#### Analysis of measure of suffering from Twitter survey On 21st July 2021, I had [tweeted](https://twitter.com/paraschopra/status/1417731217001095168) a survey where I asked people to answer the following question: > Imagine there's a perfect fix for depression. It's an instant fix, doesn't require any effort, doesn't cost money and is free from side effects. The way it works is this: you pop a pill that puts you into an unconscious coma and you wake up after a certain time period with your condition fixed. Assume we have a *highly depressed* patient (aged 25) who wants to try this. After getting this perfect treatment for his major depression, he's expected to live until 80 years which is another 55 years to go. According to you, what is the MAXIMUM amount of time (in minutes, hours, days, months or years) do you think this patient will be willing to go into a coma for to cure his major depression? If your answer is closest to the average of all answers I get in this survey, you'll get a $50 Amazon card. So try to think carefully and provide your best estimate. ##### Description of data - Got 121 responses and did following cleaning - About 50% of respondents are highly or somewhat depressed (it's concerning) - Removed two responses without any answer to the main question - Converted all answers to "equivalent number of days" (e.g. 1 year became 365 days, while 1 month became 30 days) - Combined some open ended answers in "Your current state of mind question" to different categories (e.g. "Happy" to "Not depressed") #### Summary of results ![[Screenshot 2021-07-23 at 12.45.37 PM.png]] ![[Screenshot 2021-07-23 at 12.18.56 PM.png]] - The average answer is close to ~3 years, while the median is 1 year - Four people gave the answer closest to the average of all answers. I'll split $50 reward among all the four. - In other [surveys](https://link.springer.com/article/10.1007/s11136-015-1018-3), people have responded to average of trading off 3-4 years for curing depression. It looks like a good validation but there are some differences in test set up: - In most studies: choice is between 10 years to live in diseased/suffering state v/s min X years to live in perfect health (in our case, it was 55 more years to go) - Across the groups of respondents, there is a clear difference in responses: - Median responses: - The median answer of 365 days / 1 year is surprisingly similar for different categories of respondents - **Taking median answer as a representation of the intensity of underlying condition seems justified** as it appears reliable (i.e. giving same answer upon multiple surveys) - Viewed purely from median point of view, **those who're not depressed believe depression causes slightly less suffering as compared to those who suffer from major depression** - For those who're not depressed or somewhat depressed (median answer: 365 days) v/s for those who're highly depressed (median answer: 465 days) - This analysis is corroborated by another study ([Do individuals with and without depression value depression differently? And if so, why?](https://link.springer.com/article/10.1007/s11136-015-1018-3)) that found out similar conclusion. - Although it is surprising that "Studies on valuations of somatic conditions show that somatically ill patients tend to value their states less negatively than the general population" (from the same paper linked above). - - Average responses: - Respondents who are not depressed gave a much higher number (~1500 days / 5.5 years) for how long a highly depressed person be to be willing to go into coma than respondents who're depressed (~720 days for highly depressed, ~1000 days for somewhat depressed) - If we remove the outlier response of 55 years, averages of these groups get closes (~833 days for not depressed and 720 for highly depressed and ~600 days for slightly depressed) - This suggests that people without depression are making wild, uninformed guesses of picking the highest number possible as an answer - **Removing the outlier suggests that even average is robust** across groups and can be used as a measure of suffering - Standard deviation - The standard deviation of responses for highly depressed respondents is much less than not depressed people - Ideally, we should **weigh the responses from the population who intimately understands the suffering more than those who don't understand it so well** So, in nutshell, if I were to pick an answer for the question asked above in the survey, I would go with 1 year. Comparing answers from surveys on different conditions (e.g. blindness) #### Learnings for study design - Study Time Trade Off method - Choice is between 10 years to live in diseased/suffering state v/s min X years to live in perfect health (in our case, it was 55 more years to go) - Ask for max/min rather than simply an answer - Extremes make people think carefully about tradeoffs - Experiment how asking people to give an answer v/s trying to predict everyone else's answers makes a difference in responses TODO: - Check whether differences are statistically significant - Read comments by respondents - Give rewards to people who won #inbox