[Pythonic]

I don’t doubt India’s results in this, given its PCGDP at 2-3k USD but most of these surveys are bogus. Finland has a higher suicide rate than India for example. Finnish people have it at 14/lakh while Indians have it at 11 or 12/lakh of the top of my head.
Just something to occupy a slow news day.

Whatever the methodology might’ve been the truth is Indians are a very unhappy people. Reasons might vary from abject poverty to BMW’s beaver tooth grills

[binand]

Originally Posted by SmartCat

Basically, there are a large number poor people in India, both on absolute and percentage basis. If the sample is able to capture India's financial demographics accurately, then such a low ranking is plausible.

If you read through their methodology, then you will understand that (a) they have an urban bias, (b) they don't do their survey in large, traditionally poor parts of India. For example, all of north-east is excluded from the survey. So are areas where they consider it is unsafe for their surveyors to go into.

[dragracer567]

Originally Posted by volkman10

However, the top 20 has some new entrants—Costa Rica and Kuwait—matched by the departures of the U.S. and Germany from the same ranks.

Perhaps they forgot to ask the expats in Kuwait?

Kuwait is the most miserable country in the GCC, let alone the world. Even their nationals aren't happy as their economy tanks due to cronyism, corruption and no real pivot away from oil, unlike their neighbours.

[Redex]

Hi,
yet another meaningless survey of no value. The U.K. should be less than 3. Very high energy and fuel prices combined with a very high (and increasing) cost of living has put many people on the poverty line. Most people have no money left over at the end of the month. Public services are falling apart. I have just been quoted a minimum 12 month wait to see a cardiac specialist. We don't have enough Police, Ambulance, Fire Brigade, Hospital Staff, Teachers, Armed Service Personnel etc etc. The quality of life is falling dramatically. The infrastructure is falling apart. The roads are falling apart. Crime goes unpunished. Political correctness has reached ridiculous levels. Older people have been robbed of their pensions. Pensions that are no where near enough to live on. I lost 1 year of pension payments and my wife will loose 7 years of pension. That totals to 85 lakh Rupees stolen by the Government because they have increased the pension qualifying age to 67 !!!
20% of the UK population struggle to live on a pension.
The minimum living wage is 18 lakh Rupees per person per year. Just enough to cover the cost of existing.
The UK pension is only 11 lakh Rupees per person per year.
How is an older retired person meant to survive.
The general population of the UK are not "happy". They are just managing to survive.

[Samurai]

Originally Posted by ValarMorghulis

Agreed we Indians aren't exactly the happiest but basing your results on a 1000-person survey is hilarious/ ludicrous to say the least.
I would take this survey with a stadium full of salt. My MBA surveys had more credibility than this.

Can you please elaborate on why you feel the sample size is inadequate?

After reading your post, I checked on an online sample size calculator.

For a confidence level of 99% with margin of error of 4%, just 1040 samples are adequate for even the Indian population. It almost remains same for 1000 times lesser population.

However, we can definitely find fault with how the question was asked, and what was their sampling method. However, they have not revealed that info.

Often, these surveyors are given a conclusion and asked to collect data to support that conclusion. That bias encourages the surveyor to form the question and choose samples to suit that conclusion. If you want the result to be unhappy, then you will go looking for unhappy people to ask the question.

[pandabear]

I would not blame the survey company Gallup for the botched up survey that has been sponsored by the most unbiased inclusive bully club aka United Nations but at the quality of substance smoked. 😂

Below survey shows diametrically opposites results https://www.ipsos.com/en-in/india-ra...ss-2020-survey

The level of optimism when among the poor that I see in India is unmatched amongst other nationalities. I see a lot of depression, negativity, and anxiety among Americans European and even the Chinese Who are very uncertain about their future. And yes, optimistic people = happiness is a direct no brainer correlation.

I would divide Indians into two categories one the the ruling dispensation supporter who believes that we are the next superpower and the hater who believes that we are down the dictatorial drain. We Indians might be totally divided on this, but outside of India the world is extremely bullish on the india growth story.

Ironically, when you have a sample size of just 2000, which street corner political rally you are standing with your survey can fetch you a result, curated to the needs of the sponsor.

Happiness is very subjective, and the survey throws up a lot of questions more than it answers, especially about the intent.

[Noctis]

I always question the method of measuring something extremely subjective like happiness, but the outcome for the most part seems somewhat obvious: People in wealthy and stable countries are happier than those in poorer or chaotic ones. And among the wealthy, those with a robust social safety net has even higher scores.

I'm not sure if we needed a survey to tell us that but I guess it's interesting to look at.

[binand]

Originally Posted by Samurai

Can you please elaborate on why you feel the sample size is inadequate?

The thing is, and this point ought to be intuitive, the probability that the value measured from a given sample is off from the value for the entire population decreases with sample size. This is encapsulated in the confidence interval parameter that you provide as input to your online form - reduce the confidence to 95% and you can get away with a sample of 601 too.

Since you want your sample to resemble the population as closely as possible, then you have two options:

1. Increase the sample size so that the aforementioned probability is reduced, or
2. Repeatedly sample and then average.

A one-time, 1000-member sample survey does not capture the heterogeneity of the population of India, so it is not a representative one. You have to do one of the above to make it so. It is no coincidence that the NFHS has a sample size of 6,00,000+. Their paperwork/documentation has a good description of how they do the sampling, and what the considerations and challenges of conducting a sample survey in a hugely diverse country such as India are.

Originally Posted by Samurai

After reading your post, I checked on an online sample size calculator.

In this form, you have to supply the population proportion - that is, you are telling the calculator upfront that you know 50% (the number you supplied) of the population of India are happy. Then the calculator tells you that if you were to survey 1041 people, you can be 99% confident you will find that between 49.5-50.1% of the sample will turn out to be happy.

Also note that a lot of online resources on statistics seem to be catering to the medical research fraternity; so the formulas used might have assumptions based on what works for the sample and population sizes they deal with. You have to specifically look for sociological or demographic research output.

[Samurai]

Originally Posted by binand

This is encapsulated in the confidence interval parameter that you provide as input to your online form - reduce the confidence to 95% and you can get away with a sample of 601 too.

I picked 99% CI and 4% EM to match the sample size they used.

Originally Posted by binand

In this form, you have to supply the population proportion - that is, you are telling the calculator upfront that you know 50% (the number you supplied) of the population of India are happy.

So what is the population proportion you would pick that would provide a higher sample size for 99% CI and 4% margin or error?

Or are you trying to tell us that sample size algorithm cannot be used in these kinds of surveys?

[anjan_c2007]

I will just not give my two cents such a concocted, coloured and non-representative survey. The mere percentage of the populations of the countries surveyed in most cases raises doubts about the integrity of their database. With such miniscule percentages of the demography surveyed the end result is destined to be skewed. North Korea appears to be out of bounds for their surveyors.

Like for instance there are laughter clubs in many localities of the world. If their surveyors question the laughter club participants, each one would say he's too much happy. And also croon "don't worry, be happy" Bobby McFerrin's pop hit from 1988.

This website atleast lists India as a "Regional Economic Heavyweight", which should be a consolation for us.

Meanwhile, in South and East Asia, Singapore (30th) and Taiwan (31st), are well ahead of regional economic heavyweights, China (60th), Japan (51st), and India (126th), when it comes to happiness levels.

The survey lists a nationally representative group of approximately 1000 people for the sample survey based on their methodology. We have 543 + 245 MP's (788) who represent our 130 crores population in the Lower and Upper Houses. At least if the survey team had gone to the Parliament to interview them for the happiness surveys and obtain their filled up survey forms, we could have treated the survey as representative.

Methodology: A nationally representative group of approximately 1,000 people is asked a series of questions relating to their life satisfaction, as well as positive and negative emotions they are experiencing. The life evaluation question is based on the Cantril ladder, wherein the top of the ladder represents the best possible life for a person (a score of 10/10) and on the flipside, the worst possible life (scored as 0/10). The main takeaway is that the scores result from self-reported answers by citizens of each of these countries. The results received a confidence interval of 95%, meaning that there is a 95% chance that the answers and population surveyed represent the average. As well, scores are averaged over the past three years in order to increase the sample size of respondents in each country.

Criticisms: Critics of the World Happiness Report point out that survey questions measure satisfaction with socioeconomic conditions as opposed to individual emotional happiness. As well, there are myriad cultural differences around the world that influence how people think about happiness and life satisfaction. Finally, there can be big differences in life satisfaction between groups within a country, which are averaged out even in a nationally representative group. The report does acknowledge inequality as a factor by measuring the “gap” between the most and least happy halves of each country.

Quotes from this link:-

https://www.visualcapitalist.com/a-m...oogle_vignette

[binand]

Originally Posted by Samurai

Or are you trying to tell us that sample size algorithm cannot be used in these kinds of surveys?

Yes. The formula they use is: sample size = z^2 . p . (1-p)/e^2

Here, e is the error margin you specify, 4%.
p is the population proportion.
z is the z-value for the CI (which is 2.58 for CI = 99%; you can use the calculator here or the normal distribution table found in any statistics book).

Elementary calculus tells you that the sample size has a maxima when p = 0.5. Plugging these numbers:

Optimal sample size = 2.58 x 2.58 x 0.5 x 0.5 / (0.04 x 0.04) = 1040.

[The online calculator gives it as 1041; I am assuming they round up]

The problem is, this formula holds good only for homogeneous populations - that is, you have to assume that all subgroups of the population of India are equally happy (in other words, if 50% of urban Keralites are happy then 50% of rural Biharis are too. If 50% of Koli fisherfolk are happy 50% of Sentinelese tribals are as well. And so on).

The accepted statistical technique is to divide the heterogenous population into homogeneous clusters, calculate the happiness index for each cluster (here you can determine your sample size using the above formula) and then aggregate it all using a weighted average method. From what I can make out, this is what the NFHS does (see: https://rchiips.org/NFHS/NFHS5/pdf/N...0assurance.pdf - discussion on sample selection and weights in pages 6-7).

[amitoj]

The main reason for US dropping below the top 20 is the below-30 demographic. In this age group, US ranks 62. In the above 60 age group, it is in top 10. It is very clear why such a large gap exists. People below 30 have no certainty about their future and they attribute this to several things, some of which are unaffordable housing, costly healthcare, inflation and climate change.

Originally Posted by TROOPER

Since Finns are leading I wonder why Valterri Bottas went to Australia in search of some extra happiness Leaving Mercedes completely changed him I guess.

Isn't his girlfriend Australian and a world cyclocross champion? Bottas has gotten into gravel riding big time and I follow his adventures on Strava. He and his wife even organize a big gravel race down under.

[Samurai]

Originally Posted by binand

The problem is, this formula holds good only for homogeneous populations - that is, you have to assume that all subgroups of the population of India are equally happy (in other words, if 50% of urban Keralites are happy then 50% of rural Biharis are too. If 50% of Koli fisherfolk are happy 50% of Sentinelese tribals are as well. And so on).

Ok, I get your point. But when you start looking for homogeneous groups, there would be thousands of groups in large countries like India, China, etc. It would be too expensive to conduct such surveys.

[binand]

Originally Posted by Samurai

But when you start looking for homogeneous groups, there would be thousands of groups in large countries like India, China, etc. It would be too expensive to conduct such surveys.

Exactly. Which is why from message #3 onwards I've been arguing that this survey is flawed and its results, with no sound basis. Their declared sample size (2100) is simply inadequate to arrive at any sort of pan-Indian conclusion, forget something as imprecise as "happiness". I doubt it is even sufficient to conclude anything about Bangalore, even.

[vjbox]

Originally Posted by dragonfire

I do not know when the above cited survey was held, but I can say for sure that from 2024 the majority of people in India are happy.

Almost all of the real problems our country had faced has been solved, all the historical troublemakers have been shown their place, and we are only a step away from making the current bliss last until eternity.

Why wouldn't the country be happy with the current state of things?

That I am afraid is a pegion hole perspective. There is literally no comparison with some of these Euro/ Scandinavian countries when it comes to infrastructure, standards and quality of life, access to facilities and the earning potential and note that this is common for all income groups.

Another interesting thing is, People over there especially in countries like Sweden practice a concept called Lagom which means moderation. They do everything in moderation but to utmost efficiency and perfection resulting in a life with no greed, lesser anxiety, lesser peer pressure and no needs of seeking validation leading to a happier life overall - all of which being quite the opposite for us in India.