Keywords

1 Introduction

Much has been written about how 3D printing technology is set to disrupt, if not transform, economies and manufacturing [35]. It has opened up possibilities of a shift in both production methods and consumption patterns. In terms of production, artifacts can now be made using processes and materials that are starkly different from those used for mass- manufactured products. In terms of consumption, consumers can now potentially participate in the production of usable objects – “prosumers” making, creating, and innovating using affordable 3D printers [38].

Fig. 1.
figure 1

Materials and objects used in the study (from left, PLA, SLA resin and mass-manufactured plastic)

Full size image

The bulk of the research literature on the subject so far has focused on types of applications of 3D printing (e.g., [36, 39, 46]), innovations in 3D printing technologies and materials (e.g., [28, 43]), and disruption of business models (e.g., [37]). Some research has addressed who the users of 3D printing are, and why and how they engage in the practice (e.g., [20]). However, little is understood on how people engage with 3D-printed objects, or in other words, on the outcomes of 3D printing. This paper investigates people’s perception of 3D-printed objects in terms of objective value and subjective meaning through a within-subjects lab-based study with 22 participants.

The success of mass-manufactured products is such that it has established certain expectations about objects for use. These expectations are evident in, for example, literature comparing consumer perceptions of mass-manufactured and handmade products (e.g., mass-manufactured products are perceived as less authentic [17]). 3D-printed objects are neither mass-manufactured, nor handmade. Rather, they are customizable like handmade objects but made through less intimate and perhaps production-like methods as in mass-manufacturing. Thus 3D-printed objects have unique qualities that remain to be investigated. Cost and quality level are unlikely to be the only limiting factors in the adoption of 3D printing and use of 3D-printed objects in regular, everyday scenarios for meaningful purposes. Like with other technologies, attitudes and perceptions are bound to have significant influences on meaning. We propose that 3D printing technologies today have reached a level of development that warrants investigations to understand user perception of 3D-printed objects. Such investigations can provide useful (and necessary) knowledge to inform directions for further development of different facets of 3D printing technologies, especially 3D design interfaces and support software.

Following, we first review the literature on 3D printing and the importance of materiality to situate our work within the broader context of the material turn in HCI. We then describe our study design, protocol, and results. We conclude with a discussion of the significance of our study results for future research on 3D printing design software, as well as more generally for understanding materiality in HCI.

2 Background: An Overview of 3D Printing

3D printing typically refers to additive manufacturing techniques that create objects by depositing materials layer by layer [9]. Its ability to deliver one-of-a-kind objects [32] to people with relatively low expertise in design and manufacturing has ignited the public’s imagination in recent years [49]. Beaman [5] argues that 3D printing has blurred the roles between designer and consumer, subverting the usual ‘design for manufacturing’ model to become closer to ‘manufacturing for design’.

A wide variety of materials can be utilized in 3D printing, but the choice of materials is contingent on the type of production process desired [51]. The most common production method is Fused Deposition Modeling (FDM). FDM melts filament-formed material into a heated nozzle to extrude the 3D part layer by layer [19]. Polylactic Acid (PLA) is the most easily accessible and commonly used material for FDM. PLA can provide a good level of detail, but has high stiffness and a somewhat rough texture. Stereolithography (SLA) is another production method that entails a layered manufacturing process whereby the material is cured via projecting ultra violet laser beams to turn liquid resin into a solid [42]. SLA employs photopolymer resin as material, which when cured has fine features, high stiffness, and a smoother surface finish.

Initial 3D printing technologies in the early 1990s produced objects that were visibly of lower quality, and were thus used almost exclusively for rapid prototyping of single discardable items [37]. Yet, even with tremendous advances in these technologies, the use of 3D printing for prototyping is still dominant today while final production applications (ready-for-consumption products) are much scarcer [26]. This may suggest that there are barriers to the full-fledged adoption of resultant objects from 3D printing for common and casual use. The typical production process of an object involves three key stages: design, manufacturing, and distribution [38]. Research in 3D printing has coalesced around the first two stages, with HCI addressing the use of 3D design software (e.g., [18, 21]), and literature from engineering addressing manufacturing systems and methods heavily [7, 31]. Research on the distribution and delivery stage of 3D printing can be found mostly in the discipline of business and management, and often addresses models of supply chain (e.g., [6, 8, 53]). However, for 3D printing to be effective for final production applications, it is imperative for insights to be gained as well on the consumption aspect of 3D printing. These insights can inform the prior stages of the production cycle.

3 Related Work: Consumption of 3D-Printed Objects

The most common class of users of 3D printing currently are specialized industry-based users and amateur Makers, both of whom have a relatively high level of knowledge of design and manufacturing processes. We are interested in potential users who have been called ‘casual makers’ [20] or ‘everyday designers’ [27, 47]. Baudisch et al. [4] argued that personal fabrication, “the ability to design and produce your own products, in your own home, with a machine that combines consumer electronics with industrial tools” [13], is not out of the question, although there are many barriers to overcome for it to become pervasive across societies.

We are not aware of any work to date that has directly investigated people’s perception of 3D-printed objects. The literature on the consumption aspect of 3D printing tends to revolve around how and why casual users may decide to 3D print an object, which is still relevant for our investigation. We review this literature focusing on factors that have been identified as being significant in 3D printing design and the resultant objects. Further, we also describe some work analyzing attitudes towards handmade objects as it may help to contextualize our work on 3D-printed objects.

Shewbridge et al. [40] investigated what people may want to design if they had a 3D printer at home. Their findings showed that material type and functionality are important dimensions that people consider for objects that they would like to print (e.g., wooden parts, metal objects). The authors also highlight that these potential everyday users of 3D printers lack the needed 3D design and modelling skills.

Lee et al. [27] studied the process that people use to decide what they would want to 3D-print in terms of furniture for their homes. They found that the design requirements of users for furniture include aesthetics (namely color, style and size), ergonomics (e.g., comfortability of the material) and functionality. They suggest that users need support to understand interdependencies of design decisions within printed objects.

Based on a large-scale survey of peer production communities, Moilanen and Vadén [31] found that the top five wanted features for 3D printing relate to the physical aspect of the process, specifically ‘object quality’, ‘speed’, and ‘cheaper material prices’. The most critical bottleneck that users reported seeing was ‘materials and quality’.

Overall, the literature on 3D printing motivations for casual users seems to indicate that materiality, with emphasis on certain properties, is an essential aspect of 3D-printed objects. With respect to literature on handmade objects, values seem to be more positive for this class of object as compared to mass-manufactured products. After conducting four studies investigating various production modes on perceived product attractiveness for items presented as handmade vs. machine-made, Fuchs et al. [11] found a positive effect on attractiveness for handmade products. Participants perceived products described as handmade more positively. Further, participants preferred objects if they were marketed as handmade and were willing to pay more for them. Groves [17] compared people’s perceptions of home-made food items with mass-manufactured food. She found that home-made food was seen as being more authentic. Higher prices for an item also led people to perceive the item to be of higher quality and more authentic, suggesting that consumers have high expectations for home-made food.

4 Theoretical Foundation: Materiality in HCI

To provide a basis for understanding 3D-printed objects, we adopt the material perspective in HCI. Research in HCI has traditionally emphasized use and functions in the study of objects. Spurred particularly by the increase in digital interfaces, researchers have called for a return to the study of form, materials and materiality. We describe below a few prominent perspectives in that paradigm.

Wiberg [50] suggested using a “material lens” to explore interaction design from the perspective of materials as the basic components of design. He operationalized this “material lens” into a methodology that consists of analyzing designs at four levels: materials, details, texture, and wholeness. At the level of materials, the properties, character, potential and limitations of materials are analyzed. At the level of details, attention is paid to aesthetics and quality of the object. At the level of texture, focus is on appearance and authenticity, with authenticity being defined as the true relationship between materials, material composition, and appearance. And finally, at the level of wholeness, all the different aspects of the object are brought into a composite so as to enable the study of object meaning to an observer.

Jung and Stolterman [24] conceptualized the study of artifacts in HCI as encompassing form and materiality. They suggest three perspectives on form: material, shape, and making (production method). Two perspectives are relevant for materiality: meaning (how objects are understood in personal and social life), and material ecology (connections among multiple artifacts in use).

Fuchsberger et al. [12] applied McLuhan’s theories to the study of materiality in HCI. McLuhan’s proposition that the medium is the message highlights the distinction between sensory impressions and sensory effects (seeing a comic illustration of a bang and actually feeling like you hear a bang). Thus McLuhan ascertained that the medium affects more senses than just the mode of presentation. Similarly, materials may have greater effects than what their properties suggest.

Giaccardi and Karana [14] proposed the ‘materials experience’ framework as a way of understanding materiality in HCI. They ascertain that materials are experienced only through a dynamic relationship among materials, people, and practices. And thus, the experiential qualities of materials are dependent on factors such as the properties of a material, the specific artifact in which it is embedded, the user’s previous experiences and expectations, and social and cultural values.

They propose a framework consisting of four experiential levels: sensorial, interpretive, affective, and performative. At the sensorial level, materials impact the basic human sensory system of touch, vision, smell, sound and taste. At the interpretive level, people interpret materials and attribute situated meanings. At the affective level, the qualities of the material trigger specific emotions, even if unconsciously so. And at the performative level, all levels of experience from perceptions, meanings and affects are combined to create a certain usage and behavioral pattern towards the object.

Last but not least, the notion of affordances needs to be reviewed in a discussion on materiality. Gibson [15] defined object affordances as preconditions for activity, or possibilities in the environment that spur certain kinds of interactions. Affordances are properties of the object, present in the world irrespective of users and their presence, and independent from influences such as experiences and culture [16]. Applying Gibson’s idea to HCI, Norman [33] put forth that affordances provide critical clues to a user for the operation of an object. However, he also offered a fundamentally different perspective from Gibson in that there are two kinds of affordances: real affordances are the actual properties of objects such as physical form, material, and character. Perceived affordances are what a user perceives an object about an object in terms of what it is and how to use it. Perceived affordances are dependent on one’s prior experiences and culture.

Taken together, the literature on materiality in HCI informs us of the following: (i) There are clearly two levels that need to be taken into account in the study of materiality: an object’s material properties that are more or less objective, and a more holistic understanding of the object (Wiberg’s wholeness, Jung and Stolterman’s meaning, Fuschberger’s sensory effects, Giaccarda’s interpretive); and (ii) Previous and current factors affect a user’s understanding of the object: the object’s production process, other surrounding objects, the specific situation of use, the user’s practices, and social and cultural values.

5 Study Description

5.1 Research Questions

The overarching goal of our study was to investigate how 3D-printed objects are perceived as compared to their mass-manufactured counterparts. As per the insights gained from our literature review on materiality, we examined the issue at two levels, first, at level of the basic material properties and second, at the more holistic level of object meaning or interpretation. The specific research questions were as follows:

RQ 1: a. Are there differences in how people perceive the material properties of 3D-printed objects as opposed to those of mass-manufactured objects?; b. If so, what are these differences in perception?

RQ 2: a. Do people interpret 3D-printed objects differently than mass-manufactured objects?; b. If so, what are the differences in interpretation?

5.2 Study Materials

To enable the conduct of the study, two types of objects were prepared as probes: Object A was a small white cube and object B was a white round clothing button. The objects were chosen based on the following rationale: Object A was a simple geometric object, and was thus assumed to have no obvious functionality and minimal cultural biases. Object B had a rather evident function as a fastener as per their common use in many cultures.

Each of the two objects were 3D-printed with two types of materials: (i) PLA; and (ii) SLA photopolymer resin. Mass-produced versions of each plastic object were also used in the study. In total 6 objects, shown in Fig. 1, were used as probes (2 objects X 3 materials). Mass-manufactured objects were first bought off-the-shelf, and 3D digital model of the mass-manufactured objects were subsequently created. The 3D models were then 3D printed using the SLA printing method and the FDM method using 100% infill and the highest quality printer settings. The Formlabs Form 2 printer was employed for the SLA resin objects, and the Ultimaker 3 printer with a 0.4 mm nozzle was employed to print the PLA objects. The PLA and resin objects were 3D-printed to match, as exactly as possible, the mass-produced objects in terms of color, shape, and size. The white color was chosen in order to avoid judgment based on specific colors. All of the 3D-printed objects were slightly sanded by hand after printing to remove rough edges while preserving the initial look and feel. The final study objects are shown in Fig. 1.

5.3 Study Design and Protocol

Our study had a within-subjects design with two independent variables, object and type of material. Object had 2 levels: cube, where participants engaged with the cube objects, and button, where participants engaged with the button objects. Material type had 3 levels: PLA where participants engaged with objects 3D-printed using PLA, SLA resin, where participants engaged with objects 3D-printed using the SLA method, and mass-manufactured plastic, where participants engaged with objects bought off-the-shelf.

The study involved 22 participants (13 males and 9 females with mean age = 30.2). All participants had no prior experience with 3D printing technology. Their demographics were as follows: White 45.5% (n = 10); Asian 22.7% (n = 5); Hispanic 18.2% (n = 4); Black 4.5% (n = 1); Native American 4.5% (n = 1); and Mixed Races 4.5% (n = 1). Recruitment was made through university listserv email announcements. Participants were individually contacted on a first-come, first-served basis for study sessions at mutually agreed times. Participants were asked to come to an on-campus lab for one session which lasted approximately 1 h and 30 min. A summary of the study protocol is shown in Table 1.

Table 1. Study protocol
Full size table

At the beginning of the session, participants were given a consent form to sign, were briefly introduced to the study, and were asked to fill in a pre-questionnaire that asked demographic information and prior knowledge in 3D printing or 3D tools. Then, participants were handed the 6 objects (see Study Materials section), one at a time. After the participant had examined and interacted with an object, he or she was engaged in a short interview and asked to fill in a questionnaire. The contents of the interview and question are described in the “Measures” section below. The order of presentation of the objects were randomized for each participant. After a participant had reviewed all the objects, he or she was asked to fill in a final questionnaire.

6 Measures and Data Analysis

Data about user perception was collected and analyzed both quantitatively and qualitatively. For RQ1 relating to object material properties, assessment was done by asking participants to fill a repertory grid after having reviewed all the objects. The repertory grid is a technique used to elicit a participant’s personal constructs without the biases of the researchers [45]. Although typically an interview-based technique, it has also been used in a questionnaire form [52].

Participants are asked to list adjectives or phrases to compare and contrast at least 3 objects. They are then asked to list the opposites of the descriptive terms. For example, if a participant used the term “smooth” to describe one object, he or she may put down the bipolar opposite of the term to be “rough”, thereby creating a “smooth-rough” bipolar construct. And finally they are asked to either assign a value or ranking for each object based on each construct. In our study, participants assigned rankings to the six objects for each construct that they came up with. We note here that the repertory grid could also capture aspects of RQ2 on the object meaning since it was open-ended in nature.

For RQ2 relating to the meaning or interpretation of the objects, assessment was done in four ways: (i) Participants were asked to bid for each object on a scale of 1 to 100 cents (“how much would you pay for this object out of 100 cents?”). Such a simulated bidding process has been used successfully by others to assess object value [34]. (ii) Participants were asked to handwritten notes about each object under three given hypothetical scenarios – first, giving the object as a gift to someone close; second, receiving the object from someone close; and third, presenting the object as an artifact in a museum – so that we could capture the diversity of possible interpretations; and (iii) Participants were asked to complete the Perceived Value Scale (PERVAL) by [44] that measures the following constructs using 5-point likert scales (example items are given in brackets) – “quality/performance” (e.g., Is well made); “price/value for money” (e.g., Is reasonably priced); “social value” (e.g., Would make a good impression on other people); and “emotional value” (e.g., Is one that I would enjoy).

For the ‘price/value for money’ construct, the actual (purchase/production) cost of each object at the time of the study was given (cube: $1.73 mass-manufactured, $4.93 PLA-printed, $14.81 SLA-printed; button: $0.49 mass-manu factured, $2.50 PLA-printed, $2.50 SLA-printed), and the participant was asked to indicate how much he/she agreed with the price. Prices for the 3D printed objects were the price paid in a University 3D Printing Shop, where students could request a 3D design be printed on a variety of printers and materials for a cost that scales with the duration of the print and quantity of material consumed. As the buttons required less time and less material to print, they cost the minimum charge of the shop ($2.50). (iv) And finally, participants were asked to fill the PANAS-X scale [2, 48] that measures one’s distinguishable affective reactions. The PANAS-X provides a list of 60 adjectives (e.g., “calm”, “inspired”) on which the participant rates each object using a likert scale (6-point likert scale was used in this study). The adjectives in the PANAS-X can be grouped into 2 general higher-order concepts (Negative affect and Positive affect), and 11 specific affects (Fear, Sadness, Guilt, Hostility, Shyness, Fatigue, Surprise, Joviality, Self-Assurance, Attentiveness, and Serenity).

Quantitative data from the PERVAL and PANAS-X variables, and the bidding prices given by participants for each object were collected from post-question naires. PERVAL data was aggregated by subconstruct. As per instructions on the scoring of the PANAS-X [48], responses to the adjectives making up each affect scale were summed. All quantitative data was entered into the SPSS statistical analysis software package. Repeated measures ANOVAs were conducted to test the effects of material (PLA vs SLA resin vs Mass-produced plastic) and object (Cube vs Button). Greenhouse-Geisser corrections were applied when sphericity assumptions were not met.

Qualitative data from the participants’ verbal object descriptions and handwritten notes about the objects were transcribed and imported into the qualitative analysis software, MaxQDA. A total of 396 handwritten notes were collected. A qualitative analysis was performed on the verbal descriptions to identify key concepts that the participants associated with the objects. The coding process consisted of two cycles, first using the open coding approach whereby concepts were labelled using either descriptive or in-vivo codes, and then using the axial coding approach whereby relationships were found among codes and categories [10].

The bipolar constructs from all the participants’ repertory grids were collated. Using a process similar to affinity diagramming, constructs that were semantically the same or that had close relationships to each other were grouped. For instance, “Glossy - Matte” was grouped with “Shiny - Dull”. “Classy - Unsophisticated” was grouped with “Distinct - Generic”. Each group thus indexed a dimension of meaning, and was given a unique ID. A dataset was constructed with the constructs, their associated group ID, and the ranks given on each construct for each of the objects. Friedman tests (since the data was ordinal and the study design was within-subjects) were ran on the dataset to see whether there were significant differences in ranks among the cube and then button objects.

7 Results and Findings

We present our findings by assessment or measure below. For quantitative analyses, only significant results are reported.

7.1 Repertory Grid

We found 13 groups of semantically unique constructs from the repertory grid responses. These are listed in Table 2 with the total number of instances of all constructs represented in each group.

The Friedman test was ran for each group of constructs that had 10 or more instances (see Table 2) to compare whether there was a significant different in given rankings among materials. Tests were ran for the cube and button objects separately. Out of the 6 groups tested, only 3 had a statistically significant difference in given ranks based on material for both objects. Values are shown in Fig. 2.

Table 2. Semantic groups from repertory grid
Full size table
Fig. 2.
figure 2

Repertory grid results. Note: Only groups that showed an overall significant difference are shown.

Full size image
Fig. 3.
figure 3

Repertory grid post-hoc results. Note: Only groups that showed an overall significant difference on the Friedman test are shown. Median ranks are from the Friedman test. Post-hoc results are from the Wilcoxon Signed Ranked test, and show p values. Significant values are marked with an *. The lower the rank on ‘texture’ the smoother, on ‘shine’ the shinier, and on ‘quality’ the higher the quality.

Full size image

Median ranks for each material in each of these 3 groups are given in Fig. 3. Post-hoc analyses (see Fig. 3) showed that with respect to texture for the cube object, there were no significant difference among the specific materials despite an overall significant difference in ranks. However, there were statistically significant rank differences among all the materials for the button object on texture. With respect to shine as well as quality, there were significant differences between both PLA and SLA resin, as compared to mass-manufactured plastic for both objects.

7.2 Bidding Price

There were both a significant main effect of material \((F^{(2, 42)} = 9.72, p = .000, partial \eta ^2 = .316)\) and of object \((F^{(1, 21)} = 13.40, p = .001, partial \eta ^2 = .389)\). No interaction effect was found. Post-hoc tests showed that only PLA objects (M = $0.36) were significantly different (p = .000) from mass-manufactured objects (M = $0.52).

7.3 PERVAL Scale

On ‘Quality’, there were significant main effects of both object \((F^{(1, 21)} = 5.75, \textit{p} = .026, \eta ^2 = .159)\), and material \((F^{(2, 42)} = 22.70, \textit{p} = .000, \eta ^2 = .519)\). Post-hoc tests showed that both PLA (p = .000) and SLA resin objects (p = .000) were significantly different from the mass-manufactured objects.

On ‘Price/Value’, there was only a significant main effect of material \((F^{(2, 42)} = 73.31, \textit{p} = .000, \eta ^2 = .778)\). Post-hoc tests showed that both PLA (p = .000) and SLA resin objects (p = .000) were significantly different from the mass-manufactured objects.

On ‘Social value’, there was a significant main effect of material \((F^{(2, 42)} = 9.46, \textit{p} = .000, \eta ^2 = .310)\), as well as a significant interaction effect of Object X Material \((F^{(2, 42)} = 3.95, \textit{p} = .027, \eta ^2 = .158)\). Post-hoc tests showed that both PLA (p = .004) and SLA resin objects (p = .001) were significantly different from the mass-manufactured objects.

And on ‘Emotional value’, there were both a significant main effect of object \((F^{(1, 21)} = 12.16, \textit{p} = .002, \eta ^2 = .367)\), and a main effect of material \((F^{(2, 42)} = 4.63, \textit{p} = .015, \eta ^2 = .181)\). No interaction effect was found. Post-hoc tests showed that only SLA objects (p = .007) were significantly different from the mass-manufactured objects.

Fig. 4.
figure 4

Object X Material interaction effect on Social Value

Full size image
Fig. 5.
figure 5

Means of each condition for main effects on PERVAL. Note: bidding price values indicate amounts given by participants. For PERVAL constructs, values are on a 1–5 agreement scale. The higher the value, the better on that construct.

Full size image

Figure 5 shows the mean values for each condition for bidding price and the PERVAL constructs for the main effects reported above. The means for the Object X Material interaction effect for the social value construct is shown in Fig. 4.

7.4 PANAS-X Scale

No significant differences were found on the two high-level scales of negative and positive affect. Significant differences were found for only 4 specific affects. Main effects of material were found on “shyness” \((F^{(2, 44)} = 3.30, \textit{p} = .046, \eta ^2 = .130)\), “joviality” \((F^{(2, 44)} = 6.64, \textit{p} = .003, \eta ^2 = .232)\), and “attentiveness” \((F^{(2, 44)} = 4.97, \textit{p} = .011, \eta ^2 = .184)\). A marginal main effect of material was seen on “fatigue” \((F^{(2, 44)} = 3.84, \textit{p} = .056, \eta ^2 = .149)\) Main effects of object were found only on “shyness” \((F^{(2, 44)} = 3.84, \textit{p} = .039, \eta ^2 = .179)\), and “fatigue” \((F^{(1, 22)} = 4.50, \textit{p} = .045, \eta ^2 = .109)\). No significant interaction effects were obtained.

Mean scores of the main effects are shown in Fig. 6. Post-hoc pairwise comparisons showed that there were no significant specific differences among materials for “shyness”. On “joviality”, both PLA (p = .028) SLA (p = .025) were significantly different from mass-manufactured plastic. On “attentiveness”, the only significant difference was between PLA and SLA (p = .007).

Fig. 6.
figure 6

Mean scores for main effects on PANAS-X scales

Full size image

7.5 Handwritten Notes

From our qualitative analysis of the 396 object notes collected from participants, we uncovered five themes of interest with respect to the kinds of meanings that participants associated with the objects. Descriptions of the themes are summarized in Fig. 7, together with example notes.

Fig. 7.
figure 7

Themes from qualitative analysis of the object notes

Full size image

Theme 1: Replicability is about the possibility of an object to be mass-produced or not. While the main use of 3D-printing currently is for rapid prototyping or the production of single objects, the notes associated with the 3D-printed objects showed that people perceive 3D-printed objects to have replicability potential as well. This replicability potential was expressed in terms of, for example, the object being a part of a larger set, a sample from a batch, the first of many more to come, something that can also be sent to others. The theme of replicability is highly encouraging for the possibility of 3D-printed objects to enter mainstream markets.

Theme 2: Flexibility entails the potential of an object for customization, personalization, or artistic expression. Aspects of participants’ notes that were coded as addressing the theme of flexibility included emphases on the ability of the object user to change the object’s properties like color and general appearance, in scenarios such as incorporating the objects as part of art projects (e.g. making a table out of buttons or building blocks). This theme was less pronounced in notes associated with the mass-produced objects.

Theme 3: Participation in production is about whether and how much specific individuals participated in the making of the object. In-vivo codes such as “I built this object”, “this is a handcrafted object”, and “it is part of my creative process” were indicative of this theme, which was much more prevalent in the 3D-printed objects’ notes. This theme also had associated notions of effort contributed to the creation of the objects (e.g., amount of time spent). In contrast, in notes about mass-produced objects, mentions were often made of the object having been found, bought, or given.

Theme 4: Representation relates to the fact that the object is seen as symbolizing other objects, identities or histories. This was often seen in notes for 3D-printed objects where participants perceived the objects as being about more than just the objects, for example, as being representative of historical events, a specific process, or a specific famous person.

Theme 5 Material simulation refers to the ability of the object to convey a specific material feel. Many of the notes for the 3D-printed objects specified the objects as being made of other materials such as wood, limestone, marble, plastic, resin, and synthetic material. This theme was also often based on descriptions of the object’s properties that were coded as addressing size/weight, quality, shine and texture.

8 Discussion

We were interested in how people’s perception of 3D-printed objects differ from that of mass-manufactured objects. We first summarize the results from the various conducted analyses, and then integrate them to answer our research questions. We wrap up by identifying implications for design and by listing the limitations of our study.

8.1 Summary of Results

The repertory grid indicated a list of 13 dimensions that participants found salient among the three objects compared. Only three dimensions of object rankings were significantly different (texture, shine and quality). In these dimensions, mass-produced objects were ranked as the smoothest, shiniest, and of the highest quality. No main effect on object dimension was found. Mass-produced objects were also assigned the greatest dollar amount during the bidding process among the three material types. Only the mean cost assigned to PLA objects (not SLA) was significantly different from that of mass-produced objects. This trend persisted for the results of the four constructs from the PERVAL scale. Scores for mass-produced objects were always the highest among the three materials, followed by PLA and then SLA, except on the ‘price/value’ construct where PLA objects were seen as having better price for value than SLA resin objects. Main effects of object on the PERVAL also showed that there were significant differences between the cube and the button object types on the ‘Quality’ and ‘Emotional value’ constructs, where the cube had higher scores than the button. Interestingly, there was an interaction effect on the ‘Social value’ construct whereby the mass-produced button had a higher mean score than the 3D-printed buttons, but the same didn’t apply for the cube objects.

Ratings on the PANAS-X were significantly different on material type for only three (shyness, joviality and attentiveness) out of the 11 specific affects that the scale measures. Mean scores for mass-produced objects were significantly higher that of mass-produced objects on “joviality”, and PLA objects mean was significantly higher than that of SLA-made objects on “attentiveness”. Last but not least, five themes (replicability, flexibility, participation in production, representation, and material simulation) emerged from our qualitative analysis of the notes that participants wrote about the objects in different hypothetical scenarios.

8.2 Answering Our Research Questions

RQ1a asked whether there are differences in how people perceive the material properties of 3D-printed objects as opposed to those of mass-manufactured objects. Our results showed that people perceive a clear difference in the properties of the various objects, but mostly generally between 3D-printed objects and mass-produced objects. Material property differences were not evident between the two types of 3D-printed objects (PLA- and SLA-made).

RQ1b asked what the differences in perception of material properties consist of. We know from the repertory grid results that the material property differences manifested themselves perceptibly only in terms of texture and shine. Mass-produced objects were perceived to be smoother and shinier than both 3D-printed objects, regardless of object type. Codes from the notes analysis corroborated this with these two properties being the most prominent in terms of the material simulation theme.

RQ2a asked whether people interpret 3D-printed objects differently than mass-manufactured objects. Our results provide evidence that the interpretation of 3D-printed objects vary from mass-produced objects on certain dimension of meanings but not all.

To answer RQ2b about the nature of these differences in interpretation, people valued 3D-printed and mass-manufactured objects differently at monetary, social and emotional levels, and perceived them to be different quality.

At a monetary level, the bidding price and PERVAL price construct analyses tell us that mass-produced objects are seen as being more valuable. People were willing to pay more for mass-produced objects.

Interestingly, our 3D-printed objects costed more than mass-manufactured products, mostly because of the material and time spent during their production process. It is possible that increasing people’s knowledge of the 3D printing process may lead them to assign a greater monetary value to 3D-printed objects. Other work (e.g., [41]) has found results with respect to craft objects that support this hypothesis.

At the social level, the PERVAL results suggest that not only are 3D-printed objects considered to be less acceptable than mass-produced objects in social settings, but this seems to be especially so for functional objects (the buttons in our study). This is highly indicative of the polarizing effect of mass-manufactured commodity objects.

At an emotional level, PERVAL results show that both object type and material type have an effect. People seem to enjoy the cube more than the button, perhaps because of its non-descriptive functionality. For material type however, mass-produced objects are seen as more enjoyable than SLA-printed objects, but not significantly more enjoyable than PLA objects. This finding is echoed in the PANAS-X results, which showed that people felt more ‘jovial’ towards mass-produced objects. An explanation for this may be found in the findings from the notes. SLA-made objects were assigned adjectives such as “old” “historic” and “fragile”whereas mass-produced objects had more instances of “related to high-technology”and “fun” This result is reminiscent of findings from Isbister et al.’s study [23], where they found that participants perceived smooth and round objects as “happy”or “fun”as compared to other object shapes and forms.

In terms of quality, people perceived 3D-printed objects to be worse than mass-produced objects. This finding is consistent across the Repertory Grid and PERVAL results. Despite the remarkable advancement of 3D printer technology, it appears that mass manufacturing presently remains the benchmark of quality. Beyond assessments of material properties such as durability, this quality judgment may be because of anchored societal conventions. It may be that over time, expectations will shift to allow people to perceive 3D-printed objects as more valuable without requiring them to be exactly like mass-manufactured objects.

Finally, we do see potential in 3D-printed objects to become everyday objects in the themes of replicability, representation and material simulation. Currently, the value of 3D-printed object seems to be explained largely in terms of the themes of flexibility and participation of production.

8.3 Implications for Design

From our results as a whole, we see two major implications for the HCI community:

(1) The potential for 3D-printed objects to move to everyday consumption for casual users is at least partly dependent on people’s perceptions of and attitudes towards them. There is still improvement needed for 3D-printed objects to be perceived on at least a similar status as mass-produced objects. With current 3D printing technologies, the focus for 3D-printed objects is on use and functionality, and by and large they can serve the function for which they are designed. However, our study indicates that people perceive and approach 3D-printed objects differently than mass-produced objects, especially in social settings where people find 3D-printed objects less acceptable. It is perhaps possible that attitudes towards 3D-printed objects will change as 3D printers increase in sophistication. For example, state-of-the-art printers are already able to accommodate a variety of materials such as foods [30], felts [22], and ceramics [29].

(2) Scaffolds should be designed to support materiality as a key design parameter in 3D printing customizers. It was evident from our findings that resultant 3D-printed objects are distinctively perceived. However, current 3D design software that are used for 3D printing do not scaffold any aspect directly related to materiality. Research in 3D design software and editors for 3D printing are presently heavily focused on supporting users in handling geometry. For example, Hofmann et al. [18] proposed a framework called PARTs that can accommodate users’ design intents of a 3D model by visualizing assertions of the functional geometry to meet semantic expectations, and Kim et al. [25] developed the FitMaker editor to attenuate measurement errors in design. It provides a library of standardized format for CAD models to support novice users to adjust the 3D design models to fit their needs and design intent. The focus on geometry is warranted because of the emphasis on functionality for 3D-printed objects, and functions like scaling, adjustment and placement in 3D editors are typically very challenging for inexperienced users.

We argue that in order for the creativity and 3D-printed objects to be consumed by everyone, the user’s design intent, in which materiality has a large influence, needs to be better supported. Possibilities should include the user. For example, being able to express through the interface that “I want a vintage photo frame to give to my grandmother for her birthday” leading the modeling software to recommend printing with an SLA printer with specific parameters.

8.4 Study Limitations and Future Work

Our study has a few limitations that need to be kept in mind: First, different 3D printers with the same production process, environment, and settings may still result in objects with slight variations in appearance, potentially affecting people’s perceptions of them. Second, the probe objects in our study were relatively small in size. We note that bigger objects may allow material properties to be more evident, leading to varying participant interpretations. And third, we looked only at the immediate effects of 3D-printed objects.

Longer-term effects may be different since people tend to project additional meaning on objects over time through an appropriation process [17].

9 Conclusion

This study investigated the perception of 3D-printed objects. Participants were given objects 3D-printed in PLA and SLA resin to evaluate as compared to mass-produced versions of the objects. Results showed that people distinguish between 3D-printed and mass-produced objects on specific dimensions, both at the material properties level and at an interpretive level associated with, for instance, quality, price, social, and emotion. Only small differences were found in perception between the two types of 3D-printed materials, suggesting that 3D-printed objects overall as a class of objects are distinct from mass-manufactured objects. We propose that for the vision of everyday 3D printing [1, 3] to become a reality, further research need to be done on people’s perceptions of 3D-printed objects and the notion of materiality in 3D design customizers.