Children expect others to prefer handmade items

Abstract

Although children frequently engage in creative activities (in which they make foods and objects by hand), the development and scope of children’s thinking about handmade items is largely unexplored. In the present studies, we examined whether 4- to 12-year-old children at a local children’s museum (54% girls, 46% boys; 51% White, 11% Asian/Asian-American 10% more than one group, 4% Latinx, 3% Black/African-American, 18% did not report race/ethnicity) would expect other people to prefer handmade over factory-made items, including foods and nonfoods. In Experiments 1 (n = 124) and 2 (n = 122), participants expected a child character to prefer items the character made themselves and items made by the character’s parent or a local person. However, this expectation did not persist at all costs: When considering imperfect handmade items in Experiment 3 (n = 122), children demonstrated a handmade preference when considering nonfoods made by a parent but demonstrated a factory-made preference when considering foods made by a parent. Children’s explanations were associated with their choices: When children’s explanations referred to emotions or relationships, they were more likely to select handmade items. When children referred to item features, they were more likely to select factory-made items. Across studies, we observed persistent age and gender effects: Children’s handmade preference increased with child age and girls demonstrated a more robust handmade preference than boys. These findings highlight children’s developing and nuanced reasoning about object value. At an early age, children consider who made an object as a contributor to its value.

A growing body of literature examines children’s assessments of the value of objects, including how much they are willing to pay for something (Gelman et al., 2015), whether they prefer one item over another (Echelbarger & Gelman, 2017; Maimaran & Fishbach, 2014), and the effort they invest in an object, such as how much time they spend looking for a particular feature (Gelman et al., 2016). As burgeoning consumers, children’s consideration of what objects are worth is of interest both in its own right and as a window into the processes underlying economic decisions across the lifespan (see Gelman & Echelbarger, 2019). In addition to attending to the visible properties of objects (e.g., their color, shape, decorative details), young children are also sensitive to non-visible properties of objects, such as their histories. Children consider who owns an object to be an important part of the object’s history that confers value (Nancekivell et al., 2019), leading children to report being willing to pay more for items owned by famous people or characters (Gelman et al., 2015). Some objects (e.g., attachment objects, celebrity possessions) have special non-visible properties that children do not think can be easily duplicated (Gelman & Davidson, 2016; Hood & Bloom, 2008).

One aspect of object history that may be important is an item’s creation process. A few studies suggest that children consider who made an object when evaluating that object (DeJesus et al., 2019; Marsh et al., 2018) or determining its owner (Kanngiesser et al., 2010, 2014; Kanngiesser & Hood, 2014; Li et al., 2013). However, the development and scope of this phenomenon is still largely unexplored. The present research examines children’s expectations of value depending on who made something (e.g., whether a hypothetical child would prefer an item they made themselves or bought from a factory).

One reason to consider whether children view the creation process as part of an object’s value is their own rich experience with these activities. Creative activities are part of children’s daily lives. According to a report from the Pew Research Center (Parker et al., 2015), 54% of parents with 6- to 17-year-old children reported that at least one of their children had taken music, art, or dance lessons in the past 12 months. Among parents who affirmed their child had participated in these activities, 41% of parents reported an annual family income of less than $30,000 (Parker et al., 2015), suggesting that these activities are pursued across the socioeconomic spectrum. Yet few studies have examined children’s thinking about objects they created in an experimental context (Marsh et al., 2018, in press), highlighting the need for further investigation.

The present studies include items in which children are asked to consider foods as well as other objects they often create (e.g., drawings). Making foods may be an especially important context as children frequently experience foods being made for them, and involving children in food preparation activities has been examined as a novel component to interventions aiming to increase children’s willingness to eat healthy foods (e.g., Allirot et al., 2016; Davis et al., 2011; Ensaff et al., 2015; Hersch et al., 2014; van der Horst et al., 2014). Indeed, one recent experimental study found that children eat more of a food they prepared themselves, compared to an identical food they did not prepare (DeJesus et al., 2019; cf. Raghoebar et al., 2017, for a different pattern of results).

The present research asks children to assess whether a character would prefer an item made by the character’s parent, compared to an item made in a factory. Children may be especially interested in items made by parents in light of evidence that children are more willing to accept foods that they saw modeled by their mother compared to an unfamiliar researcher (Harper & Sanders, 1975) and are generally more trusting of information provided by their mothers compared to other informants (at least among securely-attached children; Corriveau et al., 2009). Moreover, children and adults may be socialized to view handmade items as a marker of good parenting, especially of what it means to be a good mother, in which economic decisions (e.g., purchasing vs. handmade) are often characterized as evidence of care or part of a constructing a family identity (Jennings & Brace-Govan, 2014; Moisio et al., 2004; Thompson, 1996; see Moscato & Machin, 2018, for a related perspective in the context of natural foods). These patterns fall into the model of “intensive mothering,” which encourages mothers to highly invest their money, time, and energy in raising their children, even while working outside the home (Hays, 1996), including in food contexts (Brenton, 2017). Although these characterizations assume that handmade is better, whether children actually value that investment is relatively unknown – children could prefer a factory-made item, or their evaluation of their parent’s investment could depend on the output. A child may value a parent’s effort in some contexts more than others, or their evaluation could depend on the quality of what the parent made.

Finally, in addition to examining whether children’s valuations vary based on maker identity, we asked children to explain their value judgments and coded those explanations for references to socioemotional and item features. Prior studies on adults’ and children’s perceptions of objects have asserted that “labor leads to love” (Marsh et al., 2018; Norton et al., 2012), so a natural extension of investigating children’s value judgments is whether children view the process of making objects as imbued with love or other emotions. If children do value items differently depending on who made them, how do they explain these preferences? Some children could find social or emotional resonance in handmade items (highlighting concepts of pride or love, or referencing the relationships between the maker and the recipient), whereas others could focus more on the features of those items (highlighting tailoring items to one’s specific preferences or judging based on impressions of the item’s quality).

The Present Research

The present research examines children’s valuations of foods and nonfoods based on the origins of each item (whether a character made an item vs. bought it from a factory), whether the item is a food or nonfood, and the identity of the maker (the character themselves, the character’s parent, or an unrelated person in their town). We operationalized children’s value judgments by asking them which item (a handmade vs. a factory-made item) a hypothetical child would prefer. We chose this third-person method (rather than eliciting participants’ own preferences) to examine their reasoning based on different social relationships, while considering that any potential participant may not have that social relationship personally (i.e., a child with any family configuration could reason about another child’s mother or father). We recruited a wide age range for this study (4- to 12-year-old children) to examine potential developmental change in children’s thinking about handmade vs. non-handmade items (see Chaplin et al., 2020, for parallel findings that the value 3- to 12-year-old children place on experiences increases with age). The few studies that have examined children’s thinking about handmade foods and objects have focused on children age 6 and younger (DeJesus et al., 2019; Marsh et al., 2018, in press); therefore a wider age range provides additional insight into the development of reasoning about handmade items across childhood.

In all experiments, children were shown images of other children (here referred to as a “character” to differentiate between the child participant and the hypothetical child described to the participant) with two items from the same category (i.e., two foods or two nonfoods). Children were told that one item was made by a specific person whereas the other item was made in a factory. Children were then asked to predict which item the character would prefer and why. In Experiment 1, children were shown two matched items (e.g., two nice-looking cookies) and asked to compare a factory-made version to a version made by the character or their parent. In Experiment 2, parent trials were replaced by trials that described a person in the character’s town who makes that item (a specific local person, but less familiar or relevant to the character than a parent), based on prior accounts suggesting that children may especially value objects they made themselves because that making process links that object to them as an extension of their identity (Marsh et al., 2018). In Experiment 3, we varied the quality of the items – instead of two high quality items, the handmade item looked imperfect, whereas the factory-made item appeared to be of higher quality, to probe the robustness of children’s valuation of handmade items. Based on the object history literature, we predicted that children would expect a character to prefer items that the character made themselves. However, it is unknown from this literature whether this effect would differ between foods and nonfoods, how these patterns might differ developmentally, or how the social identity of the person making the item might influence children’s predictions. The present experiments allow us to test these questions directly.

Experiment 1: Items Made by Children and Their Parents

Children in Experiment 1 were asked to judge other children’s preferences for handmade vs. factory-made foods and nonfoods. Specifically, for each trial, children were introduced to a character and were asked about that character’s preference for handmade vs. factory-made items. Across trials, the handmade item was either made by the character themselves (child trials) or by the character’s parent (parent trials). We opted for a third-person design (rather than asking participants about their own preferences) to manipulate who made the item without referring to participants’ own families or relationships (for instance, we could create a character, “Rob,” whose mother made a cake and ask the participant to consider that scenario, regardless of the participants’ own family structure).

Method

Participants

Participants included 124 4- to 12-year-old children (32 4- and 5-year-olds, 32 6- and 7-year-olds, 32 8- and 9-year-olds, and 28 10- to 12-year-olds; mean age = 7.89 years, range = 4.02 – 12.92 years, 95% CI = 29.30, 34.50; 63 girls, 61 boys) tested at a local children’s museum. In terms of race and ethnicity (reported by children’s parents), 65 participants were White, 18 were Asian/Asian-American, 11 reported more than one race/ethnicity, 4 were African-American, 3 were Hispanic/Latino, 1 was Native American, and 22 reported other or did not report. Children received a small prize for participating in the study. In all experiments, a target sample of 32 children per age group was planned; if additional children completed the study, their data were included. Children participated in only one study in the present research. These studies were approved by the Institutional Review Board of the University of Michigan (HUM00110855, “Children’s food concepts”) and conducted from 2016–2018.

Materials and Procedure

Children were shown a series of images of children, foods, and objects on a laptop computer (see Figure 1, top). Images of children were selected from the Child Affective Facial Expression (CAFE) set (LoBue & Thrasher, 2015). Images were selected to be reliably rated as happy and included two Black girls (10514-happy_F-AA-02.jpg, 10518-happy_F-AA-07.jpg), two White girls (10542-happy_F-EA-15.jpg, 10547-happy_F-EA-21.jpg), two Black boys (10583-happy_M-AA-03.jpg, 10590-happy_M-AA-11.jpg), and two White boys (10604-happy_M-EA-07.jpg, 10623-happy_M-EA-34.jpg). Because the CAFE face set cannot be published, Figure 1 displays an outline of a child to stand in for a CAFE face. Images of foods were selected from the Food Pics Database (Blechert et al., 2014, 2019) where possible; the remaining images were found through Internet searching. See Open Science Framework (OSF) for study images (DeJesus, 2022): https://osf.io/pytcm/.

Figure 1.

Representative stimuli. For child trials, children were told, “This is a cookie he made himself,” while pointing to one item and, “This is a cookie he bought from the cookie factory,” while pointing to the other. In parent trials, children were told, “This is a cake her mom made herself,” while pointing to one item and, “This is a cake her mom bought from the cake factory,” while pointing to the other. In artisan trials, children were told, “This is a quilt she bought from a person in her town who makes quilts,” while pointing to one item and, “This is a quilt she bought from the quilt factory,” while pointing to the other. In Experiment 3, the nicer-looking item (left) was the factory-made item; the messier item (right) was the handmade item. The cookies and cakes (top) were actual stimuli in Experiments 1–2 and are reproduced with permission from “Food-pics: an image database for experimental research on eating and appetite,” by Blechert et al., 2014, Frontiers in Psychology, 5, 617. Copyright 2014 by Frontiers. The fish and pizzas (bottom) are representative of what children saw (see OSF for actual materials) and are edited version of freely useable images from unsplash.com or hand drawn.

For 8 trials, children were introduced to a target child (e.g., Rob) who had two items (e.g., cookies). For 4 of the 8 trials, children were told that one item was made by the character (e.g., while pointing to the item, the researcher said, “This is a cookie he made himself.”) and one item was purchased from a factory (e.g., while pointing to the item, the researcher said, “This is a cookie he bought from the cookie factory.”). These trials will be referred to as child trials. For the other 4 trials, children were told that one item was made by the character’s parent (e.g., “This is a cake her mom made herself.”) and one item was purchased from a factory (e.g., “This is a cake her mom bought from the cake factory.”). These trials will be referred to as parent trials. Then the researcher asked which item the character “likes best.” Children responded by either pointing to the item or verbally describing the item (e.g., by saying, “the one from the factory”). Children were also asked to explain why they selected that item. Example scripts for each experiment are available on OSF: https://osf.io/pytcm/.

Child trials included cookies, bread, scarves, and pictures. Parent trials included cake, pasta, mugs, and quilts. Items were selected to include a variety (i.e., avoid repeating items) and be plausibly made by children or adults (i.e., a quilt was not described as being made by a child).

Explanations Coding

To examine children’s explanations, two coders determined whether each explanation referred to item features (yes or no) and whether each explanation referred to socioemotional features (yes or no). These coding categories were developed to examine whether children viewed the choice of handmade or factory-made items as being related to the quality or composition of the item itself versus the non-visible social or emotional properties of the item.

For socioemotional explanations, children could mention an emotion someone might feel about the item (“he was probably more proud of himself for making it”) or a connection to the person who made the item (“because it was sweet for the mom to make it”). The goal was to highlight non-observable aspects of the items (rather than observable features such as taste, color, and shape), as these may be especially relevant to children’s thinking about handmade items.

For item features, our goal was to create a single category to contrast with socioemotional explanations. Thus, although a variety of qualities could be noted for different items (e.g., foods vs. nonfoods), we did not provide more fine-grained coding. In general, explanations were coded as being related to item features if they referred to the taste, quality, appearance, composition, or functionality of the item. For food trials, explanations often referred to flavors (e.g., “chocolate is better”), general taste (“it looks tastier”), appearance (“it looks good”), quality (or lack thereof, “because the other looks burnt and gross”), or a combination of these features (“thicker, cheesier, and overall looks better”). For nonfoods, explanations often referred to the appearance of the item, such as its color (“because the other one just has regular colors, this one has different colors together”), level of detail (“because it has things on it and that one’s a plain one”), and functionality (“because she thinks it’s more comfy and more warm”). These explanations may be prominent if children are not especially interested in the handmade aspect of the item, and instead are focused on its quality or observable features.

Codes were not mutually exclusive, meaning that the same explanation could be coded as referring to both item and socioemotional features (“because even though it doesn’t look as good her parents put time and effort into it”) and some received neither code (explanations that just repeated back the story content, such as “because he made it himself”; explanations that reported general liking, such as “because I think he likes that scarf”; or uninformative responses, such as “I don’t know” or “because he might chomp”). Explanations from all experiments were combined and sorted using a random number generator to create smaller sets for training, reliability, and independent coding. Coders completed two small training sets (32 explanations) and then completed a reliability set of the same 612 explanations (21% of the total; 2 codes provided per explanation). Coders were reliable (kappa = 0.81) and resolved disagreements by discussion. Coders then independently coded the remaining 2,273 explanations (coder A = 1,178; coder B = 1,095).

Design, Scoring, and Analyses

Items were counterbalanced such that half of the participants learned that one item (e.g., the chocolate cake) was handmade and the other half learned that the same item was factory-made, so that any effect would not reflect a general preference for a particular item (such as a general preference for chocolate cake over strawberry cake). Position on the screen (i.e., whether the chocolate cake was on the right or left side) was also counterbalanced across participants. Items were blocked such that children either viewed all 4 child trials and then all 4 parent trials, or the reverse. Participants either viewed a male or female character first, and items were counterbalanced such that they were paired with a male character for half of the participants and with a female character for the other half of the participants (e.g., some children heard about cookies made by a girl and some children heard about cookies made by a boy; some children heard about a quilt made by a mother and some heard about a quilt made by a father). For parent trials, the gender of the character and their parent were matched (i.e., for girl characters, the parent was their mother and for boy characters, the parent was their father). This counterbalancing plan resulted in 8 unique scripts per experiment.

Responses were scored as a 1 if participants selected the handmade item and a 0 if they selected the factory-made item. Responses were excluded if participants did not pick one item (either handmade or factory) after being asked the test question two times or if another person (such as their parent) interfered with the trial (9/992 trials).

To test whether children expected characters to prefer handmade items over factory-made items, we compared the number of handmade responses to chance (4 out of 8) using Wilcoxon signed-rank tests (two-tailed). To examine predictors of the preference for handmade vs. factory-made, we ran a binomial generalized linear mixed model (using the glmer function in lme4 package in R) with the following predictors of children’s handmade vs. factory-made item for each trial: Item type (food vs. nonfood), maker identity (child vs. parent), maker gender (female vs. male), participant gender (female vs. male), participant age (in years), and two interactions (Item type × Maker and Participant gender × Maker gender).

To examine the association between children’s explanations (item or socioemotional) and their selection of handmade vs. factory-made items, a binomial generalized linear mixed model was performed on children’s selection of the handmade vs. factory-made item for each trial, with whether the child provided a socioemotional explanation, whether the child provided an item features explanation, child age, and child gender as predictors.

For all experiments, deidentified data and analysis code in R (R Core Team, 2016) are available on OSF (https://osf.io/pytcm/); studies were not preregistered.

We hypothesized that children would generally prefer the handmade item, given their attention to object history in previous studies. We did not have any directional hypotheses regarding child age, child gender, or maker identity (the character vs. the character’s parent), but we did hypothesize that children’s expectations about handmade foods could be especially strong given evidence that they ate more of foods that they made themselves (DeJesus et al., 2019). At the same time, given children’s robust attention to object history in studies of nonfoods (e.g., Gelman et al., 2016; Marsh et al., 2018), we acknowledge the possibility that children’s expectations about handmade items could be especially robust for nonfoods, or could be similar to their expectations about foods. In terms of children’s explanations, we expected that the more children referred to socioemotional features (rather than item features), the more they would prioritize handmade items over factory-made items because of their social histories.

Results

Preference Judgments

Children expected other children to prefer handmade items (M = 5.65 out of 8, 95% CI = 5.30, 5.99), p < .001. The regression analysis revealed a significant effect of item type, child age, child gender, and maker gender (ps < .05; see Table 1 for full regression results). For the effect of item type (p = .002), children tended to select more handmade items for nonfoods (scarves, pictures, quilts, and mugs; M = 2.96 out of 4) compared to foods (cookies, breads, pastas, and cakes; M = 2.69 out of 4), see Figure 2. For the effect of child age (p < .001), older child age predicted a greater likelihood of selecting handmade items (based on the positive regression coefficient, b = 0.23). For the effect of child gender (p = .002), girls (M = 6.16 out of 8) selected more handmade items than boys (M = 5.11 out of 8). For the effect of maker gender (p = .043), children tended to select the handmade item slightly more often for female makers (collapsing across child and parent, 2.89 out of 4) than for male makers (2.76 out of 4). No significant interactions were observed.

Table 1.

Regression tables from Experiment 1. Top: Children’s selection of handmade items based on maker identity (child vs. parent), item type (nonfood vs. food), child age, child gender, and maker gender. Bottom: Children’s selections of handmade items based on child age, child gender, and response explanations (socioemotional, item features).

	Mean (95% CI)	Estimate	SE	z	P
(Intercept)		−0.35	0.46	−0.76	.450
Maker identity		0.09	0.22	0.42	.677
Item type	Nonfoods: 2.96/4 (2.76, 3.16) Foods: 2.69/4 (2.48, 2.89)	0.71	0.23	3.11	.002
Item × Maker		−0.59	0.32	−1.86	.064
Child age		0.23	0.05	4.56	< .001
Child gender	Girls: 6.16/8 (5.71, 6.61) Boys: 5.11/8 (4.63, 5.60)	−0.91	0.29	−3.15	.002
Maker gender	Female: 2.89/4 (2.69, 3.08) Male: 2.76/4 (2.56, 2.95)	−0.48	0.24	−2.02	.043
Child gender × Maker gender		0.52	0.32	1.63	.103
	Mean (95% CI)	Estimate	SE	z	P
(Intercept)		0.32	0.41	0.79	.429
Child age		0.18	0.05	3.77	< .001
Child gender		−0.49	0.22	−2.23	.026
Socioemotional explanation	Yes: .98 (.95, 1.00) No: .67 (.64, .70)	2.28	0.65	3.49	< .001
Item features explanations	Yes: .62 (.58, .66) No: .82 (.79, .86)	−0.90	0.19	−4.80	< .001

Figure 2.

Experiment 1 (error bars represent standard errors).

Explanations

Overall, children provided more item-based explanations (M = 4.28, 95% CI = 3.85, 4.71) than socioemotional explanations (M = 1.14, 95% CI = 0.84, 1.44). The number of socioemotional explanations children provided was positively correlated with child age, r(122) = .51, p < .001; item features explanations were not correlated with age, r(122) = −.11, p = .244. In the regression, we observed effects of explanation type (ps < .001). Children were more likely to select the handmade item if they provided socioemotional explanations (yes = .98, no = .67, p <.001) and were less likely to select the handmade item if they provided item features explanations (yes = .62, no = .82, p < .001); note that proportions are greater than .5 as children overall were more likely to select the handmade item than the factory-made item. This analysis also replicated the effects of child age (i.e., the older the child, the more likely they were to select the handmade item, p < .001) and child gender (i.e., girls were more likely to select handmade items than boys, p = .026).

Discussion

The results of Experiment 1 suggest that children expect other people to value handmade items (compared to items bought from a factory), including items that characters made themselves and items made by a character’s parent. We also observed associations with age (the older the child, the more likely they were to select the handmade item over factory-made item) and gender (girls demonstrated a stronger handmade preference expectation than boys), topics we return to in the General Discussion. Additionally, children’s explanations were related to their value judgments – children who made socioemotional explanations were more likely to select handmade items, whereas children who referred to item features in their explanations were less likely to select handmade items.

In Experiment 1, items were either made by the child character themselves or by someone highly familiar to the character (their parent). Participants may be especially interested in handmade items in this context given their socioemotional explanations, as they often referenced feelings of specialness, pride, and love. However, it is possible that these positive feelings toward handmade items may be exclusive to self-made items or items made by close others, rather than just knowing that a specific person made the item. In Experiment 2, we further probed this idea by replicating this study but replacing parent trials with trials in which items were made by a person in the character’s town. This local person was described as a specific individual who lived nearby the character (and thus may be somewhat familiar or from a similar cultural background). Therefore, children may also value handmade items in that context. In contrast, the local person may be viewed as less familiar or close to the character (compared to a parent); if children see objects they made as extensions of themselves (Marsh et al., 2018), perhaps they would use similar reasoning when considering a closely related person (like a parent) but not any known person (like a person in the character’s town). The local person’s productions also may not be viewed as being specifically made for the character (anyone could potentially purchase the local person’s wares). As such, children may demonstrate a reduced handmade preference when considering items made by a person in the character’s town.

Experiment 2: Items Made by Children and Local People

Children in Experiment 2 were again asked to predict other children’s preferences for foods and nonfoods that varied in maker identity. To examine whether children’s judgments would vary based on who made the item (does the maker have to be someone very close to them or just a specific, known person?), we replaced the parent trials in Experiment 1 with trials in which items were made by a person in the child’s town who makes that item. For brevity, we refer to the local person here as an artisan (this term was not used in the experiment). In addition, both items in Experiment 2 were described as bought (one from the artisan and one from the factory), rather than contrasting between making and buying (as was the case in Experiment 1).

Method

Participants

Participants included 122 4- to 12-year-old children (35 4- and 5-year-olds, 32 6- and 7-year-olds, 27 8- and 9-year-olds, and 28 10- to 12-year-olds; mean age = 7.80 years, range = 4.23 – 12.94 years; 67 girls, 55 boys) tested at a local children’s museum. In terms of race and ethnicity (reported by parents), 66 participants were White, 12 reported more than one race/ethnicity, 6 were Asian/Asian-American, 6 were Hispanic/Latino, 2 were Black/African-American, 1 was Native American, and 29 reported other or did not report.

Procedure, Design, Scoring, and Analyses

The same procedure, design, scoring, and analyses were employed as in Experiment 1. The only difference was that parent trials were replaced by artisan trials in which children heard that one item was made by a person in the character’s town (e.g., “This is a cake she bought from a person in her town who makes cakes.”) or was purchased from a factory (e.g., “This is a cake she bought from the cake factory.”). Children were presented with 4 child trials (e.g., cookies made by the character vs. bought from the cookie factory) and 4 artisan trials. Maker gender was not included in the analyses as artisans were not described with a gender.

Results

Preference Judgments

Children expected other children to prefer handmade items (M = 5.08 out of 8, 95% CI = 4.78, 5.39), p < .001. A regression analysis revealed significant effects of maker identity (self vs. artisan, p = .041) and item type (food vs. nonfood, p = .016), which were qualified by an Item type × Maker identity interaction (ps = .046; see Table 2 for full regression results), see Figure 3. To examine the interaction, a Wilcoxon signed-rank test (two-tailed) was performed to compare responses to foods and nonfoods for both child trials and artisan trials. For trials in which items were made by the character, handmade items were selected more often for nonfoods (M = 1.51 out of 2) than for foods (M = 1.33 out of 2), p = .026. In contrast, for trials in which items were made by the artisan, no difference was observed between nonfoods (M = 1.11 out of 2) and foods (M = 1.13 out of 2), p = .871.

Table 2.

Regression tables from Experiment 2. Top: Children’s selection of handmade items based on maker identity (child vs. artisan), item type (nonfood vs. food), child age, and child gender. Bottom: Children’s selections of handmade items based on child age, child gender, and response explanations (socioemotional, item features). Note that maker gender is excluded here because artisan trials did not describe gender.

	Mean (95% CI)	Estimate	SE	z	P
(Intercept)		−0.24	0.32	−0.77	.444
Maker identity	Chil d : 2.84/4 (2.67, 3.00) Parent: 2.25/4 (2.03, 2.46)	−0.41	0.20	−2.05	.041
Item type	Nonfood: 2.62/4 (2.43, 2.82) Food: 2.46/4 (2.27, 2.65)	0.51	0.21	2.42	.016
Item × Maker	Nonfood-child: 1.51/2 (1.40, 1.62) Food-child: 1.33/2 = 1.21, 1.45	−0.57	0.29	−1.99	.046
	Nonfood-artisan: 1.11/2 (0.98, 1.25) Food-artisan: 1.13/2 (1.00, 1.26)
Child age		0.15	0.04	4.29	< .001
Child gender	Girls: 5.33/8 (4.92, 5.73) Boys: 4.78/8 (4.32, 5.24)	−0.37	0.17	−2.22	.027
	Mean (95% CI)	Estimate	SE	z	P
(Intercept)		0.44	0.32	1.35	.178
Child age		0.11	0.04	3.00	.003
Child gender		−0.33	0.16	−2.00	.046
Socioemotional explanations	Yes: 0.94 (0.89, 99) No: 0.61 (0.58, 0.65)	1.61	0.45	3.60	< .001
Item features explanations	Yes: 0.58 (0.54, 0.61) No: 0.79 (0.74, 0.83)	−0.80	0.18	−4.41	< .001

Figure 3.

Experiment 2 (error bars represent standard errors).

Effects of child age (p < .001) and child gender (p = .002) were also observed. For the effect of child age, the positive estimate (b = 0.15) indicates that the older the child, the more likely they were to select handmade items. For the effect of child gender, girls (M = 5.33 out of 8) were more likely to select handmade items than boys (M = 4.78 out of 8).

Explanations

The number of socioemotional explanations children provided was positively correlated with child age, r(120) = .50, p < .001; item features explanations were negatively correlated with child age, r(120) = −.22, p = .017. Examining the association between children’s explanations and their responses, the regression revealed significant and positive effects of socioemotional explanations on children’s responses: Children who gave socioemotional explanations were more likely to select handmade items in test trials (yes = .94, no = .61, p < .001). Examples of socioemotional explanations here included, “He would be more familiar with the person in his town and would probably choose that one,” “He’s friends with the person he bought it from,” and “Because she likes the person in town and she doesn’t know the person who made the cake in the factory,” suggesting some relationship with the artisan. Children who provided items features explanations were less likely to select handmade items (yes = .58, no = .79, p < .001). This analysis also replicated the effects of child age (i.e., the older the child, the more likely they were to select the handmade item, p = .003) and child gender (i.e., girls were more likely to select handmade items than boys, p = .046).

Discussion

As in Experiment 1, children in Experiment 2 expected others to prefer handmade items compared to items bought from a factory, and this effect was stronger with age. We observed the same gender effect as in Experiment 1: Girls demonstrated a stronger handmade preference expectation (a topic which we return to in the General Discussion). We also observed an interaction between maker and item type, wherein children expected a stronger handmade preference for nonfoods compared to foods when considering items that a character made for themselves, but did not show a difference between nonfoods and foods when considering items made by a local person, suggesting that participants may especially prioritize nonfoods that someone made themselves. This finding contributes to the account that handmade items are valued, at least in part, because they are viewed as an extension of the self (Hood et al., 2016; Marsh et al., 2018; see also Diesendruck & Perez, 2015), and therefore items made by people who one is close with (e.g., a parent vs. someone from one’s town) may be especially valued.

How robust is the handmade preference expectation in the face of other considerations, such as item quality? As discussed previously, the concept of intensive mothering promotes the idea that handmade is best, particularly in a food context. Experiments 1 and 2 provide some evidence that children do value handmade creations but do so more robustly for nonfoods than foods when considering items similar in quality. In the real world, handmade items often look different from factory-made items, with discernible irregularities and imperfections, which could influence children’s value judgments. Therefore, in Experiment 3 we manipulated item quality by presenting participants with imperfect handmade items. Items looked like someone could have made them but were purposely selected to be messier and more imperfect-looking than the factory-made alternative. Handmade items could be so highly valued that children would still expect others to prefer them, regardless of the item’s appearance. Other studies have found that children prefer worn or damaged objects with sentimental value, such as their own attachment objects (Gelman & Davidson, 2016) or objects that are adorned by an image of a familiar character (Danovitch & Mills, 2017). In contrast, imperfect items may draw attention to item features, rather than socioemotional concepts, which were related to children’s handmade selections in prior experiments.

Experiment 3: Imperfect Handmade Items

In Experiment 3, children were asked to compare nice-looking factory-made items with imperfect-looking handmade items (made either by a character or that character’s parent). The procedure replicated the procedure of Study 1, but with new images.

Method

Participants

Participants in the main task in Experiment 3 included 122 4- to 12-year-old children (31 4- and 5-year-olds, 30 6- and 7-year-olds, 33 8- and 9-year-olds, and 27 10- to 12-year-olds; an additional parent did not provide a specific age but reported their child’s age as being between 8 and 12; mean age = 7.91 years, range = 4.17 – 12.72 years; 69 girls, 53 boys) tested at a local children’s museum. In terms of race and ethnicity (reported by parents), 58 participants were White, 18 were Asian/Asian-American, 16 reported more than one race/ethnicity, 6 were Hispanic/Latino, 5 were African-American, 3 were Native American, and 16 reported other or did not report.

We also tested an additional 38 children in a stimuli check (11 age 5 years and younger, 18 6- and 7-year-old, 7 age 8 or older; age not specified for two children; mean age = 6.18 years, range = 2.92 – 16.08 years; 19 girls, 17 boys, 2 not reported). In terms of race and ethnicity, 15 participants were White, 8 were Hispanic/Latino, 5 were African-American, 5 reported more than one race/ethnicity, 3 were Asian/Asian-American, and 2 reported other or did not report.

Procedure, Design, Scoring, and Analyses

The same procedure, design, scoring, and analyses were employed as in Experiment 1. The only difference was that item images were replaced so that the handmade items looked less perfect than the factory-made items (see Figure 1, bottom). Images were selected to be similar in perceptual features (e.g., two images of chocolate chip cookies, rather than two different cookie flavors), but different in how perfect vs. messy they appeared. Our focus here was to characterize quality based on the look of the item, rather than directly providing information about differences in taste or functionality. Therefore, we focused on messy or misshapen items. As the goal was to have “imperfect” items be imperfect in similar ways across foods and nonfoods, we did not manipulate imperfection based on taste or features that would be possible for only one item type (e.g., foods only). Where possible, the handmade item was selected from the images from Experiments 1 and 2. Due to copyright restrictions, the Experiment 3 images in Figure 1 (bottom left, right) are not the actual images shown in the present research but were selected for display purposes from repositories that are freely available (unsplash.com) and edited as needed to represent the stimuli in Experiment 3, or were hand drawn. Study images are available on OSF: https://osf.io/pytcm/.

To examine whether children interpreted the messy manipulation as intended, we performed two checks. First, we recoded children’s explanations in all experiments to focus on item quality. Coders looked for explanations that directly compared two items (e.g., “prettier,” “because it’s better decorated,” “because it may be a little yummier,” “looks more appetizing”), described one item as high in quality (e.g., “because it’s really pretty”, “because it’s tasty”, “because it looks warm”), or described one item as low in quality (e.g., “his looks disgusting”, “because that one isn’t the right shape,” “that one looks kind of weird”). Explanations could be coded as relevant to the task but not about quality if they referenced handmaking but not quality (e.g., “because her mom made it,” “because he made it and it’s his own work”), a feature of the item but not in a way that was obviously high or low quality (e.g., “because it has squares”), or the participant’s own preferences rather than the character’s (e.g., “I love strawberries”). Explanations were coded as not task relevant if they did not provide any additional information (e.g., “I don’t know”, “because”). Coders first completed a small training set (n = 32 explanations) to become familiar with the coding manual and then completed an overlapping set of 612 explanations (21% of the total), achieving a kappa of 0.73. Disagreements were resolved by discussion. Coders then independently coded the remaining explanations (coder A: 1,178; coder B: 1,095). To examine whether children in Experiment 3 were especially likely to mention quality, we created a binary variable (referenced quality vs. did not) and performed a binomial generalized linear mixed model with study, child age, and child gender as predictors of each explanation. We observed a significant effect of study; children were more likely to mention quality in their explanation for Experiment 3 (n = 435, or 45% of explanations) compared to Experiment 1 (n = 338, or 34%), b = −0.53, SE = 0.19, z = −2.86, p = .004, with no significant difference from Experiment 2 (n = 437, or 46%; b = 0.10, SE = 0.18, z = 0.54, p = .587). Children were more likely to mention quality with greater age (b = 0.17, SE = 0.03, z = 5.30, p < .001), with no significant effect of child gender (b = 0.13, SE = 0.15, z = 0.86, p = .391).

Second, an additional group of children (n = 38) was given a worksheet with images of the item pairs and asked to circle “which one looks better.” The study was conducted at a local children’s museum in 2021 and was approved by the Institutional Review Board of the University of North Carolina at Greensboro (IRB-FY22–59, “Community studies at a distance”). Image pairs were counterbalanced in the same manner as in Experiment 3; example worksheet available on OSF. Overall, children selected the intended item as the nicer one (M = 5.68 out of 8, 95% CI = 5.24, 6.12), Wilcoxon signed-rank test, p < .001. To examine individual items, we performed binomial tests to examine whether children selected the intended item more often than would be predicted by chance. Children selected the intended item for all four foods (cookie: p < .001, bread: p = .036, cake: p = .003, pizza: p < .001) and for three of the four nonfoods (mitten: p = .007, drawing: p = .036, mug: p = .017, quilt: p = .928). We provide an additional analysis omitting the quilt trials (as children did not differentiate between the two quilts when asked which was better; see Table 4), which yielded similar results as the analysis of the full dataset.

Table 4.

Regression tables from Experiment 3, omitting quilt trials. Top: Children’s selection of handmade items based on maker identity (child vs. parent), item type (nonfood vs. food), child age, child gender, and maker gender. Bottom: Children’s selections of handmade items based on child age, child gender, and response explanations (socioemotional, item features).

	Estimate	SE	z	P
(Intercept)	−0.70	0.46	−1.54	.124
Maker identity	−0.98	0.21	−4.69 <	.001
Item type	< −0.01	0.20 <	−0.01	> .9
Item × Maker	1.74	0.33	5.20 <	.001
Child age	0.12	0.05	2.45	.014
Child gender	−0.40	0.28	−1.44	.149
Maker gender	0.12	0.21	0.58	.564
Child gender × Maker gender	0.07	0.32	0.21	.833
	Estimate	SE	z	P
(Intercept)	< 0.01	0.39	0.01	.993
Child age	0.12	0.05	2.53	.011
Child gender	−0.19	0.20	−0.92	.356
Socioemotional explanations	2.06	0.51	4.06	< .001
Item features explanations	−1.49	0.20	−7.61	< .001

Results

Preference Judgments

Children were not significantly more likely to select imperfect handmade items compared to factory-made items (M = 4.28 out of 8, 95% CI = 3.95, 4.61), p = .079. The regression analysis revealed a significant effect of maker, which was qualified by an Item Type × Maker interaction (ps < .001, see Table 3 for full regression results),¹ see Figure 4. To examine the interaction, a Wilcoxon signed-rank test (two-tailed) was performed to compare responses to foods and nonfoods for both child trials and parent trials. No difference was observed between foods (M = 1.07 out of 2) and nonfoods (M = 1.07 out of 2) that were made by children, p = .956. For trials in which items were made by parents, handmade items were selected more for nonfoods (M = 1.49 out of 2) than for foods (M = 0.65 out of 2), p < .001. In addition to differing from each other, both responses differed from chance (1) according to Wilcoxon signed-rank tests (nonfoods above chance, foods below chance, ps < .001).

Table 3.

Regression tables from Experiment 3. Top: Children’s selection of handmade items based on maker identity (child vs. parent), item type (nonfood vs. food), child age, child gender, and maker gender. Bottom: Children’s selections of handmade items based on child age, child gender, and response explanations (socioemotional, item features).

	Mean (95% CI)	Estimate	SE	z	P
(Intercept)		−0.71	0.42	−1.71	.087
Maker identity		−0.95	0.21	−4.63	< .001
Item type		< 0.01	0.20	< 0.01	> .9
Item × Maker	Nonfood-child: 1.07/2 (0.93, 1.21) Food-child: 1.07/2 (0.94, 1.21)	2.03	0.30	6.77	< .001
	Nonfood-parent: 1.49/2 (1.38, 1.61) Food-parent: 0.65/2 (0.52, 0.78)
Child age		0.12	0.04	2.64	.008
Child gender	Girls: 4.61/8 (4.13, 5.08) Boys: 3.85/8 (3.41, 4.29)	−0.44	0.25	−1.74	.081
Maker gender		0.21	0.19	1.07	.285
Child gender × Maker gender		0.09	0.29	0.30	.766
	Mean (95% CI)	Estimate	SE	z	P
(Intercept)		0.18	0.32	0.56	.576
Child age		0.10	0.04	2.63	.009
Child gender		−0.19	0.17	−1.17	.242
Socioemotional explanations	Yes: 0.94 (0.90, 0.99) No: 0.49 (0.46, 0.53)	2.03	0.49	4.16	< .001
Item features explanations	Yes: 0.42 (0.38, 0.46) No: 0.74 (0.70, 0.79)	−1.23	0.17	−7.15	< .001

Figure 4.

Experiment 3 (error bars represent standard errors).

There was an effect of child age (with the positive estimate, b = 0.12, indicating that the older the child, the more likely they were to select handmade items), but not child gender, p = .081.

Explanations

The number of socioemotional explanations children provided was positively correlated with child age, r(116) = .41, p < .001); item features explanations were not correlated with child age, r(116) = .07, p = .456. Examining the association between children’s explanations and their responses, a regression showed effects of explanations: When children provided socioemotional explanations, they were more likely to select handmade items in test trials (yes = .94, no = .49, p < .001), and when children provided item features explanations they were less likely to select handmade items in test trials (yes = .42, no = .74, p < .001). This analysis also replicated the effect of child age (i.e., the older the child, the more likely they were to select the handmade item, p = .009), but not child gender (p = .242).

Discussion

Experiment 3 demonstrates that children do not necessarily expect others to prefer handmade items at all costs but do still consider a variety of factors in their judgments. When considering imperfect handmade items, children expected a preference for handmade nonfoods (even though they were imperfect) and factory-made foods (compared to the imperfect handmade foods). Children’s valuations of imperfect foods made by parents were especially low; children were more likely to select the factory-made foods. Overall children do appear to focus more on quality in Experiment 3 and view the items intended to be higher in quality as better looking, but their explanations also reveal a strong tendency to consider object history. Even when we intended to manipulate visible quality (and our stimuli check highlighted consistency in children’s selection of the intended item as better looking), many of children’s explanations referenced other considerations, such as pride in one’s work or the love of a parent (nonvisible aspects of the item’s history). Children who provided socioemotional explanations were again more likely to select the handmade item, whereas children who provided item features explanations were more likely to select the factory-made item. Taken together, these results suggest that objective value may be a particularly strong consideration when evaluating foods, whereas sentimental value, such as the fact that one’s parent made an object, may be a stronger consideration when evaluating objects outside the food domain.

General Discussion

The present studies reveal children’s robust expectation that handmade items are preferred over factory-made items, an expectation that increased across development. Across studies, girls tended to demonstrate a stronger handmade preference than boys, and children’s explanations tended to predict their responses. That is, children who explained their selection in terms of socioemotional concepts were more likely to select the handmade item, whereas children who explained their selection in terms of item features were more likely to select the factory-made item. We also observed nuanced patterns of children’s consideration of foods and nonfoods depending on who made the item. In Experiment 1, children tended to value handmade items more for nonfoods than for foods (with no interaction with maker identity). In Experiment 2, this pattern was observed when considering items made by characters themselves (not by artisans), whereas in Experiment 3 this pattern was observed when considering objects made by the character’s parent (but not by the character themselves).

These findings have interesting implications for how children think about the value and importance of foods compared to other items that people can make. The high value of handmade nonfoods aligns with previous findings that children highly value items based on their history (Gelman et al., 2015; Gelman & Davidson, 2016; Hood & Bloom, 2008; Nancekivell et al., 2019), including nonfood objects they made themselves (DeJesus et al., 2019; Marsh et al., 2018, in press). Although children have been found to eat more food that they prepared themselves than foods that others prepared for them (DeJesus et al., 2019) and they consider a food’s level of preparation (e.g., whether a fruit is presented whole or cut up) when making judgments about a food’s properties (Lafraire et al., 2020), these studies did not vary food quality. Children may engage in more nuanced thinking about foods when cues to quality are available, and they may consider a variety of factors, such as their expectation of what a food might taste like or a food’s safety, in addition to who made the food. For instance, a poorly made food could indicate that it is not safe to eat or that hygienic practices were not observed while cooking. Sentimental value may play less of a role in children’s judgments of foods, whereas other considerations may be a higher priority. Conversely, object history may be more critical for non-consumable goods, as signs of wear or prior handling may actually be viewed as positives for some items, like attachment objects or celebrity possessions, where such imperfections are reminders of their special (valued) history (Gelman & Davidson, 2016; Newman & Bloom, 2014). More generally, children’s valuations of consumable goods (including foods) compared to non-consumable goods is an important topic for future research. People may have stronger expectations for consumable goods to be identical every time (e.g., whether favorite chips taste the same in every bag or a favorite face cream has the same scent in every jar), whereas if non-consumable goods are kept for longer periods of time, there may be less opportunity to compare different instances of the same item. Consumable and non-consumable goods have aspects of history that can be shared, but in slightly different ways. For foods, recipes and preparation practices can be shared through families and cultures across time. Foods can be shared in the moment with other people, whereas non-consumable objects can be held over time and accumulate value. The differences between foods and non-foods observed here could be suggestive of differences based on consumability, but future studies are needed to explore these differences in more detail.

In addition, across studies we observed persistent gender effects: Girls expected other children to prefer handmade items more than boys did, and in Experiment 1 children demonstrated a slightly stronger handmade preference when considering female makers (girl characters and mothers) compared to male makers. This is an intriguing finding, given prior studies that found no gender differences in children’s actual creative abilities or interest in creative activities (Gaither et al., 2019; Gralewski & Karwowski, 2013; Marsh et al., 2018). Nonetheless, we observed a persistent gender difference in children’s expectations about what another child might like or whose creations may be more valued. One possible explanation for the gender difference here might be related to children’s explanations. Recall that the number of socioemotional explanations predicted children’s selection of handmade items across experiments. Girls also provided more socioemotional explanations than boys (girls: M = 1.17; boys: M = 0.56; Wilcoxon signed-ranks test p < .001). Experimentally manipulating descriptions of socioemotional and item features (and comparing children’s value judgments) would be an interesting future direction to examine this question. Alternatively, this task could be tapping into gender stereotypes or gendered expectations about caregiving. Although there may be no overall difference in girls’ and boys’ ability or interest in creative activities, children may nonetheless expect that creative activities are more for girls than for boys or hold gendered expectations for specific creative activities, such as knitting, cooking, baking, and drawing as more feminine versus assembling models, working with tools, and “handicraft” as more masculine (Trautner et al., 2005). Expectations regarding parents’ roles in creative activities, particularly for “intensive mothering” (Brenton, 2017; Hays, 1996; Jennings & Brace-Govan, 2014; Moisio et al., 2004; Moscato & Machin, 2018; Thompson, 1996), may also contribute to the gender differences observed here. Additional research would be needed to further examine the development of these stereotypes, in particular to assess whether the gender differences in the present studies reflect descriptive judgments (i.e., what a girl or boy usually likes), prescriptive judgments (i.e., what a boy or a girl should like), or both.

These findings open several areas for future investigation, including the level of effort put into making an item and studies of handmade items across cultures. First, manipulating the amount of effort or processing in children’s thinking about handmade vs. store-bought or factory-made items would be worthwhile to investigate further. Although it is possible that the less processing that is required, the less of a preference would be anticipated, this is not necessarily the case, given the range of aspects of processing that could be manipulated or considered. A paper airplane may not require as much effort as a detailed drawing, but children could still value their own work and prefer playing with a paper airplane they made over a similar airplane that they did not make (see DeJesus et al., 2019, supplemental material). Considering a food example, carrot sticks can be bought pre-cut at the store or cut up at home without substantial effort. Little or no preference might be expected here – the only ingredient is a carrot (so less opportunity to tailor to someone’s specific preferences) and not much time or effort would be invested to cut it up (compared to cooking or baking). In contrast, if participants heard that a character grew the carrot themselves, that could create an opportunity to demonstrate skills, invest time and energy, or do something special for one’s family, which could in turn generate more positivity toward the handmade item. The children included in the present studies may also have little context about what it means for something to be made in a factory. Just as items made at home can vary in the amount of effort required, items made in a factory may vary widely in terms of whether a specific person can be identified as the maker, how many people contributed to the making of an object (e.g., on an assembly line), or the extent to which machines or robots contributed to the object’s creation. Considering children’s thinking about items crafted by people or autonomously by machines across settings (e.g., a parent using a 3-D printer at home, a person carving an item in a factory) would be interesting future directions.

Additionally, cross-cultural research would be valuable to understand how cultural traditions may influence children’s perceptions of handmade items (see Marsh et al., in press), including imperfect ones. The present research is limited to children in the United States, but children raised in different contexts may view handmade items differently. For example, children in Japan may be familiar with practices such as kintsugi, the Japanese art of visibly mending broken pottery with gold or precious metals, and consequently may develop a different interpretation of repaired or imperfectly handmade items compared to children in the U.S. These findings highlight important opportunities to examine children’s thinking about handmade objects as window onto their object concepts, judgments about people, cultural experiences, and the development of their economic reasoning (see also Gelman & Echelbarger, 2019).