Demonstrating change from a drop-in space soundscape exhibit by using graffiti walls both before and after
DDemonstrating change from a drop-in space soundscape exhibit byusing graffiti walls both before and after
Martin O. Archer
1, 2 , Natt Day , and Sarah Barnes Space and Atmospheric Physics, Department of Physics, Imperial College London, London, UK School of Physics and Astronomy, Queen Mary University of London, London, UK Centre for Public Engagement, Queen Mary University of London, London, UK
Correspondence:
Martin O. Archer([email protected])
Abstract.
Impact evaluation in public engagement necessarily requires measuring change. However, this is extremely challeng-ing for drop-in activities due to their very nature. We present a novel method of impact evaluation which integrates graffiti wallsinto the experience both before and after the main drop-in activity. The activity in question was a soundscape exhibit, whereyoung families experienced the usually inaudible sounds of near-Earth space in an immersive and accessible way. We applytwo analysis techniques to the captured before and after data — quantitative linguistics and thematic analysis. These analysesreveal significant changes in participants’ responses after the activity compared to before, namely an increased diversity oflanguage used to describe space and altered conceptions of what space is like. The results demonstrate that the soundscape wassurprisingly effective at innately communicating key aspects of the underlying science simply through the act of listening. Theimpacts also highlight the power of sonification in stimulating public engagement, which through reflection can lead to alteredassociations, perceptions and understanding. Therefore, we show that this novel approach to drop-in activity evaluation, usinggraffiti walls both before and after the activity and applying rigorous analysis to this data, has the power to capture change andthus short-term impact. We suggest that commonly used evaluation tools suitable for drop-in activities, such as graffiti walls,should be integrated both before and after the main activity in general, rather than only using them afterwards as is typicallythe case.
Drop-in activities — short, interactive, two-way engagements — tend to form a significant fraction of all non-school publicengagement, e.g. ± of all public activities across the UK’s South East Physics Network in 2017/2018 were less than30 min in duration per individual (Galliano, 2018). Such activities though are difficult to effectively evaluate the impact of, sincethis necessitates a measure of change on participants (King et al., 2015). While surveys both before and after may be one of themost robust methods of impact evaluation in general (Jensen, 2014), these are neither appropriate for nor commensurate withdrop-in activities. This is because participants are arriving all the time, the engagement duration is so short, and surveys riskaffecting participants’ experience (Grand and Sardo, 2017). A number of evaluation tools more suitable for drop-in activitieshave been reported including feedback cards, rating cards, snapshot interviews, and graffiti walls (e.g. Grand and Sardo, 2017; a r X i v : . [ phy s i c s . e d - ph ] F e b ublic Engagement with Research team, 2019). Graffiti walls are large areas (often a wall, whiteboard, or large piece of paper)where participants are free to write or draw responses in reaction to the engagement activity or some prompt question, eitherdirectly on the area itself or by sticking responses to it. All of these evaluation methods for drop-ins are particularly useful inprocess evaluation — assessing the implementation of the activity. Under typical usages (post-activity only) though, they arelimited in their ability to routinely demonstrate change from, and thus the impact of, the engagement activity on participants ingeneral.This paper presents a novel implementation of graffiti walls for impact evaluation, integrating them into both the start andend of a drop-in activity. The activity was a soundscape experience surrounding current space science research that used geo-stationary satellite data converted into audible sound. We show that this evaluation method (through its design, data collection,and analysis) can indeed capture immediate impact — changed language and conceptions of space in this case. Appendicesinclude details of statistical and qualitative coding techniques employed throughout. A common misconception is that space is a true vacuum completely devoid of matter and thus there is no activity other thanthat of the celestial bodies, e.g. planets or asteroids. However, the universe is permeated by tenuous plasmas — gases formed ofelectrically charged ions and electrons that generate and interact with electromagnetic fields (e.g. Baumjohann and Treumann,2012). One such example is the solar wind streaming at several hundreds of kilometres a second from the Sun to the edge of theheliosphere, something which only ± of the UK adult population are aware of (3KQ and Collingwood EnvironmentalPlanning, 2015). Space plasmas are also not just limited to our solar system, with other stars having their own stellar winds(e.g. Lamers and Cassinelli, 1999) and the interstellar medium bridging the gap between these plasma bubbles in outer space(Gurnett et al., 2013).The presence of a medium in space allows for plasma wave analogues to ordinary sound (pressure waves) that occur atultra-low frequencies — fractions of milliHertz up to 1 Hz. They are routinely measured by many space missions and canhave perturbations that are significant fractions of the background values. For a further discussion of the equivalence of theseplasma waves to sound see Archer (2020a). One way in which ultra-low frequency waves are generated is through the highlydynamic solar wind buffetting against Earth’s magnetic field. This process plays a key role within space weather and thus howphenomena from space can affect our everyday lives (e.g. Keiling et al., 2016). However, the belief by the public that spaceis completely empty in turn leads many to incorrectly think that there is absolutely no sound in space, reinforced by schoolscience demonstrations such as the bell-jar experiment (see Caleon et al., 2013, for a nuanced discussion of this experimentand sound in near-vacuum conditions) or even popular culture like in the marketing to the movie ‘Alien’ (“in space no one canhear you scream”). Public engagement with this research area may help correct this fallacy.Sonification — the use of non-speech audio to convey information or perceptualise data (Kramer, 1994) — can be used toconvert satellite measurements of these usually inaudible space sounds into audible signals, simply by dramatically speedingup their playback (Alexander et al., 2011, 2014). This has already been leveraged in public engagement projects for both ntrance Undergraduate
AmbassadorsTable, post-it notes, pens, headphonesPre-soundscapegraffiti wall Post-soundscapegraffiti wallResearchers
Exit T a b l e , po s t- i t no t es , p e n s Banner standsCreative Short Films
Figure 1.
Layout and photos of the soundscape exhibit. scientific and artistic outputs (Archer et al., 2018; Archer, 2020b). Sonification in general has been applied to various scientificdatasets (Feder, 2012). Supper (2014) posits that through the public experiencing data in this way it can grip their imaginationand produce sublime experiences because of sound’s immersive and emotional nature. These arguments, however, are mostlybased on reflections from researchers and artists, rather than through the evaluation of participants’ own thoughts and feelings.This paper evaluates the short-term impact on participants of experiencing the sounds of space using graffiti walls both beforeand after a soundscape.
The space soundscape exhibit was held at the free Science Museum in London (United Kingdom) whose informal learningadopts an inclusive, accessible ‘science capital’ approach that attracts a diverse range of audiences (Science Museum Group,2017, 2020). ‘Science capital’ is defined as the total science-related knowledge, attitudes, experiences, and resources that aperson has built up over their life (Archer and DeWitt, 2017). This includes what science they know about, what they think andfeel about science, the people they know and their relation to science, and the day-to-day engagement they have with science.The exhibit formed part of the museum’s ‘Summer of Space Season’, held in celebration of the 50th anniversary of the Apollomoon landings, for which the museum solicited drop-in space-themed activities aimed at young families. It ran between thehours 12:00–16:00 during the May 2019 half-term school holiday over the course of 4 days. he purpose of the space soundscape was primarily to provide young children and their parents/carers (as key influencesupon them) an accessible and immersive experience with space research that would enable participation and spark discussion.Such experiences may, when taken in conjunction with all the other formal and informal interactions with science afforded toa young person, contribute towards developing their science identity and hence build their ‘science capital’. Using a genericlearning outcomes framework (Hooper-Green, 2004), the main intentions for the activity fell within the realms of ‘Enjoyment,Inspiration, Creativity’ and ‘Attitudes & Values’, with explicitly enhancing ‘Knowledge & Understanding’ being only a sec-ondary aim. Figure 1 shows the layout of the exhibit, which was integrated amongst the museum’s usual collections, alongwith accompanying photos. The activity worked as follows:1. Museum attendees were invited to participate at the entrance by undergraduate ambassadors. They were first askedto write or draw on a post-it note what they think space around our planet is like. Some younger children requiredfurther prompting beyond this broad question however, with ambassadors often asking “what do you think space soundslike?” The participants placed their responses on the pre-soundscape graffiti wall and were handed bluetooth wirelessheadphones playing the sounds of space.2. Participants went on a journey while listening to the sounds, following a set of coloured arrows marked out on the floor.A number of banner stands with further information about the sounds were placed along this path, though it was observedthat few people read these. This may be either because participants preferred to listen to the sounds or that it was notclear the stands were part of the experience given the exhibit’s location amongst other collections.3. Near the end of the journey, researchers took participants’ headphones and asked them to reflect on what they think aboutspace after having listened to the sounds. Participants then recorded their thoughts on post-it notes again and placedthese on the post-soundscape graffiti wall. The researchers would use what they had written or drawn to prompt a shortdialogue about aspects of the space environment around Earth and space weather research. This method was informedby the ‘science capital’ research (Archer and DeWitt, 2017), which recommends scientists use and value participants’own experiences within their engagement practice to help enable lower ‘science capital’ audiences to feel included inscience and that science is for “people like me”. These discussions provided an opportunity to solidify, or in some casesclarify, the associations that participants made from the soundscape experience in a tailored and audience-focused way(e.g. only going into an appropriate level of detail depending on the individual or group).4. Finally, researchers would change the channel on the headphones so that participants could watch on a large TV screen aseries of creative short films inspired by and incorporating the sounds (Archer, 2020b). The films also featured epiloguetext reinforcing the importance and relevance of space weather research. Surprisingly, these artistic films proved muchmore popular than anticipated.The graffiti walls were used as an open opportunity for participants to reflect upon their perceptions and associations with spaceboth before and after the soundscape, with this being intentionally left broad to elicit a wide range of possible responses andthus potential impacts. This method was chosen specifically due to its suitability for evaluating drop-in activities, ability to be ntegrated within the activity itself, and alignment with our intended overall experience for participants. While graffiti walls area common evaluation tool, we are unaware of any published public engagement activity that has captured and analysed databoth before and after a drop-in activity using them. This makes our evaluation approach for the exhibit novel.Ethical considerations in the design of the exhibit and its evaluation followed the British Educational Research Association(BERA, 2018) guidelines and were discussed with institutional funders and the Science Museum before the activity occurred.All respondents consented to providing graffiti wall responses as these were not mandatory for participation in the soundscapeexhibit. Children only participated in any of the activities when accompanied by their appropriate adult. All data collected wasanonymous and no characteristics about participants were solicited. Overall it was deemed (due to the nature of the exhibit, itsdesign, and the types of responses being collected) there was very little risk of harm arising from participation.The space soundscape was experienced by 1,003 people, recorded using a tally counter. The majority were in family groups(approximately three-quarters were children based on observations) with some independent adults also. It was observed that infamilies typically only the children contributed to the graffiti walls (with no substantive difference in respondents before andafter) and in many cases accompanying adults did not take headphones when offered, perceiving the activity as just for theirchildren. There were 535 and 446 responses (predominantly textual) on the pre- and post-soundscape graffiti walls respectively,rates of ± and ± . This is some 3–10 times greater than reported for typical graffiti walls (Public Engagement withResearch team, 2019) likely due to their integration into the overall activity here. The data captured on the pre- and post-soundscape graffiti walls are displayed in Figure 2. However, simply presenting these isinsufficient to robustly demonstrate any potential changes and thus impacts. Instead, analysis is required and two approachesare taken here, namely quantitative linguistics and thematic analysis.
Quantitative linguistics investigates language using statistical methods and has uncovered several linguistic laws that mathe-matically formulate empirical properties of languages. One of these is Zipf’s law — the frequency of words are approximatelyinversely proportional to their rank (where the more often a word is used the higher its rank, i.e. closer to 1) (Zipf, 1935, 1949).An alternative way this law is stated is that the statistical distribution of word ranks follows a power law with an exponentthat is typically quoted as − . Zipf’s law holds well for almost all languages as well as many other human-created systems(Piantadosi, 2014). The Zipf exponent, however, can vary and is a measure of the diversity of words. Baixeries et al. (2013)showed that children’s Zipf exponents become less-negative / shallower with age, demonstrating increasing variety of languageand thus linguistic complexity as they develop. However, we are not aware of Zipf’s law being exploited in public engagementevaluation before.Figure 3 shows rank-frequency plots of the textual responses to the soundscape before and after the experience. This par-ticular analysis thus omits any purely pictorial responses. Ties in ranks have been accounted for using standard competition ) Before (n=71) d) After (n=55) a) Before (n=464) b) After (n=391) Figure 2.
Wordclouds (a,b) and drawn images (c,d) from both before (a,c) and after (b,d) experiencing the soundscape. ranking (also known as “1224” ranking, where a gap is left following the tie). It is clear from these plots that the distributionsfollow broken power laws (apart from the top word which is of similar frequency before and after). Break points and expo-nents have been ascertained by a piecewise regression (see Appendix A). Interestingly, the breaks in the two datasets occurat similar ranks namely ∼ ∼ ± of words before and ± after, making the two entire distributions significantly different ( p = 8 × − in a two-sample Kolmogorov-Smirnovtest, see Appendix A). The overall effect is an increased diversity of words resulting following the soundscape. We interpretthis positive impact as signifying the participants engaged with and reflected on the stimulating experience afterwards, ratherthan continuing to draw from common associations concerning space which they likely did beforehand. We have therefore Word Rank -4 -3 -2 -1 R e l a t i v e F r equen cy -1.70 0.13 -0.85 0.25-0.36 0.36 -1.00 0.22Before (n=464)After (n=391) Figure 3.
Log-log rank-frequency plot of words before (orange) and after (blue) the soundscape. Power law exponents from a piecewiselinear regression are indicated. Uncertainties refer to standard errors. demonstrated language change in participants resulting from a public engagement activity through the novel usage of Zipf’slaw applied to graffiti wall responses.
Thematic analysis (Braun and Clarke, 2006) was used to analyse the meaning behind both textual and drawn responses. Thisfinds patterns, known as qualitative codes, in the data which are then grouped into broader related themes. Instead of usingpre-determined codes, the analysis drew on grounded theory (Robson, 2011; Silverman, 2010), allowing the themes to emergefrom the data itself as outlined in Appendix B. This more exploratory and data-driven approach enables unexpected outcomesand impacts (be they positive or negative) to come to light, rather than analysing the qualitative data only through a particularlens based on specific intended outcomes. The main themes and underlying (typically antithetical) codes determined by thefirst author are given in Table 1.We quantify the number of responses in each theme and qualitative code (cf. Sandelowski, 2001; Sandelowski et al., 2009;Maxwell, 2010) to investigate any changes from before to after the soundscape experience. These are shown in Figure 4 relativeto the total responses (panel a) and within each theme (panel b).The theme of sound is highly relevant to the activity and was commonly expressed both before and after. Responses before-hand mostly considered space to be quiet/silent ( ± within the theme). However, a non-negligible fraction thought it tobe loud, which may be due to participants second-guessing the question because of the nature of the activity and/or the phras-ing by undergraduate ambassadors. Nonetheless, the overwhelming majority ( ± within the theme) after the experience hemes Codes Description Sound Quiet Space is “silent” or “quiet”Loud Space is “loud” or “noisy”Emptiness Empty Space is an “empty” vacuum with “nothing” in itFull Space is filled with material or activity such as “wind”Dynamism Slow Space is slow (e.g. “calm” or “peaceful”)Busy Space is highly dynamic exhibiting busy movementElectricity Electrical Expressions of electrical phenomenaSpace Objects Space Objects Commonly known celestial bodies (planets, stars, meteors etc.) or artificial spacecraft
Table 1.
Themes and underlying qualitative codes in the thematic analysis. B e f o r e A ft e r B e f o r e A ft e r B e f o r e A ft e r B e f o r e A ft e r B e f o r e A ft e r P r opo r t i on i n t he m e P r opo r t i on o f t o t a l r e s pon s e s Sound
1. Quiet2. Loud
Emptiness
1. Empty2. Full
Dynamism
1. Slow2. Busy
Electricity
1. Electrical
Space Objects
1. Common a)b) G df pIJ(K) 31.08 6 <0.0001 IK(J) 11.30 8 0.1853JK(I) 4.48 6 0.612 G df pIJ(K) 38.00 6 <0.0001IK(J) 30.14 8 0.0002JK(I) 11.02 6 0.0878 G df pIJ(K) 37.86 6 <0.0001IK(J) 13.48 8 0.0964JK(I) 9.70 6 0.1379 G df pIJ(K) 10.88 3 0.0124IK(J) 0.18 4 0.9962JK(I) 0.08 4 0.9992 I = CodeJ = TimeK = Coder
Figure 4.
Comparison of qualitative themes and codes before ( n = 535 ) and after ( n = 446 ) the soundscape experience normalised bytotal responses (a) and totals within each theme (b). Error bars depict the standard error in proportions. Log-linear analysis statistics of theagreement between coders are also shown for each theme. xpressed space to be a noisy environment — a considerable change to beforehand. The perceived loudness of sound, both interms of human hearing and measurement, necessitates logarithmic scales (Robinson and Dadson, 1956). Such scales, like thedecibel, therefore require some reference base-level. For sound this is typically set at the threshold pressure for human hearingof µ Pa (Roeser et al., 2007). One must remember though that pressure fluctuations depend on the background pressure levelalso ( ,
000 Pa at sea level). Therefore, while the absolute amplitude of variations in space are clearly small, relative to thebackground they are large (as was noted in section 2) and thus one can consider space to be “noisy” in this sense. Anotherequally valid perspective is that the process of sonification has revealed the presence of sound that would otherwise not beaudible and thus participants have discovered, thanks to the exhibit, that space is “noisier” than they had previously imagined.We note that the theme of dynamism exhibits quantitatively similar results to that of sound — a clear majority ( ± withinthe theme) thought space to be slow beforehand, whereas the vast majority ( ± ) consider it highly dynamic afterwards.The dynamism of Earth’s magnetosphere is relative to the natural timescales of the system. The typical periods of oscillationsare of the order of several to tens of minutes, and the properties of the waves (and even their drivers) can significantly changewithin just a few wave periods (e.g. Keiling et al., 2016). This is unlike most sounds we are used to on Earth, which oftenremain coherent for many hundreds or even thousands of oscillations. Therefore, just like with sound, space around our planetcan be considered dynamic both relative to the properties of the environment and relative to participants’ prior expectations.The theme of emptiness (including both of its underlying codes) was quite common in responses beforehand, however itwas expressed much less often following the soundscape. The prevailing opinion before was that space is empty and thisdramatically reduced following the soundscape, both relative to the total responses (from ± to ± ) and within thetheme (from ± to ± ). In contrast, the expression of space being full was communicated a similar number of timesboth before and after. Therefore, participants that had previously thought space was empty typically went on to write words thatfell within a different theme, rather than a response signifying space as being filled with material. Since space is not absolutelydevoid of material, being permeated by tenuous plasmas, the exhibit has successfully challenged this common misconception.There was a clear increase in the proportion of responses relating to electricity following the event, from ± to ± .Electricity is of fundamental importance to the plasma state, and thus the increased realisation of this by participants is awelcome change resulting from the exhibit.Common space objects such as planets, stars, or satellites (typically expressed through drawings) may appear at first glanceof Figure 2 to be more frequent before the soundscape than after. As a fraction of the total number of responses though, thisdifference is small and not strictly statistically significant ( p = 0 . ).We checked the reliability of all these trends resulting from the qualitative coding by applying log-linear analysis to a subsetof the data additionally coded by the co-authors (see appendices for details). Using the notation that I denotes the qualitativecodes, J the time (i.e. before or after), and K the different coders, for the results to be consistent one would expect thatthe IJ ( K ) test be statistically significant, constituting the reported trends in codes with time, but the IK ( J ) and JK ( I ) interactions should not be, indicating independence from individual coders. These statistics are displayed in Figure 4 for eachtheme (apart from space objects which was less common) indicating the expected behaviour apart from in the case of emptiness.This theme showed some inconsistency between coders for the “full” code, whereas when only “empty” was considered coders ere in agreement ( G = 32 . , . , . respectively). Therefore, the main results of the paper are robust and hence we havedemonstrated a change in conceptions of space, well-aligned with the underpinning research, that resulted from this drop-inengagement activity. A challenge within public engagement is evaluating the impact of drop-in activities since this necessitates a measure of changeusing evaluative tools that are appropriate to and commensurate with the engagement (Jensen, 2014; King et al., 2015; Grandand Sardo, 2017). We have presented a novel implementation and analysis stemming from a common evaluation tool, graffitiwalls (e.g. Public Engagement with Research team, 2019). These were integrated both before and after a soundscape exhibit onspace science research using sonified satellite data. The pre- and post-soundscape graffiti walls provided data on participants’conceptions of space and, through their integration into the activity itself, had much higher response rates than is typical. Thecaptured data was analysed in two different ways.We investigated the statistical properties of the words expressed by using Zipf’s law from quantitative linguistics. This statesthat the frequency of words in languages typically follow power laws whose exponents give a measure of the diversity of words,where shallower exponents indicate greater variety. The distributions from the graffiti walls showed that the exponent for thetop ∼
10 words (constituting ± of the responses before and ± after) became significantly shallower from beforeto after, whereas the exponents were consistent for the remaining words. This demonstrates an overall increased linguisticcomplexity concerning participants’ thoughts about space following the activity. This positive result aligns with the exhibit’saims in the realm of ‘Enjoyment, Inspiration, Creativity’ (cf. Hooper-Green, 2004), since being exposed to the sounds of spaceled to stimulation, reflection, and ultimately a more diverse and creative set of words about space than had been expressedbeforehand. We are unaware of Zipf’s law being used in impact evaluation for public engagement before.We also investigated themes present in the responses, which again yielded significant and robust positive changes frombefore to after. Beforehand participants typically expressed common misconceptions of space being completely empty, silent,and with little activity. However, after experiencing the space sounds they felt space was a noisy and dynamic environmentwith electrical phenomena present. It is astounding that simply by listening to the sounds these simple aspects of the underlyingspace plasma physics were successfully and innately communicated to participants before they even spoke to the researchers.This therefore demonstrates the power of sonification for audiences. While this had been argued by Supper (2014) based onreflections from researchers and artists, here we have shown it from evaluating participants’ experiences directly. Therefore, wehave shown postive effects in the realms of ‘Knowledge & Understanding’ and ‘Attitudes & Values’ (cf. Hooper-Green, 2004)resulting from the soundscape. The measured changes in associations, conceptions, and perceptions will have been furtherreinforced by researchers drawing from participants’ own reflections in their subsequent dialogues (cf. Archer and DeWitt,2017).Overall, integrating existing evaluation tools suitable for drop-in engagement activities, such as graffiti walls, both beforeand after a drop-in activity can enable practitioners to demonstrate changes resulting from the engagement and therefore its hort-term impact. However, typically such tools are only used following activities, which limits the ability to demonstratesome measure of change and thus impact. We suggest that our approach, both in terms of data capture and analysis, should beadopted more regularly, not just for soundscape exhibits, but for a range of different drop-in activities in general. Appendix A: Statistical techniques
Statistical uncertainties in proportions are estimated using the Clopper and Pearson (1934) conservative method based on thebinomial distribution, where standard (68%) errors are shown throughout.A piecewise linear regression in log-log space was used to minimise the sum of squared error between the data and a modelmade up of a specified number of line segments whose break points could be varied iteratively. This was performed for anincreasing number of segments, each time calculating the degrees-of-freedom-adjusted R which accounts for the number ofexplanatory variables added to the model: R = 1 − (cid:0) − R (cid:1) n − n − m − (A1)where R is the usual coefficient of determination, n is the number samples, and m = 2 s − is the total number of explanatoryvariables in the piecewise linear model with s segments. The final model was selected as the first peak in R with s . Anysegments with only two datapoints are later ignored. The statistical significance of the slopes was determined by ANCOVAwith a multiple comparison procedure (Hochberg and Tamhane, 1987). The standard errors in the slopes quoted are derivedfrom a propagation of uncertainty in the proportions within the linear regression.A two-sample Kolmogorov-Smironov test is used to non-parametrically test the equality of two probability distributions. Itquantifies the distance between two one-dimensional empirical (cumulative) distribution functions F ,n ( x ) and F ,m ( x ) as D n,m = sup x | F ,n ( x ) − F ,m ( x ) | (A2)where sup is the supremum function (Massey, 1951). The critical value of this statistic is given by (cid:113) − ln ( α/
2) ( m + n ) /mn for desired significance α .Finally, log-linear analysis is employed to check the consistency of the changes in coding with time across the differentcoders. This extension of the χ test of independence to higher dimensions uses a similarly distributed statistic, the deviance,given by G = 2 (cid:88) O ijk ln O ijk E ijk (A3)for observed O ijk and expected E ijk frequencies (Agresti, 2007). Here we assess conditionally independent models denoted IJ ( K ) , which tests the two-way IJ interaction with the effects of the IK and JK interactions removed. Computationally thiscalculates G for each level of K summing the results, with G having ( n I − n J − n K degrees of freedom. heme Codes Before After Unique Total Unique Total( n = 202 ) ( n = 535 ) ( n = 190 ) ( n = 446 )Sound 1. Quiet 22 234 7 102. Loud 59 150 91 305Total 81 384 98 315Emptiness 1. Empty 36 250 4 72. Full 31 109 37 79Total 67 359 41 86Dynamism 1. Slow 33 247 8 122. Busy 74 171 111 323Total 107 418 119 335Electricity 1. Electrical 13 27 36 162Space Objects 1. Common 51 57 154 32 Table B1.
Number of responses (both unique and total) in each theme before and after the soundscape.
Appendix B: Qualitative coding
The qualitative coding process of thematic analysis drawn from grounded theory involved the following steps:1. Familiarisation: Responses (Figure 2) are studied and initial thoughts noted.2. Induction: Initial codes are generated based on review of the data.3. Thematic Review: Codes are grouped together into themes and applied to the full data set.4. Reliability: Codes are applied to a subset of data by second coders to check reliability of results.5. Finalisation: Theoretical interpretation and narrative are formulated from final coding.Table B1 shows the number of responses (both unique and total) across words and pictures in each theme and its underlyingcodes both before and after the soundscape experience. To ensure the reliability of the main qualitative coding of the entiredataset, second coders applied the thematic analysis to a subset of the data. This subset constituted the top 16 words before(58% of total responses) and 15 words after (49%), with the slightly different number of words used in the two datasets beingdue to ties in the ranking of words making it impossible to have exactly the same number in both. Table B2 shows the totals ofhow these unique words were grouped across all three coders. These results are used in the log-linear analysis to test reliability,which we note does not require equally sized datasets. The codes’ association to the raw data can be found in the supplementarymaterial, both for the main and second coders. oder 1 Coder 2 Coder 3Before After Before After Before AfterSound 1 Quiet 8 0 5 0 8 12 Loud 6 11 5 11 4 5None 2 4 6 4 4 9Emptiness 1 Empty 8 0 6 0 9 12 Full 5 3 0 1 6 7None 3 12 10 14 1 7Dynamism 1 Slow 7 0 5 0 6 02 Busy 7 11 4 12 10 11None 2 4 7 3 0 4Electricity 1 Electrical 2 6 2 7 2 6None 14 9 14 8 14 9 Table B2.
Statistical comparison of the number of unique words in each qualitative code as judged by different coders across a subset of thedata (the top 16 words before and 15 words after).
Data availability.
Data supporting the findings are contained within the article and its supplementary material.
Author contributions.
MOA conceived the project and its evaluation, performed the analysis, and wrote the paper. ND and SB assisted withthe analysis.
Competing interests.
The authors declare that they have no conflict of interest.
Acknowledgements.
We thank the researchers (Alice Giroul, Christopher Chen, Emma Davies, Jesse Coburn, Joe Eggington, Luca Franci,Oleg Shebanits) and undergraduate ambassadors (Avishan Shahryari, Cheng Yeen Pak, Christopher Comiskey Erazo, Habibah Khanom,Safiya Merali, Yinyi Liu) who helped deliver the exhibit along with all the staff at the Science Museum (including Becky Carlyle, ImogenSmall, Sevinc Kisacik). This project has been supported by QMUL Centre for Public Engagement Large 2016 and Small 2019 Awards,an EGU Public Engagement Grant 2017, and STFC Public Engagement Spark Award ST/R001456/1. M.O. Archer holds a UKRI (STFC /EPSRC) Stephen Hawking Fellowship EP/T01735X/1. ipf, G. K.: The psycho-biology of language, Houghton Mifflin, Boston, MA, USA, 1935.Zipf, G. K.: Human behavior and the principle of least effort, Addison-Wesley Press, Boston, MA, USA, 1949.ipf, G. K.: The psycho-biology of language, Houghton Mifflin, Boston, MA, USA, 1935.Zipf, G. K.: Human behavior and the principle of least effort, Addison-Wesley Press, Boston, MA, USA, 1949.