CCBC-Net Archives
Identity, Diversity, Multicultralism, Forecasting and Recommendatiosn
- Contemporary messages sorted: [ by date ] [ by subject ] [ by author ]
From: Charles Bayless <charles.bayless_at_ttmd.com>
Date: Wed, 12 Feb 2014 13:48:26 -0500
So following up on how people choose to identify themselves. I have been wrestling with what we mean in this conversation in terms of identity, diversity and multiculturalism and how that is related to the books we choose to read and the books we recommend to others.
To test the concept of identity and diversity, I looked at my own collection of readings across my various fields of interest which include history, economics, sociology; archaeology, travel, poetry, and mysteries. (I note there is very little literary fiction and that identity distributions might vary significantly by field). Without going into all the details the upshot is that women are underrepresented with only about 15-30% (by field) of what I read being by women - however, that is not because of selection on my part but because of their significant underrepresentation in those fields. When I look at my favorites in global economic development histories all five are white male, one being British, the rest American. Not particularly diverse. But when I look at Research Economists in general for whom I have the largest collection of writings it is 1 African-American, 1 Transgender, 1 19th century French and 1 20th century Austrian - can't get more multicultural and diverse than that. When I look at all historians in my collection, it is about 60% European, 30% American, 10% Asian. As a test this is obviously limited as an atypical sample of one. But it is also illuminating. Even though clearly across all fields of interests, there is a very large degree of multiculturalism and diversity, it is also true that women are underrepresented because they aren't writing as much in those fields. And if it were to be judged solely on number of African American and Hispanic authors (of which there are only one of each, both in economics) then it is incredibly non-diverse. So is it diverse and multicultural? Yes. Is it non-diverse? Yes. It depends on your preferred definition.
There seems to be a strong supposition that we should try and define people by census categories and also an inference that there is some correlation between those categories and what people are likely to be interested in reading. I haven't seen any robust research that establishes a strong correlation between census identity and reading interests. On the other hand, this research calls into question the premise that self-identity is strongly established in physical attributes: http://www.ncbi.nlm.nih.gov/pubmed/24503450?dopt=Abstract Their findings are that people self-identify based on behavior and moral decision-making far more than on physical attributes. That is consistent with research in networking effects where people tend to associate (association being a variant of self-identity) most strongly by shared experience and interests.
I suspect that we have allowed ourselves to be captured by history. Post 1965 with civil rights legislation of various sorts, it became important to measure compliance but what was conceptually easy, turned out to be challenging in practice. In 1980 census, Asian Indians were counted as White but in 1990 that was changed to Asian. A woman emigrating from Spain is White but if her parents emigrated to Argentina where she was raised and then she immigrated the US as an adult then she would be counted as Hispanic. The census keeps getting more and more detailed trying to keep up with an increasingly diverse America. But trying to stay within the straight jacket of census terms does a disservice to individuals. Census terms are useful to an extent but it is a limited extent. It made sense at one time but I suspect it is clouding the issue today.
I think we have three things occurring. First, I suspect that how people actually identify themselves is different from how we assume they identify themselves, and that their self-identity varies over time and varies by context and by circumstance. I have teased this out on a slideshare document What Do Identity, Diversity and Multiculturalism Mean?
(http://www.slideshare.net/ThroughtheMagicDoor/definitional-diversity-311313 99). I don't think you can see the notes in slideshare very well, so here they are for each slide.
Slide 2: Possible Generic Framework - What are the attributes by which individuals electively self-identify? It is hard to determine because it is often context specific, variable by circumstance, and varies over time. More fundamentally, I have not seen much direct research that allows people to freely determine their self-identifiers. Listed are some common attributes. Many potential self-identifications are omitted from this list including neighborhood, interests, hobbies, employer, education attainment, education institution, sports, music, books, instrument, etc. It is particularly unclear as to what are the attributes by which children self-identify. It is not clear that they correspond with those attributes important to adults. Even within these attributes, people often define attributes somewhat differently. For example: is class SES (social economic status), income, manners, or something else? At what level is Religion defined? Example: Christian, Protestant, or Methodist?
Slide 3: Ordinal Ranking Issue - Even if we nail down which attributes an individual wishes to use for self-identification and also nail down exactly what we mean by each attribute, the bigger issue is different ordinal rankings. Different people may identity themselves primarily by race or religion or nationality, etc. Someone early in this conversation pointed out the absence of religion in children's books (and how we identify ourselves) even though Pew recently reported that 58% of Americans identify themselves as very religious. By which attributes does a person choose to identify themselves and in what order?
Slide 4: Trade-Off Ranking Issue - Even if we nail down the ordinal ranking, there is the weighting that can differ. For one person, their number one attribute of identity might be their race. Not only that but it is very heavily weighted, say 70% of their self-identity is wrapped up in their race. A second person may have Profession as her number one ordinal form of identity but only invest 20% of her self-identity in that attribute.
Slide 5: Variability by Age Issue - There is the further complication that how we identify ourselves to ourselves and to others changes over time. When I am young, I may have a sharply limited self-concept. As a young adult, I may strongly affiliate with my university or sorority. As a mature adult, perhaps now by profession or employer. As a retiree, perhaps by hobby.
Slide 6: Variability by Context Issue - There is the further complication that how we identify ourselves to ourselves and to others changes by context. In my home town I might identify myself by the neighborhood where I live but when I am travelling internationally I might identify myself by my nationality or state.
Slide 7: Variability by Circumstance Issue - There is the complication that how we identify ourselves to ourselves and to others changes by circumstance. When I feel threatened, I am may identify more with attributes that reassure me (family or religion for example).
Slide 8: Identity and Books - So given the issue of attributes, definitions, ordinal ranking, circumstantial variability, contextual variability, and variability over time, What are the pertinent elements in a book with which I might identify? It raises the importance of knowing your target when making recommendations. Variability probably explains why most people identify sources of recommendations as being those closest to them
(family, friends, teachers, etc.). It may also explains why recommendations are hard and often hit/or miss. A librarian making a recommendation has to have an almost impossible level of nuanced knowledge of the child and circumstance. There are likely two sources of mismatch - 1) How I define myself may blind me to how others define themselves and 2) How I define myself may blind me to what others will find interesting.
Slide 9-10: Identity and Recommendations.
Second, I think we individually are not fully aware as to how our own self-identification affects how we assume other people self-identify themselves. For example, if I prioritize my familial status as how I self-identify myself (single, married, divorced, with children, with grandchildren, etc.), I likely assume other people similarly self-prioritize on familial status when in fact they may self-identify more with their profession (or some other attribute).
Third, I suspect that even when we correctly understand how others might self-identify themselves differently than we do ourselves, that we fail to adequately take that into account when we forecast what others might enjoy in terms of reading. Is a middle aged female teacher who loves literary fiction going to recommend books that young boys are really likely to enjoy or is she recommending those books closest to her own self-identity that are also most likely to appeal to boys. In other words, she is recommending from the small intersection of interests and identities rather than from the full Venn diagram of what the boys are interested in. Sure, boys are likely to enjoy Little House on the Prairie, but maybe they would enjoy 60 Seconds Over Tokyo far more.
To make it particular: which description is likely the best predictor of an individual's reading interests: 1) Black, Female, Heterosexual, (Census type categories) or 2) Research Scientist in Chemistry, Married mother of two, Baptist (in other words profession, familial status and religion). Both sets of descriptors might apply to one and the same person. Perhaps she chooses to self-identify primarily by census categories as in 1 or perhaps the more adjectival identities as in 2. Whichever she chooses to self-identify by, which one then gives some third-party a better bead on what she might most like to read? I have no research that suggests one or the other. Logic and statistics dictates that the large categories such as the census are typically less predictive simply because they are so encompassing (i.e. they generally have greater variability than a narrowly defined category).
So do we take into account how people wish to self-identify, recognizing that that is likely always in flux, or do we attempt to shoe-horn them into already specified categories with which they may or may not affiliate? I would argue that it is much more insightful to take heed of the ways in which people actually choose to self-identify rather than force fit them.
I think this conflict between generic categories and self-chosen specific attributes is the genesis of the complaint that publishers are not serving up the books people would like to see.
If I am a publisher, I have pretty cheap access to census based data so I can target promotions and marketing budgets based on census categories. However, the Census averages for Black, Female, Heterosexual do not look anything like the averages for Research Chemist, Married Mother of two and Baptist. Take income and education attainment. All you can conclude from Identity 1 is average income of about $17,000 and High School graduate. If I am a publisher, that is not a promising market. But Identity 2 tells me average household income is likely to be >$100,000 and education attainment of at least a master's degree and likely a PhD. OK - that's a market I can invest in. But getting that granular information is expensive if even feasible.
I think what is happening is that publishers are probably not able to get the information that would allow them to access a market that likely exists but is hidden behind generalities. Those that are demanding such books can see that the demand is real and are perplexed that there are not the books to meet that demand, i.e. books in which race and gender etc. are likely to be present but are not the basis for didactic messages.
So I think this leads us to a problem definition: Publishers cannot cheaply and easily identify and connect with very specialized reading interests which may be small in terms of percent of market but are real in terms of absolute dollars. Conversely, groups with highly specific interests, cannot easily signal via their purchasing patterns their interest in particular types of books. Right now, that inability to connect is the slough of despond that needs to be bridged.
Charles
Date: Wed, 12 Feb 2014 13:48:26 -0500
So following up on how people choose to identify themselves. I have been wrestling with what we mean in this conversation in terms of identity, diversity and multiculturalism and how that is related to the books we choose to read and the books we recommend to others.
To test the concept of identity and diversity, I looked at my own collection of readings across my various fields of interest which include history, economics, sociology; archaeology, travel, poetry, and mysteries. (I note there is very little literary fiction and that identity distributions might vary significantly by field). Without going into all the details the upshot is that women are underrepresented with only about 15-30% (by field) of what I read being by women - however, that is not because of selection on my part but because of their significant underrepresentation in those fields. When I look at my favorites in global economic development histories all five are white male, one being British, the rest American. Not particularly diverse. But when I look at Research Economists in general for whom I have the largest collection of writings it is 1 African-American, 1 Transgender, 1 19th century French and 1 20th century Austrian - can't get more multicultural and diverse than that. When I look at all historians in my collection, it is about 60% European, 30% American, 10% Asian. As a test this is obviously limited as an atypical sample of one. But it is also illuminating. Even though clearly across all fields of interests, there is a very large degree of multiculturalism and diversity, it is also true that women are underrepresented because they aren't writing as much in those fields. And if it were to be judged solely on number of African American and Hispanic authors (of which there are only one of each, both in economics) then it is incredibly non-diverse. So is it diverse and multicultural? Yes. Is it non-diverse? Yes. It depends on your preferred definition.
There seems to be a strong supposition that we should try and define people by census categories and also an inference that there is some correlation between those categories and what people are likely to be interested in reading. I haven't seen any robust research that establishes a strong correlation between census identity and reading interests. On the other hand, this research calls into question the premise that self-identity is strongly established in physical attributes: http://www.ncbi.nlm.nih.gov/pubmed/24503450?dopt=Abstract Their findings are that people self-identify based on behavior and moral decision-making far more than on physical attributes. That is consistent with research in networking effects where people tend to associate (association being a variant of self-identity) most strongly by shared experience and interests.
I suspect that we have allowed ourselves to be captured by history. Post 1965 with civil rights legislation of various sorts, it became important to measure compliance but what was conceptually easy, turned out to be challenging in practice. In 1980 census, Asian Indians were counted as White but in 1990 that was changed to Asian. A woman emigrating from Spain is White but if her parents emigrated to Argentina where she was raised and then she immigrated the US as an adult then she would be counted as Hispanic. The census keeps getting more and more detailed trying to keep up with an increasingly diverse America. But trying to stay within the straight jacket of census terms does a disservice to individuals. Census terms are useful to an extent but it is a limited extent. It made sense at one time but I suspect it is clouding the issue today.
I think we have three things occurring. First, I suspect that how people actually identify themselves is different from how we assume they identify themselves, and that their self-identity varies over time and varies by context and by circumstance. I have teased this out on a slideshare document What Do Identity, Diversity and Multiculturalism Mean?
(http://www.slideshare.net/ThroughtheMagicDoor/definitional-diversity-311313 99). I don't think you can see the notes in slideshare very well, so here they are for each slide.
Slide 2: Possible Generic Framework - What are the attributes by which individuals electively self-identify? It is hard to determine because it is often context specific, variable by circumstance, and varies over time. More fundamentally, I have not seen much direct research that allows people to freely determine their self-identifiers. Listed are some common attributes. Many potential self-identifications are omitted from this list including neighborhood, interests, hobbies, employer, education attainment, education institution, sports, music, books, instrument, etc. It is particularly unclear as to what are the attributes by which children self-identify. It is not clear that they correspond with those attributes important to adults. Even within these attributes, people often define attributes somewhat differently. For example: is class SES (social economic status), income, manners, or something else? At what level is Religion defined? Example: Christian, Protestant, or Methodist?
Slide 3: Ordinal Ranking Issue - Even if we nail down which attributes an individual wishes to use for self-identification and also nail down exactly what we mean by each attribute, the bigger issue is different ordinal rankings. Different people may identity themselves primarily by race or religion or nationality, etc. Someone early in this conversation pointed out the absence of religion in children's books (and how we identify ourselves) even though Pew recently reported that 58% of Americans identify themselves as very religious. By which attributes does a person choose to identify themselves and in what order?
Slide 4: Trade-Off Ranking Issue - Even if we nail down the ordinal ranking, there is the weighting that can differ. For one person, their number one attribute of identity might be their race. Not only that but it is very heavily weighted, say 70% of their self-identity is wrapped up in their race. A second person may have Profession as her number one ordinal form of identity but only invest 20% of her self-identity in that attribute.
Slide 5: Variability by Age Issue - There is the further complication that how we identify ourselves to ourselves and to others changes over time. When I am young, I may have a sharply limited self-concept. As a young adult, I may strongly affiliate with my university or sorority. As a mature adult, perhaps now by profession or employer. As a retiree, perhaps by hobby.
Slide 6: Variability by Context Issue - There is the further complication that how we identify ourselves to ourselves and to others changes by context. In my home town I might identify myself by the neighborhood where I live but when I am travelling internationally I might identify myself by my nationality or state.
Slide 7: Variability by Circumstance Issue - There is the complication that how we identify ourselves to ourselves and to others changes by circumstance. When I feel threatened, I am may identify more with attributes that reassure me (family or religion for example).
Slide 8: Identity and Books - So given the issue of attributes, definitions, ordinal ranking, circumstantial variability, contextual variability, and variability over time, What are the pertinent elements in a book with which I might identify? It raises the importance of knowing your target when making recommendations. Variability probably explains why most people identify sources of recommendations as being those closest to them
(family, friends, teachers, etc.). It may also explains why recommendations are hard and often hit/or miss. A librarian making a recommendation has to have an almost impossible level of nuanced knowledge of the child and circumstance. There are likely two sources of mismatch - 1) How I define myself may blind me to how others define themselves and 2) How I define myself may blind me to what others will find interesting.
Slide 9-10: Identity and Recommendations.
Second, I think we individually are not fully aware as to how our own self-identification affects how we assume other people self-identify themselves. For example, if I prioritize my familial status as how I self-identify myself (single, married, divorced, with children, with grandchildren, etc.), I likely assume other people similarly self-prioritize on familial status when in fact they may self-identify more with their profession (or some other attribute).
Third, I suspect that even when we correctly understand how others might self-identify themselves differently than we do ourselves, that we fail to adequately take that into account when we forecast what others might enjoy in terms of reading. Is a middle aged female teacher who loves literary fiction going to recommend books that young boys are really likely to enjoy or is she recommending those books closest to her own self-identity that are also most likely to appeal to boys. In other words, she is recommending from the small intersection of interests and identities rather than from the full Venn diagram of what the boys are interested in. Sure, boys are likely to enjoy Little House on the Prairie, but maybe they would enjoy 60 Seconds Over Tokyo far more.
To make it particular: which description is likely the best predictor of an individual's reading interests: 1) Black, Female, Heterosexual, (Census type categories) or 2) Research Scientist in Chemistry, Married mother of two, Baptist (in other words profession, familial status and religion). Both sets of descriptors might apply to one and the same person. Perhaps she chooses to self-identify primarily by census categories as in 1 or perhaps the more adjectival identities as in 2. Whichever she chooses to self-identify by, which one then gives some third-party a better bead on what she might most like to read? I have no research that suggests one or the other. Logic and statistics dictates that the large categories such as the census are typically less predictive simply because they are so encompassing (i.e. they generally have greater variability than a narrowly defined category).
So do we take into account how people wish to self-identify, recognizing that that is likely always in flux, or do we attempt to shoe-horn them into already specified categories with which they may or may not affiliate? I would argue that it is much more insightful to take heed of the ways in which people actually choose to self-identify rather than force fit them.
I think this conflict between generic categories and self-chosen specific attributes is the genesis of the complaint that publishers are not serving up the books people would like to see.
If I am a publisher, I have pretty cheap access to census based data so I can target promotions and marketing budgets based on census categories. However, the Census averages for Black, Female, Heterosexual do not look anything like the averages for Research Chemist, Married Mother of two and Baptist. Take income and education attainment. All you can conclude from Identity 1 is average income of about $17,000 and High School graduate. If I am a publisher, that is not a promising market. But Identity 2 tells me average household income is likely to be >$100,000 and education attainment of at least a master's degree and likely a PhD. OK - that's a market I can invest in. But getting that granular information is expensive if even feasible.
I think what is happening is that publishers are probably not able to get the information that would allow them to access a market that likely exists but is hidden behind generalities. Those that are demanding such books can see that the demand is real and are perplexed that there are not the books to meet that demand, i.e. books in which race and gender etc. are likely to be present but are not the basis for didactic messages.
So I think this leads us to a problem definition: Publishers cannot cheaply and easily identify and connect with very specialized reading interests which may be small in terms of percent of market but are real in terms of absolute dollars. Conversely, groups with highly specific interests, cannot easily signal via their purchasing patterns their interest in particular types of books. Right now, that inability to connect is the slough of despond that needs to be bridged.
Charles
---
You are currently subscribed to ccbc-net as: ccbc-archive_at_post.education.wisc.edu.
To post to the list, send message to: ccbc-net_at_lists.wisc.edu
To receive messages in digest format, send a message to...
ccbc-net-request_at_lists.wisc.edu
...and include only this command in the body of the message:
set ccbc-net digest
CCBC-Net Archives
The CCBC-Net archives are available to all CCBC-Net listserv members. The archives are organized by month and year. A list of discussion topics (including month/year) is available at http://www.education.wisc.edu/ccbc/ccbcnet/archives.asp
To access the archives, go to:
http://ccbc.education.wisc.edu/ccbc-net
and enter the following:
username: ccbc-net
password: Look4Posts
Received on Wed 12 Feb 2014 12:49:38 PM CST
