10/21: Anjali Adukia

[shevajia]

Comment below with questions or thoughts about the reading for this week's workshop.

Please make your comments by Wednesday 11:59 PM, and upvote at least five of your peers' comments on Thursday prior to the workshop. You need to use 'thumbs-up' for your reactions to count towards 'top comments,' but you can use other emojis on top of the thumbs up.

[Thiyaghessan]

Hi Professor Adukia,

Thank you for taking the time to share your work with us. I appreciate how succinctly the core arguments and methodology were presented. I was wondering if you had considered a few avenues for future work:

1. Occupation as an additional variable a. I am wondering if it would be a valuable area for future research to look at the sorts of occupations individuals of differing racial and gender identities are shown holding in these books. In particular, if such representations are changing over time. Image recognition in this circumstance might be challenging but you could use text analysis to see if gender and racial minorities are associated with specific occupations. i. For example, in Singapore books used in schools are mandated to include a diversity of racial groups. However, inequality perpetuates itself in the form of occupations members of racial minorities hold. For example, Indians and Malays will tend to be shown holding jobs that are considered less “prestigious” or associated with lower socioeconomic status in society

2. Gender roles and actions undertaken a. I am also wondering if in the text analysis, we could see if the adjectives or verbs used to describe male/female presenting characters in these books are systematically different i. For example, are male characters associated with more aggressive actions than female characters? b. Additionally, if the prevalence of male-presenting characters engaging in activities not commonly associated with their gender identification increasing over time and vice versa i. For example, a male character engaging in “cooking” or “cleaning”, which are not the types of actions men have been traditionally associated with ii. Of course I am wondering if this could be done via image analysis as you have done, but it also seems that image analysis would be much more challenging

[YLHan97]

Hi Professor Adukia, Thank you so much for sharing such interesting topic! I think it's a very valuable topic to relate how children learn about social norms to race and gender. Education has long been an area of discussion and research by researchers all over the world. The viewpoints in the article also brought me many new inspirations.

[nswxin]

Hi Professor Adukia,

Thank you for taking the time to share your work with us. I appreciate how succinctly the core arguments and methodology were presented. I was wondering if you had considered a few avenues for future work:

1. Occupation as an additional variable a. I am wondering if it would be a valuable area for future research to look at the sorts of occupations individuals of differing racial and gender identities are shown holding in these books. In particular, if such representations are changing over time. Image recognition in this circumstance might be challenging but you could use text analysis to see if gender and racial minorities are associated with specific occupations. i. For example, in Singapore books used in schools are mandated to include a diversity of racial groups. However, inequality perpetuates itself in the form of occupations members of racial minorities hold. For example, Indians and Malays will tend to be shown holding jobs that are considered less “prestigious” or associated with lower socioeconomic status in society
2. Gender roles and actions undertaken a. I am also wondering if in the text analysis, we could see if the adjectives or verbs used to describe male/female presenting characters in these books are systematically different i. For example, are male characters associated with more aggressive actions than female characters? b. Additionally, if the prevalence of male-presenting characters engaging in activities not commonly associated with their gender identification increasing over time and vice versa i. For example, a male character engaging in “cooking” or “cleaning”, which are not the types of actions men have been traditionally associated with ii. Of course I am wondering if this could be done via image analysis as you have done, but it also seems that image analysis would be much more challenging

COOOOOOOOOL

[nswxin]

Dear Professor Adukia, I find your research super exciting and your discovery meaningful. You mentioned that Black and Latinx people are underrepresented in the mainstream book collection for books collected had more characters with lighter skins. Do you think such underrepresentation comes from the writing (so we need to encourage more writers writing for the minority groups), publishing or the education institutions who select the books? In addition, you mentioned that in recent years, there exists a favorable trend, characters of dark skins become more frequent in books. Thus, is it possible that the current situation of light-dominance is merely an accumulation of the past and we should be more optimistic about the future?

[nijingwen]

hi, professor Education inequality is an important factor that cause further inequality. There are many reasons may cause education inequality: income, parents' education level, and race etc. It is necessary for us to think about how to make the education distributes more equally which can help the social stable status and also reduce the crime rate. Looking forwarding to seeing you on Friday.

[pranathiiyer]

Hi Professor Adukia! Thank you so much for sharing your work, which is an exhaustive and informative piece of research. While I think some of my technical queries would be clear after your session on Thursday, I was wondering more about the broader aspects of this research.

1. Do you think this study can be used in the larger context of policy, if so, how? I ask this because governments do realize how influential children's books can be in moulding opinions on policy, religion race etc. For instance, the Indian government has been heavily modifying contents of History and Political Science textbooks to erase representation and origins of certain minorities all together, and this is a huge matter of concern. And in this sense, I feel this study could be extremely relevant, but can't articulate how.
2. I know you say that the 'optimal' level of representation is out of the scope of this paper. However, I was wondering if and how this could be achieved, and what variables such a process of understanding the optimal level, might be a function of?

[DehongUChi]

Hi Professor Adukia! Thanks for sharing your work. My question is: how much of your findings are correlated with, or can be attributed to, the race/ethnicity/gender of the author. For example, do Asian authors tend to have more Asian characters in the book? Would you consider adding the race/ethnicity/gender of the author as a factor in future studies? Thank you!

[ZHE-ZHANG-0213]

Hi Professor Adukia! Thank you for sharing your work. My question about your research is: I think people tend to describe individuals they know well, so whether your findings are related to the different levels of education of people of different colors in history. For the above reasons, there has always been a small number of black writers (in the past, the right of black people to receive an education was not fully reflected), which led to the characters in the book being flooded with light skin? I look forward to seeing you on Thursday!

[MkramerPsych]

Dr. Adukia,

Thank you for sharing your research with us! I think this is an excellent application of machine learning methods to tackle an important issue in underrepresentation.

I am particularly curious as to what you believe the advantages are to using these computational methods over traditional methods in this type of work. It seems to me that one could come to the same conclusions about the underrepresentation of BIPOC characters in literature without the use of natural language processing or computer vision techniques. Beyond the ability to demonstrate these inequities in specific units, what are the other potential benefits of conducting these types of studies using machine learning? Are there questions that are specifically suited to these methods? Do the benefits of these techniques outweigh the potential costs (computation time, model interpretability, additional training time for researchers)?

[JadeBenson]

Thank you for this fascinating and important research!

I'd be curious if you can identify who is speaking in children's books and how they view the other characters in the story? Could we identify the characters that are the heroes or protagonists and how they speak or interact with other men/women in the text? There may be different ways of interacting depending on perspective and who is seen as desirable/a role model in the story. I know many children's movies include queer or ethnically coded villains which is somewhat representation but perhaps not what we would desire. Villains might also rely on stereotypes to signal that they're bad. Relatedly, heroes might change the way they speak/interact with other characters. For example, in the past stories about a male knight searching for a princess might often mention women but perhaps not in a way that's as empowering or equal as contemporary stories.

[jiehanL]

Hi Prof. Adukia, thank you for sharing this exciting work! In terms of further work, I am curious about the sentiment assigned to characters with different skin colors. For instance, as JadeBensen mentioned above, we could expect a generally different sentiment tone in depicting/describe villains (mostly negative) and heroes (mostly positive), could it be possible that sentiment assigned to different colors character also varied systematically?

[tangn121]

Hi Professor Anjali Adukia, thank you for attending our workshop and sharing your work with us. I like your point on studying the race and gender problems in children's books and think it is very important because they are the next generation. I have a question about the connection between the author and the result. I am wondering if more characters with lighter skin and males have something to do with the author of the book? It might because the majority of authors are white males, so they tend to use characters similar to themselves. If this is the case, could we extend this study to the race and gender problems among authors?

[yjhuang99]

Dear Professor Adukia, thanks for coming to our workshop and sharing your work! Your work provided us with a new way to investigate how the values of young generations are shaped through children's education, especially to quantify the culture and social norms conveyed in books. The work is also useful for assisting publishers and content creators to prospectively assess representation in the creation of new content! I really look forward to hearing the new artificial intelligence methods you introduced for systematically converting images into data.

[ChongyuFang]

Hi Prof Anajli Adukia, BIG thanks for attending this workshop! Your work is a lens through which we can take a serious look at the “black box” of education through children’s books in US schools and homes. Quantifying this is NOT easy. It was indeed excellent work. I look forward to the workshop tomorrow.

[sabinahartnett]

Hi Professor Adukia, I'm so excited to hear your presentation at our workshop! Thank you for sharing with us. I'm curious about the impact since the publication of this paper: have there been any instances of authors/libraries/other book sources responding about the inclusivity of their own contents?

[Lynx-jr]

Hi Professor Adukia, thanks for coming to our workshop! Do you expect the inclusivity of authors/ books to vary in different languages and extent?

[JunoWuu]

Hi Professor Adukia, thank you for coming to our workshop! I really like that you have studies some major social norms about gender, and race. I think it will be really interesting to learn more about other norms and how they are represented in books such as norms about SES and more inclusive cultural norms.

[hsinkengling]

Hi Prof. Adukia, Thanks for coming to the workshop! I'm really excited to hear your talk. There seem to be a lot of visual and linguistic markers for gender and race that help with automated content analysis. Do you think it's possible in the future to code for other, more subtle dimensions of representation? (eg. different personalities, emotions)

[GabeNicholson]

Thank you for coming to the workshop! Are white males overrepresented in award-winning books because the authors themselves are typically white? Also, what is the explanation for why lighter-skinned individuals are more common, conditional on race (light skin vs dark skin African American for example)?

[lguo7]

Dear Prof. Anjali Adukia, Thank you very much for sharing your research. I am very interested in the artificial intelligence method you mentioned -- the way to turn images into data. Could you please introduce this method in detail? And how did you use artificial intelligence to scan images and text in so many children's books? How do you find this large amount of electronic materials?

[JoeHelbing]

I understand the study used facial recognition technology so I'm very curious what, if any, adaptations were necessary to fit that from photos to hand drawn illustrations.

[isaduan]

Hi Professor Adukia,

Thank you for sharing your research! Really looking forward to your talk.

You mentioned in the article "there are challenges to the numeric measurement of representation," and you gave an example of trying to measure race, a multifaceted and ill-defined construct. I wonder, however, how you deal with this challenge when measuring other constructs, for example, 'old' vs. young gendered words.

Also, in general, what do you think of the worry that, computational methods bias towards easy-to-measure constructs versus hard-to-measure ones?

[xin2006]

Hi Prof. Adukia! Thanks for sharing such a meaningful paper about images and text of children's books. What people are taught in their childhood do have great impact on the formation of their values, so I am really interested in the research. But I am a little confused about the specific area of these books. Since the fraction of different racial varies from area to area, those images and text of books published in different area may not be comparable. For example, the proportion of black in South Chicago is much higher than North Chicago, which might lead to different proportion of darker skins on the images of books. Do we need to consider the case?

[Hongkai040]

Hi Professor Adukia, Thank you for coming to the workshop! You work is really inspiring. I think many findings in your research are honest reflections of reality. I am wondering have you ever considered to extend the current findings in your future work ? For example, in your paper you stated that "Relative to their growing share of the US population, Black and Latinx people are underrepresented in the mainstream collection". It's easy to understand with consideration to the historical backgrounds or even to the fact that Barack Obama was the only African-American president of United States so far. I believe a lot of stories in children's books are based on life, even if they have a lot of fictional elements. So is it possible to interpret the changes of characters from a temporal perspective, e.g. along with some major events in the real world? Maybe the event that Barack Obama became the president of the United States has inspired many writers to write stories featuring dark-skinned people.

[LFShan]

Hello Professor Adukia, Thank you for the presentation. My question is that would Parents' inclination changes the type of book their children are able to acquire? If so, could it be the case that mainstream books would never be given to some children despite books' popularity among other children? Thank you.

[afchao]

Thank you for sharing your work with our group!

This question is only barely related to the paper we read in anticipation of your presentation, but I'm interested in your perspective on the discourse surrounding the term "Latinx". As my heritage is not in Latin America, I rely on those better-informed to guide my own use of language in these cases, but here in particular I have witnessed significant disagreement among ostensibly equally well-informed sources about the meanings and implications burdening the various terms commonly used to designate this demographic group

[AlexPrizzy]

Thank you for visiting our workshop Professor Adukita. Your approach to classifying illustrations is very interesting here, considering that many racial machine learning studies use a binary classification rather than a spectrum of skin tone as you do. I know it is explicitly stated in the paper that the goal was not for this to be a normative study in optimization of representaiton. Though my question is whether you believe this method should be used for optimization studies of diversity or more so for tracking historical changes in media?

[javad-e]

Thank you for presenting at our workshop! I was wondering if you have any thoughts on books published in languages other than English? Particularly, do you know if children's books published in Northern Europe have better gender representation? Gender distribution in job market is slightly different in Scandinavian countries, so I was wondering whether we see a similar pattern in children's books.

[erweinstein]

Thank you for sharing your very interesting recent research with us! In most types of arts and media, the award-winners are not necessarily representative of what people actually watch/read/play/see/attend/etc. Critical acclaim and prestige do not guarantee widespread consumption of that media. Are there issues, then, with using ALSC-ALA Award-winners to represent "children's books in the USA"? Would finding out which books are purchased for schools and libraries (maybe a much more difficult data-collection task) be a more meaningful measure? While I find the comparison you did using Seattle Public Library data very persuasive, Seattle is far from a representative location for the entire USA. If in parts of "Red America" the school boards, parents, and local governments don't let them buy even the mainstream award-winning books for their school libraries, then the diversity and representation issues are even worse than you and your colleagues found! But maybe the situation is reversed in some places that are more culturally "Blue" than Seattle, like Evanston, IL or Berkeley, CA?

[Sirius2713]

Thanks for joining us, Prof. Adukia! Given that many characters in children's books are cartoon figures, like Peppa Pig. I'm curious about how you deal with those figures to decide their race, gender and other identities.

[borlasekn]

Thank you so much for sharing your expertise and passion for this topic with us. As someone who has worked in schools and with children of varying races and ability levels, I couldn't agree more that representation is SO important. I was excited to look at your research showing why it is so vital. I was wondering how educators should ensure that the representation they show is just that - representative? Often when it is just one type of book about a certain person, or one character, identities can become quite singular. Thanks!

[YaoYao121]

Hi Professor Adukia, thank you for sharing such an interesting and innovative research with us. Your methodology is quite creative in this research field. But I have some questions about the discussion of findings. My first question is about the identity of characters in the books. For some storybooks, there always have good guys and bad guys. I am very curious if people with some kind of skin will appear more than others, which may convey more serious inequality and discrimination opinions in children books. Besides, I also have a question about the category of books in the sample. I think if some books are history books, for example, history about the UK or Germany or the early USA, it may be plausible that the white will on average show up much more than the black and other minorities, since the black or Asian-American celebrities in the early USA are much less than the white. I think the less appearance of the black in such kind of books may not be counted as a sign of under-representativeness, since this is an objective result and is due to historical reason, not due to the authors want to select less black American or other minority in and present some inequality to children.

[kthomas14]

Thank you for sharing such compelling research with us Professor Adukia! I was wondering if you have considered researching the implications of this research, such as the longitudinal effect that exposure to representative or under-representative literature may have on children. Additionally, I was wondering if there are any public policy initiatives that you hope to support with these findings.

[kuitaiw]

Hi professor Adukia. Thank you for coming and making this presentation! I also think that Black and Latinx people are underrepresented in the mainstream book collection for books collected had more characters with lighter skins. So what do you think what trigger this kind of underrepresentation. Looking forward to your presentation.

[linhui1020]

Professor Adukia, thanks for coming to our workshop. It is really insightful to see racism through the lens of the literature book. My question is that overtime, will that kind of ideology or phenomenon encourages more participation of the underrepresented group of authors, thus further enriching the dynamic environment of the literature?

[yutaili]

Hi Professor Adukia,

Thanks for sharing! I think your work is really interesting and insightful. In the paper, you've mentioned that more characters with darker skin color appear over time, but in the mainstream award-winning book these characters are still underrepresented. Do you think in a general sense that the effect of more minority characters appearing across all books and the effect of fewer minority characters in award-winning books will counterbalance each other? Or do you think the trend that minority group characters are underrepresented in children's books is not changing/still persist?

[Raychanan]

Thanks for sharing your work with us!

[hshi420]

Hi Professor Adukia,

Thanks for sharing. Skin color is the best way to show ethnicity on book, but I think accent can be good way to identify ethnicity in videos. I was wondering the cartoon films' effects on child comparing to books' effects. Thank you!

[zhiyun0707]

Hi Professor Adukia,

Thanks for coming to the workshop! I have never thought about the books portraits of race would have impact on children before, so it's an interesting topic for me to consider in the race and ethnicity field. The book's impact on one's childhood may explain some of the "rooted racial discrimination" in people's mind. In the research paper, when you talk about the limitations of the study, I really like what you mentions that "future tools should measure how people are represented and capture the message." Again, thank you!

[cgyhumble0612]

Hi Professor Adukia,

Thank you for sharing us your excellent work and creative thoughts. I can't wait to join the workshop tomorrow!

[a-bosko]

Dear Dr. Adukia,

Thank you for sharing your work! I am interested in the field of child psychology, so it was very interesting to learn about how books can shape a child's education and understandings about society. It was particularly interesting to learn about age differences in books, especially that adults appear more often than children in children's books. Do you believe that there should be more representation of children in books?

Thank you again, and I look forward to your presentation!

[Toushirow1]

Hi Professor Adukia, Your research on education inequality is compelling and insightful! Since education is an important manifestation of human capital accumulation. I look forward to seeing your thoughts about education inequality's impact on human capital accumulation and economic growth.

[LuZhang0128]

Hi Professor Adukia, I previously read a paper that "documents changes in racial images and examines the relationship between culture, gatekeeping, and conflict in society" by using "three sets of children's books." One of the sets is books that received the Caldecott Medal or were designated Caldecott Honor, which is also used in this research. However, previous researchers need to manually code the images. It's nice to see how image processing is making things easier now.

[taizeyu]

Dear Dr Adukia, I am so appreciated that you are able to make a presentation for us. I am so surprised that how can the researchers combine the computational method into the policy and social research.

[mikepackard415]

HI Dr Adukia, Thank you for joining us at the CSS workshop. Very much looking forward to the presentation tomorrow.

[AlexBWilliamson]

Dear Dr. Adukia

Thank you for agreeing to presenting on this topic - I am looking forward to it!

[zoeyjiao1104]

Hi Professor Adukia, thank you for giving us the talk about your paper. This topic on race, gender and child's education is really interesting and meaningful. I am also very interested in your methodology of using Artificial Intelligence techniques to convert image into data. Looking forward to the talk!

[Emily-fyeh]

Hi Dr, Adukia, Thank you for sharing this working paper with us! Just curious about a small point of this paper: in the text analysis session, the words representing colors are counted and referred to as a proxy to race. However, colors in children's book may be used to describe other subjects, usually not exclusively for skin colors. Is it a common practice to make this assumption?

[ValAlvernUChic]

Hi Professor Adukia! Thank you for sharing this paper and highlighting computational methods to interrogate the topic of representation in text and images. Two questions! 1) I was wondering if it might be meaningful to have done a narrative analysis of the books analyzed - specifically, what the story was about, what the characters were doing, what they were saying, how they were saying it, what narrative role they played in the book etc. 2) The methods in the paper were focused on analysing text and images but I was wondering about such AI applications in video (movies, dramas) and the speech embedded - what new opportunities/insights might those other mediums afford us in terms of additional variables? Off the top of my head, tone of voice - attached to different demographics - could be indicative of certain social biases located in the artistic direction. Thank you!

[ShiyangLai]

Hi Professor Adukia,

I realize how many efforts you have paid on this project, combining multiple types of computational methods to one research is really inspiring. However, one concern is that excessive emphasizing on the richness and advancement of technology may make people feel that it lost the feature and value of sociology itself. In this case, how should we balance methodology and theory in the research and writing? If you can give us some insight on this question, it definitely will be very helpful. I'm looking forward to your presentation!