Role of lighting and color in microeconomics: preference and purchase intent

Lighting and color are key factors under the control of lighting designers that can significantly impact shoppers’ perceptions and actions in the context of culinary and retail environments. Previous studies have examined the effects of color and lighting on preference and purchase intent, but results are often conflicting due to confounding variables, and a strong connection between these phenomena remains unexplored. To bridge this gap, two visual experiments were conducted, revealing that lighting can indeed influence consumers’ perception of products. Participants perceived the changes in objects under different color gamut and illuminance levels. Further research is needed to better understand the mechanisms of purchase intent, particularly across different price and product types.


Introduction
In microeconomics, demand refers to the quantity of a good or service that consumers are willing and able to purchase at a particular price and time, while purchase is the act of buying a good or service at a specific point in time.Purchase intent is affected by various factors, such as consumer preferences, income, prices of related goods, advertising, and physical factors in a retails store.Conceptual frameworks of purchase intent indicate seven key factors: cleanliness, music, scent, temperature, display, lighting, and color [1].Among these seven, lighting and color are aspects that a lighting designer can influence through the rigorous planning and use of electric lighting systems.

Culinary experience and lighting
The relationship between lighting and food-related has been a focus of consumer behavior and sensory experience studies.For example, Wu et al. investigated the effect of focal vs general, and dim vs bright lighting on customers' intentions toward restaurants using images [10].Images that present focal lighting (e.g., table lighting) resulted in a boost of the restaurant's perceived attractiveness and visit intention among diners.Steele et al. explore the shelf life of fresh meat products under different lamps (LED vs. fluorescent), and found that LED lighting extends beef longissimus dorsi, semimembranosus steak and ground beef color life by up to one day [11].The study also noted that LED lighting has minimal effects on factors that cause meat color to deteriorate.Unfortunately, in these studies, lighting conditions (e.g., IOP Publishing doi:10.1088/1755-1315/1320/1/012024 2 spatial and spectral distribution of lighting) were not characterized in detail, preventing the replication of these studies.
Other research papers that focus on the effects of lighting on food perception provided better control of experimental conditions.Utilizing lighting conditions that are less-commonly found in architectural spaces, Hasenbeck et al. found that participants' willingness to consume bell peppers increased the most under yellow lighting and the least under blue, monochromatic lighting [12].Similarly, Özkul et al. investigate the effect of chromatic lighting (red, blue, green, yellow) on customers' perceptions of service quality and satisfaction in restaurants [13].Their study highlights that the perception of service quality and the level of satisfaction were higher in red and yellow ambient lighting compared to blue and green ambient lighting.
Several studies also utilized nominal white light sources that are common to architectural spaces, where consumers interact with food (e.g., restaurants, supermarkets).A market analysis of lighting conditions in Slovakia for baked goods and fresh produce found that light levels and CCTs widely varied across retail chains and for different product categories (i.e., between 2530 K and 4450 K; between 200 lx and 1800 lx) [14].A study by Wu and Wang used nominal white lighting to investigate the effects of CCT and illuminance levels in restaurants on customers' emotional states and spatial impressions [15].Their study underscores that in a restaurant 2700 K light sources have a greater tendency to "emphasize pleasurable aspects" while 5600 K causes greater "arousal."Chen et al. studies the effect of CCT (2700 K, 3300 K, 3900 K, 4500 K and 5100 K), illuminance (400 lx, 800 lx and 1200 lx), and background (wood vs. white) on baked goods, such as bread and cakes [16].High illuminance levels promoted the purchase intention for both bread and cakes, while CCT didn't impact results.More recently, Ahmadian et al. studied consumers' preference of vegetables, fruits, and packaged foods under 3000 K at 500 lx to apply the stimulus-organism-response (SOR) [17] framework in the context of retail lighting [18].The authors created an optimal spectrum for most objects except the can of Pepsi and red cabbage, providing support for the potential for an object-oriented illumination strategy.A tailored-spectra strategy supports previous studies that found similar results [19][20][21].

Retail experience and lighting
Understanding how lighting design impacts customer behavior in retail settings is a central theme explored by several research groups around the world.Lombana and Tonello investigated the perceptual and emotional effects of light distribution and object color in retail spaces in Argentina [22].Their results indicated that overhead direction of lighting improves the appearance of the less favored color conditions.Similarly, Tantanatewin and Inkarojrit studied the effects of chromaticity (warm vs. cool) and light distribution (general vs. accent) on impression and identity using digital images of a bank in a department store in Thailand [23].Warm color tones significantly enhanced positive impression and space identity, but lighting arrangement was a secondary factor.In a field study conducted in the United States, the lighting (soft vs. bright) in a centrally located retail establishment was varied over a twomonth period, and the results indicated that brighter lighting influenced shoppers to examine and handle more merchandise, though sales were not influenced [24].
In a more controlled study conducted in China, color preference and discrimination of blue jeans were the highest at 5000 K [25] Hemalatha et al. extend this discussion to retail apparel stores, examining the effects of lighting conditions (2700 K, 3000 K, 4000 K and 5700 K; 300 lx, 500 lx, 700 lx and 900 lx) on user preferences within the cultural context of India [26].Both CCT and illuminance influence the spatial impressions of a retail environment, with the most preferred scenes being 5700 K at 500 lx, 5700 K at 300 lx and 5700 K at 700 lx.Meanwhile, Park and Farr explore the impact of lighting on American and Korean consumers' emotions and behavioral intentions in retail environments under 3000 K and 5000 K, and color rendering index (CRI) of 75 vs 95 lighting [27].The results indicate that 3000 K was more pleasurable, while Koreans found 5000 K to be more "approachable" compared to Americans, highlighting cultural differences across lighting conditions.This body of work underscores the importance of lighting design in creating attractive retail environments and influencing consumers' preferences and behavior.

Customer experience and lighting
Understanding how lighting influences customer experience (product perception, price fairness, waiting times, purchase intention) in different settings is another key theme in the literature.Hutchison, Thomas, and Elias provide intriguing insights into the world of advertising by investigating the effects of leftward lighting bias in advertisements [28].Their study reveals that the direction of lighting in advertisements can significantly influence advertisement ratings and purchase intentions, confirming the positive influence of illuminating the objects from the left.Another study exploring the effect of interior colors (blue vs. orange) and lighting (bright vs soft) on price fairness and purchase intent found that blue interiors were associated with more favourable evaluations, higher store patronage intentions, and higher purchase intentions than orange interiors [29].However, "soft lights" with an orange interior generally nullified the negative effects and produced the highest level of perceived price fairness.Bilgili, Ozkul, and Koc explored the influence of light chromaticity (red, green, blue, yellow, white) on customers' waiting time perceptions, and found a significant impact of lighting on perception of the length of the waiting time, specifically, under green lighting, customers perceived the waiting period to be relatively shorter [30].Biswas et al. conducted a series of field studies at multiple locations and found that consumers tend to choose less healthy food options when ambient lighting is dim (vs.bright) and connect it to lower mental alertness caused by low light levels [31].
Similar to some of the previously discussed studies, the independent variable (lighting conditions) was not thoroughly reported in these studies.In a better controlled study, Phumchan and Tuaycharoen explored the effects of illuminance (between 100 lx and 900 lx) and CCT (2700 K, 4000 K, and 65000 K) on the emotional state and perception of customers in clothing retail stores [32].Results indicated that 4000 K lighting and an illuminance of 400 lx were optimal for the emotional state and perception of participants, including price perception and motivation.Despite their limitations, these studies collectively underscore the nuanced relationship between lighting design and customer experience, emphasizing the importance of lighting in shaping how consumers perceive and evaluate products.

Research questions
Despite the vast literature investigating the effect of lighting and color on subjective evaluations of the environment, the results of these studies are yet to be translated to concrete purchase intent and action.The outcomes of the research studies on purchase intent are at times inconclusive and challenging to compare.For example, Tešić et al. found that CCT can affect the shoppers' perception of the quality of the observed products in the Serbian market, but do not affect the price perception [33].On the other hand, Liu et al. documented that consumers' purchase intentions followed an inverted "U" curve under the change of luminance and hue (i.e., consumers' purchase intention was stronger under medium luminance and low CCT) [34].
One of the potential explanations for the conflicting results is the lack of control for confounding variables in the experimental protocols.For example, studies often do not acknowledge or account for the limitations of CCT, which is a proxy metric for the chromaticity of light source and do not provide any information about objects or the architectural environment [35].A complimentary metric Duv (the distance from the Planckian locus) [36] can provide an additional depth to light source characterization.When controlled for CCT and Duv, gamut shape and color shifts in red hue accounts for the visual perception of objects, even more so than color fidelity metrics (i.e., CRI Ra, TM-30 Rf) [37,38].This understanding leads to two research questions: 1) What is the relationship between preference and purchase intent when lighting is controlled for CCT and Duv? 2) What is the effect of lighting on purchase intent and visual quality of objects in a retail environment under more tightly controlled experimental protocols?Two pilot studies were conducted to close the gap between color preference and purchase intent by addressing these research questions.

Methods
Aiming to investigate the influence of lighting on preference, naturalness, vividness, and purchase intent, two psychophysical studies were conducted.These studies examined the impact of illuminance and color gamut on attractiveness, saliency, and purchasing intent in different contexts.Controlled experiments exposed human participants to various color gamut (ANSI/IES TM-30 Rg = 85, 90, 100, 115, 120) and illuminance levels (50 lx, 400 lx, 500 lx, 750 lx).The first experiment investigates the effect of illuminance and color gamut on preference and gaze duration.The second study explores the effect of illuminance and color gamut on purchase intent.A verbal informed consent was collected before each study.Participants were checked for normal color vision.

Study 1
A vision experiment was conducted by using eye-tracking technology to detect gaze distribution on four natural objects (pear, green apple, orange, red apple).Illuminance levels and tunable LED spectra were controlled using MATLAB ® .Six lighting conditions were shown to six American participants (3 males, 3 females) three times (18 trials per participant) in a randomized order, as shown in Fig. 1 (left).The lighting conditions varied in terms of illuminance (nominally 50 lx and 400 lx), color fidelity (nominally ANSI/IES TM-30 Rf = 70, 80, 100), and color gamut (nominally ANSI/IES TM-30 Rg = 85, 100, 115), as shown in Table 1 and Fig. 1.The conditions were generated to represent desaturating, neutral, and saturating gamut at two light levels.Participants were asked to judge the preference and (un)naturalness of the objects.Each trial lasted 20 seconds, followed by 2 seconds of darkness.The experiment duration was 10 minutes in total.Eye tracking data was collected using Pupil Labs Invisible eyeglasses, which has been used in similar studies [39,40].

Study 2
A second experiment was conducted to evaluate the impact of illuminance levels, color gamut size and shape in a miniature model of a retail space.Six lighting conditions were generated using a Rockville linear LED fixture that has five LED channels mounted on a small-scale model of a retail store.The lighting conditions varied in terms of color gamut (nominally ANSI/IES TM-30 Rg = 90, 100, 120) and illuminance (nominally 500 lx and 750 lx) to meet the IES retail lighting guidelines [41], as shown in Table 2 and Fig. 3.The trials were randomized using the William Latin square design [42].Each lighting condition was repeated three times and lasted for 45 seconds.Between trials there was a two second dark transition period to reduce afterimage effects [43].The experiment duration was 14 minutes in total.Thirteen American participants (4 males, 9 females) were asked to judge the preference and unnaturalness of the objects, as shown in Fig. 2

Study 1
Results from study 1 suggest that light source spectrum (color gamut) can impact subjective evaluations of fresh produce more than illuminance, as shown in Fig. 4.Under desaturating light source, the most preferred produce was orange likely due to the already saturated nature of the fruit.The opposite was also true where orange was the least preferred fruit under saturating lighting confitions (i.e., increased saturation made the orange look cartoonish, thus preference reduced in this lighting condition).Green and red apples under saturating lighting conditions of 400 lx were the second most preferred produce.On average, pear was the least preferred produce.While the color appearance can make an impact on preference evaluations, it is important to acknowledge other factors that can influence participants' decisions (e.g., the taste of the fruit).
Unnaturalness responses supported preference results, where the fruit orange under saturated 400 lx lighting condition was perceived to be the most unnatural fresh produce.Red apple in desaturating lighting conditions followed orange in terms of naturalness.Green apple was the least unnatural on average, indicating the robustness of near-neutral (originally less saturated) objects.
Gaze data analysis was limited due to technical difficulties (i.e., sensitivity of surface creation in video recordings).However, the limited information from the gathered data indicated that participants gazed at unnatural or preferred objects most of the time, as shown in Fig. 5.

Study 2
The second study results indicate an increase in willingness to purchase as the lighting level and color gamut increased, as shown in Fig. 6.The higher light level resulted in a higher purchase intention at the same color gamut.The difference between Rg values at 750 for purchase intent was significant (p = 0.003, and η2 = 0.084).The lowest purchase intent was for desaturating spectra at 500 lx.There was also a significant difference between light levels.However, the effect of illuminance should be taken with a grain of salt due to the limited number of illuminance ranges that may cause range bias [43,44].

Conclusions
In the realm of microeconomics, the physical properties of the environment can significantly impact observers' actions, particularly their intent to purchase goods.Lighting plays a pivotal role in influencing the perceived quality of objects, with illuminance and color being the primary stimuli involved.Therefore, lighting designers hold direct control and responsibility over these key factors, enabling them to shape shoppers' visual perceptions of goods.Although past studies have separately explored the influence of color gamut on preference and purchase intent, a strong connection between these two related phenomena has not been established due to a lack of controlled conditions.This research aims to bridge this gap and establish a robust link between color gamut, preference, and purchase intent, shedding light on the intricate interplay between lighting, perception, and consumer behavior.Visual experiments were conducted to examine the effect of lighting on purchase intent.The findings indicate that lighting has a significant influence on consumers' perception of products.Participants in these studies demonstrated the ability to discern changes in color gamut and illuminance, highlighting the sensitivity of their visual perception to lighting conditions.However, it is important to note the limitations of the studies, including a small sample size, a restricted range of illuminance levels, and a limited visual field of view in the second experiment.To gain a more comprehensive understanding of the mechanisms underlying purchasing intent, further research is warranted, particularly exploring different price ranges and product types such as fresh produce versus manufactured goods.Further research in this field will enable a deeper exploration of the complex relationship between lighting, consumer perception, and purchase behavior.

Figure 2 .
Figure 2. (Left: Study 1) Eye tracking and subjective judgments of fresh produce.(Right: Study 2) Purchase intent and subjective judgments of model fruits shown in a miniature retail setting.

Figure 4 .
Figure 4. Participants' choices of the fresh produce under desaturating (top), neutral (middle), and saturating (bottom) spectra for 50 lx (light blue columns on the left) and 400 lx (dark blue columns on right).

Figure 5 .
Figure 5. Participant SP's gaze counts were the highest for the most unnatural object (pear) followed by the most preferred object (green apple) under 50 lx Rf = 100 and Rg = 100 (neutral condition).The error bars show the interquartile ranges.

Table 1 .
The lighting conditions used in Study 1

Table 2 .
The lighting conditions used in Study 2