Content area
Air traffic control (ATC) operations are critical to the U.S. aviation infrastructure, making ATC training a critical area of study. Because ATC performance is heavily dependent on visual processing, it is important to understand how to screen for or promote relevant visual processing abilities. While conventional wisdom has maintained that such abilities are largely innate and stable (i.e., not impacted by training or experience), recent research has begun to question this assumption. For example, intelligence has been thought to be stable, yet intelligence scores on the Ravens Progressive matrices (RPM) change over time. Because the RPM, like ATC, relies on visual pattern recognition, one hypothesized reason for these changes is the increased exposure to visually intensive technologies like videogames. Other constructs such as field-dependence/independence (FD-I) which, like the RPM, rely on visual pattern detection and are thought to be stable, might therefore also be affected by visual technology exposure. Two studies sought to examine the role of videogame play on FD-I and ATC simulation performance by ATC students at a Midwestern university. No benefits of videogame play were found. However, more fourth-year ATC students were field-independent (FI) than were first-year students and the general population. Findings suggest that exposure to the ATC training program over 4 years may result in more FI students, and that FD-I may not be stable. Possible explanations and implications are discussed.
http://crossmark.crossref.org/dialog/?doi=10.1007/s12528-015-9099-0&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1007/s12528-015-9099-0&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1007/s12528-015-9099-0&domain=pdf
Web End = J Comput High Educ (2015) 27:196214
DOI 10.1007/s12528-015-9099-0
Richard N. Van Eck1 Hongxia Fu2
Paul V. J. Drechsel3
Published online: 23 August 2015 Springer Science+Business Media New York 2015
Abstract Air trafc control (ATC) operations are critical to the U.S. aviation infrastructure, making ATC training a critical area of study. Because ATC performance is heavily dependent on visual processing, it is important to understand how to screen for or promote relevant visual processing abilities. While conventional wisdom has maintained that such abilities are largely innate and stable (i.e., not impacted by training or experience), recent research has begun to question this assumption. For example, intelligence has been thought to be stable, yet intelligence scores on the Ravens Progressive matrices (RPM) change over time. Because the RPM, like ATC, relies on visual pattern recognition, one hypothesized reason for these changes is the increased exposure to visually intensive technologies like videogames. Other constructs such as eld-dependence/independence (FD-I) which, like the RPM, rely on visual pattern detection and are thought to be stable, might therefore also be affected by visual technology exposure. Two studies sought to examine the role of videogame play on FD-I and ATC simulation performance by ATC students at a Midwestern university. No benets of videogame play were found. However, more fourth-year ATC students were eld-independent (FI) than were rst-year students and the general population. Findings suggest that exposure to the ATC training program over 4 years may result in more FI students, and that FD-I may not be stable. Possible explanations and implications are discussed.
& Richard N. Van Eck [email protected]
1 School of Medicine and Health Sciences, University of North Dakota, 1301 N. Columbia Rd.,
Stop 9037, Grand Forks, ND 58202-9037, USA
2 College Research and Evaluation Services Team (CREST), Mary Lou Fulton Teachers College, Arizona State University, 1000 S Forest Mall, Suite 204, Tempe, AZ 85281, USA
3 College of Aerospace Sciences, University of North Dakota, 4251 University Avenue Stop 9023, Grand Forks, ND 58202-9023, USA
http://crossmark.crossref.org/dialog/?doi=10.1007/s12528-015-9099-0&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1007/s12528-015-9099-0&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1007/s12528-015-9099-0&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1007/s12528-015-9099-0&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1007/s12528-015-9099-0&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1007/s12528-015-9099-0&domain=pdf
Web End = Can simulator immersion change cognitive style? Results from a cross-sectional study of eld-dependenceindependence in air trafc control students
123
Can simulator immersion change cognitive style? 197
Keywords Games Videogames Cognitive style Field-independence Field-
dependence Air trafc control
Introduction
The U.S. economy depends on safe, reliable, and affordable air transportation; ensuring our air transportation infrastructure is robust and effective is therefore critical. There are more air travellers than ever at the same time that planes are getting smaller, which increases the number of planes in the sky for the air trafc control (ATC) system to handle. (Edwards and Poole 2010). Many experts predict signicant problems with ATC as increased air trafc outstrips the capacity of available facilities. The system by which we train ATC personnel must as effective as possible.
Because ATC is highly dependent on visual processing and pattern recognition, it is important to understand not only the best way to train for ATC skills (i.e., procedural knowledge, reaction time), but also for abilities related to ATC (i.e., innate traits like pattern recognition) that are typically thought to be stable. Traditional thinking in this latter regard suggests that we can screen ATC applicants to nd those who will be most successful. However, screening can only, at best, ensure that those we retain will be among the best, leading to fewer, better ATC operators. At worst, screening may eliminate others with potential to be successful
Fig. 1 Theory linking social and developmental change. Printed with permission by Patricia M.
Greeneld
123
198 R. Eck et al.
in spite of lower abilities in a specic area. But what if we could actually train people in ways that could change these supposedly stable abilities? We could then increase both the number and quality of ATC operators. Is there any evidence that such abilities can be changed through training?
Flynn (1987) examined a documented increase in IQ scores of approximately three points every 10 years as measured primarily by Ravens Progressive Matrices, or RPM (Raven 1936). By statistically controlling for shifting socio-economic status, access to healthcare or education, or other likely contributors to increased IQ scores, he argued that this change, since dubbed the Flynn Effect, had to have some other explanation. Greeneld (2009) has expanded on this proposition with her proposed model linking shifting social patterns with corresponding changes in human development, mediated by changing learning environments and possibly by cultural values (see Fig. 1). Among other things, this model accounts for changes in cognitive abilities as a result of exposure to technology (screen time), an argument advanced by others as well (e.g., Johnson 2006).
Videogames and simulations are good examples of visually intensive technologies that have been shown to promote a wide range of skills, including spelling, reading, physics, health, biology, computer science, surgical skills, knowledge mapping (Tobias et al. 2011) and mathematics (Van Eck and Dempsey 2002). Several meta-analyses have found advantages for games over conventional instruction (Clark et al. 2014; Hays 2005; Randel et al. 1992; Vogel et al. 2006). While not every study nds an advantage for the use of games, evidence suggests that contradictory results may be due as much to quality of design of the game-based instruction (Clark et al. 2014; Leemkuil et al. 2003; ONeil et al. 2005).
Asking whether videogames and simulations can produce changes in how we think (skills) rather than what we know (abilities), however, is a fundamentally different question with profound implications. The impact of these screen technologies on xed abilities/traits such as those measured by the RPM has been less studied, yet what results exist suggest a similar benet of such technologies, including spatial visualization and divided attention (Tobias et al. 2011) and visual processing (Haier et al. 2009). Greeneld et al. (1994b), for example, observed that expert videogame players had faster reaction times (RTs) for low-probability stimuli locations in target acquisition tasks on computer screens. A second, follow-up study further suggested that 5 h of action videogame play (Robotron) ameliorated the difference between novices and experts. This causal effect for videogame exposure received further support from another study Greeneld and her colleagues did later that same year (Greeneld et al. 1994a). Just over a decade later, Green and Bavelier (2006) found similar results: action videogame players (those who played at least 5 h of action videogames per week) could count objects on a screen more accurately, track moving objects better, and do so faster than non-videogame players. Like Greeneld and colleagues had done 12 years earlier, they sought next to determine whether this was the result of self-selection or the effect of videogame play. They trained non-videogame players how to play the action videogame Medal of Honor, Allied Assault (EA Games 2002) until each had a basic level of mastery (determined by shothit ratios, number of successful kills, etc.). Researchers then had them play the game for an average of 1 h per day for 10 out of 13 days and
123
Can simulator immersion change cognitive style? 199
compared them to action videogame players using the same protocol as the prior study. They found that non-videogame players had the same visual processing skills as videogame players, providing evidence that it is the videogame play that produces changes in visual processing skills. Because ATC training is so heavily dependent on visual skills and often makes signicant use of simulated ATC experiences using technology, it may be a good venue for studying such effects.
If these theorists and researchers are correct, the question becomes what other xed abilities might, in fact, be changeable through exposure to simulations, videogames, and other screen time devices? One likely area is the cognitive style of eld-dependence/independence (FD-I). The test for FD-I, the Group Embedded Figures Test (GEFT; Witkin et al. 2002) shares a common underlying nonverbal methodology with the RPM: identifying patterns in complex images. In the GEFT, individuals are presented with a simple geometric shape and asked to nd and trace that same shape embedded within a larger, more complex image (see Fig. 2. In the RPM, individuals are presented with an incomplete pattern (a pattern with a missing section) and several possible candidates that could complete the pattern and are asked to select the best match (see Fig. 3).
These two tests rely on similar visual processing abilities. In fact, some research has found that the more eld independent (FI) you are, the more likely you are to score higher on the RPM (Tramer and Schuldermann 1974). Researchers have therefore suggested that FD-I may be an important variable to study in ATC operators (e.g., Hopkin 1980; Maliko-Abraham 2004).
FD-I is one of the few cognitive styles that has been found to be stable over a lifetime (Panek et al. 1980). The GEFT comprises 18 simple and complex gures embedded in other shapes that must be traced by the participants in a 5-min time period. Those who can nd the simple gures in the more complex gures are more FI. Those who are more FI do better in tasks that require similar disembedding of visual information from its surrounding eld.
FD-I has also been correlated with a wide range of academic performance indicators. In their study of 213 college freshmen, Ates and Cataloglu (2007) found that FI students scored higher on problem-solving skills than eld-dependent (FD) students. Others have found that those who are FI tend to do better in all areas of academics (Tinajero and Paramo 1997). Unfortunately, efforts to impact FD-I via interventions have been inconclusive (Ates and Cataloglu 2007), meaning that those who are FD may face a lifetime of academic challenges compared to their FI counterparts. Yet if RPM scores may change as a result of screen-time exposure, is it possible that GEFT scores, with their similar underlying testing methodology,
Fig. 2 Sample item from the
GEFT. Individuals must nd the rectangular shape at the left within the larger, more complex image at the right
123
200 R. Eck et al.
Fig. 3 Sample item from the RPM. Individuals must select one of the six patterned shapes at the bottom that best matches the missing pattern in the image at the top
may also change? Both are primarily visual pattern recognition tests and, as has been found with other visual abilities, may thus be inuenced by videogame play (Green and Bavelier 2006; Greeneld et al. 1994a; Greeneld et al. 1994b).
Purpose of the study
The purpose of this study was to examine whether exposure to technology (e.g., videogame play) could improve scores of FD-I on the GEFT as has been shown with other visual pattern recognition tests (e.g., the RPM and the tests of visual processing used by Green and Bavelier 2006). Specically, we sought to replicate the test protocols of Green and Bavelier (10 h of play of the action videogame Medal of Honor: Allied Assault; (EA Games 2002) with ATC students at a Midwestern university in the United States. In addition, because ATC students regularly practice and demonstrate visual pattern recognition skills such as those found by Green and Bavelier to increase as the result of videogame play (e.g., counting on-screen objects, tracking moving objects, speed of pattern recognition), we also sought to measure whether ATC students improved their performance on ATC tasks performed in simulators.
ATC students at this university spend approximately 15 h per week in a 360 ATC tower simulator. Students rotate between radar, pilot, and tower operations in classes and during open practice times. Radar simulation operators track simulated air trafc on a radar screen and control trafc by voice commands to pilots. ATC tower operators interact with radar simulation operators regarding the same
123
Can simulator immersion change cognitive style? 201
simulated air trafc, but do so inside a hexagonal room built of rear-projection video screen. Six projectors work in sync to project one-sixth of a 360 moving simulation of the eld of view within an ATC tower. The images are seamlessly combined via computer software to create simulated live images of air trafc at hundreds of airports under any conditions (e.g., night, day, rain, snow, sleet, hail, fair weather). Other ATC students play the role of pilots who control their planes based on what the radar and tower operators tell them. Performance during these simulations are graded by ATC experts. Because the skills used in simulated ATC operations are similar to those used in the Green and Bavelier study, we sought to examine whether, in addition to any changes in FD-I, ATC student performance in the simulators improved. Our research questions were (1) will replication of the Green and Bavelier (2006) protocol result in changes in FD-I? and (2) will replication of the Green and Bavelier protocol result in increased ATC simulator performance?
Study one
Participants
In the summer of 2006, we took a convenience sample from two ATC courses at a Midwestern university in the United States. One was a beginning course whose students had not yet taken any ATC simulation courses (courses in which they regularly experience use of a 360 virtual ATC tower). The other course was an advanced ATC course that is typically taken as one of the last three classes in the ATC program. These two courses were the only two courses taught during this summer term. With the exception of the number of ATC courses taken, the groups were equivalent on all other measures and were thus treated as a single pool. Because our research questions focused the effect of videogame play on GEFT score and ATC simulator performance, it was necessary to control for prior videogame play experience; only those who were considered to be non-videogame players (NVGPs) would be eligible. Students were classied as either action videogame players (VGPs) or NVGPs based on the cutoff score of 5 h of action videogame play per week used by Green and Bavlier (2006).
The remaining NVGPs were then randomly assigned to control (no videogame play) or experimental [10 h of videogame play following the Green and Bavelier (2006) protocol]. This yielded an initial sample of 27 control condition students and 24 experimental condition students. However, failure to complete all pre- and posttest measures resulted in small, unequal group sizes for experimental (n = 10) and control (n = 26). Unequal group sizes thus precluded meaningful analyses within- and between-groups. The experimental conditions required signicantly more time commitment from the students (an additional 1 to 2 h of game play for training and 10 h of game play over 15 days outside of class as per the Green and Bavelier protocol) than did the control condition (no out-of-class time required), which may explain the higher rate of attrition by missing data for the experimental conditions. The numbers lost from the experimental group came equally from the beginning (n = 8) and advanced (n = 6) students.
123
202 R. Eck et al.
The decision was made to recruit additional participants for the experimental condition. For reasons of scheduling and availability, these students were not successfully recruited until the summer of 2008. These new students were enrolled in an advanced class and were classied as VGPs and NVGPs using the same procedure as before. Because the experimental condition was the smallest group, the new sample of NVGPs (n = 10) were all assigned to the experimental condition. While the lack of random assignment creates a potential confound, because the students were identical in age, sex, major, and videogame experience, and were all from the same university, the same advanced class, and during the same term as the students from 2006, it was determined that their inclusion did not represent a signicant challenge to the assumption of group equivalence. The nal conditions comprising students from 2006 and 2008 consisted of 46 control (n = 26) and experimental (n = 20) participants, with 11 beginning students and 35 advanced ATC students in these groups (see Table 1). Participants were all between 18 and 24 years old and were recruited at the beginning of the course in which they were enrolled. None of the students had participated in Study One (those in 2006 had either graduated, left the program, or elected not to participate; those in 2008 were all advanced students and had graduated by the start of study two).
Procedure
The researchers developed a research protocol document detailing how each part of the research would be conducted, and all researchers and graduate research assistants were trained on the protocol. This included administration of the GEFT as described in the GEFT training materials and manual and additional elements such as the use of stopwatches to record completion times for each section of the GEFT. The protocol document also included all procedures and data collection methods and scripts for the experimental (game) conditions. All students completed a videogame survey for the purposes of classication as VGP or NVGP as well as the paper-and-pencil GEFT administered by the researchers with assistance from graduate research assistants during the rst 30 min of class. Initial data on ATC simulator performance were collected next, before the NVGPs in the experimental condition played 10 h of the game Medal of Honor, Allied Assault (EA Games 2002) over 13 days, replicating the research intervention protocol of Green and Bavelier (2006) described earlier. The GEFT was then administered again by the researchers and simulator performance data were collected.
Table 1 Participants by ATC experience, year of recruitment, and condition assignment
ATC exp. 2006 2008 Final N
Control Exp. Control Exp. Control Exp. Total
Beginning 9 2 0 0 9 2 11
Advanced 17 8 0 10 17 18 35
Total 26 10 0 10 26 20 46
123
Can simulator immersion change cognitive style? 203
Results
Between and within-subject analysis indicated that neither simulator performance nor FD-I were affected by videogame play, failing to extend Green and Baveliers ndings (2006) to FD-I or real-world tasks involving similar visual pattern recognition skills. However, we unexpectedly found that our sample contained more FI students than would be expected given the average of FI in the general population.
Further analysis revealed that this difference was localized to the advanced ATC students (those who were at the end of the program). Advanced students had a mean score of 15.34 out of 18, placing them in the fourth quartile of the distribution (FI). Accordingly, their distribution of GEFT scores was highly negatively skewed (skewness = -1.859, kurtosis = 4.36, p \ .001). The general population, in contrast, is distributed equally across all four quartiles. Figures 4 and 5 show the distribution of GEFT scores for beginning and advanced ATC students. A one-way ANOVA showed that the difference between beginning and advanced ATC student GEFT scores was statistically signicant: F = 12.42 (1, 44), p = .001.
However, the skewness of the advanced student GEFT scores and the unequal group sizes make interpretation of an ANOVA problematic. A WilcoxonMann Whitney test, which relies on rank scores rather than means, also revealed that
Fig. 4 Number of beginning (white bars) and advanced (patterned bars) ATC students in each quartile of the GEFT distribution as measured in 2006 and 2008
123
204 R. Eck et al.
Fig. 5 Distribution of GEFT scores for beginning and advanced ATC students as measured in 2006 and 2008
advanced students GEFT score distribution was signicantly different from that of
beginning students: U = 139.50, z = -2.17, p \ .016.
Finally, based on the cutoff scores from the normative data for the GEFT, we classied our participants into four quartiles by their scores. We then compared their score distribution to the normed distribution for GEFT scores. Chi square analysis conrmed that advanced students differed from the normal population: X2 (3, N = 36) = 23.33, p \ .001. Beginning students did not.
Study two
Participants
To examine whether our sample of beginning and advanced students at the institution was atypical, we invited the entire population of all ATC majors to take the GEFT between October 20 and November 3 in 2009 (see Table 2 for breakdown of total ATC students by year of experience, or cohort). This sample potentially included some of the beginning students from the 2006 sample; all of the advanced students from 2006 graduated prior to this.
123
Can simulator immersion change cognitive style? 205
Table 2 Number of students by academic year of program (Cohort) in ATC Program
A total of 155 out of 205 possible ATC majors participated, yielding an initial participation rate of 75.6 %. Ages ranged from 18 to 24. Eight students did not have complete data and were deleted from the data set. Of the remaining 147 participants, 116 (23 females) were non-transfer students: 31 (4 females) were transfer students. Unless otherwise noted, all tests were run on non-transfer students only to control for curriculum training from outside of the university.
Procedure
Procedures for administering the GEFT were identical to study one. No additional measures or interventions occurred.
Results
GEFT scores for rst-year students were once again normally distributed (M = 12.76, SD = 4.26). The distribution of GEFT scores for fourth-year students was once again negatively skewed (M = 13.88, SD = 3.81; skewness = -1.86, kurtosis = 3.18, p \ .016). The distribution of GEFT scores for intermediate students (those who were in their second and third year of ATC courses) was also negatively skewed, although not to the same extent as the fourth-year students (skewness = -.42, kurtosis = 1.29, p \ .016). With the exception of the second-year ATC students, the GEFT scores for each year of the ATC study (beginning through advanced students) were not normally distributed (KolmogorovSmirnov p = .034, .141, .006, and .049), making the use of ANOVA unadvised. Means did not appear to differ signicantly, but further statistical analysis using SPSS boxplots with outliers revealed four extreme cases in the advanced student group (none in the other groups). Removing these extreme cases resulted in a much higher mean for advanced students (M = 15.14, SD = 1.63). In addition, the distribution of GEFT scores appeared to differ between beginning and advanced students. As with study one, GEFT scores were once again categorized by quartile based on the normative data for GEFT scores as recommended in the manual (i.e., 25 % of the general population scores between zero and nine on the test, so ATC student scores in this range were categorized as Quartile One). Figures 6, 7 and 8 present GEFT score quartile distribution by cohort.
As can be seen from the gures, fourth-year (advanced) ATC students are more likely to show up in the third and fourth quartiles than in the rst or second, and
Cohort Frequency Percent
Year One 27 18.4
Year Two 37 25.2
Year Three 34 23.1
Year Four 49 33.3
Total 147 100.0
123
206 R. Eck et al.
Fig. 6 Number of beginning (white bars) and advanced (patterned bars) ATC students in each quartile of the GEFT distribution as measured in 2009
beginning students are more likely to show up in the rst and second quartiles. It also appears that advanced students have higher GEFT scores. To examine whether any of these perceived differences were signicant, we ran Chi square analyses of all years by all quartiles. Results approached statistical signicant (p = .052), and closer examination of the table of scores revealed that 96 % of the Year four students were in the third and fourth quartiles. Analysis of the adjusted standardized residuals (ASR) in this initial Chi square suggested that Years One through Three were normally distributed across all four quartiles (i.e., no ASRs were greater than an absolute value of two). Year four appeared to be disproportionately represented in the rst (ASR = -2.8) and third quartiles (ASR = 2.4).
Because the differences appeared to be localized predominantly in the Year four students, we examined the data using other accepted classication systems for GEFT scores, including using the mid-point score as a cutoff (i.e., the 50th percentile) and comparing extreme scores (Quartile 4 vs. Quartile 1). For the 50th percentile cutoff score, results indicated that Year four students were more likely to be FI than FD, v2 (3) = 13.078, p = .004. The effect size for this difference,
Cramers V, was moderate, .34 (Cohen 1988). ASRs for Year four were -3.6 (FD) and 3.6 (FI). No other ASRs were greater than an absolute value of two. Results using the 25th and 75th percentiles (extreme FD-I) were also signicant, v2
123
Can simulator immersion change cognitive style? 207
Fig. 7 Number of ATC students at all levels of program experience (Years OneFour) in each quartile of the GEFT distribution as measured in 2009
(3) = 8.525, p = .036, indicating that Year Four students were not only more likely to be classied as FI than FD but that their scores tended to be much higher. The effect size was once again moderate at .36.
To replicate our analysis from study one, we next compared rst-year to fourth-year students, once again using Chi square analysis and variations in the GEFT score classications as described above. Results were once again signicant using the 50th percentile cutoff score, v2 (1) = 12.325, p \ .0001. Phi was larger in this case, .48, indicating a medium effect size. ASRs were 3.5 and -3.5 for both Year
One and Year Four students, with the direction indicating that there were fewer FI Year One students and more FI Year Four students than expected by chance. Similar results were found using the 25th and 75th percentile GEFT scores, v2 (1) = 7.446, p \ .006. The effect size was once again medium at .49.
To examine whether differences could be attributed to localized experiences (i.e., the curriculum and environment at this university alone), we further compared the scores of transfer and non-transfer students (see Table 3 for transfer students in each cohort). The analysis indicated there were no differences between transfer and non-transfer students in Years One through Three using either the 50th percentile or the
123
208 R. Eck et al.
Fig. 8 Distribution of GEFT scores for ATC students at all levels of program experience (Years OneFour) as measured in 2009
Table 3 Number of transfer students by academic year of program (Cohort) in ATCProgram
25th/75th percentiles (extreme FD-I). Chi square analysis of Year Four transfer and non-transfer students using the 50th percentile cutoff GEFT score, however, were statistically signicant, v2 (1) = 10.233, p = .001, with a medium effect size of .48.
Examination of the ASRs indicated that non-transfer students were more likely to be FI (ASR = 3.2) than were transfer students (ASR = -3.2). Similar results were obtained when using the 25th/75th percentiles (extreme FD-I), v2 (1) = 9.938, p = .002, with a medium effect size, .66.
While the rst two studies suggest that ATC students become more FI over time by some means other than self-selection into majors or higher rates of retention for those who are FI than those who are FD, the cross-sectional nature of this study precludes any interpretation of causality. The authors hoped that a third,
Cohort Frequency Percent
Year One 2 6.5
Year Two 3 9.7
Year Three 9 29.0
Year Four 17 54.8
Total 31 100.0
123
Can simulator immersion change cognitive style? 209
longitudinal study to track the progress of Year One students from 2009 through 2013 would provide more denitive evidence of such a change. Unfortunately, signicant attrition rendered that study inconclusive. With each passing year, the students became less interested in completing the GEFT again (there was no course credit given and no perceived benet to them). This loss was equally distributed across students classied as FD and FI. The authors hope to replicate the longitudinal study in the near future.
Discussion
Because of the cross-sectional nature of this study, all that can be said with certainty is that videogame play of 10 h over 13 days did not seem to improve eld independence or simulator performance, that ATC students appear to be more FI than the general population, and that this difference appears to be localized primarily to advanced ATC students at this specic university (i.e., non-transfer students).
However, our data do raise interesting questions. What is it that accounts for the high number of FI students in their fourth year of ATC study? One answer could be that FI students are disproportionately drawn to ATC study, yet beginning students in our study did not differ from the general population in terms of FD-I classication. Another possible explanation is that FI students are retained in the program longer because of the advantage they have over FD students. However, several factors combine to make this explanation less likely. First, beginning students in our study were no more FI than the general population in either 2006, 2008, or 2009 (three cohorts). It is possible that the 2006 cohort of beginning students (those who would be in Year Four by the fall of 2009) were more FI than the general population and than the 2006 sample in our rst study and the 2007 (our Year Two students in 2009) and 2008 (our Year Three students in 2009) cohorts. This cannot be ruled out, of course, but there is no evidence that any other cohort has been more FI than the general population.
It is also possible that FI students were merely retained at a higher rate than FD students, yet we would expect a smaller class size in Year Four (as FD students were selected out of the program), which was not the case. Likewise, it is possible that FI students from the 2006 cohort were retained at a higher rate than FD students, yet the 2009 Year Four student class size was similar to previous Year Four classes as reported by program faculty and there were actually more Year Four students than Year One students (see Table 2). Further, when the number of transfer students are removed from Year Four, the class sizes are equivalent (see Table 4). It therefore seems less likely that the higher proportion of Year Four FI students can be easily explained by self-selection.
While only a longitudinal study can answer these questions specically, it would seem that there is sufcient evidence to at least allow for the possibility that FD-I may, in fact, be something that can be changed by environmental factors. Accordingly, and at the risk of speculation, we explore this possibility in the nal section of this paper. If FD-I is not as stable as once thought, what mechanisms
123
210 R. Eck et al.
Table 4 Number of non-transfer students by academic year of program (Cohort) inATC Program as measured in 2009
might explain such changes? It is not possible to account for all the possible environmental factors that could contribute to a change in FD-I, of course. Yet, an examination of the nature of ATC training and real-world performance yields some interesting possibilities.
ATC students in the studies described here take classes in the same building as all other aviation majors, including commercial pilots. Prior research has shown that college students are no more or less FD than the general population (e.g., Witkin et al. 2002). It makes sense, then, to examine what parts of ATC training are unique to ATC majors (since other students experience education in typical college classrooms without impacting FD-I). As described earlier, ATC majors at this university spend up to 2000 h in ATC simulators (radar screens, 360 tower simulators). No other students ever access these simulators. The simulation training is designed to mimic ATC performance in the real-world, post graduation. What evidence do we have that this kind of exposure could impact visual processing and cognitive abilities in a way that might lead to changes in FD-I?
Previous research suggests that changes in visual processing, brain structure, and cognitive function may result from exposure to complex digital environments like simulations and videogames (Greeneld et al. 1994a, b; Green and Bavelier 2006). Given the amount of simulation experiences these students are exposed to, it is not unreasonable to expect some changes in visuospatial and cognitive reasoning: a conclusion bolstered by the Flynn Effect. Given the similarity of the GEFT and the RPM and the positive correlation of eld independence and intelligence scores on the RPM (Tramer and Schuldermann 1974), one cannot rule out a change in FD-I as a result of ATC experience and simulation. Of course, another factor that could explain differences in the current studies is that most ATC students at this university are also commercial pilot and aviation majors. They also spend many hours in ight simulators and in planes, which could possible affect FD-I. On the other hand, prior research has shown that pilots are no more FI than the general population (Carretta 1987). We hope to further explore this in a future study of pilot and commercial aviation majors.
If intensive simulation exposure played a part in Year Four student increases in eld independence, what might account for such a change? Simulations are not the same thing as games, of course, yet the simulation in this study supports and requires the same kinds of visual processing outcomes found to increase with videogame play (e.g., tracking moving objects, counting objects, reaction time). Greeneld (2009) explicates an argument for human developmental changes more specically to account for, among other things, an increase in visual processing
Cohort Frequency
Year One 27
Year Two 35
Year Three 25
Year Four 32
Total 119
123
Can simulator immersion change cognitive style? 211
skills that result from our societal shifts, including technology exposure. Green-elds model further suggests that such changes are incremental and take place over a long time. While the 4 years ATC students spend in their program may not be equivalent to the 10 years it takes to produce changes in the RPM in different populations, it is nonetheless a longer time frame than we often use in studies like this. And the length of time may not be as critical as the rate of exposure; the students in this program spend one-third of their entire academic time in this immersive simulator. Yet, if simulator exposure plays a role in promoting eld independence, and if that experience reects what ATC operators experience on the job, there should be evidence that ATC operators in general are more FI and that this difference is related to their time on the job (i.e., their exposure to real-world radar screens and tower operation environments).
Maliko-Abraham (2004) compared military ATC operators to non-ATC operators in the military to test the hypothesis that ATC operators were more FI. She found that the number of ATC operators classied as FI was statistically different from non-ATC operators. Specically, 74 % of Air Force and 68 % of naval ATC operators were FI, while only 47 % of non-ATC operators in the same services were FI. This was attributed to the similarities between ATC tasks and the kinds of visual processing at which FI learners are good: The core of what an [air trafc control specialist] does on a day to day basis is to distinguish (disembed) from the totality of information available to them, the information that is essential to their taskthat of determining the exact coordinates (direction, projects ight path, etc.) of any given aircraft in their airspace in order to maintain safe separation (Maliko-Abraham 2004, p. 1213).
So there is some evidence that ATC students are more FI by graduation and that military ATC operators are also more FI. While neither our current study nor that of Maliko-Abraham (2004) can make any claim of causation (e.g., both were cross-sectional studies, and no evidence to rule out self-selection was presented by Maliko-Abraham), we believe that the possibility that intensive ATC training and experience may increase FI warrants further study. There is little evidence that FD-I changes over time except with age, as people in general become more FD (Panek 1985), so any changes would most likely be due to environmental factors.
First, it may suggest that such changes may be highly inuenced by intensive, cognitively engaging technology experiences. ATC students are immersed in this high-delity, authentic simulator as part of their professional practice; because they have chosen (and are paying for) this degree as a path to their careers, it can be presumed that they are fully engaged and highly motivated when in this environment. More typical technology screen exposure that comes from computers, television, etc., may not produce the same effects, or perhaps not at the same rate. Previous studies of technology and human development may have neglected to account for the required amount/rate of exposure needed. As a result, this study may suggest a threshold effect for media exposure; 2000 h of concentrated exposure to immersive technology with full cognitive engagement over a 36-month period is a standard not easily accomplished. Simulations like those used in ATC may be particularly efcient in promoting FI, which suggests the need for future targeted studies.
123
212 R. Eck et al.
It may be just as likely that it was not the simulator itself so much as its ability to replicate real-world ATC environments, with similar cognitive components and requirements. Military ATC operators were found to be more FI than non-ATC operators (Maliko-Abraham 2004), and only one of the 19 had a bachelors degree (that is, almost all of them were trained on the job).
Whether it is simulator experience, ATC task experience, a combination of the two, or some other unknown factor, the possibility remains that FD-I is not as stable as we have assumed. If so, this has signicant implications for education and society. Interventions that may promote FI in FD individuals can make signicant emotional and nancial differences in the lives of at least 25 % of the population. Future research should replicate this study in longitudinal designs and with exposure to other intensive technologies for similar amounts of time.
Limitations and future research
Several factors in the current study limit generalizability and interpretation of ndings. The cross-sectional nature of the study precludes any causal inferences. Despite the evidence that suggests a change over time due to environmental factors, no inferences can be made and our speculations about causal factors such as ATC environments and simulation exposures, however well-reasoned, are just that: speculation. In addition to a longitudinal study design, future research should examine pilots and commercial aviation majors who are also exposed to signicant simulation experience. Longitudinal studies of federal and military ATC operators should also be conducted to further examine the role of ATC experience (e.g., non-simulated) on FD-I. Finally, cognitive task analyses of real-world and simulated ATC performance should be conducted to examine what factors may be most aligned with both the RPM and GEFT tests to better understand any potential mechanism for changes in FD-I.
Compliance with ethical standards
Conict of interest The authors declare that they have no conict of interest.
Ates, A., & Cataloglu, E. (2007). The effects of students cognitive styles on conceptual understandings and problem-solving skills in introductory mechanics. Research in Science and Technological Education, 25(2), 167178.
Carretta, T. R. (1987). Field dependenceindependence and its relationship to ight training performance(Technical Report). Brooks Air Force Base, TX: Air Force Research Laboratory.
Clark, D. B., Tanner-Smith, E. E., & Killingsworth, S. (2014). Digital games, design, and learning: A systematic review and meta-analysis. Menlo Park, CA: SRI International.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: LawrenceErlbaum.
EA Games. (2002). Medal of Honor: Allied Assault, (software). Redwood City, CA: Electronic Arts. Edwards, C., & Poole, R. W. Jr. (2010). Airports and air trafc control. Washington, D.C.: The CATO
Institute. Available from http://www.downsizinggovernment.org/transportation/airports-atc
Web End =http://www.downsizinggovernment.org/transportation/airports-atc .
References
123
Can simulator immersion change cognitive style? 213
Flynn, J. R. (1987). Massive IQ gains in 14 nations: What IQ tests really measure. Psychological Bulletin,101, 171191. doi:http://dx.doi.org/10.1037/0033-2909.101.2.171
Web End =10.1037/0033-2909.101.2.171 .
Green, S. C., & Bavelier, D. (2006). Enumeration versus multiple object tracking: The case of action video game players. Cognition, 101, 217245.
Greeneld, P. M. (2009). Linking social change and developmental change: Shifting pathways of human development. Development Psychology, 45(2), 401418.
Greeneld, P. M., Brannon, C., & Lohr, D. (1994a). Two-dimensional representation of movement through three-dimensional space: The role of videogame expertise. Journal of Applied Developmental Psychology, 15, 87103.
Greeneld, P. M., deWinstanley, P., Kilpatrick, H., & Kaye, D. (1994b). Action videogames and informal education: Effects on strategies for dividing visual attention. Journal of Applied Developmental Psychology, 15, 195223.
Haier, R. J., Karama, S., Leyba, L., & Jung, R. E. (2009). MRI assessment of cortical thickness and functional activity changes in adolescent girls following three months of practice on a visual-spatial task. BMC Research Notes,. doi:http://dx.doi.org/10.1186/1756-0500-2-174
Web End =10.1186/1756-0500-2-174 .
Hays, R. T. (2005). The effectiveness of instructional games: A literature review and discussion. Technical Report 2005-004. Naval Air Warfare Center Training Systems Division. http://oai.dtic.mil/oai/oai?verb=getRecord&metadataPrefix=html&identifier=ADA441935
Web End =http://oai.dtic. http://oai.dtic.mil/oai/oai?verb=getRecord&metadataPrefix=html&identifier=ADA441935
Web End =mil/oai/oai?verb=getRecord&metadataPrex=html&identier=ADA441935 .
Hopkin, V. (1980). The measurement of the air trafc controller. Human Factors, 22(5), 547560. Johnson, S. (2006). Everything bad is good for you. New York: Riverhead.
Leemkuil, H., de Jong, T., de Hoog, R., & Christoph, N. (2003). KM quest: A collaborative internet-based simulation game. Simulation and Gaming, 34, 89111.
Maliko-Abraham, H. (2004). A study of eld independence and United States military air trafc controllers. In Proceedings of the Human Factors and Ergonomics Society 48th Annual Meeting, New Orleans, LA, September 20 24, 2004, 12131217.
ONeil, H. F., Wainess, R., & Baker, E. (2005). Classication of learning outcomes: Evidence from the computer games literature. The Curriculum Journal, 16(4), 455474.
Panek, P. E. (1985). Age differences in eld-dependence/independence. Experimental Aging Research,11(2), 9799.
Panek, P. E., Funk, L. G., & Nelson, P. K. (1980). Reliability and validity of the group embedded gures test across the life span. Perceptual Motor Skills, 50 (3, part 2), 11711174.
Randel, J. M., Morris, B. A., Wetzel, C. D., & Whitehill, B. V. (1992). The effectiveness of games for educational purposes: A review of recent research. Simulation and Gaming, 23(3), 261276. Raven J. C. (1936). Mental tests used in genetic studies: The performance of related individuals on tests mainly educative and mainly reproductive. MSc Thesis, University of London.
Tinajero, C., & Paramo, M. F. (1997). Field dependenceindependence and academic achievement : A reexamination of their relationship. British Journal of Educational Psychology, 67(2), 199212. Tobias, S., Fletcher, J. D., Dai, D. Y., & Wind, A. P. (2011). Review of research on computer games. In S.
Tobias & J. D. Fletcher (Eds.), Computer games and instruction (pp. 127221). Charlotte, NC: IAP. Tramer, R. R., & Schuldermann, E. H. (1974). Cognitive differentiation in a geriatric population.
Perceptual and Motor Skills, 39, 10711075.
Van Eck, R., & Dempsey, J. (2002). The effect of competition and contextualized advisement on the transfer of mathematics skills in a computer-based instructional simulation game. Educational Technology Research and Development, 50(3), 2341.
Vogel, J. F., Vogel, D. S., Cannon-Bowers, J., Bowers, C. A., Muse, K., & Wright, M. (2006). Computer gaming and interactive simulations for learning: A meta-analysis. Journal of Educational Computing Research, 34, 229243.
Witkin, H. A., Oltman, P. A., Raskin, E., & Karp, S. A. (2002). Group embedded gures test manual.Menlo Park, CA: Mind Garden.
Richard Van Eck is the founding Dr. David and Lola Rognlie Monson Endowed Professor in Medical Education, and Associate Dean for Teaching and Learning at the University of North Dakota School of Medicine and Health Sciences. He has a BA in English and Psychology, an MA in English, and a PhD in instructional design. He has been studying games since his doctoral studies at the University of South Alabama, where he worked on Adventures in Problem Solving (Texas interactive Media Award, 1999), and Ribbits Big Splash (Gulf Guardian Award 2002). His recent work includes serving as an evaluation
123
214 R. Eck et al.
and game design consultant on several other games, including PlatinuMath (mathematics for preservice teacher education), Project NEO (science for preservice teachers), Project Blackfeather (programming for middle school students), Contemporary Studies of the Zombie Apocalypse (mathematics for middle school students). He is a frequent keynote speaker and author on the educational potential of video games. He also publishes and presents on intelligent tutoring systems, pedagogical agents, authoring tools, and gender and technology.
Hongxia Fu received her PhD in Teaching and Learning from the University of North Dakota, with a focus on Instructional Design and Technology. Her research interests include instructional design, educational technology integration, and program evaluation. She is a Research Analyst in the Institute for the Science of Teaching and Learning, Arizona State University.
Paul Drechsel is an Associate Professor of aviation at the University of North Dakota. Paul has over 25 years of academic and Air Trafc Control (ATC) experience. Since September 1998, Pauls responsibilities have included teaching and air trafc control program development. He is the Assistant Chairman for Air Trafc Control for the ICAO Air Trafc Control Training Program and Federal Aviation Administration FAA, ATC-Collegiate Training Initiative (CTI) approved program at the University of North Dakota. His areas of expertise are air trafc control and instructional design and technology. His current research efforts include cooperative use of airspace with manned and unmanned aircraft, and cognitive style as a function of visual training in air trafc control students.
123
Springer Science+Business Media New York 2015