As part of a longitudinal research effort that examines the influence of social capital on differential persistence and retention among undergraduate engineering majors, this study examines how engineering degree-related social capital differs for first-year engineering students by gender and ethno-racial groups. Social capital is operationalized as a person's network of relationships with individuals who hold influential positions (e.g. parent, teacher, advisers) and access to resources that support persistence in engineering programs. Our data comprise survey responses from 2186 first-year engineering students, from eleven diverse colleges and universities, who provided information about their participation in engineering-related courses, activities, and programs while in high school as well as the individuals who influenced their decision to pursue an engineering major. We found few differences in social capital between men and women, but found many differences among ethno-racial groups, which suggests that different levels of social capital could influence students' persistence and retention.
Keywords: Social capital; persistence; retention; ego-networks; minority/majority groups
In the United States, women and minorities continue to be underrepresented in science, technology, engineering, and mathematics (STEM) fields compared to the population as a whole. The National Science Foundation (NSF [
Numerous studies suggest that while academic preparation and socio-economic status are contributing factors, the low numbers of women and URM in STEM are likely due to implicit factors unrelated to aptitude and interest (Margolis, Fisher, and Miller [
Previous research suggests that students' entering levels of social capital may influence academic outcomes. For instance, URM students might not know that academic preparation for the sciences and engineering may need to start by middle school (Hilton et al. [
Tyson, Smith, and Ndong ([
Likewise, experiences of engineering majors during their first years in the program are important in persistence (Tyson, Smith, and Ndong [
STEM department cultures that did not espouse student-teacher relationships because of the 'weed-out system' appeared to be the biggest difficulty for women in their early years. Seymour and Hewitt ([
Collectively, the research cited above reveal factors beyond academic preparation that influence outcomes for women and URM, which suggest that they must acquire social capital and receive guidance and strategies from more knowledgeable persons to navigate their engineering programs and address challenges that are encountered. Because of this, it is important to understand the role that social capital plays in students' experiences and how they acquire additional social capital as they engage in their engineering programs. Thus, the research question guiding this study is, what systemic differences in social capital acquired prior to enrolling as engineering majors are there between women and URM and dominant groups?
To understand factors in persistence that include social connections and the environment, STEM researchers are increasingly studying social capital (see for example, Foor and Walden [
The first conceptualization of social capital emphasizes the networks of social ties in which an individual is embedded. Coleman ([
The second conceptualization of social capital emphasizes the resources accessible to an individual through their participation in organizations and/or institutions that have as part of their mission making available such resources. For example, in many U.S. high schools, school activities and programs, clubs, study groups, tutoring, and so on constitute resources that students can choose to access and activate. This form of social capital is less about the particular ties that a person has and more about the resources that are available through social group memberships. Small ([
This study adheres to Lin's ([
This study is part of a longitudinal research effort that examines how social capital contributes to the retention and degree attainment of women and URM majoring in engineering. Our research combines network-based social capital with participatory social capital. We believe that resources must be available, accessible, and, finally, activated or taken advantage of in order for positive outcomes to occur and for an individual's social capital to develop. Thus, we examine the role played by an individual's network in the provision of available, accessible, and activated resources while they were high school students as it relates to their gender and ethno-racial group membership. A major premise of our work is that success in any endeavor requires appropriate knowledge and resources. This study addresses how women and URM may differ, as compared to dominant groups (e.g. White and Asian men) with respect to availability, access, and activation of engineering resources. Specifically, we ask about the availability, access, and activation of resources contingent on participation and assess students' personal social capital by the associates they view as influential on their career choice.
We administered an online survey that inquired about respondents' participation in STEM-related activities and programs and their experiences in STEM-related courses while in high school. We also asked respondents who they felt were influential on their decision to select engineering as a major. In the technical language of social network analysis, each respondent is an 'ego' and the persons they name as influential are 'alters' with the connections from ego to his or her alters constituting an 'ego-network' (Perry, Pescosolido, and Borgatti [
The sample is a cohort of engineering students who responded to the survey from 11 universities serving diverse student populations (7 predominantly white institutions, 3 Hispanic-serving institutions, and 1 historically Black college and university). We used purposive sampling to gather information from diverse engineering programs and contexts. In this study, gender is operationalized as a binary variable and ethno-racial identity choices mix conventionally defined race categories (Black, Asian), ethnic categories (Hispanic), and national origin categories (Japanese, Cuban). We acknowledge the scholarship that denies the binary nature of gender (Oakley [
Table 1. Ethno-racial grouping and gender of respondents*.
Ethno-racial Grouping Woman Man Total % % % White 335 15.4 736 33.8 1071 49.2 Black/African American 55 2.5 76 3.5 131 6.0 Hispanic 158 7.3 371 17.0 529 24.3 Cuban 9 0.4 31 1.4 40 1.8 Mexican 12 0.6 18 0.8 30 1.4 Puerto Rican 97 4.5 238 10.9 335 15.4 Other Hispanic 40 1.8 84 3.9 124 5.7 Asian 114 5.2 236 10.8 350 16.1 Asian Indian 32 1.5 80 3.7 112 5.1 Chinese 46 2.1 78 3.6 124 5.7 Filipino 9 0.4 13 0.6 22 1.0 Japanese 0 0.0 8 0.4 8 0.4 Korean 8 0.4 11 0.5 19 0.9 Vietnamese 5 0.2 8 0.4 13 0.6 Other Asian 14 0.6 38 1.7 52 2.4 Other 35 1.6 62 2.8 97 4.5 American Indian/ Alaskan Native 2 0.1 6 0.3 8 0.4 Hawaiian/Othera 1 0.0 2 0.1 3 0.1 MidEast/Arabb 11 0.5 32 1.5 43 2.0 Other 21 1.0 22 1.0 43 2.0 Missing 3 - 5 - - - Total 700 32.0 1486 68.0 2178 100.0
*Eight cases NA on ethno-racial grouping.
Table 2 presents the three social capital variables used to analyze the data. Set I includes variables that measure participatory social capital on the dimensions of availability, accessibility, and activation. The first subset asks about participation in specific programs and if there was no participation, whether that was due to lack of availability or lack of activation. The second subset of variables asks about the availability of various general types of activities (rather than specific programs). The third subset probes the level at which a respondent activated his or her participation in available activities. When available, participation depends on other factors, such as having the finances to access the activities/programs and related resources. Finally, when a program is available and accessible (i.e. no barriers to participation), the student may choose to not activate participation or may participate at lower versus higher levels. Set II consists of variables that measure the ego-network of a respondent in terms of the statuses the respondent identifies as influential in their decision to pursue engineering (alters), including variables intended to probe how alters were influential in providing resources and advice.
Table 2. variables used in analyses.
Set I IA. Program participation (Responses: Yes, No (not available), No (available), Uncertain if available) 1. Advanced Placement (AP) 2. AVID (Advance in Individual Determination) 3. Dual Enrollment 4. Duke University Talent Identification Program (Duke TIP) 5. Engineering courses (e.g. Science Technology Engineering & Mathematic [STEM] academies) 6. Gear Up 7. International Baccalaureate (IB) 8. MESA (Math, Engineering, Science Achievement) 9. Project Lead The Way 10. Talent Search 11. Upward Bound 12. Other programs, please specify IB. Activity availability (Responses: Yes, No, Don't know) 1. Math or science camp 2. Math or science club/competition/fair 3. Math or science study group of any kind 4. Math or science tutoring (as a tutor or tutee) 5. Robotics club/competition 6. Reading science books/magazines 7. Visiting science museums, planetariums, or environmental centers 8. Making industry tours and visits 9. Oral and/or writing skills development 10. Visiting STEM-related web sites (e.g. Gizmo, NASA) 11. Other activities, please specify IC. Activation as frequency of participation [Responses: Never (0), Rarely (1), Sometimes (2), Frequently (3), or Very Frequently (4)] 1. Math or science camp 2. Math or science club/competition/fair 3. Math or science study group of any kind 4. Math or science tutoring (as a tutor or tutee) 5. Robotics club/competition 6. Reading science books/magazines 7. Visiting science museums, planetariums, or environmental centers 8. Making industry tours and visits 9. Oral and/or writing skills development 10. Visiting STEM-related web sites (e.g. Gizmo, NASA) 11. Other activities, please specify Set II For each alter, respondent was asked about: gender, ethno-racial identity, whether alter is/was an engineer, the number of years ego knew the alter, whether the alter has worked with engineers 1. Parent (or guardian) influenced decision 2. Sibling (or other family member) influenced decision 3. Peer influenced decision 4. High school teacher influenced decision 5. High school counselor influenced decision 6. Club or organization contact influenced decision 7. Other influenced decision Set III Ways in which the alters influenced the respondents 1. Take STEM courses 2. Participate in STEM extra-curricular activities 3. Information about engineering schools or disciplines 4. Encourage to do my best 5. Serve as role model or mentor 6. Checked on academic progress 7. Recognized engineering talents 8. Bought engineering toys, books, etc. 9. Find internships, jobs, scholarships etc. 10. Suggest application for college, financial aid 11. Study for SAT/ACT 12. Visited a college 13. Information about engineering work 14. Advised about overcoming an obstacle 15. Introduced me to someone who gave advice
We compared gender and ethno-racial identity groups with respect to network-based and participatory social capital using t-tests, ANOVA, and χ
Table 3. Program participation by gender, frequencies.
Program Men Women Yes No/NAa No/NPb Unsurec Yes No/NA No/NP Unsure Advanced Placement (AP) 1187 158 90 34 575 80 19 19 Engineering courses (e.g. STEM academies) 408 709 76 196 186 355 58 71 Dual Enrollment 304 438 524 131 159 194 272 48 International Baccalaureate (IB) 131 966 131 155 75 485 59 54 Project Lead the Way 86 977 49 262 29 510 19 109 Other programs, please specify 85 613 32 173 32 274 15 63 Duke University Talent Identification Program (Duke TIP) 79 988 85 224 32 483 49 102 Talent Search 55 974 68 274 20 501 28 116 MESA (Math, Engineering, Science Achievement) 28 1012 48 287 10 518 18 120 AVID (Advance in Individual Determination) 23 1010 141 200 8 511 75 73 Upward Bound 20 977 79 296 6 496 46 119 Gear Up 13 1059 46 252 6 535 13 110
Notes: Differences in distribution of responses by gender over the Yes, No, na; No, np categories are not significant (p <.05) with two exceptions: program 1, men underrepresented in Yes and overrepresented in No, NP and in program 2, men overrepresented in Yes and underrepresented in No/NP.
Table 4. Frequencies of available activities by gender.
Activity Men Women Yes No Unsure Yes No Unsure Math or science club/competition/fair 1097 269 99 524 134 35 Math/Science tutoring (tutor or tutee) 1027 325 95 506 147 40 Oral and/or writing skills development 1008 310 122 500 153 35 Math or science study group of any kind 896 407 148 409 224 55 Reading science books/magazines 889 417 138 439 179 67 Robotics club/competition* 819 548 75 354 292 41 Visiting science museums, planetariums, or environmental centers*** 706 566 170 383 248 58 Visiting STEM-related web sites (e.g. Gizmo, NASA) 536 678 222 261 332 93 Making industry tours and visits 450 783 201 192 410 83 Math or science camp 426 713 300 184 382 121 Other activities, please specify 46 506 237 23 209 95
*A significantly (p <.05) larger proportion of men reported availability of this activity (omitting those unsure). ***A significantly (p <.05) larger proportion of women reported availability of this activity (omitting those unsure).
Table 5. Activity activation by gender.
Activity Mean Frequency of Participation Cohen's d Men Women Math or science club/competition/fair 2.53 1090 2.65 521 1.67 0.09 Math/Science tutoring (tutor or tutee) 2.51 1021 3.06 503 7.53* 0.41 Oral and/or writing skills development 2.97 1004 3.38 497 6.05* 0.33 Math or science study group of any kind 2.51 889 3.00 408 6.50* 0.39 Reading science books/magazines 2.69 888 2.82 438 1.84 0.11 Robotics club/competition 2.05 816 1.74 352 −3.50* 0.22 Visiting science museums, planetariums, or environmental centers 2.54 703 2.79 382 3.77* 0.24 Visiting STEM-related web sites (e.g. Gizmo, NASA) 2.91 532 3.00 261 0.96 0.07 Making industry tours and visits 2.57 448 2.69 190 1.17 0.11 Math or science camp 2.01 421 2.51 182 4.39* 0.40 Other activities, please specify 2.97 34 2.24 21 −1.84 0.51
*Statistically significant, p <.001.
Men and women differ very little on the measures of availability and access with respect to participatory social capital though they differ somewhat more on activation. Only two of the 12 program participation variables show significant differences between men and women, with men less likely to participate in AP courses and more likely to participate in engineering courses. Similarly, men and women differ in only two of the 11 available activities variables, with men being more likely to participate in 'robotics club/competition' and less likely to participate in 'visiting science museums, planetariums, or environmental centers.' However, significant gender differences are far more prevalent in the activation variables. Women reported higher frequencies of participation in five of the 11 activities ('math/science tutoring [tutor or tutee]', 'oral and/or writing skills development', 'math/science study group', 'visiting science museums, planetariums, or environmental centers', and 'math/science camp') while men reported higher participation in only one of the 11 activities ('robotics club/competition'). The effect sizes associated with these statistically significant differences are small in conventional terms.[
Unlike with gender, we found many significant differences for ethno-racial identity on the availability, access, and activation items. Tables 6–8 tabulate responses to the variables in Table 2 that ask about program participation (Set IA), available activities (Set IB), and measure activation in terms of frequency of participation (Set IC). Expected levels of availability are calculated from the marginal distributions of availability and do not refer to levels expected on the basis of substantive considerations related to different social positions of the groups. Table 6 shows that Asian respondents participate more than expected based on the marginal distribution of participation in five of the 12 programs and less than expected in one. White respondents participate more in four of the 12 and less than expected in another four. Black/African American respondents participate more than expected in three of 12 and less than expected in none. Hispanic respondents participate more than expected in three of 12 and less than expected in five of the 12 programs. Table 7 shows that six of the 11 activities are more available than expected for Asian and White respondents while two are less available than expected for Black respondents and eight of 11 are less available than expected for Hispanic respondents.
Table 6. Program participation by ethno-racial grouping: summary
Program/Activity Findings Advanced Placement (AP) 232.6, Asian and White students reported more participation than expected, Hispanic students less AVID (Advance in Individual Determination) 51.1, Black and Hispanic students reported more participation than expected, Asian and White students less Dual Enrollment 197.6, Black and White students reported more participation than expected, Hispanic students less International Baccalaureate (IB) 68.1, Asian and Black students reported more participation than expected, Hispanic students less Engineering courses (e.g. Science Technology Engineering & Mathematic [STEM] academies) 47.9, White students reported more participation than expected, Hispanic students less MESA (Math, Engineering, Science Achievement) 8.2, ns Gear Up 17.9, White students reported less participation than expected Talent Search 21.8, Asian and Hispanic students reported more participation than expected, White students less Upward Bound 25.9, Asian and Hispanic students reported more participation than expected, White students less Other programs, please specify 14.7, ns Project Lead The Way 12.2, ns Duke University Talent Identification Program (Duke TIP) 17.2, Asian and White students reported more participation than expected, Hispanic students less
Table 7. Available activities by ethno-racial grouping: summary
Program/Activity Findings Math or science camp 37.6, More available for Asian and White students than expected, less for Black and Hispanic students Math or science club/competition/fair 21.0, More available for Asian students than expected, less for Black and Hispanic students Math or science study group of any kind 19.6, More available for Asian students than expected, less for Hispanic students Math or science tutoring (as a tutor or tutee) 9.5, More available for Asian and White students than expected, less for Hispanic students Reading science books/magazines 52.2, More available for Asian and White students than expected, less for Hispanic students Visiting science museums, planetariums, or environmental centers 2.9, ns Making industry tours and visits 6.3, ns Oral and/or writing skills development 9.5, More available for White students than expected, less for Hispanic students Visiting STEM-related web sites (e.g. Gizmo, NASA) 17.7, More available for White students than expected, less for Hispanic students Other activities, please specify 4.6, ns Robotics club/competition 76.3, More available for Asian and White students than expected, less for Hispanic students
Table 8. Activity activation by ethno-racial grouping: summary.
Program/Activity Significant group differences ( ANOVA effect size (partial eta squared) Math or science camp 7.12, Asian, Black, and Hispanic student groups have higher scores than White student group 0.05 Math or science club/competition/fair 19.39, Asian and Hispanic student groups have higher scores than White student group 0.05 Math or science study group of any kind 5.50, Hispanic student group has a higher score than White student group 0.02 Math or science tutoring (as a tutor or tutee) 2.26, ns 0.006 Reading science books/magazines 1.84, ns 0.006 Visiting science museums, planetariums, or environmental centers 21.39, Hispanic student group has a higher score than all other groups 0.07 Making industry tours and visits 30.28, Hispanic student group has a higher score than all other groups 0.16 Oral and/or writing skills development 2.64, Hispanic student group has a higher score than Asian and White student groups 0.007 Visiting STEM-related web sites (e.g. Gizmo, NASA) 3.24, Hispanic student group has higher score than White student group 0.02 Other activities, please specify 1,10, ns 0.08 Robotics club/competition 2.48, Asian student group has a higher score than White student group 0.008
These findings suggest Asian and White students are advantaged (i.e. activities are more available to them), when compared to Black/African American and Hispanic students. However, when we examine activation with respect to the same set of activities in Table 8, Hispanic respondents are more engaged with the activities than other groups. They report more frequent participation in seven of the 11 activities than at least one other group and in two cases all other groups ('visiting science museums, planetariums, or environmental centers' and 'making industry tours and visits'). In nine of 11 activities, White students report less frequent engagement as compared to at least one other group (Hispanic students seven times, Asian students three times, and Black students once). The values of partial η
To elicit an ego-network of influencers, respondents were asked to 'THINK BACK to when [they] were considering engineering as a college major while in high school or middle school and indicate the person(s) who influenced [their] decision-making process at that time in some way.' Respondents could pick one or more alters from the following list: (
Table 9 displays the top ten types of ego-networks reported by the respondents. There are 128 = 2
Table 9. Top ten types of ego-networks of influencers.
Ego-network Type Frequency Parent 466 Parent|Sibling|Peer|Teacher|Counselor|Club Contact 214 Parent|Sibling 192 Parent|Teacher 187 Teacher 108 Parent|Sibling|Teacher 84 Sibling 72 Other 65 Parent|Sibling|Peer|Teacher 65 Parent|Peer 64
Tables 10–12 compare the ego-networks of men and women with respect to which alter-types they include, their size, and the top ten profiles in each gender group. As in the analysis of participatory social capital, gender has little effect. Women mention peer as influential more often than expected and men less often than expected. For both men and women, the most frequently cited alter is parent/guardian (cited by 79.3% of all respondents), with other contact as the least frequently cited alter (cited by 8.8% of all respondents). There is no significant difference in the size distributions of the ego-networks, and both gender groups have almost same top 10 ego-network profiles with parent identified as the most common single influencer. Eight of the profiles in the top 10 are the same for both groups. In fact, the first five are identical just differently ordered. The profiles 'other' and 'parent|sibling|peer|teacher' are in the top 10 for men but not the in the top 10 for women, while the types 'peer' and 'parent|sibling|peer|teacher|counselor' are in the top 10 for women but not the top 10 for men.
Table 10. Influential alters by gender.
Alter-status Men Women Yes No Yes No Parent/guardian 1187 299 557 143 High School Teacher 683 803 348 352 Sibling 622 864 309 391 Peera 530 956 219 481 High School Counselor 341 1145 149 551 Club/Organization Contact 280 1206 131 569 Other 143 1343 50 650
Table 11. Network size by gender (N
Size Men Women 1 561 233 2 380 210 3 201 112 4 93 41 5 57 27 6 153 64 7 41 13
Table 12. Top ten types of ego-networks of influencers by gender (N
Men Women Ego-network Type Ego-network Type Parent 337 Parent 129 Parent|Sibling|Peer|Teacher|Counselor| Club contact 151 Parent|Sibling 76 Parent|Sibling 116 Parent|Teacher 74 Parent|Teacher 113 Parent|Sibling|Peer|Teacher|Counselor| Club contact 63 Teacher 67 Teacher 41 Other 52 Parent|Sibling|Teacher 33 Parent|Sibling|Teacher 51 Sibling 24 Parent|Sibling|Peer|Teacher 51 Peer 17 Sibling 48 Parent|Peer 16 Parent|Peer 48 Parent|Sibling|Peer|Teacher|Counselor 15
Tables 13 and 14 compare ego-networks of alters by ethno-racial grouping. With respect to ethno-racial grouping, there are significant differences for three of the seven alter-types. Asian respondents mention school-related alter-types less often than expected as compared to the Black and Hispanic respondents. For four of the alter-types, including the most often mentioned influencer (parent/guardian), there are no differences between ethno-racial groups. Finally, there are significant differences in the mean size of the ego-networks such that Asian respondents have a smaller network on average when compared to the Black/African American, Hispanic, and White respondent groups. There are no significant differences between pairs of these latter three groups.
Table 13. Influential alters by ethno-racial grouping, frequencies.
Alter-status Asian Black Hispanic Other White Parent/guardian 283 99 418 76 861 High School Teacher*** 125 69 245 45 545 Sibling*** 134 63 268 39 426 Peer 111 45 206 31 352 High School Counselor*** 54 40 150 18 227 Club/Organization Contact* 52 35 106 17 200 Other 19 10 56 7 101
*Significantly different at.05 level with Asian students underrepresented and Black and Hispanic students overrepresented. ***Significantly different at.001 level with Asian students underrepresented and White students overrepresented with respect to High School Teacher; Asian students and White students underrepresented and Black and Hispanic students overrepresented with respect to High School Counselor.
Table 14. Frequency of network size by ethno-racial grouping
Size Asian Black Hispanic Other White 1 162 42 172 35 379 2 88 38 136 31 294 3 39 17 83 13 161 4 18 6 29 4 77 5 11 3 31 4 35 6 28 22 61 9 96 7 4 3 17 1 29 Total 350 131 529 97 1071
Respondents were asked to describe whether each alter-type (e.g. parents) influenced them in 15 different areas listed in Table 2, Set III. We measured the amount of alter-type's influence by determining the total number respondents who answered yes for each areas of influence for the whole sample and by gender and ethno-racial categories. Figure 1 shows that the parent/guardian alter-type that has the most influence on average for the entire sample. We found significant differences by gender and by ethno-racial grouping in the amount of influence that alter-types have. Tables 15 and 16 report these results along with Cohen's d for the t-tests and partial η
Graph: Figure 1. Distribution of Amount of Influence by Alter-types in Table 2, Set III.
Table 15. Comparison of alter influence by respondents' gender.
Alter t-statistic Significant group differences ( Cohen's d Parent −0.099, ns 0.005 Sibling/Other Family Member 3115, Women score higher than men 0.22 Peer 0.961, ns 0.08 High School Teacher −0.865, ns 0.06 High School Counselor −0.725, ns 0.07 Club Contact 0.030, ns 0.03 Other −0.358, ns 0.06
Table 16. Comparison of alter influence by respondents' ethno-racial grouping.
Alter Significant group differences ( ANOVA effect size (partial eta squared) Parent 3.146, None 0.007 Sibling/Other Family Member 1.255, ns 0.005 Peer 2.466, None 0.013 High School Teacher 3.804 Hispanic student group scores higher than Asian student group and White student group 0.015 High School Counselor 2.244, ns 0.018 Club Contact 0.349, ns 0.003 Other 0.767, ns 0.037
Consistent with other researchers, we assert that factors other than academic preparation influence outcomes for women and URM in engineering programs (Margolis, Fisher, and Miller [
Although the focus on social capital is a useful framework for our study, some researchers believe such interpretation of social capital inherently focuses on the 'deficits' of women, and in particular, of members of URM population (Yosso [
Earlier, we discussed two possible types of social capital. The first, network-based social capital, highlights the social ties in which a person is embedded (Coleman [
In contrast to network-based social capital, we frame participatory social capital as an approach to make needed resources, not readily accessible from a persons' social network, accessible through their participation in organizations and institutions that have as part of their mission to make available such resources. Such an approach acknowledges the historical and social development of culture that may have excluded individuals from certain groups and allows for proactive measures to support more inclusive and thoughtful practices, thereby providing a means for obtaining the social capital needed to thrive engineering programs. This approach is less reliant on a person's social network and more focused on their promise or potential by intentionally providing opportunities to uncover needed social capital. Participatory social capital expands the social capital of groups by focusing on areas of common interest, growth, and development. Through participation in STEM-related activities, for example, groups of individuals from different socioeconomic and sociocultural backgrounds can access information and resources that can support their decision making and success.
Although men and women differed little on availability and access to STEM-related resources, we found significant differences in activation based on the ethno-racial group. These differences may influence the ability of URM to obtain participatory social capital through engagement in these activities, many of which are designed to provide the needed support to thrive academically and/or generate interest in STEM. This suggests a need to attend to influences on activation as interventions are developed and implemented to support STEM interest, recruitment, and retention and determine approaches for encouraging the engagement of all students.
In the prior section, we assert that engaging in academic and STEM-related activities may be a more productive approach to providing women and URM access to social capital that they may not develop simply from their social networks. To determine the efficacy of such an approach, research is needed to test this hypothesis and examine what types of social capital women and URM gain from participating in activities, particularly those designed to engender interest or success in STEM. Further, many interventions are designed to engender students' interest in STEM and to provide academic supports to encourage their success. Our findings related to students' influencers suggest that the audience for STEM-related interventions should be expanded to include the individuals most likely to influence students' decisions, which include parents and school-based personnel.
We compared women and URM to more the dominant group in engineering, White and Asian men. Although this provided some significant results, the effect sizes were small and therefore the findings are inconclusive. Additional research is needed to examine differences between and among ethno-racial groups. For example, it might be important to compare the experiences of each group independently. Rather than consider White women and URM as a collective group, it might be important for researchers to examine social capital differences between each group against the dominant group as well as differences between White women and URMs. Similarly, researchers should explore differences between women and URM from low- and high-SES backgrounds. Such comparison might highlight nuanced differences that are not visible when the groups are combined. Our results highlight the differences among groups, but do not explain why such differences exist. Research is therefore needed that examines causes behind those differences, particularly as related to activation of available and accessible resources.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
No potential conflict of interest was reported by the authors.
Rebecca Campbell-Montalvo
By John Skvoretz; Gladis Kersaint; Rebecca Campbell-Montalvo; Jonathan D. Ware; Chrystal A. S. Smith; Ellen Puccia; Julie P. Martin; Reginald Lee; George MacDonald and Hesborn Wao
Reported by Author; Author; Author; Author; Author; Author; Author; Author; Author; Author