Exploring EWB-USA Members’ Descriptions of Self, Engineers, and their Fellow Members

Engineers Without Borders-USA (EWB-USA), an organization with over 13,800 members involved in engineering service projects around the world, presents a large population for research on an engineering service activity. Although anecdotal evidence claims that EWB-USA may be helping to change the conversation about the engineering field, additional research is needed. This exploratory study asked EWB-USA members to describe themselves, engineers, and fellow members at regional EWB-USA conferences around the country and analyzed the results (n = 505) using qualitative coding techniques to compare thematic overlap among the responses and to consider gendered differences. Findings showed that engineering and EWB-USA are nearly equivalent components of members’ self-descriptions, and that females showed more overlap with EWB-USA than males. These results can inform future studies on the organizations’ role in impacting changed perceptions of the engineering field and broadening participation. Index Terms – Engineering education, Engineers Without Borders, gender


INTRODUCTION
Engineers are a necessity for addressing the increasing challenges to support the world's growing population, 1 and a lack of gender diversity in science and engineering challenges both the quantity of engineers as well as the problem-solving capabilities of the field; 2 therefore, understanding how to attract additional engineers is critical.To do this, the National Academy of Engineering (NAE) 3 suggests "changing the conversation" when describing what engineering is and what engineers do to include an emphasis "that engineering and engineers can make a difference in the world."In addition to changing the emphasis of what engineers do, Cosentino de Cohen and Deterding 4 recommend "widening the net" for female recruitment by increasing students' exposure to engineering.In order to move away from the dominant perception of engineers as Caucasian males, 5 the profession must consider new ways to recruit a more gender diverse population.
Anecdotal evidence suggests that the organization Engineers Without Borders-USA (EWB-USA) is helping to "change the conversation" by reconnecting engineers to the profession's roots of helping society 6 and to "widen the net" in terms of engaging more female engineers.This 13,800+ member engineering service organization has both student and professional volunteers who have worked on 389 engineering service projects in 47 countries. 7With a focus on engineering service projects for developing communities, and with over 40% female membership, 8 which contrasts with the typical 20% or lower female presence in the field 9 that has been stagnant for over a decade, 10,11 EWB-USA exhibits desirable characteristics that address adjusting the image of the profession and aiding gender diversity within engineering.This exploratory research aims to better understand the members involved with this organization, including their high female involvement.

LITERATURE AND RESEARCH QUESTIONS
Except for a few small-scale studies of individual EWB-USA chapters (e.g. 12,13), EWB-USA's membership has not been well studied.Some reports have made anecdotal claims about EWB-USA members, such as one by UNESCO 1 which states that, "[m]any EWB members are engineering evangelists who are passionate about their profession…" and goes on to say that EWB members are role models, strong communicators, and are capable of interdisciplinary work.Similarly, EWB-USA's website 6 uses words such as "passionate" and "responsible leaders" to describe its members.5][16][17][18] Recently, researchers have begun to bring evidence to anecdotal claims about learning gains from EWB-USA, 19 but the field still lacks evidence about who EWB-USA members are and their perspectives on themselves as engineers.
Common descriptions of engineers tend to highlight stereotypes 3,20,21 which the NAE 3 is trying to improve through changing the messaging around the engineering profession to one "which emphasizes that engineering and engineers can make a difference in the world."This kind of messaging aligns with the roots of the profession as civil servants 7 and with the service work of EWB-USA members, and it suggests that activities such as EWB-USA can help adjust perceptions about engineers.Klotz et al. 22 showed that associating "improving quality of life" and "saving lives" with engineering can broaden participation in engineering, which suggests that engagement with EWB-USA can also broaden participation in engineering.Additional research has shown that even one experience, such as an internship or organizational involvement, can be enough to encourage or discourage students in their engineering identity and ultimately to continue in, or leave, engineering. 23herefore, this study is interested in understanding the alignment of engineering and EWB-USA identities as a means to adjust perceptions of the profession and to broaden participation.Identity has been shown to influence persistence in the field, [24][25][26][27][28] and the alignment of engineering and self identities has been found to be critical for continued participation, particularly for females. 28,29The goal of this exploratory study is two-fold.Our first goal is to understand EWB-USA members' descriptions of self, engineers, and EWB-USA members to better understand which of these three personas overlap with each other.This helps us address whether or not EWB-USA engagement may facilitate changed engineering perceptions.Our second goal is to determine if there are gender differences in these overlaps, which may begin to explain female identification with engineering and/or EWB-USA.Specifically, our research questions ("RQ's") ask: RQ1.How do EWB-USA members describe themselves, engineers, and fellow EWB-USA members?RQ2.Do these descriptions overlap?And if so, how?RQ3.Are there gender differences in terms of how EWB-USA members describe themselves, engineers and fellow EWB-USA members?

DATA COLLECTION METHODS
Data collection took place at five EWB-USA regional conferences held in the fall of 2011.These conferences gathered both student and professional EWB-USA members from chapters within designated regions (Great Lakes held in Champaign, IL; Midwest held in Rolla, MO; Mountain held in Tucson, AZ; Northeast held in New York, NY; and Southeast held in Jacksonville, FL) and offered an opportunity to gather information from a large, widespread collection of EWB-USA members.Among the five conferences, 740 members attended, which represented about 6% of the organization's total reported members.
At the conference, all members present at the large group session were asked to participate in the study, those who agreed were provided a sheet of paper with colorful, blank boxes on it.During the conference, a workshop facilitator would ask an open-ended prompt and would ask participants to provide their response to that prompt within a particular colored box.Members were then given two to three minutes to write an open-ended, unique response that was not constrained by pre-determined answers, as in a typical survey.The bottom of the sheet collected basic demographic information (gender, year of birth, organizational membership, and professional or student status).This paper analyzes the responses to three prompts ("P's"), which asked respondents to: P1.Describe yourself P2.Describe a typical engineer P3.Describe a typical EWB-USA member We refer to these three prompts as the three "personas" to identify the type of person that participants were describing in their responses: the self, the engineer, and the EWB-USA member.

ANALYSIS METHODS, RESULTS & DISCUSSION
After the responses were collected, they were transcribed into a spreadsheet for each of the five conferences.We used two main techniques to analyze the data after qualitative coding-(1) response percentages and (2) thematic overlap analyses.This second technique, developed by the authors for this study, used two sub-techniques to compare the themes used to describe the three personas-one for the number of people that had overlapping themes and another for the number of themes with overlap.Each of these techniques is described below and is followed immediately by its results and discussion.Figure 1 provides an illustration of the research methods for clarity along with corresponding results and research questions.

Qualitative Coding
The spreadsheets from each conference were imported into qualitative coding software, QSR NVivo 10. 30 NVivo allowed us to code the responses based on emergent themes.First, we read all participants' responses to each prompt and created initial themes based on responses with at least five occurrences.Responses with similar meanings were coded into the same theme.For example, participants often described themselves as "motivated," which we grouped with similar responses such as "hard-working" or "ambitious."For consistency between spreadsheets and researchers, a coding dictionary was kept throughout the coding process.The first author coded the questionnaires through three iterations to ensure that all responses were coded using the final coding dictionary.Meetings with two other members of the research team were held after each iteration to discuss the dictionary and any coding uncertainties.Once saturation was reached, the themes were reduced to those that had emerged both at least ten times in one of the three prompts and at least once in one of the other two prompts.In the end, there were 27 themes for P1, 28 themes for P2, and 30 themes for P3.To ensure reliability, a research assistant then performed an inter-coder reliability check on one of the five data sets.Out of the 85 themes, the lowest percent of coding agreement was 92.8% with 77 themes (> 90%) having over 98% agreement.These results provided adequate coding agreement to continue the analysis.
Among all three personas, 24 themes were common, excluding the "other" theme.Because this research focused on the comparison of themes by persona, we used only the 24 themes common among all three personas for comparisons, and responses for any themes that were not common among all three personas were added into the "other" theme to account for all responses in the calculations.One exception was that-in addition to "engineer" being one of the three personas analyzed-a response theme of "engineer" also emerged from the data.This theme was used to collect responses in which respondents described their self explicitly as an engineering student or professional, or when they described an EWB-USA member as an engineer explicitly.Naturally, this theme did not emerge as a response for the persona of an engineer; however, because we were focused on comparisons among these three personas, the use of "engineer" as an explicit response was an important theme to include in the analysis.

Response Percentages
After the data were coded, the researchers' first analytic technique used a unit of analysis of the respondent to determine and compare the percentage of respondents that had indicated at least one response to each theme for self, engineer or EWB-USA member.The percentage was calculated by dividing the number of people responding to that theme by the total number of people who had responded to the prompt.Table I lists the final themes, their response percentages for each of the three personas, and examples of key words coded into each theme.Results in the table are ranked by the percentage for the first persona, the self.The top three responses for the description of self were "motivated," "humanitarian," and "passionate."The top two response themes for self-descriptions, "motivated" and "humanitarian," were reported by nearly twice as many people as other themes.The top three responses for the description of the engineer were "analytical," "smart," and "problem solver."While other themes such as "creative" and "motivated" were frequently reported, "analytical" stood out from the other themes nearly twice as much.The top three responses for the description of an EWB-USA member were "humanitarian," "motivated," and "passionate."These were the same top themes as the description of self; however, the "humanitarian" theme for an EWB-USA member was reported by nearly twice as many people as for the self.

TABLE I PERCENTAGE OF RESPONDENTS WITHIN CODING THEMES FOR THREE PERSONAS (%)
Three themes-"motivated," "passionate" and "goal-oriented"-were commonly used by EWB-USA members to describe themselves and convey a similar type of work ethic and energy.These themes also align with existing literature that notes the passion of these engineers 1 and EWB-USA's website. 6Another important description was "humanitarian," which captured these members' sense of kindness and compassion for other people.Compared with the descriptions of the engineer and the EWB-USA member, both of which showed one dominant theme, respondents' descriptions of self were much more evenly spread, indicating a wider range of self-descriptions than descriptions for the other personas.
EWB-USA members generally described the typical engineer in a positive manner, using themes such as "analytical," "smart," "creative," etc.However, some people described engineers in the "narrow" theme, which captured a wide variety of responses such as nerdy, being socially awkward, being closed-minded, etc.These descriptions align with a stereotypical view of engineers that is common in society. 3With only a fifth of the respondents using this theme, most EWB-USA members described a more positive view of engineers-one that highlighted their intellect and skills.
In describing the typical EWB-USA member, respondents focused on the "humanitarian" theme.This response stood out among the other themes indicating some agreement among the group.The other top themes of "motivated" and "passionate" stress the energy and work ethic of these members, and the "global perspective" theme captures the global context and interest within this international organization.Interestingly, only "motivated" overlapped as a response theme in the top five responses across all three personas indicating the importance of this theme.

Thematic Overlap
For the second analytic approach, we compared the responses for each theme among all three personas for each respondent.To do this, we first created a table that listed each of the 505 respondents in the first column and listed each of the themes as subsequent column headers.A dichotomous coding scheme was used for each theme, where each respondent received a code of "1" if he or she had given a response within that theme or "0" if he or she did not.For example, one respondent, said that she was "giving, organized, and liked math," so she had a "1" in three themes-"humanitarian," "organized," and "analytical"-and a "0" in the remaining 22 themes.If respondents answered with two similar words that had been coded to the same theme (such as hardworking and driven), the cell for that theme still contained a "1," indicating that at least one response was provided for that theme.
Similar tables were created for the responses to the other two prompts, description of engineer and description of EWB-USA member.Each cell was similarly dichotomized for presence or absence in a theme; however, in the table for the description of the engineer, all cells were coded with a "2" (instead of "1") and, in the table for the description of the EWB-USA member, the cells were coded with a "4" (instead of a "1").The values of the cells were then summed for each respondent across the three descriptions in a fourth "overlap" table.The dichotomous coding values within each description table were chosen so that the sums would indicate distinct overlap between the three personas for each respondent.Using this method, the overlap table indicated whether or not respondents described themselves using similar themes as they used to describe engineers or EWB-USA members.For example, a "3" in the overlap table would indicate that the particular individual had given a description falling into the same particular theme for both himself and an engineer (a "1" in the self table plus a "2" in the engineer table).A visual of the overlapping personas and values is illustrated in Figure 2. The four values of most interest were those that indicated overlap between personas: 3 (self and engineer), 5 (self and EWB), 6 (engineer and EWB), and 7 (self, engineer and EWB).These four are referred to as the "overlap types."

FIGURE 2 VENN DIAGRAM INTERPRETATION OF OVERLAP
This method allowed us to determine if individuals described themselves in similar ways to how they described an engineer or an EWB-USA member.Because respondents often wrote descriptions falling into multiple themes, an individual would often have multiple overlap types.For example, a respondent could have indicated that she is "smart, organized, and motivated," that an engineer is "smart, analytical, and curious," and that an EWB member is "smart, curious, and humanitarian."In our coding structure, she would have a "1" in the "motivated" and "organized" themes (self only), a "2" in the "analytical" theme (engineer only), a "4" in the "humanitarian" theme (EWB only), a "6" in the "curious" theme (engineer and EWB), and a "7" in the "smart" theme (self, engineer and EWB).With these results, she would have two types of overlap: (1) between her perceptions of an engineer and an EWB-USA member and (2) between herself, engineering, and EWB-USA.
For each of the four overlap types, a unique spreadsheet column was created that used a logic statement to count individuals based on whether or not they had written overlapping descriptions.For example, for the overlap of the self and engineer, we could not simply count the number of individuals with any "3's" because this would not account for those who had overlap between all three categories (a "7") or those who had described themselves explicitly in the "engineer" theme.Therefore, the logic statement assigned a "1" to individuals that had a "3" or "7" in any of the 24 main themes or had a "1" in the "engineer" theme, indicating that he had described himself as an engineer explicitly.This combination accounted for all of the potential ways that individuals could have indicated an overlap between their self and engineer descriptions without double counting an individual in the results.A similar pattern of logic was used for the other three types of overlap, which is summarized in Table II.Figure 3 presents a visual summary of the method.Once all four overlap types were analyzed, we used two different approaches to count the overlaps.The first counted the number of people with overlap (the "breadth" of overlap) as described above.For this, the total number of individuals with overlap was divided by the total of 505 participants to determine the percentage of participants that expressed each type of overlap.The second approach counted the number of themes with overlap (the "depth" overlap) using a similar approach to the logic shown in Table II, but counted the number of unique themes rather than assigning a dichotomous "1" or "0" for presence or absence.The total numbers of themes for each of the overlap types were divided by the number of respondents to determine an average number of themes with overlap per person.The results from both methods are combined in Table III.

FIGURE 3 REPRESENTATION OF THEMATIC OVERLAP METHOD (FOR PERSON COUNTS)
The first two rows of results in Table III show the results from the number of people that had overlapping responses in their descriptions.Nearly 60% of the respondents showed overlap between their descriptions of self and their descriptions of an engineer.A similar amount showed overlap between their self-descriptions and the descriptions of an EWB-USA member.Approximately 40% showed overlapping themes for their descriptions of an engineer and an EWB-USA member, and 44% showed overlap among all three personas.
The high percentage of people with overlap between their self and engineer descriptions makes sense given that the respondents are all current or future engineers.Because a similar percentage of respondents wrote overlapping themes between the descriptions of self and an EWB-USA member, it suggests that EWB-USA members identify with these two groups at a similar rate; however the results from the response percentages indicate that the respondents describe these overlaps using different themes.Several participants also indicated overlap between an engineer and an EWB-USA member, which makes sense given the active engineering practice EWB-USA members perform and may suggest EWB-USA's role in helping to bridge the connection between engineering and the self.The third and fourth rows of results in Table III show the average number of response themes with overlap per person.As a comparison, the average person described him or herself in 2.65 unique themes, which means about 30% of the themes in their descriptions of self overlapped with their descriptions of an engineer (0.79 divided by 2.65), and about 30% of the themes in their descriptions of self overlapped with their descriptions of an EWB-USA member.About 9% of the themes in their self-descriptions overlapped with both an engineer and an EWB-USA member.
Focusing on the number of themes with overlap as opposed to the number of people with overlap accounts for people who indicated multiple overlapping themes rather than dichotomizing a person with having overlap or not.In this second approach EWB-USA members once again indicated an equal amount of self-overlap with both an engineer and an EWB-USA member.This strengthens the claim that these two personas may hold equal weight for the respondents.Further research should confirm this, but it may speak to the high importance of EWB-USA in the lives of these engineers.

Gender Differences in Response Percentages
To determine variations based on gender, results from the two approaches were reanalyzed for men (n = 275) and women (n = 215).Anyone who did not report his or her gender was excluded from this analysis (n = 15).For both approaches, significant differences between genders were tested using chi-squared tests of proportions; however, because the frequency counts come from qualitatively coded data, which do not satisfy the assumptions implicit in the statistical tests, the statistical results should be interpreted with caution.We share all of the actual percentages in an attempt to limit the emphasis on statistical differences, and only use asterisks to indicate themes with statistical differences to facilitate discussion.It should be noted that because we ran multiple chi-square tests, the possibility of a Type I error greatly increases; however we did not reduce the cutoff of the p-value because of the limitations mentioned about using qualitatively coded data, and results should be interpreted accordingly." Results from the respondent percentages are separated by gender in Table IV.For the description of self, males and females listed the same top five themes, however, significantly more females responded with themes of "motivated" and "humanitarian" than males in addition to themes of "global perspective" and "narrow."Significantly more males described themselves as "engineer" and "introvert" than females.
For the description of an engineer, the top two themes of "analytical" and "smart" were the same for both genders despite more females using those themes.Significantly more females also used the themes "motivated," "creative," "well-rounded," and "introvert," while significantly more males described an engineer as "goal-oriented," "curious," "adventurous," "organized," and "involved." For the description of an EWB-USA member, the top three themes of "motivated," "humanitarian," and "passionate" were both the top responses for males and females; however, significantly more females used all three themes than males in addition to "well-rounded."More males used themes of "respectable" and "optimistic" than females.Results from the response percentages showed that gendered differences existed.Of the three personas, the description of the self had the most themes with significant gender differences (n = 5) while the description of the EWB-USA members had the most significantly different results.Interestingly, for all three personas, females responded in the "motivated" theme significantly more than males, indicating the importance of this theme for females.This finding makes sense in light of literature that has shown females studying engineering tend to have higher GPA's than their male classmates, 25,28,31 which may highlight females' strong work ethic for their studies.
Significantly more females than males described themselves in the "humanitarian" theme, which aligns with previous research.For example, in an ethnographic study of men and women in engineering, Faulkner 32 showed that men were drawn to the technical side of the profession and women to the social side (similar to other studies 33 ), and she encouraged the profession to blend social and technical aspects better in order to draw more women.These studies suggest that women would be drawn to the altruistic or humanitarian aspects of EWB-USA, and our results indicate that women note these aspects of both themselves and EWB-USA members more than men.Future work is needed to make a causal connection between humanitarian interests and motivations to join EWB-USA, but the results suggest that altruism could contribute to women's engagement with the organization.
In order to view the overlaps visually, Figure 4 presents the results using the Venn diagram shown in Figure 2.For this figure, any theme that was mentioned by at least 10% of the respondents for at least two of the three personas is written in the corresponding overlapping area of the Venn diagram.Most overlapping themes were the same between both genders except for "curious" and "engineer," which were both only over 10% for males.Although "engineer" was not a possible theme for the engineering persona, it is included in the figure to illustrate that males self-identified with this persona more so than females.No themes uniquely overlapped with an engineer and EWB-USA member apart from those themes that overlapped with all three personas.
The center of Figure 4 shows that there are multiple themes that these EWB-USA members acknowledge as common traits among the three personas.In addition, these members commonly describe both themselves and EWB-USA members as humanitarian, passionate, and outgoing, which are not acknowledged as common traits of engineers.In literature, these are some of the traits that are seen as important to future engineers, 34 including the related abilities to lead, innovated, and communicate, and as some of the traits needed to address the worlds' most pressing engineering challenges. 1 In addition, Figure 4 shows that two of the three themes commonly overlapping between the self and the engineer are unique to males, which stresses females' stronger overlap with EWB-USA than engineering.

Gender Differences in Thematic Overlap
In the analysis of the number of people with overlapping descriptions, all four overlap types showed females with greater overlap than males (Table V), and the differences in overlap between the self and an EWB-USA member and between all three personas were significant.One possible explanation for the higher percentages of female overlap is that the women in the study wrote more in their open-ended responses.On average, across all three personas, women's responses were coded to 8.3 themes while the men's responses were coded into 7.6 themes.This difference may have contributed to the women's higher overlap; however, this difference alone could not be responsible for the results because, if little alignment existed in these women's minds, their responses, even if longer than the men's, would still have indicated a lack of overlap.The significant differences between the self and EWB-USA and between all three themes, further align with literature showing women's higher connection to the altruistic nature of engineering.As in Table III, the percentages in the last row of Table V are based on the average number of themes with overlap per person divided by the average number of themes that males or females used to describe him or herself (2.71 for females and 2.60 for males).In the analysis of the number of overlapping themes, there were no significant differences; however in the three overlaps involving descriptions of EWB-USA members, females showed more overlapping themes than males, whereas males showed more overlapping themes between the self and the engineer than females.These results further suggest that engagement with EWB-USA may uniquely assist women align their self identities with engineering.Others have noted that, in order to increase women's representation in engineering, "[a]n emphasis should be placed on the direct human benefits of engineering practice." 35While additional work is needed to further unpack why women are so drawn to the organization and how their involvement impacts their engineering identity, some literature suggests that the reasons could be due to increased exposure to engineering, 36 extra-curricular activities, 37,38 or strong social networks 24 such as teamwork 39 or mentorship, 5 all of which are present within EWB-USA.

CONCLUSION
This study presented results from an exploratory study trying to better understand EWB-USA members and their perceptions of themselves, engineers, and their fellow members.We used established qualitative coding methods as well as a novel method for analyzing thematic overlap.Results from this study showed that EWB-USA members described themselves in ways that overlapped with their descriptions of engineers and EWB-USA members with relatively equal percentages using unique themes.In addition, results pointed to the ways in which EWB-USA may help to change perceptions about the engineering profession by showing that engineers and engineering work can be humanitarian, passionate, and outgoing.Furthermore, gender differences pointed to females' stronger connection to EWB-USA than to engineering, whereas males showed more balanced overlap.Further research is needed to show that engagement with EWB-USA enhances a connection to engineering, but these results provide initial evidence to prior anecdotal claims that such a connection exists.
As with any study, there are limitations that arise from both the data collection and analysis methods employed.Because the data were collected at EWB-USA conferences, the sample population was limited to EWB-USA members, and, even more so, the population of members who were able to attend the conferences.This eliminated any potential comparisons between engineers involved and not involved in the organization and may have sampled from more active members that choose to attend these conferences.In addition, thematic overlap analysis was dependent on the 24 main themes that emerged during coding.If more themes had been developed, less overlap would have occurred, while fewer themes would have led to additional overlap.Although the number of themes bounded the potential of overlap occurrences, we believe that our rigorous method of creating themes from emergent data, combined with larger research team reviews, aided the unbiased construction of 24 themes.Future work should confirm these exploratory findings by using the results as a theoretical framework to test hypotheses about a broader sample of EWB-USA engineers and their peers not involved with the organization.
Based on this work, we believe that EWB-USA can help change perceptions about engineers to broaden perceptions to include humanitarian service work and that many females are uniquely attracted drawn to this aspect of engineering because they see that work in alignment with themselves.Sheppard et al.'s 40 study on engineering pathways recommended that practitioners "help students identify what excites them about engineering, and advise them in their planning for coursework, extracurricular activities and co-op/internship experiences that allow them to realize that excitement," and we see EWB-USA engagement as one such activity that male and even more so female engineers identify with.We encourage future research into EWB-USA and other similar organizations to better understand differences between student and professional members, how these organizations and programs may influence engineering identities of its members, and the extent to which such engagement alters the pathways of its members.Further understanding of such a rapidly growing, internationally engaged, and gender balanced organization can continue to give insight in better engaging engineers that can solve the world's complex problems.

TABLE IV RESULTS FOR RESPONDENT PERCENTAGES BY GENDER
32,33