-High School	−1.596	−1.148	−0.268	−0.078	−0.754	−0.731
High School-High School	---		---		---

^a Reference group = ‘varied’ social network. ^b UG Degree denotes those who have received an undergraduate degree or higher. * p < 0.05; ** p < 0.01; *** p < 0.001.

). As the results displayed in the table show, both parental occupational status and parental educational status show statistically significant predictive relationships for social network groups. First, compared to those with narrow social networks, those with varied social networks are much more likely to have parents who are either cross-status or medium status in their occupations than they are to have parents who were both of low occupational status. However, there is no predictive statistically significant difference between narrow and varied networks for those with two parents of high occupational status and low occupational status. While there are no statistically significant results between broad and white-collar networks for parental occupational status, there are similar, although slightly less pronounced between parental status and blue-collar networks, where those with cross-status parents or two parents of middle status are much more likely to have varied social networks than primarily blue-collar social networks compared to those with two low status parents.

With regard to parental educational status, compared to those with two parents of lower occupational status, those whose parents have mixed educational status or two associates degrees appear to be much more likely to have varied social networks than narrow ones, although this relationship is only statistically significant for the cross-status relationships of high school-associate degree and high school-undergraduate degree. It is again noteworthy that the highest homogenous parental status category, those whose parents both have undergraduate degrees, again does not show as strong or significant a relationship as the other combinations, save for the lowest homogenous status category. However, this highest status group shows strong statistically significant results to be more likely to have primarily white-collar social networks than varied social networks and varied social networks than primarily blue-collar networks.

With both parental occupational status and educational status showing significant results for predicting social network variety, we then included both variables to directly compare how predictive they are when controlling for each other (see Table 2). When controlling for each other, however, and while two high education parents is still highly predictive of white-collar social networks, it is parental occupational status that remains predictive of varied compared to narrow social networks (while statistically significant results for parental education fall away). We therefore included occupational status within a full regression model to see if other respondent socioeconomic or demographic variables mitigated these parental effects.

When other socioeconomic and demographic variables are included in a regression model, which are variables that in theory may also have an influence on the social network variety of respondents, we generally find that the statistically significant results of parental occupational status have been replaced by stronger factors, while the general non-statistically significant predictive direction holds (see

Table 3. Log odds ratios from regression analysis of social network groups including both parental occupational status and relevant socio-economic and demographic variables. ^a

	Narrow	Primarily	Primarily
		White-Collar	Blue-Collar
Parental Occupational Status
High-High	0.766	0.086	0.124
Medium-High	−0.685	0.069	−0.546
Low-High	−0.396	−0.018	−0.529
Medium-Medium	−1.101 *	0.222	−0.519
Low-Medium	−0.410	−0.367	−0.446
Low-Low	---	---	---
Education
Graduate degree	−0.520	1.521 ***	−0.748
Undergraduate degree	−0.608	0.768 *	−0.680 *
Associate’s degree	−1.190	0.661	−0.165
High school degree or less	---	---	---
Family income
$130,000+	−0.969	1.318 **	−0.518
$90,000–129,999	−2.440 ***	0.256	−0.850 *
$60,000–89,999	−1.338 **	0.238	−0.104
$40,000–59,999	−0.501	−0.029	0.171
$25,000–39,999	−0.809	−0.134	−0.412
Less than $25,000	---	---	---
Race
Other	1.713 **	1.232 *	1.190 *
Black	−1.238 *	−0.149	−0.395
White	---	---	---
Sex
Male	0.516	−0.084	−0.132
Female	---	---	---
Age
18–24	0.681	−0.215	0.804
25–34	−0.500	−0.551	−0.005
35–49	−0.956	−1.415 ***	−0.205
50–64	−0.905	−0.892 *	−0.300
65+	---	---	---

^a Reference group = ‘varied’ social network. * p < 0.05; ** p < 0.01; *** p < 0.001.

). The one exception to this is that having parents that both have medium occupational status is still predictive of varied versus narrow networks compared to those with two parents of low occupational status. Beyond this, family income and race are most predictive of varied versus broad social networks, with those of upper-middle family income, those of ‘Black’ races and ethnicities, and ‘White’ versus ‘Other’ races and ethnicities more likely to have varied compared to narrow social networks. Finally, those of ‘Other’ races and ethnicities, relatively lower education, and undergraduate education were more likely to have primarily blue collar than varied networks while elevated levels of education, income, age and ‘Other’ races and ethnicities were more likely to have primarily white collar than varied social networks.

5. Discussion

Through the research in this article, we have endeavored to contribute to the existing literature on social networks and social capital by investigating a novel potential influence upon the social capital resource that is social network variety. This potential influence was that of parental status, both occupational and educational. To this end we asked two research questions. First, are parental cross status ties predictive of increased social network variety? Second, if this is the case, is this relationship mitigated by socio-economic or demographic effects?

In answer to our first question, we found that parental status generally was predictive of more varied social networks compared to more narrow networks. While for parental educational status this was true only at the cross-status level, for parental occupational status this was also true at the intra-status level of ‘medium’ occupational status. The results for parental education rather suggest that intra-status connections at the highest levels of education may contribute to network closure and the reproduction of more narrow social networks of higher average status than contribute to the forming of diverse, bridging social connections. In this way the results around parental education end support to Bourdieu’s [19] assertions of the social inequality reproduction function of social capital, even as described in this paper more diverse social networks may ultimately convey more social capital. We, however, also mostly focus on occupational status here as it proved the more predictive variable when both parental occupational and educational status were controlled for each other in the analysis of this paper.

Although we do not have information about parental social networks from this data, the medium intra-status parental connection may suggest that this is an important starting point from which one can make bridging social connections across occupational status in both directions without the symbolic or spatial difficulties that traversing larger prestige gaps may confer. Likewise, children growing up with medium status parents may likewise not feel or experience symbolic or spatial boundaries with those of other status groups. Also, since those from more medium (compared to lower) status backgrounds are both more able to be upwardly socially mobile and more able to form bridging social ties [44], this also suggests support for Stephany’s [42] findings regarding the intergenerational inheritance of social capital contributing to social mobility. Similarly, these results suggest support for Lin’s [10] arguments around the difficulty in forming bridging social connections across large social inequality gaps.

The importance of either medium or cross status connection is further supported by the findings of this paper that both high and low intra status parental connections have less varied social networks. This comes in the form of either narrow networks, primarily either white collar networks or blue collar networks. Therefore, cross status parental relationship, themselves signifying bridging forms of social capital and connection, may indeed be an important factor in their children’s ability to form bridging forms of capital as defined by Putnam [8], and heterophilous networks [11], of which network diversity is a defining feature.

In answer to our second research question, much of the effects of parental status upon social network variety is mitigated by personal socio-economic and demographic factors. This is true except for where intra-status parental occupational status was still a prominent driver of varied compared to narrow social networks, although still not as predictive as certain income levels ($60,000–$129,999) and racial categories (Black vs. White and White vs. Other). While personal education was predictive of having either primarily blue collar or white collar social networks, education was not a strong predictor of narrow rather than varied networks. The results of the research of this paper may therefore be suggestive of the more inclusionary role of income in accessing diverse spaces where bridging connections are made and bridging forms of social capital cultivated, while also suggest a more ‘homophilous’ (e.g., [11]) role for personal education within social networks. Ultimately, these results suggest that it is rather certain personal resources and characteristics (i.e., income and race), rather than bridging bonds among parents, which contribute most to bridging social capital and the formation of such social network connections for individuals.

Finally, our findings echo Lin’s [33] insights into the complex relationship between social and human capital, where the presence of diverse parental ties does not guarantee enhanced social capital in children. Our study’s limitations preclude a definitive understanding of the mechanisms at play, particularly concerning how parental human capital might complement social capital in status attainment [53,54].

The limitations of this paper include an inability to fully understand the precise mechanisms by which parental status might convey, at least to some degree, the necessary conditions and behaviors to increase the likelihood of eventual network variety of their children. Conversely, where we see other socio-economic and demographic variables being more predictive of social network composition than parental status, we are likewise unable to fully attribute mechanistic reasons for this. Finally, although using ten occupations is a common stopping point for survey programs that do ask about occupational social networks, any study of social network variety would benefit most greatly by being able to use the largest variety possible for ascertaining the composition of these networks.

6. Conclusions

In conclusion, we sought to add insight into the field of social networks and social capital by examining the influence of parental status on social network variety. The findings suggest that parental status, particularly occupation medium and cross-status connections, can play a significant role in fostering diverse social networks. However, personal socio-economic and demographic factors largely mitigate the effects of parental status on social network diversity. These results highlight the complex interplay between parental ties and individual characteristics in shaping bridging social capital. Ultimately, our research highlights the potentially impactful role of parental bridging connections but underscores the primary importance of personal resources and attributes, especially income and race, in facilitating or limiting diverse social network connections and the cultivation of bridging social capital.