Posted : 30/9/20
Photophoretic levitation of engineered aerosols for geoengineering
David W. Keith1
Energy and Environmental Systems Group, University of Calgary, 2500 University Drive NW, Calgary AB, Canada T2N 1N4 Communicated by James G. Anderson, Harvard University, Cambridge, MA, June 30, 2010 (received for review June 15, 2009)
Aerosols could be injected into the upper atmosphere to engineer the climate by scattering incident sunlight so as to produce a cool- ing tendency that may mitigate the risks posed by the accumulation of greenhouse gases. Analysis of climate engineering has focused on sulfate aerosols. Here I examine the possibility that engineered nanoparticles could exploit photophoretic forces, enabling more control over particle distribution and lifetime than is possible with sulfates, perhaps allowing climate engineering to be accomplished with fewer side effects. The use of electrostatic or magnetic mate- rials enables a class of photophoretic forces not found in nature. Photophoretic levitation could loft particles above the strato- sphere, reducing their capacity to interfere with ozone chemistry; and, by increasing particle lifetimes, it would reduce the need for continual replenishment of the aerosol. Moreover, particles might be engineered to drift poleward enabling albedo modification to be tailored to counter polar warming while minimizing the impact on equatorial climates.
atmospheric physics ∣ solar radiation management ∣ climate change
The possibility of increasing the earth’s albedo to offset CO2-driven warming has been a subject of speculation for decades (1). Over the last few years, more systematic research and debate on the topic has emerged spurred by Crutzen’s (2) call for systematic analysis of geoengineering in response to the con- tinued acceleration of anthropogenic CO2 emissions (3) and the threat of abrupt climate change. Most research has focused on the possibility of injecting sulfur into the stratosphere (2, 4–8), although more elaborately engineered aerosols (9) and space- based solar scattering systems have also been proposed (9, 10). Here I examine the possibility that particles might be engineered to exploit photophoretic forces (11–16), enabling the manipula- tion of particle distribution and radiative forcing in ways that could not be achieved with sulfate aerosol.
Limitations of Sulfate Aerosols
The salient advantage of sulfate aerosols as a means to modify the earth’s albedo is that nature has already performed relevant experiments in the form of volcanic injections of sulfur, such as the 1991 eruption of Mount Pinatubo, which deposited ∼9 Mt sulfur in the stratosphere creating sulfate aerosols that cooled the earth by ∼0.5 C within a year (17).
As a tool for climate engineering, sulfates are, however, a blunt instrument. Disadvantages of sulfates include the following: First, it is difficult to produce sulfate aerosol with an appropriate size distribution. The mass-specific scattering efficiency of a sulfate aerosol (or similar dielectric sphere) is strongly dependent on its radius. The scattering efficiency peaks at ∼0.3 μm diameter and decreases rapidly for larger or smaller droplets, yet when aerosols are generated by continuous injection of SO2 the result- ing size distribution tends to be substantially larger than optimal because most of the added sulfur is deposited on existing parti- cles. This substantially limits the radiative forcing produced by large sulfur injections and can make it difficult to produce a ra- diative forcing sufficient to offset the radiative effect of a CO2 doubling (4). The problem is compounded if sulfate aerosols reach the warmer temperatures in the upper stratosphere where
the vapor pressure of H2SO4 is sufficiently high to enable vapor- phase transfer of mass from smaller to larger particles, com- pounding the difficulty of maintaining a suitable aerosol size distribution.
Second, a significant fraction of the light scattered by sulfate aerosols is scattered in the forward direction, increasing the ratio of diffuse-to-direct insolation at the surface. The Pinatubo erup- tion, for example, increased the amount of clear sky diffuse sun- light reaching near surface by more than a factor of 2 (18). The increase in diffuse radiation will in turn tend to whiten the visual appearance of the daytime sky and produce side effects ranging from alteration of ecosystem productivity (19) to the reduction in the output from concentrating solar power systems (20).
Finally, sulfates in the lower stratosphere provide reactive surfaces that can accelerate the catalytic removal of ozone by accelerating the conversion of chlorine from reservoir species to ClO. This effect may be more serious if water vapor concen- trations in the lower stratosphere increase with increasing global temperatures (21).
The problems of particle size control, forward scattering, and interference with ozone chemistry would apply, in varying degrees, to other nonsulfate aqueous aerosols in the lower strato- sphere (22).
Given the disadvantages of sulfate aerosols as a tool for climate engineering, it is worthwhile to explore the possibility of designing scatters that might enable perturbations in radiative forcing to be more precisely tailored or to be achieved with less severe side effects.
Photophoretic Forces on Aerosols
Photophoretic forces arise when radiation produces a tempera- ture difference between an aerosol particle and the surrounding gas (11–13). The forces are most significant when the mean free path of the gas molecules is large compared to the particle’s size. Two kinds of photophoretic force have been described. The thermal gradient (ΔT) force arises from a radiatively driven temperature inhomogeneity: Gas molecules leave the warm sur- face with, on average, higher energy than molecules leaving the cool surface, thus generating an average force away from the hot- ter side of the particle (Fig. 1A). The accommodation coefficient (Δα) force arises from variation of the thermal accommodation coefficient across the particle, where the accommodation coeffi- cient, α, is the probability that a impinging molecule of gas will thermally equilibrate with the surface (Fig. 1B). Accommodation coefficients in air vary widely with reported values ranging, for example, from ∼0.95 for Al to 0.55 for Pt for temperatures near 315 K (23). The Δα force is body-fixed and can act on isothermal particles as long as there is a temperature difference between the particle and the gas.
Author contributions: D.K. designed research; D.K. performed research; and D.K. wrote the paper.
The author declares no conflict of interest.
Freely available online through the PNAS open access option.
This article contains supporting information online at http://www.pnas.org/lookup/suppl/ doi:10.1073/pnas.1009519107/-/DCSupplemental.
16428–16431 ∣ PNAS ∣ September 21, 2010 ∣ vol. 107 ∣ no. 38
Downloaded by guest on September 30, 2020
Radius ~10 m
50 nm Fe3O4
Al2O3 Al BaTiO3
Fig. 1. Schematic illustration of photophoretic forces. A and B show, respectively, thermal gradient (ΔT) and accommodation coefficient (Δα) forces acting on idealized spherical particles. Note that for ΔT, the photophoretic force (Fp) is aligned with the asymmetry in the radiation field that drives the thermal gradient, in this case the incident solar radiation (insolation); whereas for Δα, the vector is oriented along the asymmetry in the accommodation coefficients and rotates with the particle. C illustrates design of the disk described in the text, whereas D illustrates the fields and forces acting on the disk as described in SI Text.
The Δα force can produce a net upward (levitating) force when a particle is aligned by gravitational torques. The resulting time- average force is called gravito-photophoretic (13). Such forces may alter the dynamics of natural aerosols in the upper atmo- sphere contributing to the formation of thin layers of aerosol observed in the mesosphere (12, 15, 16) and perhaps loft soot from the stratosphere to the mesosphere (14, 16).
Brownian motion randomizes a particle’s orientation as the or- ienting torque becomes small compared to the particle’s thermal energy. Particles with radii above 1 μm are readily aligned by gravitational torques, but because the photophoretic force is pro- portional to surface area the upward force per unit mass declines as r−1. For smaller particles, the gravitational torque and hence the time-average upward force is proportional to r4, so the net gravito-photophoretic force per unit mass declines as r3 (SI Text). The upshot is that for approximately spherical particles, the grav- ito-photophoretic Δα force can only levitate particles with sizes around 1 μm—far larger than the optimum for mass-efficient scattering—and even then the levitation is only effective above about 10 km, the height at which the mean free path of air molecules approaches the particle size.
Model and Results
A particle with a permanent electric or magnetic dipole moment would experience an orienting torque in the terrestrial electric or magnetic field that would, as with gravito- photophoresis, produce a time-averaged force parallel to the orienting field. One might, by analogy, call the resulting forces electro- or mag- netophotophoretic.
Magnetic or electrostatic torques can greatly exceed gravita- tional torques for small aerosols allowing particles that were en- gineered to exploit electro- or magnetophotophoresis to exhibit dynamical properties not found in natural aerosols. For example, electro- or magnetophotopheresis could be used to enable a particle to levitate in the lower atmosphere and could be used to levitate particles smaller than 0.1 μm.
Engineered particles need not be spherical. The most mass- efficient geometry for a scattering is a thin disk with radius larger than the wavelength of light. Such a design would produce mini- mal forward scattering when the disk’s radius is substantially lar- ger than the wavelength of light, eliminating the diffuse radiation problem encountered with sulfate aerosols. The use of electric or magnetic materials allows disks to be oriented horizontally, and the use of material with contrasting accommodation coefficients allows for levitation.
An Idealized Example. As a specific example, consider a thin disk with radius ∼5 μm and thickness 50 nm composed of three layers: 5 nm aluminum oxide, 30 nm of metallic aluminum, and finally
15 nm of barium titanate (Fig. 1C). The thickness of the Al layer is chosen so that it has high solar-band reflectivity and is nearly transparent to outgoing thermal infrared so as to produce a large mass-specific negative radiative forcing (cooling) (9). The Al2O3 layer serves to protect the Al layer from oxidization. The thick- ness of the BaTiO3 is chosen so that the electrostatic torque from the atmospheric electric field is sufficient to orient the disk hor- izontally against torques arising from reasonable asymmetries in thickness or α across the disk (24). Assuming a relatively small, and therefore conservative, 15% difference in α between the two materials (23), the photophoretic force on the disk would exceed 2 times its weight under diurnally averaged illumination at altitudes in the middle stratosphere or mesosphere assuming it absorbed only 10% of the solar flux (Fig. 2 and SI Text).
The Δα force is proportional to the temperature differential between particle and ambient air which, in turn, depends on the difference between the particle’s radiative equilibrium tem- perature and the ambient temperature. The resulting force is therefore strongly dependent on the particle’s optical properties. In a particle with a large εS∕εT, the ratio of solar-spectrum to thermal-spectrum emissivity will be warmer than middle atmo- sphere air and will be lofted to the mesopause, whereas a particle
|0.15 0.05 stratopause mesopause 0.15 0.15 0.10 0.05 A B|
|Key: S T|
PNAS ∣ September 21, 2010 ∣ vol. 107 ∣ no. 38 ∣ 16429
0 20 40 60 80 100 120 Altitude (km)
Fig. 2. Photophoretic force versus altitude. The upward photophoretic force, FP, is normalized by the downward gravitational force, FG, so that a particle is stably levitated at FP ∕FG 1⁄4 1 (dashed line). Downward forces (FP ∕FG < 0) can occur near the stratopause when the particle’s radiative equi- librium temperature is less than ambient; forces decline to zero above the mesopause as particle temperatures become radiatively controlled. Particles may be trapped below the stratopause at “A” or below the mesopause at “B.” Forces are computed for three combinations of solar and thermal band emissivity by solving the energy balance equation shown in SI Text, assuming a disk with h 1⁄4 50 nm, ρ 1⁄4 3;000 kgm−3, α ̄ 1⁄4 0.7, and δα 1⁄4 0.15. Calculations assume the 1976 Standard Atmosphere (30) and TE 1⁄4 255 K.
Downloaded by guest on September 30, 2020
Levitating-force/gravity (Fp /Fg)
APPLIED PHYSICAL SCIENCES
with a smaller εS∕εT that has a radiative temperature below that of the stratopause will be confined below about 45 km (Fig. 2). It might be possible to tailor particle properties so that levi- tated particles are trapped either below the stratopause or below the mesopause as shown by points “A” or “B” in Fig. 2. In either case, such particles might avoid two of the disadvantages of stratospheric sulfates: the increase in diffuse radiation and the increase in reactive surface area in the lower stratosphere, which can accelerate ozone depletion (2). Photophoretic levitation would enable longer atmospheric lifetimes, and by lofting aero- sols out of the stratosphere it would reduce the most significant
impact on ozone chemistry.
Controlling the Latitudinal Distribution of Radiative Forcing. Photo- phoretic forces might be used to control the latitudinal distribu- tion of particles. Most simply, particles confined near the stratopause (point “A” in Fig. 2) would be transported poleward by the Brewer–Dobson circulation.
Alternatively, the addition of a weak magnetic dipole will break the rotational symmetry of a particle that is levitated by gravito- or electrophotopheresis, partially orienting it along the terrestrial magnetic field and so inducing a magnetophotophore- tic effect that could be used to make the particle drift toward one of the magnetic poles while being stably levitated near the stratopause or mesopause. For example, the addition of a weak magnetic moment perpendicular to the plane of the disk de- scribed above would tilt the disk toward one of the terrestrial magnetic poles producing a meridional component of the Δα force (Fig. 1D). A single ∼0.5 μm magnetite nanoparticle adher- ing to the disk would provide sufficient torque. The resulting poleward velocities are of order 102 m∕s for particles confined beneath the stratopause, whereas for particles confined near the mesopause, velocities can exceed 10 m∕s (SI Text).
At the stratopause, the meridional velocities achievable with the tilted disk design appear to be slower than the average trans- port velocities in the Brewer–Dobson circulation (25), but the meridional drift velocity might still allow some additional control of zonal radiative forcing. Velocities may well be large enough at the mesopause, but it is difficult to estimate mean velocity be- cause particle altitude, velocity, and mesospheric winds all have strong diurnal cycles, and the time–average transport will likely depend on the relative phase of the three cycles.
More complex particle designs could achieve substantially higher meridional velocities. Many alternative combinations of forces and orienting torques might be exploited to produce a net poleward force; indeed, poleward migration might be accom- plished without electrostatic or magnetic forces (SI Text).
Poleward migration of scattering particles might allow the reduction in solar radiative forcing to be concentrated near the poles perhaps counteracting the ice-albedo feedback that amplifies CO2-driven polar warming while minimizing the impact on mid-latitude and tropical climates (26). Such an ability might be relevant in the event that warming triggers rapid deglacia- tion (27).
The use of particles engineered to exploit photophoretic forces may enable more selective geoengineering with fewer adverse effects than would the use of sulfate aerosol. Further, the ability to orient particles and to achieve long atmospheric lifetimes might enable deployment of more spectrally selective scattering systems such as patterned gratings or multilayer coatings (9).
Disadvantages and Uncertainties. Long particle lifetimes may be a disadvantage as it makes geoengineering less easily reversible than is the case with short-lived particles. This risk could be mitigated by choosing particle designs with shorter lifetimes for testing and initial deployment. Given the presence of hard-UV flux and reac-
tive atomic oxygen found in the mesosphere, it should be easy to design short-lived particles; design and fabrication of long-lived particles will likely be the engineering challenge.
Unresolved issues include (i) the ability to fabricate and deploy such particles at reasonable cost, (ii) the ability to predict the dis- tribution of particles given the complex interplay of photophore- tic motion, gravitational settling and advection by the winds of the middle atmosphere, and, of course, (iii) predicting the effec- tiveness and risks of climate engineering however the alteration in radiative forcing is achieved.
Finally, although use of engineered particles may offer substan- tial advantages over sulfate aerosols as a means to manipulate radiative forcing of climate, there is a corresponding disadvan- tage: We lack the direct natural analog provided by volcanic injection of SO2. This lack of analog means that we should be more concerned about unexpected side effects, unknown un- knowns, and consider how a careful progression from testing to monitored subscale deployment could constrain the risks.
The Cost of Engineered Particles. Is it possible to fabricate such particles at sufficiently low cost? Any definitive answer would, of course, require a sustained broad-based research effort. The following argument serves only to suggest that one cannot dis- count the possibility: Approximately 109 kg of engineered parti- cles similar to the example described above would need to be deployed to offset the radiative effect of CO2 doubling. Assuming a lifetime of 10 years, the particles must be supplied at a rate of 108 kg∕yr. A plausible upper bound on the acceptable cost of manufacture can be gained by noting that the monetized cost of climate impacts and similarly the cost of substantial reductions in greenhouse gas (GHG) emissions are both of order 1% of global gross domestic product (GDP) (28). Suppose one de- manded that the annualized cost of particle manufacture be less than 1% of the cost of abating emissions, that is 10−4 of the ∼60 × 1012 global GDP. Under these assumptions, the allowable manufacturing cost is 60∕kg. Many nanoscale particles are cur- rently manufactured at costs significantly less than this threshold. Silica-alumina ceramic hollow microspheres with diameters of 1 μm (e.g., 3M Zeeospheres) can be purchased in bulk at costs less than 0.3∕kg. Moreover, bulk vapor-phase deposition methods exist to produce monolayer coatings on fine particles, and there are rapid advances in self-assembly of nanostructures that might be applicable to bulk production of engineered aerosols.
These scaling calculations certainly do not prove that nanos- cale particles for climate engineering could be successfully man- ufactured and deployed. They do suggest that the possibility of doing this over the coming decades cannot be dismissed.
Any climate engineering scheme will only partially and imper- fectly compensate for the climatic impacts of GHGs, and it will likely impose significant risks of its own. Given common estimates of the monetized cost of climate damages, the value of reducing climate change by geoengineering could exceed 1% of GDP. It is, therefore, plausible that the costs of geoengineering will be all but irrelevant to decisions about deployment, which will focus on the risk-to-risk trade-off between the risk of geoengineering and the risk of climate damages; assuming, of course, that the direct costs of geoengineering are limited to a small fraction, say 0.1%, of GDP. Thus it might be that the cost of engineered particles could approach 1;000∕kg before the costs of manufac- ture played a significant role in deployment decisions.
Implications for Research on Solar Radiation Management The pos- sibility that engineered aerosols may offer capabilities and risks beyond those arising from injection of sulfate suggest that a ba- lanced research effort on climate modification using stratospheric scatterers should not be limited to exploring the effects of sulfate aerosols—the focus of essentially all current research—but should explore a wider range of technological options for altering
Downloaded by guest on September 30, 2020
the atmospheric radiation budget. Research on understanding the environmental risks and benefits of such geoengineering must be coupled with a more systematic exploration of the technical options that might be employed to alter the radiative forcing of climate.
Limitations of Solar Radiation Management. Whether or not the ideas proposed here prove to be fruitful, it seems likely that some method will eventually allow humanity to manipulate the solar forcing of climate at low cost and with few side effects arising directly from the system used scatter solar radiation. Such an abil- ity would provide a powerful tool to reduce the risks of climate change, but it cannot possibly solve all the problems arising from the growing burden of anthropogenic CO2 (8). Although a reduc- tion in insolation can compensate for the effect of increased CO2 on the global-average surface temperature, it will necessarily
- Keith DW (2000) Geoengineering the climate: History and prospect. Annu Rev Energ Env 25:245–284.
- Crutzen PJ (2006) Albedo enhancement by stratospheric sulfur injections: A contribu- tion to resolve a policy dilemma? Climatic Change 77:211–219.
- Canadell JG, et al. (2007) Contributions to accelerating atmospheric CO2 growth from economic activity, carbon intensity, and efficiency of natural sinks. Proc Natl Acad Sci USA 104:18866–18870.
- Heckendorn P, et al. (2009) The impact of geoengineering aerosols on stratospheric temperature and ozone. Environ Res Lett 4:045108.
- Wigley TML (2006) A combined mitigation/geoengineering approach to climate stabilization. Science 314:452–454.
- Rasch PJ, Crutzen PJ, Coleman DB (2008) Exploring the geoengineering of climate using stratospheric sulfate aerosols: The role of particle size. Geophys Res Lett 35:L02809.
- Trenberth KE, Dai A (2007) Effects of Mount Pinatubo volcanic eruption on the hydrological cycle as an analog of geoengineering. Geophys Res Lett 34:L15702.
- Tilmes S, Müller R, Salawitch R (2008) The sensitivity of polar ozone depletion to proposed geoengineering schemes. Science 320:1201–1204.
- Teller E, Wood L, Hyde R (1997) Global Warming and Ice Ages: I. Prospects for Physics Based Modulation of Global Change (Lawrence Livermore Natl Laboratory, Livermore, CA) p 20.
- AngelR(2006)FeasibilityofcoolingtheEarthwithacloudofsmallspacecraftnearthe inner Lagrange point (L1). Proc Natl Acad Sci USA 103:17184–17189.
- Ehrenhaft F (1918) Die photophorese. Ann Phys Berlin 361:81–132.
- Orr C, Jr, Keng EYH (1964) Photophoretic effects in the stratosphere. J Atmos Sci 21:475–478.
- Rohatschek H (1996) Levitation of stratospheric and mesospheric aerosols by gravito- photophoresis. J Aerosol Sci 27:467–475.
- Pueschel RF, et al. (2000) Vertical transport of anthropogenic soot aerosol into the middle atmosphere. J Geophys Res Atmos 105:3727–3736.
- Cheremisin AA, Vassilyev YV, Horvath H (2005) Gravito-photophoresis and aerosol stratification in the atmosphere. J Aerosol Sci 36:1277–1299.
- Wurm G, Krauss O (2008) Experiments on negative photophoresis and application to the atmosphere. Atmos Environ 42:2682–2690.
reduce radiant energy fluxes at the surface and this will, in turn, reduce the export of latent heat resulting in a climate with less precipitation and less evaporation than the preindustrial climate. Moreover, the mere knowledge of such methods might itself be problematic if it weakened the commitment to curb emissions (1). Despite this, the continued acceleration of anthropogenic emis- sions (3) coupled with growing concern about the possibility of dangerous nonlinear responses to climate forcing (27) argue for more systematic exploration of the feasibility and risks of geoengineering (2, 29).
ACKNOWLEDGMENTS. Three anonymous reviewers provided helpful com- ments. The work was supported by the Climate Decision Making Center through a cooperative agreement between the U.S. National Science Foun- dation (SES-0345798) and Carnegie Mellon University.
19. 20. 21.
22. 23. 24. 25.
26. 27. 28.
Soden BJ, Wetherald RT, Stenchikov GL, Robock A (1992) Global cooling after the eruption of Mount Pinatubo: A test of climate feedback by water vapor. Science 296:727–730.
Olmo FJ, Tovar J, Alados-Arboledas L, Okulov O, Ohvril HO (1999) A comparison of ground level solar radiative effects of recent volcanic eruptions. Atmos Environ 33:4589–4596.
Gu L, et al. (2003) Response of a deciduous forest to the Mount Pinatubo eruption: Enhanced photosynthesis. Science 299:2035–2038.
Murphy DM (2009) Effect of stratospheric aerosols on direct sunlight and implications for concentrating solar power. Environ Sci Technol 43:2784–2786.
Kirk-Davidoff DB, Hintsa EJ, Anderson JG, Keith DW (1999) The effect of climate change on ozone depletion through changes in stratospheric water vapour. Nature 402:399–401.
Blackstock JJ, et al. (2009) Climate Engineering Responses to Climate Emergencies (Novim, Santa Barbara, CA).
Ho CY (1989) Thermal Accommodation and Adsorption Coefficients of Gases (Hemi- sphere Publishing Corp, Washington, DC).
Fong DD, et al. (2004) Ferroelectricity in ultrathin perovskite films. Science 304:1650–1653.
Jones DBA, Andrews AE, Schneider HR, McElroy MB (2001) Constraints on meridional transport in the stratosphere imposed by the mean age of air in the lower strato- sphere. J Geophys Res Atmos 106:10243–10256.
MacCracken MC (2009) On the possible use of geoengineering to moderate specific climate change impacts. Environ Res Lett 4:045107.
Bambera JL, Alleyb RB, Joughinc I (2007) Rapid response of modern day ice sheets to external forcing. Earth Planet Sci Lett 257:1–13.
Metz B, Davidson O, Swart R, Pan J, eds. (2001) Climate Change 2001: Mitigation: Contribution of Working Group III to the Third Assessment Report of the IPCC (Cambridge Univ Press, Cambridge, UK).
Keith DW, Parson E, Morgan MG (2010) Research on global sun block needed now. Nature 463:426–427.
Anonymous (1976) U.S. Standard Atmosphere (U.S. Government Printing Office, Washington, DC).
The Aversion to Tampering with Nature (ATN) Scale: Individual Differences in (Dis)comfort with Altering the Natural World
People differ in their comfort with tampering with the natural world. Although some see altering nature as a sign of human progress, others see it as dangerous or hubristic. Across four studies, we investigate discomfort with tampering with the natural world. To do so, we develop the Aversion to Tampering with Nature (ATN) Scale, a short scale that is the first to directly measure this discomfort. We identify six activities that people believe tamper with nature (geoengineering, genetically modified organisms, pesticides, cloning, gene therapy, and nanoparticles) and show that ATN scores are associated with opposition to these activities. Furthermore, the ATN Scale predicts actual behavior: donations to an anti‐tampering cause. We demonstrate that ATN is related to previously identified constructs including trust in technology, naturalness bias, purity values, disgust sensitivity, aversion to playing God, and environmental beliefs and values. By illuminating who is concerned about tampering with nature and what predicts these beliefs, the ATN Scale provides opportunities to better understand public opposition to technological innovations, consumer preferences for “natural” products, and strategies for science communication.
Rapid technological advancements have improved our ability to produce food, manage diseases, and solve complex engineering problems (Busch, 2008; Helbing et al., 2017; Henneman, Timmermans, & Wal, 2006). Despite these benefits, laypeople perceive many modern technologies as risky, even when experts do not (Beck, 1992; Rosa, Renn, & McCright, 2013; Tenbült, De Vries, Dreezens, & Martijn, 2005). For example, many members of the public reject genetically modified organisms (GMOs), even though scientists and policy experts extoll their benefits (Tenbült et al., 2005).
Many factors influence how members of the public intuitively assess risk. Emergent technologies are perceived as riskier the more their impacts are perceived as unknown, unobservable, uncontrollable, or potentially catastrophic (Bassarak, Pfister, & Böhm, 2017; Slovic, 1987). An additional factor may be the extent to which technologies allow humans to interfere with the natural world (Hansen, 2006; Sjöberg, 2000). Researchers have proposed that resistance to scientific and technological developments may arise from discomfort with tampering with nature, and that some people experience this discomfort more than others (Corner, Parkhill, Pidgeon, & Vaughan, 2013; Dragojlovic & Einsiedel, 2013; Vandermoere, Blanchemanche, Bieberstein, Marette, & Roosen, 2010; Wolske, Raimi, Campbell‐Arvai, & Hart, 2019).
Despite scholarly interest in aversion to tampering with nature (ATN), no valid and reliable scale exists to measure this construct. To address this deficit, we developed a scale to assess how perceptions of tampering with nature may affect public perceptions of emerging technologies, such as gene therapy and climate engineering (Corner et al., 2013; Grady, 2017). Building on prior research, we aim to demonstrate that individuals vary in their reported discomfort with tampering with the natural world and that this generalized discomfort predicts their reactions to a wide range of scientific and technological activities. To this end, we first review past work on ATN and related constructs before introducing the four studies used to develop and validate the proposed scale.
1.1 Aversion to Tampering with Nature
We use the term “aversion to tampering with nature” to refer to discomfort with human activities that alter some aspect of the natural world in a way that invites risk. This limits our set of relevant activities, as not all risks involve tampering with nature (e.g., driving a car), and not all tampering is perceived as risky (e.g., constructing a building). Although no previous scales directly measure ATN, other literature suggests that altering nature may provoke such discomfort. Indeed, elements of ATN are reflected in trust in science and technology, naturalness bias, purity and contamination, religion and hubris, and environmental concern. Therefore, we conceptualize ATN as being related to the following constructs.
1.1.1 Trust in Science and Technology
To the extent that people are averse to tampering with nature, we also expect them to distrust scientific advancements and technologies that are perceived to interfere with nature. These might include activities that pose risks to natural ecosystems (e.g., geoengineering), those that mimic natural processes (e.g., cloning), as well as technologies that disrupt the natural order (e.g., GMOs). Further, extensive work by Siegrist and colleagues (Connor & Siegrist, 2010; Siegrist, 2000; Siegrist, Cousin, Kastenholz, & Wiek, 2007) shows that distrust in entities responsible for the development, use, and regulation of technology is correlated with higher perceived risks and greater perceived benefits of emerging food technologies. We believe that this distrust may extend to all new technologies—not just food‐related ones—and perhaps to the scientific process more generally, as people who are averse to tampering with nature will be more resistant to any developments that could potentially interfere with nature in the future.
1.1.2 Naturalness Bias
We expect people who are uncomfortable with tampering with nature to show a bias for naturally produced outcomes over human‐produced outcomes. Rozin (2005) and Rozin et al. (2004) find a robust preference for natural products and distrust in items that have been adulterated by humans. This naturalness bias is mirrored in research on risk perception. For example, equivalent hazards caused by humans are seen as more severe than those caused by nature (Siegrist & Sütterlin, 2014). We suspect that people most concerned about tampering with nature are particularly prone to such biases. If people are worried about human alteration of natural states, they should prefer products and outcomes that are not the results of human intervention.
1.1.3 Purity and Contamination
Beliefs that tampering with nature contaminates an otherwise pure entity may underlie general biases toward the natural, due to both instrumental and ideational reasoning. Instrumental (aka, consequential) nature‐biased beliefs assume that natural products or processes are innately superior in terms of health or effectiveness and that human interference will necessarily degrade them (Rozin et al., 2004; Rudski, Osei, Jacobson, & Lynch, 2011). Ideational (aka, moral) beliefs evoke nature as inherently pure or representing a normative standard (Li & Chapman, 2012; Rozin et al., 2004). Although some research finds a moral basis for opposition to genetic modification (Scott, Inbar, & Rozin, 2016), other research suggests opposition to technologies that interfere with natural products is based on both instrumental and ideational beliefs (and that these beliefs are interrelated) (Li & Chapman, 2012; Rozin et al., 2004).
Other theorists argue that people may create elaborate explanations for anti‐science views based on philosophical stances because they are embarrassed to admit that their reactions are really based on fears about contaminants (Hornsey & Fielding, 2017). Contamination of an otherwise pure system is also tied to feelings of disgust (Rozin, Lowery, Imada, & Haidt, 1999; Scott et al., 2016). Whether the purity is that of the natural world or one’s natural body, fear and disgust of contamination seems to be a powerful driver of anti‐tampering beliefs.
1.1.4 Religion and Hubris
Some (but not all) religious beliefs may also be tied to ATN. Integral to many recent arguments against tampering are concerns about scientists “playing God” by altering God’s creation or violating a sacred realm (Hartman, 2017; Link, 2013; Vandermoere et al., 2010). Thus, we expect that people who are disdainful of “playing God” should be similarly uncomfortable with tampering with nature.
1.1.5 Environmental Concern
Finally, to the extent that people value the natural world, feel connected to nature, and see the environment as fragile, we expect them to also worry about efforts to modify nature. Indeed, ATN has been hinted at in measures of environmental worldviews. For example, the New Environmental Paradigm (NEP) Scale encompasses concerns about the negative environmental effects of modernity, the dependence of humans on nature, and the vulnerability of nature to human intervention (Dunlap & Van Liere, 1978; Dunlap, Van Liere, Mertig, & Jones, 2000). Although some NEP items capture ATN, for example, “When humans interfere with nature it often produces disastrous consequences,” these statements have not been identified nor used as measures of ATN. Other popular measures of environmental concern do not include any reference to altering the natural world (e.g., Mayer & Frantz, 2004; Schwartz, 1977). Therefore, we hypothesize that any ATN Scale will be related to measures of environmental concern, and that this relationship will be particularly strong for NEP.
1.2 Current Research
Although each of the above constructs is expected to relate to ATN, none of the scales used in prior work directly assesses whether individuals are averse to altering the natural world in a way that may invite risk. Furthermore, time‐limited surveys may preclude measuring these potential correlates as a proxy for ATN. Measuring ATN is important, however, as many emergent technologies may be seen as tampering with nature despite their potential benefits to society. An ATN Scale offers a means to directly and efficiently gauge what factors may influence overall public support.
- H1: ATN will positively correlate with opposition to activities that tamper with nature (Studies 2 and 3), including donations to anti‐tampering causes (Study 4).
- H2: ATN will correlate negatively with trust in technology and science (Studies 2 and 4).
- H3: ATN will correlate positively with a preference for natural‐caused over human‐caused outcomes (Study 4).
- H4: ATN will correlate positively with purity values and disgust sensitivity (Study 3).
- H5: ATN will correlate positively with aversion to playing God (APG), but not with general religious beliefs (Study 3).
- H6: ATN will correlate positively with environmental concern, biospheric values, and worldviews tied to protecting the environment. This connection should be particularly strong with NEP, as this scale includes some tampering‐related items (Studies 2–4).
To the extent that H1 is supported, this will also offer evidence of the ATN Scale’s criterion validity (its ability to predict related behavioral outcomes). To the extent that H2–H6 are supported, this will offer evidence for the ATN Scale’s convergent validity, or its relationship to theoretically related constructs.
2 STUDY 1
We first identified activities that people associate with tampering with nature and measured opposition to these activities. This allowed us to assess how the ATN Scale we developed correlates with opposition to relevant activities.
Participants were 100 American adults recruited from Amazon’s Mechanical Turk (MTurk) in exchange for $1.00. Seven were removed for failing attention checks1(remaining N = 93; see Table A1 in the Supporting Information for sample demographics across all studies).
2.1.2 Procedure and Measures
We collected data for all studies using Qualtrics survey software. Participants were presented with descriptions of activities and asked to rate how much each tampers with nature (1 = Not at all to 5 = Very much; see Table I for full activity descriptions and descriptive statistics). These included eight activities that we expected to be seen as tampering with nature: GMOs, antibacterial soap, pesticides, geoengineering, gene therapy, vaccines, cloning, and nanoparticles. Four additional activities were related to nature but not tampering: wood‐burning stoves, childbirth, vegetable gardens, and visiting national parks. Four other items were unrelated to nature: video games, college degrees, car repair, and baseball. Participants also reported how much they supported or opposed each of the eight tampering with nature activities (1 = Strongly oppose; 7 = Strongly support).
|Category||Topic||Item Wording||Rating of Tampering Mean (SD)||Support for Activity Mean (SD)||Tampering/Support Correlation (r)|
|Tampering with nature activities||Geoengineering||The use of geoengineering, which is the deliberate large‐scale manipulation of the earth’s climate, in an attempt to counteract the effects of global warming.||4.29 (1.02)||4.15 (2.04)||−.26*|
|GMOs||The development of genetically modified organisms (GMOs), in which genes from the DNA of one species are extracted and inserted into the genes of an unrelated plant or animal.||4.17 (1.02)||3.62 (1.96)||−.31**|
|Pesticides||The use of chemical pesticides to protect crops against pests such as insects, weeds, and fungi.||3.98 (0.96)||3.55 (1.77)||−.40**|
|Cloning||Research on cloning, in which new cells or entire organisms are created that are genetically identical to another cell or organism.||3.84 (1.39)||3.89 (1.95)||−.22*|
|Gene therapy||The use of gene therapy in medicine, in which normal genes are transplanted into cells in place of missing or defective ones in order to correct genetic disorders.||3.80 (1.27)||5.20 (1.56)||−.01|
|Nanoparticles||The use of nanoparticlesin medicine. Nanoparticles are particles that are so small that they are able to pass into human cells for the targeted delivery of drugs.||3.33 (1.30)||4.94 (1.55)||−.05|
|Vaccines||The use of vaccines to inoculate against infectious diseases.||2.76 (1.28)||5.60 (1.57)||.00|
|Antibacterial soap||The use of antibacterial soap, a type of cleaning product that contains chemical ingredients to kill bacteria.||2.75 (1.15)||5.01 (1.72)||−.38**|
|Nontampering nature activities||Wood‐burning stoves||The use of wood‐burning stoves and fireplaces. Wood is often burned to warm rooms or cook food.||2.29 (1.26)|
|Vegetable garden||Planting a home vegetable garden, in which edible produce is grown to provide seasonal food.||1.68 (1.00)|
|National park||Visiting a national park, in which people often engage in recreation such as hiking, canoeing, camping, or sightseeing.||1.61 (0.93)|
|Childbirth||The act of childbirth, or the process by which human infants are born.||1.29 (0.76)|
|Nonnature activities||Video games||Playing video games. People often play video games through their computer, TV system, or mobile phone.||1.49 (0.95)|
|Car repair||Do‐it‐yourself car repair, in which people repair and maintain their own vehicles rather than taking it into a car mechanic or dealership.||1.48 (0.83)|
|Baseball||The sport of baseball, sometimes called “America’s pastime,” a game played between two teams of nine players each, who take turns batting and fielding.||1.46 (0.93)|
|College degree||Getting a college degree, a title conferred on students by a college, university, or professional school on completion of a program of study.||1.35 (0.96)|
- Note: Descriptions with bolded words seen by participants. Ratings of tampering ranged from 1 (not at all) to 5 (very much); support for each activity ranged from 1 (strongly oppose) to 7 (strongly support). Support questions were only asked of tampering with nature activities.
- *p < .05; **p < .01.
2.2 Results and Discussion
2.2.1 Descriptive Statistics
Six of the eight tampering with nature activities had means above the midpoint of the five‐point scale, indicating that participants rated them as more than “somewhat” tampering with nature. Two activities—vaccines and antibacterial soap—fell below the midpoint of the scale and were excluded from further analysis. All eight nontampering with nature items had means below the scale midpoint.
Correlations between participants’ assessments of how much each activity tampered with nature and their support were weak to moderate, according to Cohen’s criteria for small (.1), moderate (.3), and large (.5) effect sizes (Cohen, 1988, 1992). Thus, participants did not base their assessment of tampering purely on how much they approved of these activities.
2.2.2 Factor Analysis
Although the overall reliability of the six remaining tampering with nature activities was good, principal axis exploratory factor analysis (EFA) with direct oblimin rotation yielded two factors (eigenvalues = 3.00 and 1.24). The first consisted of medicine‐related items (nanoparticles, gene therapy, and cloning) and the second of environment‐related items (pesticides, geoengineering, and GMOs). Therefore, we included separate composites of the environment‐related (α = .72) and medicine‐related (α = .83) items, as well as the composite of all tampering activities (α = .80) in each of the following studies. Notably, the environment‐related activities were rated as tampering more with nature than were the medicine‐related activities (Table I).
Thus, in Study 1, we identified six activities that people see as tampering with nature: geoengineering, GMOs, chemical pesticides, cloning, gene therapy, and nanoparticles. Next, we set out to create a measure of individual differences in ATN that could predict opposition to these activities.
3 STUDY 2
In Study 2, we sought to create a short and reliable measure of discomfort with tampering with the natural world, the ATN Scale. We then compared this scale to measures of related constructs. We hypothesized that ATN would correlate negatively with support for tampering activities (H1) and trust in technology and science (H2), and positively with connectedness to nature, beliefs that nature is fragile and requires protection, and valuing the biosphere (a “biospheric value orientation”) (H6). We had no predictions for other value orientations commonly measured in conjunction with nature beliefs (e.g., egoistic, altruistic, and openness values). We measured political ideology but did not expect it to correlate with ATN scores.
Participants were 201 American adults recruited via MTurk for $1.00. Twenty participants dropped out of the study or failed attention checks (remaining N = 181).
3.1.2 Procedure and Measures
The research team generated 22 statements to align with our definition of ATN (Table A2 in the Supporting Information). These statements were informed by existing scales, extensive reading of related research, and discussion among the authors.
The items were developed with content validity in mind (i.e., with the goal of capturing the full range of facets that make up beliefs about tampering with nature). For example, items measured consequentialist (i.e., instrumental) beliefs about the potential consequences of human intervention (e.g., “Human progress depends on finding ways to improve upon natural processes” [reverse‐coded]), moral (i.e., ideational) beliefs about whether there is an implied moral imperative to leave nature alone (e.g., “Human beings have no right to meddle with the natural environment”), beliefs about whether altering nature is fundamentally good or bad (e.g., “Altering nature will be our downfall as a species”), and a preference for a separation between human and natural processes (e.g., “I would prefer to live in a world where humans leave nature alone.”)
Our goal was to create statements that would capture key beliefs about tampering with nature, while still being broad enough to reasonably correlate with a variety of activities perceived to tamper with nature. Participants indicated their agreement with each statement (1 = Strongly disagree; 7 = Strongly agree), with higher scores indicating greater discomfort with tampering with nature.
Participants reported their support or opposition for the six tampering with nature activities identified in Study 1 (GMOs, pesticides, geoengineering, gene therapy, cloning, and nanoparticles) using a seven‐point scale (1 = Strongly oppose; 7 = Strongly support).
22.214.171.124 Individual difference correlates
Participants also completed measures of potential correlates of ATN. Unless otherwise indicated, response options ranged from 1 (Strongly disagree) to 7 (Strongly agree).
Trust in science and technology. We included two measures of trust in which higher scores indicate more trust in science/scientists or technology. A 22‐item Trust in Science and Scientists Scale (Nadelson & Jorcyk, 2014) measured trust in the scientific process and scientists (e.g., “When scientists change their mind about a scientific idea it diminishes my trust in their work” reverse‐coded). A five‐item trust in technology scale (Achterberg, 2014) captured beliefs about technological innovations (e.g., “Technological advances can be used to solve future problems”).
Environmental beliefs. We included two measures of environmental beliefs: the 15‐item New Ecological Paradigm (NEP; Dunlap et al., 2000) and the 14‐item Connectedness to Nature Scale (CNS; Mayer & Frantz, 2004). NEP measures a pro‐environmental worldview (e.g., “Humans are severely abusing the environment”), including items related to ATN. CNS measures personal connection to the natural world (e.g., “I often feel a sense of oneness with the natural world around me”). This measure does not address issues of tampering with nature, so we expected it to be less closely related to ATN than NEP.
Values. We measured participants’ values using subscales (biospheric, altruistic, egoistic, and openness to change) first identified by Schwartz (1977) and modified by Stern, Dietz, and Guagnano (1998). Participants rated each item in terms of how important it is as a guiding principle in their life (1 = Not at all important; 5 = Extremely important) with an option of Opposed to my values (scored as −1). Each value was measured with three items. Although we were primarily interested in the biospheric value subset because it is directly tied to protection of nature, we included the other values because they have also been discussed in reference to environmental outcomes (Dietz, 2015).
Finally, participants reported demographics (age, gender, education level, income, and race/ethnicity), political affiliation, and political ideology (1 = Very liberal; 7 = Very conservative).
3.2.1 Item Selection
Our goal was to create a short single‐factor scale that could be used by researchers in time‐limited surveys. Therefore, we culled the full 22‐item pool of items using principal axis EFA with direct oblimin rotation. Examination of the scree plot and eigenvalues showed that this initial EFA resulted in a five‐factor model with one large factor (eigenvalue = 9.52) and four smaller ones (eigenvalues ranging from 1.70 to 1.02). Using an iterative process, we then removed items that loaded highest on the smallest factors, those with low communalities (below .3), and those that contributed to high nonredundant residuals in the reproduced correlation matrix, followed by additional EFAs (Meyers, Gamst, & Guarino, 2013; Tabachnick & Fidell, 2001). Remaining items captured the breadth of our definition of ATN and possible reasons for this aversion.
The resulting five‐item ATN Scale—including factor loadings from all studies—is shown in Table II. Sampling adequacy and sphericity results supported the use of factor analysis (Kaiser‐Meyer‐Olkin (KMO) Test = 0.85, Bartlett’s Test = p < .001). An EFA indicated that a single factor explained 62% of item variance (eigenvalue = 3.12). The eigenvalue of the next highest factor (0.63) and the scree plot confirmed a one‐factor solution. Inter‐item reliability was excellent (α = .85); corrected item‐total correlations exceeded .45 for all items. Scores clustered close to the midpoint of the scale (M = 4.43, SD = 1.33) and were normally distributed (skew and kurtosis <2).
|Study 2||Study 3||Study 4|
|Item||Item‐Total r||EFA Factor Loadings||Item‐Total r||CFA Factor Loadingsa||Item‐Total r||CFA Factor Loadingsa|
|1. People who push for technological fixes to environmental problems are underestimating the risks.||.55||.59||.41||.49||.47||.51|
|2. People who say we shouldn’t tamper with nature are just being naïve.b||.54||.58||.30||.36||.59||.65|
|3. Human beings have no right to meddle with the natural environment.||.72||.81||.66||.78||.81||.87|
|4. I would prefer to live in a world where humans leave nature alone.||.76||.86||.71||.84||.80||.81|
|5. Altering nature will be our downfall as a species.||.71||.79||.67||.77||.73||.65|
- a STDYX standardized model results.
- b Reverse‐coded.
3.2.2 Individual Difference Correlates
See Table III for descriptive statistics and correlations between all psychological and activity‐support constructs.
|1. ATN (five‐item)||4.43 (1.33)||.85||−||−0.65||−0.73||−0.42||−0.42||−0.07||0.74||0.50||0.52||0.20||−0.12||−0.03|
|2. TN acts total||4.31 (1.19)||.80||−.54**||−||1.14||1.08||0.55||0.30||0.36||−0.33||−0.35||−0.04||0.12||0.14|
|3. TN acts env.||3.69 (1.46)||.74||−.58**||.88**||−||0.67||0.42||0.10||−0.58||−0.47||−0.45||−0.11||0.29||0.32|
|4. TN acts med.||4.92 (1.28)||.76||−.34**||.84**||.50**||−||0.57||0.45||−0.02||−0.09||−0.14||0.04||−0.08||−0.09|
|5. Trust in tech||4.94 (1.04)||.83||−.35*||.45**||.33**||.45**||−||0.32||−0.32||0.11||−0.13||−0.04||0.16||0.36|
|6. Trust in science||4.94 (0.90)||.92||−.06||.26**||.08||.38**||.28**||−||0.28||0.02||0.03||0.25||−0.33||0.07|
|7. NEP||4.78 (1.12)||.91||.65**||.31**||−.48**||−.02||−.28**||.26**||−||0.42||0.56||0.35||−0.39||−0.04|
|8. CNS||4.61 (0.82)||.80||.41**||−.26**||−.36**||−.07||.09||.02||.36**||−||0.78||0.30||0.14||0.47|
|9. Biospheric value||3.72 (1.07)||.82||.43**||−.28**||−.35**||−.11||−.11||.03||.48**||.63**||−||0.62||0.03||0.48|
|10. Altruistic value||4.04 (0.96)||.69||.15||−.03||−.08||.03||−.03||.20**||.28**||.22**||.47**||−||−0.12||0.51|
|11. Egoistic value||2.60 (0.99)||.66||−.09||.09||.20**||−.06||.12||−.26**||−.30**||.10||.02||−.08||−||0.45|
|12. Openness value||3.59 (0.86)||.54||−.02||.07||−.03||.16*||.24**||.05||−.03||.31**||.32**||.31**||.27**||−|
|13. Political ideo.||3.38 (1.70)||−||−.05||−.22**||−.08||−.31**||−.02||−.27**||−.31**||−.02||−.00||−.12||.17*||.06|
- Note: Numbers below the diagonal reflect uncorrected correlations; correlations above the diagonal are corrected for reliability, except for political ideology (which was a single‐item measure).
- *p < .05; **p < .01.
- NEP = New Ecological Paradigm; CNS = Connectedness to Nature scale.
A moderate negative relationship emerged between ATN and trust in technology, matching the predicted pattern, but not between ATN and trust in science/scientists. This lack of relationship between ATN and trust in science/scientists may be an artifact of the scale used rather than the underlying construct, as the measure includes both assessment of trust in the scientific process and in scientists themselves. ATN correlated positively with measures of environmental beliefs, including a strong correlation with NEP and moderate correlations with CNS and biospheric values. There was a weak correlation between ATN and the measure of altruistic concern, and no relationship between ATN and openness or egoistic values. Female gender correlated weakly with ATN scores (r = .15, p = .050); there were no correlations with other demographic and political variables (age: r = .00, p = .997; education: r = −.06, p = .450; income: r = .01, p = .903; political ideology: r = −.04, p = .635).
Because ATN correlated strongly with NEP, regressions tested whether the strength of the relationships between ATN and other individual difference measures changed when controlling for NEP (Table A3 in the Supporting Information). Controlling for NEP slightly dampened the magnitude of the relationships between ATN and items directly related to nature (i.e., CNS and biospheric values), although these remained statistically significant. For altruistic values, controlling for NEP led a weak relationship to become even weaker. For other outcomes (i.e., trust in science, egoistic values, and political ideology), controlling for NEP increased the magnitude of relationships with ATN and led to them becoming statistically significant. All other patterns remained unchanged.
3.2.3 Tampering Activities
ATN was negatively related to support for activities that alter nature, particularly environment‐related activities (Table III).
As with the individual difference correlates, regressions tested the effect of ATN on activity support while controlling for NEP (Table A3 in the Supporting Information).2 Controlling for NEP slightly weakened the relationship between ATN and support for environmental activities, although the effect remained strong and statistically significant. Including NEP strengthened the negative relationship between ATN and support for medicine‐related activities. Tests of correlated correlations showed that the correlation between ATN and the composite of all tampering activities was significantly stronger than the correlation between NEP and these activities (Table A3 in the Supporting Information; Steiger, 1980). This was also the case for medicine‐related activities, and marginally so for environmental activities.
Trust in technology and trust in science/scientists also correlated with support for tampering activities. These constructs had stronger relationships with support for medicine‐related activities than ATN, but the pattern was reversed with environmental activities and for the composite of all activities. To test whether ATN is distinct from these constructs, we ran a separate series of regressions (Table A4 in the Supporting Information). When controlling for these trust measures, the strong relationship between ATN and opposition to tampering activities dampened only slightly. Steiger’s test showed that the relationship between ATN and all tampering activities was significantly stronger than that between trust in science/scientists and tampering activities, although this was not significant for trust in technology. The correlation between ATN and environment‐related activities was significantly stronger than those between trust in science/scientists or trust in technology and those activities. No significant differences emerged between correlations with medicine‐related activities.
In Study 2, we created a short and reliable ATN Scale that successfully predicted opposition to a variety of activities that alter the natural world (supporting H1). Furthermore, this construct correlated negatively with trust in technology (supporting H2) and positively with general environmental worldviews and values (supporting H6). As expected, ATN was closely related to NEP, but ATN was still predictive even when controlling for environmental worldviews. Thus, although ATN is related to environmental beliefs and values, it appears to be a distinct construct.
We also expected ATN to correlate negatively with trust in science (H2), but only found evidence for this relationship when controlling for NEP. Trust in science also failed to correlate with support for environment‐related tampering activities, suggesting that trust in science is not a major component of ATN. The trust in science questions—which were largely about the scientific process and the people who conduct this process—may have been distal to concerns about its applications. It is also possible that a more nuanced measure of trust in scientists that captured assessments of warmth or competence (Fiske, Cuddy, & Glick, 2007) would have uncovered richer relationships between that construct and ATN. We also found no relationship between ATN and ideology, suggesting ATN is not exclusive to one political ideology.
4 STUDY 3
Study 3 sought to confirm the structure of the ATN Scale identified in Study 2 and to test its relationships with additional psychological constructs. As before, we expected ATN would predict opposition to tampering‐related activities (H1). We also tested associations between ATN and purity judgments (H4) and religious beliefs (H5).
We hypothesized that ATN may be related to fears about contamination and thus disgust (H4) and that people who consider purity a moral value would be particularly upset about tampering with nature. We therefore included measures of harm, fairness, loyalty, authority, and purity values as outlined by Moral Foundations Theory (Graham et al., 2011).
We had no specific prediction about whether people who are generally religious would score high on ATN but predicted that people high in ATN would eschew acts in which humans seem to play God (H5). Thus, we included measures of APG, general religiosity, and divinity values. Because divinity has been tied to disgust and purity (Rozin et al., 1999), we predicted that this ethical value might also relate to ATN. We had no specific predictions for other ethical values (autonomy or community).
Participants were 510 American adults recruited through Social Survey International. We used quotas to match U.S. Census proportions of age, gender, and education level. Forty‐nine participants failed attention checks and were removed from the data set (remaining N = 461).
Participants completed the five‐item ATN Scale identified in Study 2 as well as measures of individual difference correlates and support for tampering activities.
126.96.36.199 Individual difference correlates
The 20‐item Moral Foundations Questionnaire measures five systems of moral judgments: harm/care, fairness/reciprocity, in‐group/loyalty, authority/respect, and purity/sanctity (Graham et al., 2011). Participants responded to two questions for each category, first indicating each factor’s relevance to judgments of morality (0 = Not at all relevant; 5 = Extremely relevant) and then their agreement with statements exemplifying these judgments (0 = Strongly disagree; 5 = Strongly agree).
Disgust sensitivity. Participants reported their tendency to experience disgust using the eight‐item version of the Disgust Sensitivity Scale (DSS; Haidt, McCauley, & Rozin, 1994; Inbar, Pizarro, Knobe, & Bloom, 2009), with response options from 0 (Not disgusting at all) to 4 (Extremely disgusting).
Ethical values. The ethical values of autonomy, community, and divinity were measured using the 12‐item version of the Ethical Values Assessment (Padilla‐Walker & Jensen, 2016). Participants rated the importance of each item (four for each subscale) from 1 (Not at all important) to 5 (Completely important).
Aversion to playing God. Participants completed the seven‐item APG Scale (Waytz & Young, 2019). Participants responded to various statements (1 = Strongly disagree; 6 = Strongly agree), with higher scores indicating greater aversion.
Environmental beliefs. As in Study 2, we measured environmental beliefs with the NEP.
Religiosity and political ideology. Participants reported their political ideology as before and how much they considered themselves to be a religious person (1 = Not at all; 6 = Extremely).
188.8.131.52 Tampering activities
Participants reported their support or opposition for the six activities identified in Study 1.
A confirmatory factor analysis (CFA) using full information maximum likelihood tested the five‐item ATN Scale (Table II). All items were sufficiently normally distributed and the overall fit for the measurement model was good. The χ2goodness of fit test (df = 5) was 13.5, p = .019, indicating that the observed and implied covariance matrices significantly differed. However, this value can be affected by sample size (Hoyle, 2011). All other indices indicated adequate fit according to recommended cutoffs (Hoyle, 2011), including the χ2/df ratio (1.5), CFI (0.99), and RMSEA (0.06, 90% CI [0.02, 0.10], p close‐fit = .275). Inter‐item reliability was again good (α = .77). Item‐total correlations indicated one item that was below .45 (item‐total r = .30), likely due to that question being reverse‐coded.
4.2.1 Individual Difference Correlates
See Table IV for descriptive statistics and correlations between all measures. ATN had weak positive correlations with purity and disgust sensitivity. ATN was also weakly correlated with the moral foundations of harm and fairness, but not loyalty or authority.
|1. ATN (five‐item)||4.70 (1.09)||.79||−||−0.29||−0.32||−0.19||0.21||0.35||0.29||0.00||0.11||0.24||0.40||0.09||0.19||0.18||0.72|
|2. TN acts total||4.08 (1.12)||.79||−.23**||−||1.15||1.10||−0.21||0.03||0.07||0.01||−0.05||−0.14||−0.33||−0.18||0.03||−0.05||−0.15|
|3. TN acts env.||3.56 (1.35)||.73||−.24**||.87**||−||0.64||−0.15||−0.11||−0.09||0.03||−0.01||−0.12||−0.28||−0.06||−0.08||−0.01||−0.30|
|4. TN acts med.||4.61 (1.26)||.76||−.15**||.85**||.48**||−||−0.24||0.18||0.22||0.00||−0.07||−0.15||−0.32||−0.28||0.14||−0.10||0.04|
|5. MFQ purity||3.22 (1.02)||.72||.16**||−.16**||−.11*||−.18**||−||0.48||0.52||0.86||0.95||0.55||0.58||0.66||0.32||0.63||−0.17|
|6. MFQ harm||3.77 (0.86)||.70||.26**||.02||−.08||.13**||.34**||−||0.99||0.51||0.46||0.38||0.27||0.21||0.62||0.54||0.34|
|7. MFQ fair||3.84 (0.77)||.67||.21**||.05||−.06||.16**||.36**||.68**||−||0.45||0.51||0.30||0.26||0.26||0.64||0.55||0.25|
|8. MFQ loyal||3.14 (0.98)||.63||.00||.01||.02||−.00||.58**||.34**||.29**||−||1.06||0.41||0.43||0.55||0.31||0.67||−0.39|
|9. MFQ auth.||3.27 (0.96)||.69||.08||−.04||−.01||−.05||.67**||.32**||.35**||.70**||−||0.45||0.57||0.64||0.31||0.70||−0.37|
|10. DSS||2.48 (0.76)||.74||.18**||−.11**||−.09||−.11*||.40**||.27**||.21**||.28**||.32**||−||0.29||0.61||0.43||0.28||0.01|
|11. APG||5.10 (1.23)||.86||.33**||−.27**||−.22**||−.26**||.46**||.21**||.20**||.32**||.44**||.23**||−||0.72||0.17||0.41||0.00|
|12. EVA div.||3.39 (1.40)||.96||.08||−.16**||−.05||−.24**||.55**||.17**||.21**||.43**||.52**||.51**||.65**||−||0.67||0.36||−0.30|
|13. EVA aut.||4.40 (0.54)||.71||.14**||.02||−.06||.10*||.23**||.44**||.44**||.21**||.22**||.31**||.13**||.55**||−||0.30||0.23|
|14. EVA comm.||3.86 (0.75)||.68||.13**||−.04||−.01||−.07||.44**||.37**||.37**||.44**||.48**||.20**||.31**||.29**||.21**||−||−0.07|
|15. NEP||4.69 (0.82)||.83||.58**||−.12**||−.23**||.03||−.13**||.26**||.19**||−.26**||−.26**||.01||−.27**||−.00||.18**||−.05||−|
|16. Political ideo.||3.98 (1.71)||−||−.10*||−.14**||−.08||−.16**||.32**||−.11*||−.11*||.32**||.37**||.06||.36**||.38**||−.03||.09||−.35**||−|
|17. Religiosity||3.41 (1.73)||−||.01||−.09*||−.01||−.16**||.46**||.08||.14**||.36**||.43**||.19**||.54**||.76**||.06||.33**||−.23**||.35**|
- Note: Numbers below the diagonal reflect uncorrected correlations; correlations above the diagonal are corrected for reliability, except for political ideology and religiosity (which were single‐item measures).
- *p < .05; **p < .01.
- NEP = New Ecological Paradigm; MFQ = Moral Foundations Questionnaire; EVA = Ethics Values Assessment; DSS = Disgust Sensitivity Scale; APG = Aversion to Playing God.
ATN was moderately and positively related to APG, but unrelated to religiosity and the ethical value of divinity. Weak correlations emerged between ATN and values of autonomy and community. Replicating Study 2, ATN was strongly correlated with NEP.
Unlike in Study 2, liberal ideology (r = −.10, p = .035), lower income (r = −.13, p = .004), and less education (r = −.12, p = .009) were all weakly correlated with ATN, and gender was not (r = −.06, p = .186). As before, age was unrelated to ATN scores (r = −.04, p = .353).
As in Study 2, regressions tested the effects of ATN while controlling for NEP (Table A3 in the Supporting Information). For the ethical value of autonomy, including NEP caused an already weak relationship with ATN to become weaker (and nonsignificant). Controlling for NEP dampened the magnitude of the relationships between ATN and the moral foundations of harm and fairness, but these effects remained significant.
For three other outcomes—the moral foundations of loyalty and authority and the ethical value of divinity—controlling for NEP led ATN to become a significant predictor when it had not been before. Controlling for NEP also strengthened existing relationships between ATN and the moral foundation of purity, disgust sensitivity, the ethical value of community, and APG. The most notable change was that the relationship between ATN and political ideology was reversed, going from ATN being associated with more liberal ideology to ATN being associated with more conservative ideology when including NEP in the model. This is likely due to the politicization of environmental beliefs. Overall, these findings add further evidence that ATN is a useful predictor of related constructs above and beyond general environmental concern.
4.2.2 Tampering Activities
As in Study 2, ATN was negatively associated with support for activities that tamper with nature (Table IV). These effects were smaller than in Study 2, but were again significantly stronger than the relationships between NEP and these activities (Table A3 in the Supporting Information). For medical activities, the APG Scale was a stronger correlate of these activities than ATN, although ATN again outperformed NEP (Tables A3 and A4 in the Supporting Information). For environmental activities, the ATN Scale was the strongest correlate among all of the measures, although this correlation was not significantly different than that between NEP and environmental activities.
Controlling for NEP had no effect on the relationship between ATN and these activities (Table A3 in the Supporting Information). Controlling for APG in regressions (Table A4 in the Supporting Information) erased the weak (but significant) effect of ATN on medicine‐related activities. It also slightly dampened the relationships between ATN and environmental activities and ATN and the composite of all tampering activities, although these effects remained significant. Thus, an aversion to playing God may play a central role in discomfort with altering the natural world through medical interventions, but less so in other domains.
Study 3 confirmed the structure of the ATN Scale and helped illuminate how this construct fits among existing psychological phenomena. Most notably, ATN again predicted opposition to activities that tamper with nature (H1). Replicating Study 2, ATN was most closely related to NEP, a measure of environmental worldview (supporting H6). However, ATN still added explanatory power above and beyond that contributed by NEP.
This study also unpacked how ATN is related to various moral and religious beliefs. We confirmed that ATN was associated with disgust sensitivity and purity values (supporting H4) as well as beliefs that one should not change God’s creation (supporting H5). However, it was not associated with general religious values (religiosity or valuing divinity).
We also found weak relationships between ATN and several nonhypothesized moral and ethical categories, including fairness, autonomy, and community. We are hesitant to interpret these findings given that they were exploratory and some changed when covariates were included in the regression models. Further research should test whether these effects are replicable and what may drive them.
5 STUDY 4
Studies 2 and 3 found that ATN correlated with self‐reported opposition to tampering activities, whereas Study 4 tested whether ATN predicts actual behavior in the form of monetary donations to a relevant cause. Given that ATN captures worry about tampering with nature (rather than excitement over its possibilities), we hypothesized that ATN would better predict donations to anti‐tampering causes than to pro‐tampering causes.
Second, we tested whether ATN correlates with an overall naturalness bias (H3). Finally, Study 4 sought to replicate the relationships found in Study 2 between ATN and technology beliefs (H2) and environmental beliefs (H6).
American adults (N = 302) were recruited from MTurk in exchange for $1.00. Of these, 23 were removed from the data set for failing attention checks, leaving 279 participants.
Participants completed the same ATN, NEP, and values scales as in Study 2. Several new measures were also included, as detailed below.
184.108.40.206 Naturalness bias
Following DiBonaventura and Chapman (2008), naturalness bias was assessed with a one‐item question describing two versions of a hypothetical flu medication that are chemically identical but derived from either natural or human processes. Participants indicated which medication they preferred on a five‐point scale (1 = Strongly prefer Drug X; 5 = Strongly prefer Drug Y). Higher scores indicate bias toward the nature‐derived drug.
220.127.116.11 Technology beliefs
The 14‐item Technology Attitudes Scale (TAS; Francis, Schwartz, & Inbar, 2016) captured feelings of aversion to technological advances (e.g., “The benefits of new technology are often grossly overstated”). Response options range from 1 (Strongly disagree) to 7 (Strongly agree).
18.104.22.168 Environment beliefs
This study included an alternative measure of nature beliefs: the Inclusion of Nature in the Self scale (INS; Schultz, 2002). This one‐item scale has seven pictorial response options with varying degrees of overlap between “Nature” and “Self,” with higher scores indicating more INS. Value orientations and NEP were measured as in Study 2.
22.214.171.124 Donation behavior
At the end of the survey, participants were told that they would receive a $0.50 bonus and were given the option to donate some or all of that bonus to a GMO‐related charity. Two real charities were described to participants. The first (pro‐GMO) cause, Biology Fortified, was described as a group that “supports GMOs and raises money to help promote them.” The second (anti‐GMO) cause, Center for Food Safety, was described as a group that “is against GMOs and advocates for strict regulation and labeling of GMO food so that people can avoid it if desired.” Participants could choose to donate to either of these charities in increments of $0.10 up to the full bonus amount ($0.50).
Finally, participants reported their demographic and political information.
CFA again confirmed the structure of the five‐item ATN Scale. All items were normally distributed, and model fit was excellent: χ2 goodness of fit (df = 5) = 7.19, p = .207; CFI = 0.997; χ2/df ratio = 1.43; RMSEA = 0.04 (90% CI [0.00, 0.10], p close‐fit = .536). Inter‐item reliability was good (α = .86), and all corrected item‐total correlations exceeded .45.
5.2.1 Individual Difference Correlates
Table V shows the descriptive statistics and correlations between all individual difference measures. As predicted, ATN was strongly and positively correlated with naturalness bias and moderately so with technology aversion. As before, ATN strongly correlated with NEP scores. ATN was moderately correlated with biospheric values, as in Study 2, as well as INS. There was a moderate positive correlation between ATN and altruistic values and no relationship between ATN and openness values, again replicating Study 2. Unlike Study 2, weak correlations emerged between ATN and egoistic values. Matching Study 3, but not Study 2, liberal political ideology was weakly correlated with ATN scores.
|1. ATN (five‐item)||4.37 (1.35)||.85||−||0.59||0.81||0.56||0.37||−0.15||0.14|
|2. Tech aversion||3.82 (0.97)||.89||.51**||−||−0.31||0.30||0.06||−0.13||−0.09|
|3. Naturalness bias||3.85 (1.09)||−||.30**||.33**||−|
|4. NEP||4.73 (1.08)||.91||.71**||−.28**||.20**||−||0.57||0.47||−0.38||0.19|
|5. INS||4.66 (1.50)||−||.34**||.24**||.14*||.36**||−|
|6. Biospheric value||3.73 (1.10)||.82||.47**||.26**||.25**||.49**||.50**||−||0.61||−0.11||0.44|
|7. Altruistic value||3.86 (1.22)||.78||.30**||.05||.06||.40**||.16**||.49**||−||0.05||−0.33|
|8. Egoistic value||2.46 (1.09)||.75||−.12*||−.11||−.08||−.31**||−.16**||−.09||.04||−||0.34|
|9. Openness value||3.60 (0.99)||.72||.11||−.07||.10||.15*||.14*||.34**||−.25**||.25**||−|
|10. Political ideo.||3.48 (1.88)||−||−.21**||.15*||.10||−.49**||−.05||−.26**||−.37**||.28**||−.17**|
- Note: Numbers below the diagonal reflect uncorrected correlations; correlations above the diagonal are corrected for reliability (for multi‐item measures).
- *p < .05; **p < .01.
- NEP = New Ecological Paradigm; INS = Inclusion of Nature in the Self.
Although education (r = −.06, p = .332) and income (r = .02, p = .738) did not correlate with ATN, age (r = .11, p = .059) and female gender was positively correlated with ATN in this study (r = .19, p = .002), partially replicating the pattern found in Study 2.
As before, regressions controlling for NEP scores found that ATN’s relationships to other constructs go above and beyond environmental worldview (Table A3 in the Supporting Information). Including NEP somewhat strengthened the relationships between ATN and technology aversion and somewhat dampened the relationship between ATN and INS, but the pattern of significance did not change. As in Study 3, including NEP led the ATN to switch from being associated with liberalism to conservatism.
5.2.2 Behavioral Outcome
We also assessed the donations to pro‐ or anti‐GMO charities. A composite measure was created with a range of −50 to 50, in which negative scores (−50 to −1) reflected donations (in U.S. cents) to the pro‐GMO cause, a score of 0 indicated no donation, and positive scores (1–50) reflected donations to the anti‐GMO cause (M = 4.73, SD = 14.66). This measure was highly kurtotic due to large proportion of people who did not donate to either charity (74.8%).
Therefore, we used multinomial logistic regression to test the effect of ATN on the probability of donation to either charity. Nondonation was treated as the reference category (0), with any donation to the pro‐GMO category coded as −1 and any donation to the anti‐GMO category coded as 1. For each unit increase in ATN, there was a 60% increase in the odds that participants donated to the anti‐GMO charity rather than not donating, B = 0.47, SE = 0.12, Wald χ2(1) = 14.50, p < .001, odds ratio = 1.60, 95% CI [1.26, 2.04]. In contrast, ATN did not predict donations to the pro‐GMO charity versus no donations, B = −0.16, SE = 0.20, Wald χ2(1) = 0.64, p = .422, odds ratio = 0.85, 95% CI [0.57, 1.26], Nagelkerke Pseudo‐R2= .08. By changing the reference group to the pro‐GMO cause, we found that higher levels of ATN also predicted greater likelihood of donating to the anti‐GMO cause than to the pro‐GMO cause, B = 0.63, SE = 0.23, Wald χ2(1) = 7.79, p = .005, odds ratio = 1.88, 95% CI [1.21, 2.93]. Thus, ATN predicted behavior asymmetrically: it was a better predictor of support for anti‐tampering than pro‐tampering activities.
A linear regression tested the amount donated among participants who had chosen to donate to either charity (N = 73; removing all nondonations led the composite measure to be normally distributed). As anticipated, people with higher ATN scores donated more to the anti‐GMO organization, b = 6.34, SE = 2.33, t = 2.72, p = .008, R2 = .09. Thus, not only did ATN predict whether or not people donated to tampering‐related causes, it also predicted the amount donated.
Study 4 demonstrated that ATN not only predicts self‐reported opposition to activities that tamper with nature (Studies 2 and 3), but also real behaviors opposing these activities (supporting H1). Furthermore, this study allowed us to unpack the linear relationships between ATN and activity support, showing that this relationship is asymmetrical: ATN predicts opposition to activities that tamper with nature but does not predict support for them.
Study 4 further found that ATN is related to an overall naturalness bias (H3). It also replicated Study 2 by showing that ATN is related to beliefs about technology using the TAS (H2), and environment‐related measures of NEP, biospheric values, and INS (H6).
Furthermore, Study 4, as with Study 3, showed a weak negative relationship between ATN scores and political ideology, hinting that liberals are somewhat more concerned with tampering with nature than are conservatives.
6 GENERAL DISCUSSION
Across four studies, we created and validated a new measure—the ATN Scale—which is the first efficient and direct measure of discomfort with altering the natural world. Most importantly, ATN correlates with opposition to activities that tamper with nature (H1). It is also related to various measures of theoretically related constructs, including beliefs about technology (H2), general naturalness bias (H3), purity and disgust concerns (H4), religious hubris beliefs (H5), and general environmental concerns and worldviews (H6). Thus, the ATN Scale provides a useful tool for researchers interested in the role of tampering‐related beliefs in issues such as moral judgments, public support for technology, and science communication.
6.1 Opposition to Tampering Activities
Supporting H1, ATN predicted support for a wide range of activities that are perceived to tamper, including both medicine‐ and environment‐related activities. The ATN Scale not only correlated with self‐reported opposition to these activities (Studies 2 and 3), but also predicted whether and how much of their money people donated to an anti‐tampering charity (Study 4).
Importantly, this effect seems to be unipolar rather than bipolar: ATN predicted donations to an anti‐GMO cause but not donations to a pro‐GMO cause. Thus, the absence of ATN does not necessarily predict support for tampering activities, but the presence of discomfort predicts opposition to them. Support for tampering may require factors beyond comfort with altering the natural world, such as a predilection for early adoption of technologies or enthusiasm about human innovation.
Although ATN consistently correlated with opposition to tampering activities, we found that this effect was stronger for environmental than medical activities (Studies 2 and 3). This trend could be an artifact of how much participants thought these behaviors tampered with nature: the activities that made up the environmental composite (geoengineering, GMOs, and pesticides) were all rated as tampering more than any of the medicine‐related items (nanoparticles, gene therapy, and cloning) (Study 1). Alternatively, the perceived benefits of medical activities may override tampering concerns; the benefits of environmental activities may seem less direct, and thus may present a weaker counterpoint to ATN. This explanation would fit research showing that the acceptability of biotechnology differs by the domain in which it is used (e.g., food vs. medicine; Frewer, Howard, & Shepherd, 1997). Future research might explore the role of perceived benefits in relation to tampering concerns across domains.
6.2 Relationships with Related Constructs
We found that the ATN Scale correlates to several constructs that are theoretically related to discomfort with altering the natural world. As hypothesized, ATN correlated negatively with trust in technology (Study 2) and positively with technology aversion (Study 4), supporting H2. Thus, at least part of the driving force behind ATN seems to be a distrust of the technology that results in tampering. However, we did not find the hypothesized relationship between ATN and trust in science or scientists (H2; Study 2). One explanation is that people who think scientists have good intentions or that the scientific process is sound might still believe those actions will cause harm. As we note in Study 2, scales assessing other perceptions of scientists (such as warmth or competence judgments) might also reveal more of a relationship between these perceptions and ATN.
ATN was also positively related to a bias toward natural over equivalent human‐made products (H3; Study 4). People may eschew human‐made products in part to avoid supporting underlying processes that interfere with nature. The ATN Scale also significantly and positively correlated with measures reflecting purity concerns and disgust sensitivity (H4; Study 3). These results mirror research findings that “nature” is associated with such terms as “pure” and “uncontaminated” (Rozin, Fischler, & Shields‐Argelès, 2012).
In addition to purity concerns, we confirmed that ATN correlated with fears about “playing God” (H5; Study 3). This pattern emerged despite the fact that ATN was unassociated with general religiosity or valuing divinity (Study 3). Thus, only the specific elements of religious beliefs that admonish humanity’s attempt to supersede godly authority predict ATN. Although some religious groups oppose particular forms of tampering (e.g., stem cell research; Squires, 2014), general religiosity appears to be unrelated.
Finally, we confirmed that ATN correlates positively with concerns about the natural world, including environmental worldviews and feelings of connection to nature (H6). As predicted, the relationship between ATN and environmental beliefs was particularly strong when environmental concern was measured with the NEP. This high correlation is likely due to inclusion of several tampering‐related items in the NEP. Yet ATN remained a significant correlate of a wide variety of beliefs and behaviors even when controlling for NEP. The distinction between these two measures may be due to the fact that NEP focuses on threats to the fragility and balance of natural systems (Dunlap et al., 2000), whereas the ATN Scale characterizes tampering activities without reference to vulnerability.
Furthermore, ATN related to other constructs in different ways than NEP (see Tables III–V). ATN was negatively related to medical activities that tamper and positively related to disgust sensitivity, whereas NEP was not significantly correlated with these variables. Similarly, ATN is positively related to APG and the moral foundation for purity, while NEP is negatively correlated with these constructs. Perhaps most notably, NEP correlates significantly with variables strongly associated with polarization around environmental issues in the United States (political conservatism, egoistic values, and two of the binding moral foundations). By contrast, ATN has much weaker and often nonsignificant relationships with these variables. Taken together, these results suggest that ATN is a more direct and inclusive measure of ATN than NEP. ATN encompasses some of the purity, religious hubris, and disgust concerns often associated with tampering while at the same time avoiding the polarizing environmental and political views that the NEP may trigger.
As seen in Tables A5–A7 in the Supporting Information, measuring several constructs together can also predict opposition to tampering activities; our five‐item ATN Scale makes a unique contribution by capturing ATN via an efficacious and direct alternative to a series of individual questionnaires. The existing measures we tested each get at ATN obliquely, and so in combination can achieve the same effect. Yet in time‐limited studies, researchers need to capture such a construct more efficiently—thus we believe that the ATN Scale offers a tool that complements existing resources.
6.3 Limitations and Future Directions
Our goal with this article is to spur future work in the area of perceptions of tampering with nature, rather than claim this is a definitive examination of this construct and measure. As such, we note a number of limitations to the present research that suggest areas for further exploration.
The items for the ATN Scale were developed by the research team by performing an in‐depth review of the relevant literature and writing items with the goal of capturing the full range of the construct. We did not use cognitive interviews as part of our initial scale development; future research may use this approach to test the generalizability of the items and ensure that they are understood as intended by diverse audiences.
A second measurement issue has to do with our use of Likert‐style response options (strongly disagree to strongly agree), as these can lead to acquiescence bias and require more cognitive effort than item‐specific responses (Krosnick & Presser, 2009). In generating the question bank in Study 2, we included an equal number of normally coded and reverse‐coded items in an attempt to account for this possibility; however only one reverse‐coded item made it into our final measure. Therefore, it is possible that the average score of ATN is higher than it would be if we had used more item‐specific response options.
In regard to the relationship between ATN and other measures, in Studies 1–3, we measured support for tampering activities via self‐report rather than objective behavior. We did measure actual donation behavior in Study 4 for one tampering activity (GMOs). However, the full pattern of ATN‐predicted opposition may look different if opposition were measured via real advocacy behavior rather than self‐reports.
Another area of future study is the relationship between ATN and political beliefs. Although we did not have any a priori hypotheses about political ideology, we found that liberals showed a slightly greater degree of ATN in Studies 3 and 4 (but not in Study 2). This result may be driven by the tendency for U.S. liberals to exhibit greater environmental concern than other political orientations (Hamilton & Saito, 2014). Consistent with this explanation, when NEP was controlled, the relationship between ATN and political ideology switched direction. Further research should clarify this relationship and identify the mechanism that underlies discomfort with tampering among those on the political left, as well as whether this trend extends cross‐culturally.
It may also be interesting to measure how stable these beliefs are over time by assessing the test–retest reliability of the ATN Scale. We suspect that while baseline levels of ATN remain fairly consistent over time, hearing about tampering‐related activities might make these beliefs more salient and thus may temporarily provoke more aversion in response.
Finally, future research should explore contexts in which people are more open to tampering with the natural world or particular people who may be more excited about doing so. For example, we found that that ATN was a better predictor of environmental outcomes than medical ones; future research might explore how ATN relates to other domains such as lab‐grown meat. We also found that the ATN was a better predictor of opposition to tampering than support for tampering. Yet some groups of people (i.e., biohackers; developers of emergent technologies) feel very positively about these activities and some forms of tampering (i.e., in material and health sciences) may be more universally supported than those tested here. Exploring the factors and contexts that predict enthusiasm about emerging technologies is equally as important as those that predict aversion to them, and suggests a robust area for further research.
The present study is the first to offer a validated scale that can be used to measure ATN. ATN provides opportunities for researchers and practitioners to better understand public opposition to technological innovations, predict individual preferences for nature and “natural” products, and refine strategies for science communication. As such, the ATN Scale is useful for risk, communication, and psychology scholars with interests in areas such as technology, medicine, and the environment.
This research was funded by a Beyond Carbon Neutral grant from the University of Michigan Energy Institute.
- 1 Attention checks varied across studies. In Study 1, one check was embedded within activity‐support questions and instructed participants to “Click on ‘Somewhat oppose’ for this question.” The second told participants “If you are reading this question, please leave it blank.” We excluded those who failed either attention check from the analyses. Study 2 added a check in the NEP bank that instructed participants to “Answer ‘Strongly agree’ for this question.” Participants who failed any of the three checks were excluded. Study 3 had three embedded questions: in the NEP bank, the activity‐support bank, and another in the Disgust Sensitivity Scale items (“Answer ‘slightly disgusting’ for this question”). Participants missing two out of three were excluded. Study 4 had the “leave it blank” question and the NEP question; missing either resulted in exclusion.
- 2 Tables A5–A7 in the Supporting Information show results of regressions that control for all individual difference measures for each study.
- 2014). The changing face of public support for hydrogen technology explaining declining support among the Dutch (2008‐2013). International Journal of Hydrogen Energy, 39(33), 18711– 18717. https://doi-org.uoelibrary.idm.oclc.org/10.1016/j.ijhydene.2014.08.053 (
- 2017). Dispute and morality in the perception of societal risks: Extending the psychometric model. Journal of Risk Research, 20(3), 299– 325. https://doi-org.uoelibrary.idm.oclc.org/10.1080/13669877.2015.1043571 , , & (
- 1992). Risk society: Towards a new modernity. London, UK: Sage Publications. (
- 2008). Nanotechnologies, food, and agriculture: Next big thing or flash in the pan? Agriculture and Human Values, 25(2), 215– 218. https://doi-org.uoelibrary.idm.oclc.org/10.1007/s10460-008-9119-z (
- 1988). Statistical power analysis for the behavioral sciences ( 2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates. (
- 1992). A power primer. Psychological Bulletin, 112(1), 155– 159. https://doi-org.uoelibrary.idm.oclc.org/10.1037/0033-2909.112.1.155 (
- 2010). Factors influencing people’s acceptance of gene technology: The role of knowledge, health expectations, naturalness, and social trust. Science Communication, 32(4), 514– 538. https://doi-org.uoelibrary.idm.oclc.org/10.1177/1075547009358919 , & (
- 2013). Messing with nature? Exploring public perceptions of geoengineering in the UK. Global Environmental Change, 23(5), 938– 947. https://doi-org.uoelibrary.idm.oclc.org/10.1016/j.gloenvcha.2013.06.002 , , , & (
- 2008). Do decision biases predict bad decisions? Omission bias, naturalness bias, and influenza vaccination. Medical Decision Making, 28(4), 532– 539. https://doi-org.uoelibrary.idm.oclc.org/10.1177/0272989X08315250 , & (
- 2015). Environmental values. In T. Bosch & D. Sander (Eds.), Oxford handbook of values (pp. 329– 349). Oxford: Oxford University Press. (
- 2013). Framing synthetic biology. Science Communication, 35(5), 547– 571. https://doi-org.uoelibrary.idm.oclc.org/10.1177/1075547012470707 , & (
- 1978). The “New Environmental Paradigm.” The Journal of Environmental Education, 9(4), 10– 19. https://doi-org.uoelibrary.idm.oclc.org/10.3200/JOEE.40.1.19-28 , & (
- 2000). Measuring endorsement of the New Ecological Paradigm: A revised NEP scale. Journal of Social Issues, 56(3), 425– 442. https://doi-org.uoelibrary.idm.oclc.org/10.1111/0022-4537.00176 , , , & (
- 2007). Universal dimensions of social cognition: Warmth and competence. Trends in Cognitive Sciences, 11, 77– 83. https://doi-org.uoelibrary.idm.oclc.org/10.1016/j.tics.2006.11.005 , , & (
- 2016). Development of the Technology Attitudes Scale. Unpublished manuscript. , , & (
- 1997). Public concerns in the United Kingdom about general and specific applications of genetic engineering: Risk, benefit, and ethics. Science, Technology & Human Values, 22(1), 98– 124. https://doi-org.uoelibrary.idm.oclc.org/10.1177/016224399702200105 , , & (
- 2017). F.D.A. approves first gene‐altering leukemia treatment, costing $475,000. The New York Times. Retrieved from https://www.nytimes.com/2017/08/30/health/gene-therapy-cancer.html?mcubz=3&_r=0 (
- 2011). Mapping the moral domain. Journal of Personality and Social Psychology, 101(2), 366– 385. https://doi-org.uoelibrary.idm.oclc.org/10.1037/a0021847 , , , , , & (
- 1994). Individual differences in sensitivity to disgust: A scale sampling seven domains of disgust elicitors. Personality and Individual Differences, 16(5), 701– 713. https://doi-org.uoelibrary.idm.oclc.org/10.1016/0191-8869(94)90212-7 , , & (
- 2014). A four‐party view of US environmental concern. Environmental Politics, 1– 16. https://doi-org.uoelibrary.idm.oclc.org/10.1080/09644016.2014.976485 , & (
- 2006). Tampering with nature: ‘Nature’ and the ‘natural’ in media coverage of genetics and biotechnology. Media, Culture & Society, 28(6), 811– 834. https://doi-org.uoelibrary.idm.oclc.org/10.1177/0163443706067026 (
- 2017). Climate engineering and the playing God critique. Ethics & International Affairs, 31(3), 313– 333. https://doi-org.uoelibrary.idm.oclc.org/10.1017/S0892679417000223 (
- 2017). Will democracy survive big data and artificial intelligence? Scientific American. Retrieved from https://www.scientificamerican.com/article/will-democracy-survive-big-data-and-artificial-intelligence/ , , , , , , … (
- 2006). Public attitudes toward genetic testing: Perceived benefits and objections. Genetic Testing, 10(2), 139– 145. https://doi-org.uoelibrary.idm.oclc.org/10.1089/gte.2006.10.139 , , & (
- 2017). Attitude roots and Jiu Jitsu persuasion: Understanding and overcoming the motivated rejection of science. American Psychologist, 72(5), 459– 473. https://doi-org.uoelibrary.idm.oclc.org/10.1037/a0040437 , & (
- 2011). Structural equation modeling for social and personality psychology. Thousand Oaks, CA: SAGE Publications, Inc. (
- 2009). Disgust sensitivity predicts intuitive disapproval of gays. Emotion, 9(3), 435– 439. https://doi-org.uoelibrary.idm.oclc.org/10.1037/a0015960 , , , & (
- 2009). Question and questionnaire design. In J. D. Wright& P. V. Marsden (Eds.), Handbook of survey research ( 2nd ed., pp. 263– 314). San Diego, CA: Elsevier. , & (
- 2012). Why do people like natural? Instrumental and ideational bases for the naturalness preference. Journal of Applied Social Psychology, 42(12), 2859– 2878. https://doi-org.uoelibrary.idm.oclc.org/10.1111/j.1559-1816.2012.00964.x , & (
- 2013). Playing God and the intrinsic value of life: Moral problems for synthetic biology? Science and Engineering Ethics, 19(2), 435– 448. https://doi-org.uoelibrary.idm.oclc.org/10.1007/s11948-012-9353-z (
- 2004). The Connectedness to Nature Scale: A measure of individuals’ feeling in community with nature. Journal of Environmental Psychology, 24(4), 503– 515. https://doi-org.uoelibrary.idm.oclc.org/10.1016/j.jenvp.2004.10.001 , & (
- 2013). Applied multivariate research ( 2nd ed.). Thousand Oaks, CA: Sage Publications. , , & (
- 2014). I just don’t trust them: The development and validation of an assessment instrument to measure trust in science and scientists. School Science and Mathematics, 114(2), 76– 86. https://doi-org.uoelibrary.idm.oclc.org/10.1111/ssm.12051 , & (
- 2016). Validation of the long‐ and short‐form of the Ethical Values Assessment (EVA). International Journal of Behavioral Development, 40(2), 181– 192. https://doi-org.uoelibrary.idm.oclc.org/10.1177/0165025415587534 , & (
- 2013). The risk society revisited: Social theory and governance. Philadelphia, PA: Temple University Press. , , & (
- 2005). The meaning of “natural”: Process more important than content. Psychological Science, 16(8), 652– 658. https://doi-org.uoelibrary.idm.oclc.org/10.1111/j.1467-9280.2005.01589.x (
- 2012). European and American perspectives on the meaning of natural. Appetite, 59(2), 448– 455. https://doi-org.uoelibrary.idm.oclc.org/10.1016/j.appet.2012.06.001 , , & (
- 1999). The CAD triad hypothesis: A mapping between three moral emotions (contempt, anger, disgust) and three moral codes (community, autonomy, divinity). Journal of Personality and Social Psychology, 76(4), 574– 586. https://doi-org.uoelibrary.idm.oclc.org/10.1037/0022-35126.96.36.1994 , , , & (
- 2004). Preference for natural: Instrumental and ideational/moral motivations, and the contrast between foods and medicines. Appetite, 43(2), 147– 154. https://doi-org.uoelibrary.idm.oclc.org/10.1016/j.appet.2004.03.005 , , , , , , & (
- 2011). Would you rather be injured by lightning or a downed power line? Preference for natural hazards. Judgment and Decision Making, 6(4), 314– 322. Retrieved from http://www.sas.upenn.edu/~baron/journal/11/10809/jdm10809.pdf , , , & (
- 2002). Inclusion with nature: The psychology of human‐nature relations. Psychology of Sustainable Development, 61– 78. https://doi-org.uoelibrary.idm.oclc.org/10.1007/978-1-4615-0995-0_4 (
- 1977). Normative influences on altruism. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 10, pp. 221– 279). New York, NY: Academic Press. (
- 2016). Evidence for absolute moral opposition to genetically modified food in the United States. Perspectives on Psychological Science, 11(3), 315– 324. https://doi-org.uoelibrary.idm.oclc.org/10.1177/1745691615621275 , , & (
- 2000). The influence of trust and perceptions of risks and benefits on the acceptance of gene technology. Risk Analysis, 20(2), 195– 203. https://doi-org.uoelibrary.idm.oclc.org/10.1111/0272-4332.202020 (
- 2007). Public acceptance of nanotechnology foods and food packaging: The influence of affect and trust. Appetite, 49(2), 459– 466. https://doi-org.uoelibrary.idm.oclc.org/10.1016/j.appet.2007.03.002 , , , & (
- 2014). Human and nature‐caused hazards: The affect heuristic causes biased decisions. Risk Analysis, 34(8), 1482– 1494. https://doi-org.uoelibrary.idm.oclc.org/10.1111/risa.12179 , & (
- 2000). Perceived risk and tampering with nature. Journal of Risk Research, 3(4), 353– 367. https://doi-org.uoelibrary.idm.oclc.org/10.1080/13669870050132568 (
- 1987). Perception of risk. Science, 236(4799), 280– 285. https://doi-org.uoelibrary.idm.oclc.org/10.1126/science.3563507 (
- 2014). Pope Francis denounces euthanasia as “sin against God.” The Telegraph. Retrieved from http://www.telegraph.co.uk/news/worldnews/the-pope/11233347/Pope-Francis-denounces-euthanasia-as-sin-against-God.html (
- 1980). Tests for comparing elements of a correlation matrix. Psychological Bulletin, 87(2), 245– 251. https://doi-org.uoelibrary.idm.oclc.org/10.1037/0033-2909.87.2.245 (
- 1998). A brief inventory of values. Educational and Psychological Measurement, 58(6), 984– 1001. https://doi-org.uoelibrary.idm.oclc.org/10.1177/0013164498058006008 , , & (
- 2001). Using multivariate statistics ( 4th ed.). Boston, MA: Allyn and Bacon. , & (
- 2005). Perceived naturalness and acceptance of genetically modified food. Appetite, 45(1), 47– 50. https://doi-org.uoelibrary.idm.oclc.org/10.1016/j.appet.2005.03.004 , , , & (
- 2010). The morality of attitudes toward nanotechnology: About God, techno‐scientific progress, and interfering with nature. Journal of Nanoparticle Research, 12(2), 373– 381. https://doi-org.uoelibrary.idm.oclc.org/10.1007/s11051-009-9809-5 , , , , & (
- 2019). Aversion to playing God scale and moral condemnation of technology and science. Philosophical Transactions of the Royal Society B. https://doi-org.uoelibrary.idm.oclc.org/10.1098/rstb.2018.0041. , & (
- 2019). Public support for carbon dioxide removal strategies: The role of tampering with nature perceptions. Climatic Change, 152(3–4), 345– 361. https://doi-org.uoelibrary.idm.oclc.org/10.1007/s10584-019-02375-z , , , & (