THE IMPACT OF INDIVIDUAL QUALITIES ON MESSAGE COMPREHENSION AND LIKING IN THE DEFENSE INDUSTRY

Professionals working in corporate communications within the defense industry have cited a need to send the same message multiple times in many different ways to maximize the effectiveness of the correspondence. At the present time, communication specialists rely on guesswork to ensure that the workforce has the relevant knowledge needed to be effective at work. It would be far more effective if these specialists had greater knowledge of the most accurate way to reach each sector of their workforce, particularly in light of the Covid-19 pandemic that caused an unprecedented increase in telework. By understanding the medium preference, generational cohort, military status, age, and telework frequency in relation to message comprehension and liking, this study aims to create new knowledge of how best to communicate with a diverse and specialized workforce with unique information needs. The study ultimately yielded largely inconclusive results because of sampling size, but further explorations is warranted. The study found evidence that receiving a video treatment results in less message comprehension compared to text or image. There also was evidence of a strong relationship between age, military status and working in the defense industry with liking. These findings will need to be substantiated by further research with a larger sample.


LIST OF TABLES
specialists rely on guesswork to ensure that the workforce has the relevant knowledge needed to be effective at work. It would be far more effective if these specialists had greater knowledge of the most accurate way to reach each sector of their workforce, particularly in light of the Covid-19 pandemic that caused an unprecedented increase in telework. By understanding the medium preference, generational cohort, military status, age, and telework frequency in relation to message comprehension and liking, this study aims to create new knowledge of how best to communicate with a diverse and specialized workforce with unique information needs.

C. Justification for and Significance of the Study
Properly replacing and training the next generation of employees is critical to the long-term health of any organization. Effective communication is a key aspect in this process, as it is important that new employees are getting the information they need to be successful in their jobs. Message effectiveness as a concept, however, is as vast as it is intricate. It entails elements of purpose, social and professional connectedness, emotional response, comprehension, and liking, among other concepts. For the purposes of this study, only comprehension and liking will be examined in relation to the qualities of the media and qualities of the study participants. These two were specifically chosen because they arein theoryrelatively easy to measure yet essential parts of an effective message. Whether or not a message can be understood is a foundational aspect of communications, while pleasing messages are more likely to elicit a favorable response. Further research should explore other elements of communication effectiveness to develop a more well-rounded understanding of the concept, which is relatively understudied with respect to media richness in organizational communication.
This study also will provide further evidence as to what effect the Covid-19 pandemic has had on this process. It is a problem Jeff Prater often has pondered over the last few years. Prater is a retired surface warfare officer in the United States Navy and currently the Director of Corporate Communications for the Naval Undersea Warfare Center (NUWC) Division Newport. This organization is the Navy's full spectrum research, development, test and evaluation, engineering and fleet support center for submarine warfare and other systems associated with the undersea battlespace. In addition to his experience at NUWC Division Newport, Prater previously served as the Public Affairs Officer for Naval Surface Warfare Center Panama City Division in Panama City, Florida. During an interview conducted March 10, 2022, Prater identified a number of issues he has noticed throughout his career and contrasted them with more recent developments. As for the present research, the most relevant development he has noticed is with respect to different messaging needed for different generations. In his previous work doing communications on Navy surface ships, the age gap was around 20-3 45 years old and most were on the same page with respect to messaging as a result of Navy training and a relatively smaller number of people. That changed when he got to the warfare centers, as now he needed to reach more than 6,000 employees with an age gap of 15 years old for interns all the way to the most senior employees in their 80s.
"Early on, I learned to get proper penetration you need to hit them with it five times, five different ways," Prater said. "You have to recognize that there are different generations and that these different generations communicate very differently." In addition to the repetition and variety needed, timing also matters, Prater said. He found different generations respond differently to certain modes of communication, which is why his team uses a variety of methods, such as face-to-face, telephone, email, digital signage, posters, articles and livestreaming videos.
Prater's approach is rooted in communications theory, yet the literature does not support this exact problem in the defense industry or provide a targeted approach like this study proposes. It is unknown if this population is unique within the field of communications as the amount of studies are limited, representing a gap in the literature.
Those working in the field of defense communications, like Prater, have cited observable communications differences between the general population. It is the intent of this research to investigate whether or not there is a quantifiable difference. This study also takes a unique approach to defining generational cohorts by examining employees who began working before, after and during the Covid-19 pandemic. By exploring variables that in totality are not typically grouped together, it will attempt to fill a gap in the current literature while simultaneously providing greater knowledge about a real-world problem.
By understanding the medium preference, generational cohort, military status, age, and telework frequency in relation to message comprehension and liking, I hope to answer following question: RQ1: How do the qualities of the reader and medium predict message comprehension and liking?

Review of Literature
The approach for this project is rooted in medium theory, intergenerational communications, and telework research. Medium theory provides an overarching framework for this study. Meyrowitz (1995) asserted that media can differ from one another in a variety of ways, but medium theory explores the relatively fixed characteristics of media content. This includes two levels of analysis. The micro level explores the consequences of choosing one medium over another in a particular situation, while the macro level looks beyond these individual uses. An example of this would be examining how a new technology affects society as a whole. Most importantly, medium theory provides proper justification for the proposed method in this study as the theory does not present a cause-and-effect view. Rather, it describes how a medium encourages or discourages the ways in which people communicate (Meyrowitz, 1995).
This is just what this study seeks to investigate, that video, images and text either encourage or discourage communications liking and comprehension. Yet, before this proposal can move forward and address independent variables, it is important to address relations with supervisor. While generally accepted as a sound method for analyzing communication satisfaction in an organization, some have taken issue with aspects of the CSQ over the years. DeConinck et al. (2008) replicated the CSQ with retail store buyers and suggested that the factors should be condensed from eight to five (DeConinck et al., 2008). This research is just one element of message effectiveness and shows just how wide-reaching it can be as a concept. Furthermore, aspects such as liking and comprehension are factors when it comes to the use of messages (Hornikx, 2017). In their study of how different cultures react to different messages, Hornikx and le Pair used a semantic differential scale to test their hypotheses (Hornikx, 2017).
It also is important to acknowledge that the term "generation" needs to be defined.
In a systematic literature review published by Han in 2022, he found that intergenerational differences in communications are understudied variables in the field as opposed to other factors like gender or sex. Even when they are the focus, it is often a family perspective and rarely workplace-focused (Han, 2022). Han's work provides a valuable roadmap for properly defining generations. As Han notes, this is no easy task given that plenty of authors have taken different approaches to establishing margins. For the purposes of this study, generations will be viewed through two perspectives. The first is outlined by the Pew Research Center, which classifies Americans both by their "place in the life cycle" and "membership in a cohort of individuals who were born at a similar time" (Dimock, 2019). In turn, the current working class according to Pew is broken down into five generations: Silent (born 1928-1945), Boomers (1946-1964), Generation X (1965-1980), Millennials (1981-1996) and Generation Z (1997-2012 (Dimock, 2019).
Other research has established similar margins (Han, 2022 (Ota et al., 2012). Polat and Yilmaz (2020) found that personal, relational and managerial factors represent the greatest barriers to successful intergenerational communications. The most common barriers were seeing experience as power, lack of motivation and poor communication. Malek and Jaguli (2018) reported similar findings in their study on women in Malaysia. Woodward and Vongswasdi (2017) in their study of 191 company executives found that technological advances are better viewed as complements to traditional communications practices rather than replacements. They cite this as evidence of the importance of understanding the preferences of different generations in the workplace. Myers and Sadaghiani (2010), and Mehra and Nickerson (2018) in their respective studies found a positive workplace communications environment minimizes the gap in intergenerational communications while increasing productivity and efficiency. Wok and Hashim (2013) A key discovery with respect to this research proposal is they found when people are satisfied with the communications channel, they are typically satisfied with their jobs.
They also noted since personality type affects job satisfaction, it is important for employers to understand the personalities of their employees to optimize their working environment. These findings also demonstrate the importance of employers making available a wide array of communications channels to teleworkers, and recommend future studies investigate the generational effects of teleworking. The concept of telework communication satisfaction is particularly understudied in the defense industry, yet not totally nonexistent. Monroe and Haug (2021) carried out a similar approach to the one proposed here for their assessment of telework at a federal agency, using a questionnaire to explore the opinions of 46 federal agency workers to determine mainly positive attitudes about telework, the agency's mission and work environment. This was done in light of common concerns about telework, such as disconnection, reduced trust leading to less effectiveness and lack of control for managers. Despite the small sample size, the framework of distributing a questionnaire like this was sound, which is one the reasons this research takes a similar approach.
Given the transformative effect digital technologies has had on telework, it is important to consider the confidence people have in using these developments at workparticularly in light of the Covid-19 pandemic. In their study of employees required to work from home, McGloin et al. (2022) found that less traditional or formal methods of communication may be required to maintain rapport between supervisors and employees.
This was evidenced in their survey results by a significant relationship between texting and rapport, defined by the authors as "employees' subjective perceptions of outcomes of interactions with their supervisors, including the viewpoint that one has been justly heard and treated" (McGloin et al., 2022, 45

Research Strategy
The first step was to determine the best possible way to answer RQ1: How do the qualities of the reader and medium predict message comprehension and liking? Survey research was identified, and proper instrumentation was developed. Once completed, a sampling and recruitment strategy needed to be established. Leveraging contacts within the defense industry, two sampling methods were developed: blast email through a defense industry advocate group and LinkedIn nonprobability sampling. After limited success, a third method utilizing probability sampling by way of financial incentives through Amazon Mechanical Turk was employed. Data was collected, cleansed and organized for preliminary analysis. Further investigation revealed the appropriate analysis tool to be used and a need to amend RQ1 to fit the data collected. The following describes in detail this process.

Experimental Design
The design of this survey was inspired, in part, by research from Hornikx and le Pair (2017) on advertisement liking and complexity. After all, in many ways, organizational messaging to employees shares plenty of similarities to the conventional relationship between advertisers and consumers. This survey, which can be viewed in Appendix A, was built and administered through Qualtrics. In this instance, actual organizational messaging provided by NUWC Division Newport was utilized. In it, the same information and wording was conveyed in three different treatments: text, image and video. These were designed to be as simple as possible in attempt to limit any further 13 confounding variables. The idea was they would ask the viewer to remember basic recall information relating to the celebration of an organization's 150 th anniversary. In practice, organizational communications messages often are far more comprehensive and involve a request for action. To ensure homogeneity, participants were randomly assigned which treatment they received by Qualtrics. The treatments can be seen in Appendix B. Once participants received their treatment, they were administered a series of questions. The first two were multiple choice questions designed to gauge comprehension. Each had one correct and three incorrect responses. Questions were simple and brief, but required participants to recall the information they had just been shown. A semantic differential scale to gauge liking followed with five poles: appealing-unappealing, good-bad, powerful-weak, exciting-boring, helpful-unhelpful. Some of these measures were adapted from Hornikx and le Pair (2017). Participants were asked to rank these liking measures on a scale of one to seven based on the medium they had just seen. At the urging of the principal investigator, this format was repeated a second time to add validity to the results. In total, each participant received two treatments, and answered four comprehension questions and 10 liking questions, as well as demographic questions to address the independent variables age, hire date, military status, defense industry and telework status. Questions on gender and race also were included. Given each set of questions addressed a specific treatment set, the degree to which these results could be aggregated was limited.

Recruitment of Subjects
For this project, three separate recruitment efforts were undertaken. Each initiative provided insight into the difficulty associated with surveying defense industry workers outside of organizationally sanctioned efforts.

Institutional Agreements
First, an agreement with the Southeastern New England Defense Industrial Alliance (SENEDIA) was formed to distribute the survey to its member organization via email. Given that 134 companies have working arrangements with SENEDIA, even meager responses from that base would yield statistically relevant results. It quickly became clear, however, that this method would not work. The researcher directly contacted an affiliated defense industry organization's human resource department to distribute the survey to its workforce, but the representative explained that they could not do that per their company's guidelines. "I'd love to support your request, however, it would create some difficulty and possible issues. If I were to distribute it, it becomes a corporate data request, which in turn needs an available charge code. This wouldn't be something billable to either overhead or direct charge codes for each contract, so it just becomes a bit problematic from that perspective." In addition to email solicitations often being ignored, this was further evidence that most companies would have a similar response. Thus, it became clear that what originally had thought to be a very wide net would have to increase not only in scope, but also in intention. The researcher sought similar distribution efforts from the National Defense Industrial Alliance (NDIA), an organization similar in makeup to SENEDIA only at the national level with many more members, but those pleas were ignored.

Nonprobability Sampling
The second strategy utilized a nonprobability method, as the researcher launched a purposive sample on LinkedIn. The researcher already had an account and on Jan. 21, 2023, he signed up for a free, one-month trial of LinkedIn Premium. An adapted version of the initial recruitment letter was posted to the researcher's main page. This message was pinned to the top of the page, and follow-up posts were made one and two weeks later, respectively. They then selected two target companies, one a current employer and one a former employer. Both companies are so large that there were very few people contacted whom the researcher knew before they were solicited for participation.
While this study acknowledges there are limitations to nonprobability sampling and possibilities for bias, it is a necessary risk to reach a population like the defense industry. Those who work in this business regularly undergo training to spot Phishing attacks, operational security risks or other cybersecurity vulnerabilities. As such, they are inherently more guarded against unsolicited requestsparticularly those asking them to click a linkthan other members of the general population.
To establish legitimacy, the researcher sent connect messages to the first 100 employees following the Indus Technology Inc. LinkedIn page. During this selection process, followers that seemed illegitimate were omitted (those with little to no information immediately listed in their profile). This message had to be carefully crafted, as it needed to establish legitimacy, include the link to the survey and adhere to the 300character limit. The following was sent to potential participants: "Hey there, fellow Indus employee here. Working on my master's thesis, a survey of defense industry employees. Hoping you could help me out and take this anonymous, 5-minute survey.
Open to all in the defense industry, please share: https://uri.co1.qualtrics.com/jfe/form/SV_8pQJN3v06oSpunk." This method allowed those interested to directly reach out with questions or comments. A number of respondents were "happy to help," while others empathized with the survey process, having completed their master's degrees previously. There also was skepticism; one respondent said they would be happy to complete the survey only if it came directly from an Indus email address. The majority ignored the message altogether, although some accepted the connection without taking the survey.
Eight days after the initial data call, the process was repeated with the first 100 While there are hesitations with using a method like thisyou are surveying people motivated to take surveys who may not be representative of the general population there are certain advantages. As a simple random sample with replacement, this is a more reliable method as an example of probability sampling. This also provides access to a much larger pool of participants, as there were as many respondents within an hour of the survey being live as there were in three weeks of the LinkedIn method. Given the difficulty involved with gaining trust of defense industry employees outside their own organization, a very large pool of participants is necessary. This process was repeated once the survey closed after two weeks, and ultimately yielded 112 responses.

Data Cleaning and Preliminary Analysis
Once data was collected, it was imported into SPSS for scrubbing. Categories were renamed, where necessary, to more accurately describe the variables.

Independent Variables
The independent variables defense industry, military status, hire date, age, telework status, gender and race did not require any recoding.

Type of Treatment: Text, Image or Moving Image
First, the data for treatment was recoded, as it was imported in separate categories since Qualtrics assigned treatments at random. Those receiving text treatment were coded as one, image as two and video as three. This process was then repeated for treatment two. In addition to specific instructions, a 30-second timer was installed on the video treatment to discourage people from skipping ahead in the video. The time each participant spent on a given page of the survey also was recorded, although it was not analyzed in this study because of research deadlines.

Dependent Variables: Comprehension and Liking
Comprehension data was recoded to represent either a correct answer (one) or an incorrect answer (zero). Results from the two comprehension questions were added together to give each participant a total comprehension score of questions answered correctly (zero, one or two). This process was then repeated for treatment two.
The liking dependent variable required combination and aggregation as well to make it more readily digestible. For treatment one, there were five semantic differential questions graded on a seven-point scale representing different adjectives associated with liking. Adding these numbers up resulted in raw liking scores of five through 35. There were two problems with this data. First, the way it was recorded in Qualtrics, lower numbers represented more liking while higher numbers indicated less liking. Not only is this counterintuitive, but also is directionally opposite of the comprehension data. The 5-35 scale also presents challenges in data analysis as the scale is nowhere near the other data collected. This was then recoded so that 30-35 would become zero, 24-29 would be one, 17-23 would be two, 11-16 would be three and 5-10 would be four. These data points were chosen to correspond with a five-point scale of strongly dislike, dislike, indifferent, like and strongly like, respectively. Once recorded, this process was repeated 20 for the data in treatment two. This data was simplified to account for knowledge shortfalls on behalf of the author, as the extent of his practice in statistics is limited.

Data Analysis
First, descriptive statistics were recorded to generally describe the population of the survey. This also was done for the comprehension and liking data using descriptive statistics and cross tabs. The next step was to determine the most appropriate analysis This is a question that can be answered with analysis of variance (ANOVA), which is used to determine if there is a statistical difference in means between two groups (variables). To properly answer RQ1', 12 hypotheses will need to be tested: H1: There is a statistically significant relationship between the type of media and comprehension.
Null: There is no statistically significant relationship between the type of media and comprehension.
H2: There is a statistically significant relationship between the type of media and liking.
Null: There is no statistically significant relationship between the type of media and liking.
H3: There is a statistically significant relationship between age and comprehension.

22
Null: There is no statistically significant relationship between age and comprehension.
H4: There is a statistically significant relationship between age and liking.
Null: There is no statistically significant relationship between age and liking.
H5: There is a statistically significant relationship between generational cohort and comprehension.
Null: There is no statistically significant relationship between generational cohort and comprehension.
H6: There is a statistically significant relationship between generational cohort and liking.
Null: There is no statistically significant relationship between generational cohort and liking.
H7: There is a statistically significant relationship between military status and comprehension.
Null: There is no statistically significant relationship between military status and comprehension.
H8: There is a statistically significant relationship between military status and liking.
Null: There is no statistically significant relationship between military status and liking.
H9: There is a statistically significant relationship between defense industry and comprehension.

23
Null: There is no statistically significant relationship between defense industry and comprehension.
H10: There is a statistically significant relationship between defense industry and liking.
Null: There is no statistically significant relationship between defense industry and liking.
H11: There is a statistically significant relationship between telework frequency and comprehension.
Null: There is no statistically significant relationship between telework frequency and comprehension.
H12: There is a statistically significant relationship between telework frequency and liking.
Null: There is not no statistically significant relationship between telework frequency and liking.
For each hypothesis, the same procedure was followed. First, a one-way between subjects ANOVA was conducted for the dependent and independent variables. In each hypothesis, this process needed to be done twice, once for the first treatment and another time for the second treatment. Once the ANOVA was conducted, the F and p values were recorded. If the p value was greater than .05, the result was recorded as not statistically significant and the null hypothesis was accepted. If the p value was less than .05, the variables were reported as statistically significant.

Findings Statistically Significant Results
Twenty-four analysis of variance (ANOVA) models were generated to test H1-12, of which 11 yielded statistically significant relationships. Before conducting these tests, however, the descriptive statistics for message type, comprehension and liking were examined. The most interesting finding from this exercise was participants receiving the text treatment were nearly twice as likely to get all questions correct than those receiving the video treatment. While not as significant, those that got the image treatment also were more likely to get a perfect comprehension score than those that watched a video.
ANOVA backed this finding up, but only to an extent. There was no statistical significance between message type and comprehension for the first treatment, but there was for the second. This represented a trend throughout the research, as there was no statistically significant relationship between media, age, military status, defense industry and telework status when set against comprehension in the first treatment, yet there was significance between all the independent variables and comprehension in the second.
There was, however, congruity between both treatments for age, military status and defense industry with liking. ANOVAs revealed statistically significant relationships in each of these pairings for both treatment one and two. When an employee began working at a company with respect to pandemic seems to have no bearing on comprehension or liking. Results and some possible explanations why are offered further in this chapter.

Demographic Information
The combined paid and unpaid surveys elicited 147 responses, which ultimately

Comprehension and Liking Data
Data was analyzed in two parts, the first of which examined general results of comprehension and liking data among the two treatments, particularly with respect to medium type received. Treatment one (text, image or video) corresponded to the first set of two comprehension questions and five liking questions, which were graded on a sevenpoint semantic differential scale. The same analysis was conducted with treatment two.
To properly analyze the data, some of it needed to be condensed into single categories.
Comprehension questions were recoded into either correct or incorrect, with a zero being recorded for incorrect answers and one for correct answers. This was re-labeled, and participants received a score of zero, one or two (correlating to the total number of questions they got correct for each treatment The results of the comprehension data were then analyzed in Cross Tabs according to which media message type they received. The results also were combined into one table as a means to discern comprehension levels. It is worth noting that this information is not representative of a cumulative score for each individual participant. Rather, this shows which treatments yielded all incorrect scores, all correct scores or half correct scores. This information is available in the table below. For the most part, the results were fairly unremarkable with the majority of participants getting one question correct. The most striking data from this was participants who received the text treatment were nearly twice as likely to get all four questions correct than those that received the video treatment (43-23). Even the image treatment to an extent showed a greater propensity for higher achievement, thus it is not unreasonable to conclude there is less comprehension associated with receiving a video treatment.
As for the liking responses, these required considerable more data scrubbing as a result of the survey design. First, data were combined into one category for each treatment. Results were simply added together for all questions, meaning liking scores ranged anywhere from five to 35 with lower scores indicating greater liking. This presented an issue, however, as both the scale was much larger and directionality of feelings reversed when compared to the comprehension data. Graphs of the raw numbers are included below. Descriptive Statistics for Liking 2 The charts show fairly favorable results with respect to liking, as most of the participants liked or were indifferent to the media type they received. A cross tabs examination of this information revealed minimal differences between media type received and liking. These tables are located in Appendix C. As noted, this data needed to be recoded, though, so that lower numbers indicated less liking just as lower numbers indicated less Liking2 30 comprehension. The scale also needed to be reduced, so a five-point scale was implemented. The recoded data is in alignment with earlier results, however, it provides a much clearer picture. For both treatment one and treatment two, there was no real preference on media type and responses to each type of treatment, whether it was text, image or video, was more favorable than unfavorable. Cross Tabs results for each treatment are below.   Interestingly enough, these results did not hold when a one-way ANOVA was run between treatment two and comprehension two. Results of the ANOVA did not show a significant difference between type of media (text, image or video) and comprehension; F (2,131) = .887, p .415. A number of factors could explain the difference in results, including sampling or design error on the part of the researcher. It also could be test fatigue as well, as being asked to do a similar task twice could lead to lapses in attention or concentration. Thus, with respect to H1, the results are inconclusive. For treatment one, since the p value is less that the alpha value (.05), the null hypothesis that there is not a statistically significant relationship between the type of media and comprehension is rejected. For treatment two, however, the researcher fails to reject the null hypothesis.
I also wanted to understand if subjects' liking or format preference was associated with media type, so one-way between subjects ANOVAs were conducted with type of media as the independent variable and liking as the dependent variable. Results in each instance revealed no statistical significance between media type and liking, confirming the null hypothesis that there is no statistically significant relationship between the type of media and liking. Findings were nonsignificant with treatment one and liking F (2,131) = .384, p .682 and treatment two and liking, F (2,131) = 1.279, p .38.

b. Age
I also wanted to understand if subjects' age was associated with comprehension or liking. As it was for H1, there were split results when testing H3. A one-way between subjects ANOVA was run for age and comprehension one that revealed no statistical significance between variables; F (4,129) = 1.057, p .380. When the same process was repeated for comprehension two, however, the results were statistically significant; F (4,129) = 9.141, p <.001. One interesting finding is that Scheffe post-hoc analysis revealed that younger participants between 11 and 26 years old (n = 49, M = 1.7551, SD = .52164) had significantly more comprehension on the second treatment than both middle age participants between 27 and 42 years old (n = 42, M = 1.0714, SD = 0.71202) and older participants between 43 and 58 years old (n = 32, M = .9688, SD = 0.73985).
Middle age and older participants are not significantly different from each other. Results are listed in the tables below: This result mirrors the outcome from treatment type vs. comprehension previously discussed where there was a difference in treatment one vs. treatment two, although the difference is more likely because of sampling error in this instance. It is plausible that younger survey participants have greater attention or stamina to stay on the task despite being asked to do two similar assignments in succession. Possibly, they are less susceptible to test fatigue than their more senior cohorts. However, a deeper look at these results reveal sampling error as the likely cause. Cross tabs of age vs. treatment two and age vs. comprehension two were performed; the results are in the tables below. A closer look at the numbers show when participants were randomly assigned treatments by Qualtrics, the order in which participants by age took the survey was just so that younger partakers received significantly more text and image treatments than video treatments.
As this study and previous research have shown, there is a statistically significant relationship between text and image treatments and greater comprehension. The two Cross Tabs back up this data, as younger participants received significantly more text and image treatments and displayed a significantly higher level of comprehension than their peers in other age groups.  1997-2012 1981-1996 1965-1980 1946-1964 1928-1945 Total Text  20  13  9  3  0  45  Image  20  13  8  2  1  44  Video  9  16  15  4  1  45  Total  49  42  32  9  2  134  Questions  Correct  Zero  2  9  9  2  1  23  One  9  21  15  4  0  48  Two  39  12  8  3  1  63  Total  49  42  32  9  2  134 Two one-way between subjects ANOVAs were run for age and liking to test H4.
Results in each instance revealed strong statistical significance between age and liking; F comprehension, reveals sampling error is not the likely cause of statistical difference between groups. In both treatment onewhich is more normally distributed than treatment two with type of medium receivedand treatment two, younger participants showed predominately like or dislike regardless of treatment received. Middle-and older-age participants were significantly more indifferent or dislike. Results are listed in the tables below.  1997-2012 1981-1965 1965-1980 1946-1964 1928-1945 Total   1997-2012 1981-1965 1965-1980 1946-1964 1928-1945 Total Strong dislike  0  2  7  0  0  9  Dislike  5  13  4  3  0  25  Indifferent  9  16  13  3  1  42  Like  14  8  4  1  1  28  Strong like  21  3  4  2  0  30  Total  49  42  32  9  2  134  Treatment  Text  13  13  9  3  0  45  Image  19  13  8  2  1  44  Video  17  16  15  4  1  45  Total  49  42  32  9  2   As it was with age, the difference between comprehension one and comprehension two is more likely because of sampling error as opposed to actual differences in the groups or survey fatigue. For military status, there were 12 more non-veterans than veterans that took the survey. A Cross Tabs analysis shows that while there were drastic differences in treatment type between groups, those that got text or image are fairly equivalent (44-45).
However, there were 28 non-veterans that received the video treatment as opposed to 17 veterans that received the treatment. These numbers are backed up by the comprehension two data, as 35 veterans got all questions correct as opposed to just 25 non-veterans. A review of the Cross Tabs data for treatment and comprehension two backs up this assertion and reasoning as to why there is no statistical significance in this data for military status. The media viewed were far more evenly distributed in the first treatment, which resulted in a normal distribution of comprehension results.  Total  Text  20  26  46  Image  21  25  46  Video  20  22  42  Total  61  73  134  Treatment 2  Text  18  27  45  Image  26  18  44  Video  17  28  45  Total  61  73  134  Comprehension 1  Zero  1  6  7  One  44  44  88  Two  16  23  39  Total  61  73  134  Comprehension 2  Zero  8  15  23  One  15  33  48  Two  38  25  63  Total  61  73  Tabs analysis backs up these results, as those who are veterans showed considerably more liking than non-veterans, whose predominant response was indifference. A likely explanation to this result is the content itself. The messaging and visuals are centered on military history, which naturally lends itself to be more popular among those who have served than those who have not. These results are recorded in the tables below.     promising results in terms of percentages, but was labor and time intensive. A large team of actual people would be needed to gather meaningful data using this method, as you could not just use bots. Given the skepticism of the population, any detection of using artificial intelligence to gather results could severely compromise the integrity and distribution ability of the survey. This being nonprobability sampling also carries bias risk. As a probability sampling method, the MTurk method proved more reliableto an extent. It was certainly a faster and more effective sampling method, however, this is a population specifically motivated to take surveys. There is an element of "the squeaky wheel gets the grease" with this group. There also is the financial element of doing research through MTurk. The costs for this survey were kept within reason, however, using this method could get expensive pretty quickly.
More quantitative research needs to be conducted if the field of communications is going to progress. This study was an earnest attempt to do just that. The purpose was to take a readily observable problem in the field and try to solve it through data analysis While the intent of this study was to investigate an internal, organizational communications issue, the potential implications for this type of research are widereachingparticularly in Rhode Island. As a defense contractor in corporate communications for more than five years, I have had the opportunity to observe how government installations interact with academia and industry. In recent years, there has been a large push from all parties to collaborate more on research. These groups working together is not necessarily new, but their relationship often was limited by a number of barriers. A common refrain from industry and academia was that it was too difficult to work with the government because of the amount of security protocols that were in place, which is understandable given the sensitivity often associated with military technology development. The government side rightfully has a number of concerns about allowing consequences could be substantial. What if, though, there are differences in communications patterns between these groups based on their individual qualities?
Would knowing them create better cohesion? Understanding the most effective ways to conduct environmental communications based on the qualities of the individuals involved could have long-lasting implications.
The overarching point here is the beauty and complexity in communications resides in nuancenuance in the message, nuance of the communicator. It is these minute distinctions that can make such a large difference in message understanding and liking. A misinterpreted eyebrow raise or a word with different connotations based on culture can drastically affect these variables. At the same time, some communications patterns can make no difference. There are universal languages that are understood no matter the qualities of the individual. This is what makes explorations of nuance utterly critical to communications professionals. It is not merely enough to just explore this phenomenon, though. It must be properly researched, documented and, perhaps most importantly, quantified to create new and lasting knowledge. According to NOAA, more than 80% of the world's oceans are "unmapped, unobserved and unexplored" (NOAA, January 2023). It is not a stretch to consider how human beings communicate to be equally unmapped, unobserved and unexplored.

APPENDICES Appendix A -Telework Communications Satisfaction Survey
You have been invited to participate in a special survey. The purpose is to create new knowledge of how best to communicate with employees working in the defense industry. On the next page, you will need to read and agree to participate in the research.

Informed Consent
You are being asked to take part in a research study. The purpose of the research study is to create new knowledge of how best to communicate with employees working in the defense industry. Please read the following before agreeing to be in the study. If you agree to be in this study, it will take you approximately five minutes to complete this survey. Questions will be asked about message comprehension and liking. There are no known risks or benefits.
Your responses will be strictly anonymous. The responses may be used in the research paper titled, "The Impact of Individual Qualities on Message Comprehension and Liking in the Defense Industry." The decision to participate in this study is entirely up to you. You may refuse to take part in the study at any time without affecting your relationship with the investigators, the University of Rhode Island (URI), Southeastern New England Defense Industry Alliance (SENEDIA) or your employer.
Your decision will not result in any loss of benefits to which you are otherwise entitled. You have the right not to answer any single question, as well as to withdraw completely from the survey at any point during the process.
You have the right to ask questions about this research study and to have those questions answered by me before, during or after the research. If you have questions about the study, at any time feel free to contact Renee Hobbs from the Harrington School of Communications at the University of Rhode Island (URI) at (401)  If you would like to keep a copy of this document for your records, please print or save this page now. You may also contact the researcher to request a copy.
By clicking on the arrow below to be taken to the survey, you indicate that you have read and understand the above and volunteer to participate in this study.

Treatment One
You will be looking at two types of messages. The first will appear after this screen and be followed by three questions. (See Figure 3 for three treatments) You will now be shown another message, followed by another set of questions. (See Figure 4 for treatments) 4. What three things did the station make major strides with during the time period mentioned in the second example of workplace messaging? (check one) ❏ Guncotton, electricity and missiles ❏ Gunpowder, missiles and electricity ❏ Torpedoes, steam engines and guncotton ❏ Torpedoes, electricity and guncotton

Figure 3 Text only treatment
Commemorating NUWC Division Newport's 150 th anniversary Everyone does their part to support the war effort as work at the Naval Torpedo Station becomes utterly critical from 1925-1950. Visit NewPortal for the latest Command news!

Image treatment
Please study the image below carefully.

Moving image (video) treatment)
You will now be shown a video. Please click play and watch carefully. The "next" button will appear at the bottom of this page after the video finishes playing. Link to video: https://www.youtube.com/watch?v=y-5VWBlFwQg

Figure 4 Text only treatment
Commemorating NUWC Division Newport's 150 th anniversary Station makes major strides with guncotton, electricity and torpedoes between 1879 and 1889 Visit NewPortal for the latest Command news!

Image treatment
Please study the image below carefully.

Moving image (video) treatment
You will now be shown a video. Please click play and watch carefully. The "next" button will appear at the bottom of this page after the video finishes playing. Link to video: https://www.youtube.com/watch?v=CUGqTC9unQc