The purpose of this script is to analyze the text of the perspective taking Senders in the pilot study of the Perspective taking trust game yoked design.

For more background on this study, see: https://docs.google.com/document/d/1uQNRJugb29BB_QcRGpLFRB_dqDK3aQkA7VNZGHk1DNs/edit?usp=sharing

This is the prompt:

We know that description probably is not much to go on. But pause for a moment – and do your best to take this particular Prolific respondent’s perspective as a Receiver in this activity.

Step into their shoes. Consider how this person might feel in the Receiver role. Imagine what they are thinking and feeling about this game … and about how much money to return to you, the Sender.

In the space below, please step into your Receiver’s shoes and write at least two sentences about what they’re thinking and feeling as they think about this game and consider how much to return to you as the Sender.

Just text

First, let’s just take a look at their responses. I’ll order it by the amount they decided to forward. This might just give us an indication as to whether more “competitive” perspectives forward less money. Let’s see.

Forward PT_openend
0 If I was the receiver and the sender sent money I would return what was most equitable. I would not expect to see any money sent to me though.
0 They are thinking that work is tough and they don’t really like it. If they could get extra money and get away with it, they will.
0 I hope they send me a lot. Then, I’ll decide how much to send back.
2 There is really no way at all to do that. They could be thinking anything. They could be thinking that they hope I send everything so they can return have and we both win. They could be thinking they hope they send everything so I can return nothing and get the most. There is absolutely no way to step into anyones shoes and figure that out. There is no real information to go on. If anything I would think they would not be sending anything back, even though I personally would, because really there are no consequences for their actions and this is just a 1 time thing with some anonymous person on the internet.
2 I evaluate my choice of focus, with my valuable time. I chose to take a Prolific survey this morning, and the researchers, Professor Daniel Ames at Columbia University et al - seem to have underestimated the amount of work time needed to complete the survey. Further, this survey is using a “Lottery bonus” which has at least 117+ participants, according to Prolific’s data. The researchers state, “When our survey is done, we’ll randomly pick one Sender-Receiver pair and pay them in Prolific bonuses according to their decisions” - This means only one pair out of 100+ participants, will get the bonus, or less than a 1% chance, of actually being paid, for the effort we put into this bonus choice. I feel a bit disappointed, because ethical researchers negotiate fair time for money work trade, and socio economically, demographically, promising a bonus, that less than 1% of participants have a chance of obtaining, may be manipulative, psychologically, and class biased. Statistically, the majority of Prolific and Mturk workers, live at poverty level income, according to two different studies. I’ll adjust my expectations, be satisfied with the base pay, although may or may not choose to take future Columbia University studies, based on their design choice. I love working with people that value my time obviously, and offer direct and fair negotiated pay. I’m unsure that I feel a strong way, regarding how I will be paid by the Sender. The design of this study, makes it more the researcher’s culpability, rather than another participant, who may also be earning less than minimum wage, to focus on Columbia University’s work.
3 I think that they think by saying that they have made me sympathetic towards them
4 They might be hopeful that I will be nice and generous. They hope that I am as kind as they are.
4 I think they’re thinking that hopefully I send a good amount of money and not the smallest amount or none at all. They’re thinking they will return a small amount back.
4 They may be more curious about how it works and what they will benefit from
5 They’re thinking, probably like me, to make this an equal-share game…how can both of us do well and benefit from this?
5 I hope I make some decent money from this. I like to help people and will give the amount that was tripled back to the sender.
5 They seem to be a person who cares about other people. This person is charitable. I think they would return a good amount of money to me and would be fair.
5 I don’t know what they are thinking. They probably want to maximize their profit
5 I am hoping that the Sender follows an equitable decision and decides to split their share of money in their account. I would consider evening up the balances in both accounts so we are equitable.
5 I think they are feeling hopeful that the Sender (me) will act fairly and share some of the money with them. It would really reaffirm faith in people to be treated fairly.
5 they like to create their own things, so they probably don’t have much cooperation with other people. this person might be more selfish
5 I like to work with my hands and work hard. I’m diligent and see where hard work pays off. I will definitely return some of the money, at least $5.
5 I saw the mention of being a trader, so this gives me a lot of information. They’re thinking about how much money they can personally make off this as a return, without worrying about giving back. They wanna benefit off me the best they can and walk away with as much money as possible, maybe they have the idea of giving back a small amount of what’s sent.
5 I feel as if they are thinking of the best way to maximize the earnings both of us will receive through the game. They want us both to be as successful as possible.
5 They would consider sending some back based on maintaining a profit.
5 Well, it’s my hope that they’re happy and hoping to share the wealth as it may be. I think they’d just be happy with half the bonus money, same as me.
5 The Receiver may be wondering if I will send any money at all. They may be feeling, “If I am sent any money that triples, then I will send back some money as a gesture of thanks.” Or, they may just keep all of the money and move on, we will see.
5 The receiver is thinking about who is going to be in the sender role and how much they are going to send. They are hoping we work together so we can make the most profit.
5 I think they are prepared to get zero, but will hope it will be more than that. I think they would try to be fair
5 The receiver is hoping that I, as the sender, decide to share some of the money with them. The receiver is hoping that I’ll be altruistic and pro-social rather than selfish.
5 Just doing the most to help each other and doing our best to do what we need to do.
5 I sense optimism from the Receiver. I believe they see the world through a “glass-half-full” lens. This individual likes to benefit society and others. I think they are excited about this game. I believe they think the Sender (me) will split some of their $10 with them. I believe if I send them money, they will return a little bit of it.
5 I think they are enjoying this game as much as I am. Enjoying the new experience and as a sender I feel like giving half back.
5 a calm coordinated and rational individual who believes in equity
5 I hope my sender is fair. We can both make money if we only share.
5 I believe this person would be grateful to be the recipient and would be eager for an opportunity to show his or her generosity.
6 This person seems to be helpful and empathetic, and so I imagine they would feel inclined to return a fair/generous proportion of the money back to me. I think they may be somewhat relived they are the receiver too, since they don’t have to make the initial decision of choosing the amount of money to send.
6 Best guess… they are thinking that their “sender” is not going to send them anything. It’s their lucky day.
7 The receiver knows that initially, I have the power in our little relationship. If I choose to send money to the receiver, though, the balance of power shifts 180 degrees. I think the receiver is aware of that balance and is curious about what I will choose to do.
7 Probably thinking it might be nice to get a bit of extra money. They are probably trying to relax after a long tedious day at work so they may be a little less giving.
7 As a person who helps address daily challenges in the marketing/business world I’m sure it might be tough to rely on the kindness of a stranger. Also it is probably difficult to not have initial control.
7 They are probably are expecting me to be fair, even generous in my decisions towards them. I think they have already considered and decided with certainty how to they plan to respond to my choices.
7 I wonder about what the sender will do. I will also keep an open mind sense the world can move in unexpected ways
7 This is an interesting notion. As a role of the sender, I would equally distribute funds for the receiver - in order to achieve a state of equilibrium. An order of balance. If I were to empathize with the receiver - I would not be concerned because as the role of the sender I would choose the fair option.
8 they are hoping that the sender will be generous and they will most likely return more based on how much they’re given
8 I would imagine that they are considering to give a small portion to the Receiver initially, so they can have a guaranteed bonus. But would be unsure if that would be too much or too little.
9 I think they seem honest and want to help others. I feel I should do what helps us both out the most as I feel they are helpful and want to do good. I think they are creative and open minded as well.
10 I think i would be feeling like i want to get the best outcome for both of us.If i could do that then i would feel like we did the best job that we could and i would feel a sense of accomplishment.
10 I would want to return at least half of everything after it triples. It would only be fair that way.
10 Thanks for giving me this money, I will return a good amount.
10 Given their response about helping people, they are probably hoping I trust them to give me a fair amount of money back.
10 They are thinking they want to be trusted so we can both maximize our potential earnings, and they want to show me they are careful, skilled, etc. Being skilled with ones hands is generally associated with trustworthiness.
10 They are probably worried I won’t send them any money. They might be thinking they won’t send me any money. They may be anxious.
10 I think splitting evenly is the right thing. I plan on returning half.
10 I should keep all of this money. The sender will never know who I am because this is anonymous.
10 I am too trusting
10 I think that the receiver has a positive view of people and has a reasonable amount of trust in others. I think that they will see this game as cooperative and will try to split the money fairly between me and them.
10 I would try to do things fairly.
10 I wonder if the other person will even send anything at all. They might just keep the full $10. If I do get any, it’s probably fair to even out our earnings, if I get enough.
10 I would think they are really hoping that i send them money. If I were them I would do my best to sound like a nice person and someone that could be trusted to send money back.
10 They are wondering if I’ll send them anything at all. They think it’ll be a pittance like a dollar and they’ll just keep the three dollars
10 I believe that the Sender is a mentor to me. I will trust in their judgment as my mentor that they will send me an appropriate amount that will benefit us both.
10 I hope the sender is fair. I hope we both end up better off than we were.
10 I sure hope the receiver is filled with the spirit of cooperation. They should understand that we will both hope to gain more by helping each other rather than only looking out for number one.
10 I think they’re hoping that I’ll send them money. I have no idea if they’ll send me money back. These surveys, sadly, often exploit us workers who, again often, are in dire financial situations. I would not blame the other person if they kept all of the money I sent, even though I need it quite a bit.
10 They are probably thinking about how fairly the sender has treated them and what would be fair to return. If the sender decided to send a large amount they may be thinking that it’s fair to split the money evenly.
10 They understand the value of work and honesty so they will do the right thing.
10 I think that these type of studies happen fairly often and if enough people do not send back some part after receiving an amount, then the “senders” will eventually stop sending anything. It would be in my self interest to give back a “fair” amount
10 I am imagining that the receiver is wondering how much, if anything I will forward to them as the sender. I also think the receiver is also thinking that if they do receive an amount of money from me, as the sender, if they should send anything back to me at all, as the sender and further if they decide not to send anything back, they are probably wondering how they might be viewed as a person.
10 I think they are very interested in what I might consider doing, and are probably calculating the likelihood of receiving a certain amount. They might be nervous that I may not send them anything at all.
10 Receiver is likely doing the math about how much they can receive and what they would return. I believe this due to their attitude and enjoyment of puzzles.
10 If the sender is kind enough to send a lot of money, it means they are probably a kind, sensitive, creative person as well. In that case, even if I desperately need more money for my life to get by, I will return half of that money to them (which has been tripled) to repay them for the kindness of being generous up front with what they sent me so that we can both have more money together.
10 They would be interested in how much I would trust them in sending them money. I would assume they hope I know they’re trustworthy.
10 The receiver is hoping I will send them all the money so it doubles and they would split it in half with me.

hmm, I’m not sure. A bunch of these have elements of trust/distrust, but I don’t know if NLP will pick this up. Let’s just try to see if some of these texts including more trusting words than others.

GloVe: Global Vectors fot Word Representation

Alright, let’s find vectors for these words and n-grams (up to 5-word sequences). We do this based on word co-occurrence in the entire corpus of Wikipedia words.

library(text2vec)
set.seed(08022023)
text8_file = "~/text8"
if (!file.exists(text8_file)) {
  download.file("http://mattmahoney.net/dc/text8.zip", "~/text8.zip")
  unzip ("~/text8.zip", files = "text8", exdir = "~/")
}
wiki = readLines(text8_file, n = 1, warn = FALSE)

# Create iterator over tokens
tokens <- space_tokenizer(wiki)

# Create vocabulary. Terms will be ngrams (1 to 4 tokens).
it = itoken(tokens, progressbar = FALSE)
vocab <- create_vocabulary(it,ngram = c(ngram_min = 1,ngram_max = 4))

vocab <- prune_vocabulary(vocab, term_count_min = 3L)

# Use our filtered vocabulary
vectorizer <- vocab_vectorizer(vocab)
# use window of 5 for context words
tcm <- create_tcm(it, vectorizer, skip_grams_window = 5L)

glove = GlobalVectors$new(rank = 100, x_max = 10)
wv_main = glove$fit_transform(tcm, n_iter = 10, convergence_tol = 0.01, n_threads = 8)

wv_context = glove$components
word_vectors = wv_main + t(wv_context)

# create a data frame with all words
all_words <- data.frame(word_vectors) %>% 
  mutate(word = rownames(word_vectors)) %>% 
  select(word,everything()) %>% 
  `rownames<-`( NULL )

# organize our open-text
df_ourwords <- df_pilot %>% 
  select(PID,PT_openend) %>% 
  filter(PT_openend != "") %>% 
  unnest_tokens(word,PT_openend) %>% 
  left_join(stop_words, by = "word") %>% 
  filter(is.na(lexicon)) %>% 
  select(-lexicon)

df_ourwords <- df_ourwords %>% 
  left_join(all_words,by = "word")

write.csv(df_ourwords,"~/Google Drive/My Drive/perspective taking/pt yoked trust game - pilot data/df_ourwords.csv",row.names = F)

Now that we have that vocabulary, let’s get cosine similarities of each word with the word TRUST

alright. Now let’s see if average cosine similarity with trust correlates with amount forwarded. But before then, let’s check out the distribution of these trust scores.

ok. Is it correlated with amount forwarded?

r = 0.11
p = 0.392

(#tab:unnamed-chunk-6)
**
Predictor \(b\) 95% CI \(t\) \(\mathit{df}\) \(p\)
Intercept 6.80 [6.01, 7.60] 17.13 66 < .001
Trust 5.33 [-7.01, 17.68] 0.86 66 .392

hmm, pretty weak. I think it’s just too noisy. I need to figure out how to do doc2vec, rather than word2vec. If we pick up the sentiment of the entire exercpt, we might be able to get something more meaningful.