The purpose of this study is to understand children's perceptual cognitive factors and processes during book selection. Moreover, this study aims to explore how the perceptual cognitive factors can be described in a metadata schema for children's libraries. The study contributes to the improvement of Knowledge Organization Systems (KOS) reflecting children's perceptual cognition. Current Knowledge Organization Systems (KOS) in most school and public libraries are minimally specialized for children. Most school and public libraries have been using the standard KOS, e.g., AACR2, DDC, LCSH, Sears Subject Headings. These KOS usually focus on describing characteristics of information and targeting adult user groups. However, children perceive, categorize, and name information in a different way than adults do. Due to the lack of consideration of children's cognition in KOS, children have trouble in searching for information. Therefore, it is necessary to examine how well current KOS consider children's cognition. In KOS, metadata elements correspond to conceptual aspects of information. Consequently, the study suggests how conceptual aspects of information based on children's perception can be represented in metadata elements.
The findings of this study identify 14 emergent facets of information through various factors of information during the perceptual cognitive process in book selection. These factors and facets are re-analyzed within a Knowledge Organization theoretical context. This study suggests multiple characteristics of perceptual cognitive processes: 1) two different types of processes: PAL (Paying Attention by Looking) and BAR (Being Aware by Recalling); 2) direct and indirect factors of resources; and 3) concrete or abstract factors. Lastly, this study discusses how these perceptual cognitive factors can be represented in metadata schema.
Knowledge Organization (KO) has paid less attention to children as a user group in the library environment than to adult user groups. Current Knowledge Organization Systems (KOS) in most school and public libraries are minimally specialized for children. Most school and public libraries use the same KOS such as AACR2, DDC, LSCH, or Sears subject headings, which are targeted mainly for adults. For example, the Children's and Young Adults' Cataloging (CYAC) program, which is originated from the Annotated Card program at the Library of Congress (LC), and the School Catalogue Information Service (SCIS) for Australian and New Zealand school libraries use AACR2 as the main descriptive cataloging standard.
These KOS are created in a resource-centered view rather than in consideration of children's cognition. The resource-centered view means describing characteristics of resources. KOS created with a resource-centered view cover bibliographical description, physical or technical description, subject description, and so on. These descriptions originate from resources themselves. However, previous studies for children's libraries indicate that children have different information needs and different information seeking or searching behaviors than do adults. Due to the lack of consideration of children's cognition in KOS, children have trouble searching for information (Beheshti et al., 2010; Bilal, 2000, 2001, 2002; Borgman et al, 1995; Borgman, Chignell, & Valdez, 1989; Large, Beheshit, & Rahman, 2002; Pejtersen, 1986; Sanlian, 1994; Walter, Borgman, & Hirsh, 1996).
A relatively small number of the studies about children's information seeking or searching behaviors use a cognitive approach. The term, cognition, is used in various ways in different disciplines. Cognition or cognitive process in the Library and Information Science field usually refers to cognitive factors or cognitive information processing behaviors for search relevance, effectiveness, or efficiency assessment. In terms of a cognitive approach in KO, particularly children's knowledge organization, the aspects of cognition are very confined. The topics of studies applying the cognitive approach usually deal with classification or taxonomy for designing children's web portals (Bar-Ilan et al., 2005; Beheshti et al., 2010; Bilal et al., 2005; Borgman et al., 1989; Cooper, 2002a, 2002b, 2004, 2005; Large et al., 2002, 2005, 2009). Before discussing how classification or taxonomy is related to cognitive processes, it is necessary to define the cognitive processes in KO.
This study delineates children's cognitive process into three different stages: perception, categorization, and naming. Perception means those conceptual aspects of information children are seeking through external and internal stimuli of sensory register. For example, when children look at a bottle of water, they might perceive water as liquid to drink, the water's quality, a bottle's shape, a bottle's material, or the design on a bottle. Depending on what aspects of information children perceive, the manner of organizing information can vary. Categorization indicates how children develop conceptual structures like conceptual hierarchy. Lastly, naming is tied to linguistic expression or terminology. Therefore, naming concerns what kinds of terms children use in order to express information needs.
There are many domains in KO. Each domain is represented by two main constituents; representation of information and expression of users. The representation of information consists of various KOS, e.g., metadata schemes, classification, thesaurus, controlled vocabularies, and taxonomy. On the other side, users express their information needs. As pointed out above, users' expression takes place in three types of cognitive processes. When it comes to the relationship between representation of information and expression of users, metadata elements correspond to conceptual aspects of information. Existing metadata elements tend to represent facets of information, that implies that metadata elements are able to reflect users' cognition related to perception. Secondly, results of categorization in cognitive processes can be depicted by classification, thesaurus, taxonomy, etc. Lastly, a naming process suggests that, to some extent, controlled vocabularies like Sears subject headings use child-friendly terminology.
1. Metadata Schema Creation
How is a metadata schema created? For what reasons, does a metadata schema include or exclude certain metadata elements? These questions allow us to consider certain aspects between users and resources that a metadata schema focuses on describing. There is a lack of literature addressing these questions.
This issue of metadata schema creation should be distinguished from metadata record creation. Many studies pay attention to the evaluation of the quality of metadata records (Bruce & Hillmann 2004; Greenberg et al., 2001; Ma et al., 2009; Moen & Stewart 1998; Shreeves et al., 2005; Tolosana-Calasanz et al., 2006; Wilson 2007; Zeng et al., 2004). These studies deal with metadata records created by catalogers or metadata professionals after the implementation of a metadata schema. The major limitation of studies about evaluation of metadata record creation is that extant metadata records cannot criticize non-existent metadata elements. That is, if a metadata schema does not consist of suitable metadata elements describing characteristics of resources and users' perspectives, then the evaluation of metadata records created by this metadata schema is meaningless.
2. Cognitive approaches used for the studies on children's libraries
In general, cognitive processes related to KO consist of three levels: Perception, Categorization, and Naming. A perception process and a categorization process are closely associated with each other. Categorization occurs after children perceive information. A perception process is more likely to focus on aspects of perceived information, whereas a categorization process deals with developing conceptual structures within one aspect of perceived information.
Previous studies using cognitive approaches for children's libraries tend to be concerned about categorization and naming processes rather than perception processes. The studies about categorization for children's libraries examine how children structure concepts on web directories or libraries (Bar-Ilan & Belous 2007; Bilal & Wang 2005; Cooper 2002a, 2002b, 2004; Large et al., 2007). Bar-Ilan and Belous (2007) and Bilal and Wang (2005) give children specific concepts selected by researchers, whereas Cooper (2002b) leads children to suggest what a library should include. Participants in Cooper's study provide individual words such as cat, Harry the Dirty Dog, dictionary, etc. These words do not include any relationships or conceptual structures. However, Cooper's approach helps us understand “what the children typified as important in their concept of a library” (Cooper, 2002b, p. 1225).
The studies addressing categorization are helpful to design classification, taxonomy, thesaurus, or browsing search interfaces. However, without consideration of the perceptional cognitive process, KOS cannot fully meet children's information needs.
Last but not least for cognitive processing, the children's naming process has also been studied in KO (Abbas 2005; Brown 1994; Cooper 2002a, 2002b). By examining keyword searching terms generated by children, these studies illuminate how children label or name certain concepts and analyze terminology from a linguistic viewpoint. Abbas (2005) claims that KO's improvement is not from the change of controlled vocabulary structures such as broad terms, narrow terms, or related terms, but from the change of reflections of user-generated terms. Brown's (1994) findings show that: “unsuccessful search terms were improved by changing from a compound to a simple term and from an abstract to a concrete term, and by choosing a subsequent term which named a related concept rather than a different level of specificity of the same term” (p. 348). Brown's study has strong points in terms of analyzing children's search terms by different age groups and by syntactic approach. As these studies suggest, controlled vocabularies can reflect more child-friendly terminology and syntactical structures through the analysis of children's naming process.
3. Perceptual process and children's book selection behaviors
Reuter's study (2007) is the most recent and comprehensive research study in terms of children's book selection behavior. Reuter claims that “behaviors are motivated by cognitions and cognition often results in behavior” (p. 16). As Dervin and Nilan (1986) address external behaviors and internal cognitions, Reuter's study analyzes children's book selection behaviors by actions and cognitive factors. Cognitive factors in her study are based on relevance assessment (p. 135) rather than perceptual process delineated by this study. Reuter's study (2007) finds 13 facets such as contents, reading experience, gestalt judgment, surface features, familiarity, and so on (p. 183).
The literature review of this study has addressed the lack of studies on metadata schemes considering children's perception. Studies of children's information behavior like Reuter's study (2007) can convey evidence or reasons why certain metadata elements should be created for users. Although Reuter's study (2007) suggests behavioral factors related to children's book selection, it doesn't make any connection with metadata elements. Therefore, this study aims to understand children's perceptual cognitive process in book selection by replicating Reuter's study. The findings of this study will embody the perceptual factors related to children's book selection. Furthermore, this study will discuss what the perceptual factors mean to creating metadata elements for children's libraries.
This study focuses on children's perceptional cognitive process during the interaction with physical books at the library. Reuter (2007) describes “the children's actions progressed through several steps-from shelf interaction, to external examination, to internal examination” (p. 133). In terms of external examination, children have limited information because of the way that books are shelved. Children mostly interact with shelved books at the first encounter. In other words, information that children can perceive through the shelved books is usually limited by spine titles, author's names, publishers, or series titles. These physical settings can have impact upon children's perceptional cognition.
In addition, participants are recruited from academic colleagues. In other words, either the mother or father of the participants has a Master's degree or a Ph.D. More on the background of participants will be explained in the method section. No fathers directly participated in the study.
The purpose of this study is to understand perceptual cognitive factors during children's book selection in order to interpret what perceptual factors suggest for creating metadata elements. Children's book selection behaviors have been studied by Reuter (2007). Therefore, this study aims to examine whether the findings of Reuter's study can be augmented, and to reanalyze children's behaviors in light of the perceptual cognitive process. In addition, this study considers issues between perceptual factors and metadata elements. To do this, this study has the following research questions.
1.What factors and facets do children consider during book selection?
2.How can the perceptual factors and cognitive process related to children's book selection be categorized?
3.What do the perceptual factors tell us about metadata schemes?
Generally speaking, this study replicates Reuter's study. Therefore, the study adopts data collection methods consisting of observation, interview, and diary. Although the main data collection methods are similar, the procedures of data collection are not identical. Children and their parents were required to visit the library three times during November and December, 2011. The parents of the participants scheduled the date and time for library visit for their convenience. They also decided the library to visit. Each library visit took about an hour. During the library visit, the researcher used a voice recording machine with parents' permission. Participants were given cookies as a token of thanks.
Participants: Three male children and three female children ranging from age of 6 to 9 (K-4 grade) participated in this study. Participants were recruited by a snowball sampling method in that children were recruited from academic colleagues. All participants went to the public library at least once a week. All participants can speak English and some of them can speak Korean, Chinese, or Afrikaans. One participant speaks Korean during the interview. Participants are not visually impaired.
Procedure:Figure 2 shows the procedure of data collection during the library visit and outside library setting.
Initial meeting: Before the actual data collection, the researcher and participants had an initial meeting at a place where participants preferred. The researcher introduced the procedure of the study to participants and their parents. The parents of the participants scheduled three library visits. In addition, the parents were given a notebook for diarying at an initial meeting.
Observation: Observation was undertaken with the full knowledge of the child participants and their parents. Participants were allowed to interact with books. They could either select books to check out or browse books to read them in the library. The researcher observed participants by positioning herself near book shelves where the participants browsed, so that the researcher could hear their voices. Observation included children's movement and monologue in interactions with books, and interactions with their parents, other people at the library, and the researcher. The researcher tried to be involved in children's interaction with books as little as possible. However, when a participant did not make any comments to explain why he or she picked up or looked at a book, the researcher asked questions such as “Why do you look at this book?,” “Why did you put down this book?,” or “What kind of books are you looking for?” By asking these questions at the moment rather than asking them in a later interview, not only do children have less cognitive work process for recalling memories, but also responses are more genuine and intuitive.
Interview: Interviews were conducted at three points: before the observation (Pre-interview), during the interview, and after the observation (Post-interview). The Pre-interview was to make children feel comfortable with the researcher and to understand children's information needs by asking “What kinds of books do you want to read today?” or “Do you have specific books to want to read?” During the observation, the interview was minimized not to interrupt children's interaction with books. However, the researcher encouraged participants to think aloud. The Post-interview allowed the researcher to understand children's profound thoughts or feelings toward book selection. The participants were asked to explain why they selected books.
Diaries and diary-interview with parents: The researcher gave participants' parents a notebook including prestructured questions and open-ended questions. Parents were asked to record their observations of their children's information needs and behaviors for book selection during the study. This method provides supplementary data of children's book selection behaviors outside the library setting. Parents brought a notebook to each library visit. After the children's observations and interviews, parents, children, and the researcher had a brief diary review meeting to discuss the diaries together.
I. Emergent factors and facets in children's book selection
The findings of this study show results similar to Reuter's study in general. Reuter (2007) categorized many factors related to children's book selection into 13 facets. The facets and factors reinforcing Reuter's findings are contents, genres, familiarity, difficulty, and emotional interest or arousal. Table 1 shows factors of book selections and facets categorizing the factors.
Factors are specific references or indications addressed by participants or observed by the researcher during their book selection process. Facets emerge from these factors. Facets are aspects that participants want to perceive through factors. Therefore, factors and facets are not mutually exclusive. In fact, many factors and facets are intertwined.
Because of the lower number of participants than participants in Reuter's study and the nature of a qualitative study, individual factors in the two studies are not identical. In addition, a different data coding scheme and different perspective of this study suggest slightly different facet categories. This paper does not explain all factors and facets, rather selectively articulates some factors and facets for clarification.
Table 1. Factors and facets related to children's book selection
Plot (Actual plot and Imaginary plot by readers)
Setting: When & Where
Table of content
Summary or review from back of books
Specific components or parts of contents
Titles of series
Characters of series
Subjects or themes of series
Illustrational patterns in series
Name of characters
Appearance of characters
Gender of characters
Personality of characters
Occupation of characters
Illustrations or particular icons on book covers
Characters or objects
Formats of illustrations
Texture on illustrations
Difficulty, Reading level, Understandability
# of words or pages
Words inside books
Book shelves' location in the library
Internal factors of person
External factors of books
Belief or value
Hidden letters in illustrations
Folded pages or pictures
Rhythmical narrative styles
Basic bibliographical information
Contents in non-fiction or fiction stories are one of the main facets of book selection. Participants perceive content topics or story themes through titles, summary or reviews written in the back of books, illustrations, or specific components or parts of books. For instance, participants perceive contents in non-fiction by looking at titles indicating topical terms, like “tadpoles.” However, for fiction, information on the back of books plays a role as a reference in understanding the plot of a story. In addition to text-based perception, participants tend to depend on illustrations rather than text in order to have a sense of plot or story. Imaginary plot created by readers based on illustrations may be different from the actual story. Even though the factor that participants look at is illustration, what they want to know or perceive relates to content or story matter.
Participants also pay attention to specific components or parts of books. For example, Child D likes to read books about reptiles in “Eye to Eye with Animals” series, because the books of this series visualize information about animals' lifecycles. These books are not about animal life cycles, rather they are just about animals. However, parts or specific components of contents such as animal life cycle influence participants' book selection.
Reuter (2007) considers series is the most prominent factor relating to familiarity (p. 80). Participants refer to series titles, themes of series, or character's names in series. Series titles often include a character's name like “Scoobydoo comic storybook series.” Therefore it is hard to separate a main character's name from a series title. When participants perceive series, there are two types of perceptual cognitive processes: recalling memory or discovering new information. In other words, when participants refer to factors related to series, they often recognize the factors because they have known about the series. Therefore, the factors are recalled from their prior knowledge or memory, which causes familiarity, rather than used for discovering new information. Further explanation about two different types of perceptual processes will be addressed in the second section of results.
Reuter (2007) does not scrutinize what aspects of series participants perceive so that they feel familiar with books. In this study, children recognize series through titles, characters, subjects or themes, or illustrational patterns. For example, “Eye to Eye with Animals” series has a unique icon or mark on each book. The icon is illustrated with an animal's footprint and shows the title of the series on the image. It is always located in the right corner of the book cover. These illustrational patterns make children readily able to recognize books as a series.
Participants are likely to choose or not to choose books because of characters in a book. In Reuter's study (2007), characters are not considered as a significant and independent facet of book selection. Reuter briefly mentions that characters are broadly associated with contents (p. 75). However, this study emphasizes the aspect of characters more than does Reuter's study.
There are many factors or aspects of characters that participants concern. For example, participants usually remember a character's name. Many children's series like “Angelina Ballerina” include main characters in a their titles. Participants also mention a character's appearance by stating, “He is fat.” Perception of a character's appearance has a strong relationship with the aesthetic expression of illustrations. Besides a character's name or appearance, participants also address characters' gender, personality, or occupation. All these factors assist children in envisaging characters, which is connected to familiarity or emotional interest.
Participants look at illustrations both on book covers and inside books. Illustrations mean visual or aesthetic representation. The factors of illustrations range from colors and aesthetic quality (i.e., detail or reality) to formats of illustration (i.e., drawing, cartoon, or photos).
Emotional interest or arousal of the emotions refers to psychological states or experiences of feeling, e.g. happy, sad, challenged, and interested. Emotional interest occurs either internally or through the interaction with books. In other words, internal emotional interest has been developed by self-directed curiosity on diverse factors such as topics, setting of story, or characters.
Internal emotional interest tends to suggest what kinds of books children want to read. For example, Child B shows strong interest in geography so that the cihld often explains his information needs with specific subjects such as Alaska, China, or Russia. In addition to internal emotional interest described by topics, participants also express emotional interest simply from self-directed feeling. Participants mention that they like challenging books, which means books at a higher reading level. Although participants often feel challenged after looking at the text size or density, sometimes they choose books by their feeling without any external input.
On the contrary, external stimuli come from various factors in books such as illustrations, contents, characters, or difficulty. For example, in the following interview with Child D, an illustration like a squid image is an external factor which causes a scared feeling (see transcript 1).
Transcript 1. Child D's interview on December 28, 2011
[A librarian gives her books with the squid image on a book cover]
(D: Chid D / R: Researcher)
R: Do you like this book?
D: Yes, but it sounds [looks] like scary, so I am not sure I like it
R: What makes you scared? Pictures?
D: Yes. Look. [pointing out the picture inside of the book] Can you put it back? It is little too scary.
An important point of emotional interest is that emotional interest not only plays an important role in making a decision of book selection, but also is perceived by many external factors of books and internal factors of person.
Personal connection bears on personal belief or value, hobbies, or personality. Child B believes that reading fiction makes him stupid. Therefore, Child B always chooses non-fiction. Child D also likes fairy tales because she believes in magic. Child A's hobby is writing letters. Child A likes the book with letters in the story because she can adapt words like “Dear” from the book. In this case, the content's narrative factor influences the participant to choose the book.
This study finds out two new factors: engaging elements and additional materials in a book. Engaging elements refer to when a book has elements that make children engage in interaction with the book. For example, children's book series titled “A to Z Mysteries” includes hidden letters in illustrations. A participant in the study mentions that she likes to find hidden letters in pictures (see transcript 2). In this case, hidden letters in illustrations play a role in making children engage in reading a book.
Transcript 2. Child A's interview on December 14, 2011
(A: Child A / R: Researcher)
A: Because I like reading these kinds of books and the pictures you see a word like … like right there, right there it is a T. [pointing at a hidden alphabet T in an illustration]
R: Oh ok, so you like the pictures here to find the hidden letters. Can you find it?
A: I don't know if I can find it. I can't find it.
R: Let me try. Let me see.
A: It is like a really small letter.
R: Small but it is not letter?
A: It has to be a letter.
R: Letter… . H?
A: H … [Shaking a head]
R: No? What about this?
A: Yeah. It's A.
In addition, through the latent analysis, this study reveals that participants show strong engagement in reading, when they can interact with their mom or the researcher. For instance, factors like rhythmical narrative style encourage participants to interact with others.
II. How can children's perceptual aspects in book selection be categorized?
Categorization of the perceptual aspect by two components in KO
There are two components in KO: resources and users (See Figure 1). This study also views children's perceptual cognitive processes within a context of KO. Facets identified in previous sections are categorized into three parts: resource-centered, user-centered, and a combination of resource and user (See Table 2 and three categories).
First, facets in a resource-centered category have more to do with more resources per se (Section A in Table 2). Factors inducing facets in a resource-centered section tend to be driven from elements of books. Section C includes facets having more strong association with users rather than resources. Participants perceive some factors out of books or more self-directed factors to choose books. For example of personal connection, a hobby of Child A does not originate from a book. A hobby itself is connected to the participant, whereas a narrative style of book is related to a book, which is resource-centered view. Lastly, some facets are affected by factors of both resources and users (Section B in Table 2). Emotional interest is basically considered to be closer to a user-centered view. However, some emotional interest occurs because of factors deriving from resources. In this case, both user and resource perspectives are combined and influence perception.
Table 2. Categorization of perceptual aspect by two components in KO
Multiple characteristics of perceptual cognitive process
Paying attention by looking (PAL) and being aware by recalling (BAR)
Direct and indirect relationship with books
Concrete/visual factors and abstract factors
Participants select books either by paying attention or by being aware of factors related to book selection. General concepts of these two processes come from an information processing model of memory (Ornridm, 2012). Figure 3 shows participants' perceptual cognitive process in book selection. Figure 3 is modified from the dual store model of memory suggested by Ornridm (2012, p. 160).
There are many sensory inputs when participants interact with books. During book selection, participants receive usually visual inputs by looking or reading and store them in the sensory memory. However, only some elements of books are selectively perceived and moved to working memory in which active thinking occurs. To do so, participants need to pay attention to information stored in the sensory memory. This study defines this first cognitive process as PAL (Paying Attention by Looking).
The other cognitive process is awareness of factors. To be aware of some factors, previous knowledge or experiences play a role as a cue. For instance, factors referring to a familiarity are highly associated with prior knowledge or experience. Many participants want to re-read the same books or other books in the same series. In this case, participant's perceptual process occurs through recalling memory. When participants recall their prior knowledge about books, they mention titles, series, or a character's name, or they describe a theme or plot of the story. Therefore, this study also defines the second cognitive process as BAR (Being Awareness by Recalling).
The factors identified in this study can be considered by the relationship with books. Some factors have direct relationship with books, whereas others have indirect relationship with books. Direct factors from books mean that factors come from books or have a strong relationship with books. They range from content, characters, and illustration to basic bibliographical information such as author, title, or language. On the other hand, indirect factors do not originate from books. Indirect factors have external referents out of books. For example, Child F recognized a book because of its illustration on a book cover. The book cover has an image of the character Iron Man. Child F had played with an Iron Man jigsaw puzzle, and this experience allows Child F to feel familiar with the book depicting the same character. In this case, a character on the book is a factor that will have a direct impact on the participant's book selection, whereas prior experience of playing with a jigsaw puzzle of Iron Man serves as a mediator to indirectly connect the participant with the book.
Lastly, the factors are either concrete or abstract. Participants perceive concrete or visual factors by usually looking or reading. However, abstract factors are perceived by feeling or thinking. That is, concrete factors appear on the book front, back covers, or inside of books. For instance, additional materials like a map in a travel guide book are concrete factors, where emotional interest is abstract. However, this distinction between concrete or visual factors and abstract factors is not decided by dichotomy. Some perceptual factors of books can be considered both in concrete and abstract ways, depending on different types of perceptual cognitive process. When participants read the back of a book in order to understand the story of a book, content is more likely to be perceived in a concrete way. However, when Child A makes up a story by herself based on illustrations or wants to read a book taking place during Christmas, content is perceived in an abstract way.
Reuter's findings are addressed by prominences. Prioritizing perceptual factors does not embody in creating metadata elements, but enables a browsing search interface to display more important aspects of books in a limited interface space. Given that the ultimate purpose of this study is to consider how the perceptual cognitive factors can be reflected in metadata elements for children's libraries, factor's prominences are not examined in this study.
Perceptual cognitive processes seem to differ with various factors. Age and ability to read text influence children's perception during book selection. Older participants who have higher skills to read more often depend on abstract perceptual factors, whereas younger participants are likely to look at concrete perceptual factors such as book covers, characters, or inside illustrations of books. In addition, older participants tend to have more reading experiences, so that they have more perceptual factors serving as cues to retrieve their memory through the BAR process.
Two different cognitive processes between PAL and BAR, suggest slightly different perceptual factors. During the PAL process, concrete or visual factors such as illustrations or the number of pages, text size and density are more considered. However, during the BAR process, participants use information stored in their sensory memory. When long-term memory is stored, people first encode information. How they encode information indicates what aspect of information is perceived. Participants usually recall three factors: series titles, themes of series, or characters' names in order to refer to book series. It is hard to exactly distinguish which factor among three factors participants perceive to refer to series, because often series titles are character's names or imply theme. However, it is clear that participants perceive series information by various factors through either PAL or BAR process.
Moreover, depending on fiction or non-fiction, participants perceive different factors during book selection. For example, in order to understand content in non-fiction, participants consider the table of contents more than information in the back of books. On the other hand, participants read the summary of short description written in the back of a book to understand a story or theme in fiction.
Regardless of the variation and ambiguity, it is more important to consider all factors perceived during the both the PAL and BAR processes for the sake of creating child-appropriate metadata elements.
As Figure 1 shows, the perceptual cognitive process relates to metadata schemes. The first part of the results identifies the emergent factors and facets of the perceptual process in children's book selection. The second part of the results explains multiple characteristics of the perceptual cognitive process in book selection by gathering data on how children perceive factors and facets. In terms of creating a metadata schema for children's libraries or web portals, these findings suggest what aspects of information need to be provided as information access points.
Table 2 suggests that participants perceive not only resource-centered factors, but also different aspects of information that traditional metadata schemes provide. Traditional metadata schemes focus on describing bibliographical information such as title, author, publication, or subjects, which is based on resource-centered view. On the other hand, the user-centered view emphasizes users' perceptual cognitive process. Metadata schemes developed from a user-centered perspective need to provide access points reflecting factors in section B and C of Table 2.
A resource-centered view tends to provide information directly related to resources. However, the finding of this study raises a question about how well current metadata schemes for children's libraries describe factors in light of resource-centered perspective. Section A in Table 2 includes factors derived from elements of books. However, these factors have been perceived by participants. Therefore, these factors cover more information beyond that which metadata schemes describe. For instance, with AACR2, school libraries are not able to describe detailed information about characters. AACR2 is limited to the character's name in the form of a subject heading like “Potter, Harry (Fictitious character).” However, participants in this study perceive various aspects of information such as characters' gender, appearance, personality, or occupation rather than only characters' names. Los Angeles Public Library's (LAPL) traditional catalog system (not the new version) partially provides character information including name, gender and occupation. However, this information is created by a third party of the library industry, not by catalogers from LAPL. In addition the information is not available for keyword searching or browsing, which may mean that there are no independent metadata elements for specifically describing character information. The importance of having independent metadata elements in a relationship with information retrieval is also addressed in Beak and Olson's study (2011). By creating various and independent metadata elements, information is more likely to be categorized by its specific facets. Children's libraries readily utilize a browsing search interface by providing access points like a character's gender or a book cover's color. It prevents information that does not fit into existing metadata elements from being described in a note area. Information described in a general note area like 500 field in MARC is searchable by keywords, but not effectively searchable in a browsing search. Given that many studies claim that children prefer browsing to keyword searching, creating independent metadata elements in light of users' perceptual cognitive processes will increase recall and precision of retrieved results.
The purpose of this study is to understand children's perceptual cognitive factors and processes during book selection. Moreover, this study aims to explore how the perceptual cognitive factors can be described in a metadata schema for children's libraries.
First, this study has identified perceptual factors related to children's book selection and grouped the factors by emerging facets. Second, these factors are reanalyzed by two components of knowledge organization and perceptual cognitive process. The study finds that children's perceptual cognitive processes in book selection include resource-centered factors, user-centered factors, and factors falling between resource-centered and user-centered information. These factors are differently perceived depending on multiple characteristics of perceptual cognitive process.
The study suggests that these perceptual factors need to be considered when it comes to creating a metadata schema for children's libraries. Factors imply access points in a metadata schema, which suggests perception-friendly categories in browsing. Consequently, this study contributes to the improvement of KOS reflecting children's cognition.
This study does not examine the differences between age, gender, or book genres. In addition, participants' interaction with physical books was observed in a physical library setting rather than interaction with library catalog or ebooks. These different environments may cause different perceptual factors. Therefore, a future study needs to consider these variables. In addition, this study does not suggest how to implement the findings in practice. Therefore, a future study requires operationalizing children's perceptual cognitive factors in a metadata schema and tests its effectiveness by children.
The author would like to thank Dr. Richard P. Smiraglia at University of Wisconsin-Milwaukee. Dr. Smiraglia has been incredible with his mentorship and has offered great guidance and comments. Additionally I would like to thank all participants in this study.