Yij = β0j + β1jX1ij + β2j Zij + Eij
Reader and commenter Alexander Cheezem issued an energetic objection to my post about another happiness study, which you can read, along with my rebuttal, in the comment threads to the post, here. His main two complaints were that I didn’t read the study itself, and that I unfairly called it policy advocacy disguised as objective social science research.
I didn’t read the study itself because the only link the Post provided was not accessible without joining a service I didn’t care to join, or take the time trying. Alexander kept referring to a “direct link,” an unfortunate and misleading description of a link that goes to a page with a link to the study that doesn’t respond when you click on it, and are directed to “register.” [ CORRECTION: This is what I thought at the time. It has been pointed out to me that the first time the reporter linked to “research,” it wasn’t the study she was writing about, but another, behind a paywall. The second link on “research” did go to a live link to the actual study. Having been frustrated once, I assumed that the second link would also be to the same inaccessible link. My error—though I’m furious at the Posts’s incompetence—and I apologize to Alexander.]
Other Bill, who flagged the Washington Post headline and story initially, has provided a free and direct link It is here.
I am relieved to find that reading the entire study revealed nothing that I didn’t discern from what the Post reporter wrote, and checking the accessible links she provided. (Obviously, it would have been preferable to read the whole study initially, and I would have, if a functioning link was provided, as it should have been.). Let me take that back a bit: the study itself was worse than I thought.
Here’s why:
1. It is full of jargon like this: “However, while sociologists have provided important insight into the mediating role that stress exposure plays in the association between parenthood and mental health, they focus on proximate stressors and ignore distal sources of these stressors.”
2. It constantly relies on the alleged authority of other happiness studies—which, I guess, a reader should read and check to see if they are valid and properly represented, right, Alexander? But then, this would make most readers unhappy, if not suicidal.
3. The study’s methodology never described or defined “happiness,” either for readers or for those surveyed. That’s a disqualification.. Happiness is not a concept that has any universally accepted definition that every individual would subscribe to.
4. Despite that fact, here are equations, presented in all seriousness, to “explain” the study’s analysis:
(1) Yij = β0j + β1jX1ij + β2j Zij + Eij where
Yij = happiness of individual i in country j
X1ij = parental status (1 if parent)
Zij = vector of individual attributes (age, gender, education, income decile,
marital and employment status, etc.)(2) β0j = λ0 + λ1X1j where
X1j = vector of variables representing each country in the analysis set with U.S as referent
(3) β1j = α0 + α1X1j where
X1j = vector of variables representing each country in the analysis set with U.S as referent
This represents a common con in social science research, whereby assigning values to factors that cannot reasonable be measured or quantified is use to imply accuracy and certitude where none exists. That is why so much of the field of social science is dishonest, and why so many studies are authoritative-sounding junk. I don’t trust any social scientist who uses or publishes any equation in which Yij = happiness of individual i in country j, since we know that happiness is not reducible to a number, unless one is trying to prove the unprovable.
5. The study, contrary to Alexander’s protestations, is even more of a bootstrapping policy advocacy exercise than I (unfairly, Alexander says) assumed. From page 7 on, the study sounds like a Bernie Sanders position paper. Since the assumption that a lack of what the researchers regard as adequate government support for parents is the underlying catalyst for the study, one must assume that this bias affected the direction, construction, results and conclusions.
6. Why yes, getting free stuff from the government that people who don’t even know you have to pay for tends to make many people happy, whatever that means, though some people, those who have been taught that you should be self-sufficient and not make others take care of you, might feel sad, embarrassed, or inadequate, which might, ironically, make them happier in the long run by motivating them to take responsibility for solving their own problems and not expect the government to do it. It might make society happier in the long run by motivating those who are unhappy to teach their children to be accountable and not look to Big Brother as the ultimate savior.
Such worthless mental and ideological academic masturbation is relatively harmless, as long as policy-makers and journalists don’t use it as authority to gull citizens who quite understandably find that their eyes glaze over reading paragraphs like..
“In the mixed-effects procedure, happiness was regressed on policy and economic variables (country level) and sociodemographic controls (individual level) using a mixed-effects multilevel estimation procedure (xtmixed in Stata; see Rabe-Hesketh and Skrondal 2008). Within each model, both country sociodemographic controls were supplemented by one country-level policy variable and by all individual-level controls. Moreover, parenthood was specified as a random covariate in order to accommodate the country-level variability in parenthood effects as demonstrated by the fixed-effects analyses. Because likelihood ratio tests consistently favored the random-coefficient model over a model constraining the effect of parenthood to be equal across countries, a random-coefficient specification is consistently used (Chi-square p < .001).”
This kind of writing, I must note, is itself unethical. Either it is incompetent as communication, or it is intentionally deceptive, using jargon to create “authoritative” justification for policies, knowing that the jargon will be translated to mean anything policy advocates need it to mean.
Ethics Alarms 
Jack said, “here are equations, presented in all seriousness, to “explain” the study’s analysis:”
Considering their use of math, I thought you might enjoy this.
What Makes 100%? What does it mean to give MORE than 100%? Ever wonder about those people who say they are giving more than 100%? We have all been to those meetings where someone wants you to give over 100%. How about achieving 103%? What makes up 100% in life?
Here’s a little mathematical formula that might help you answer these questions:
If:
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z is represented as:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26.
Then:
H-A-R-D-W-O-R-K
8+1+18+4+23+15+18+11 = 98%
K-N-O-W-L-E-D-G-E
11+14+15+23+12+5+4+7+5 = 96%
B-U-L-L-S-H-I-T
2+21+12+12+19+8+9+20 = 103%
E-X-P-E-R-I-E-N-C-E
5+24+16+5+18+9+5+14+3+5 = 104%
A-S-S-K-I-S-S-I-N-G
1+19+19+11+9+19+19+9+14+7 = 118%
G-O-O-D-A-T-T-I-T-U-D-E
7+15-15+4+1+20+20+9+20+21+4+5 = 141%
B-U-L-L-S-H-I-T-D-E-T-E-C-T-O-R
2+21+12+12+19+8+9+20+4+5+20+5+3+20+15+18 = 193%
A-S-S-K-I-S-S-I-N-G-D-E-T-E-C-T-O-R
1+19+19+11+9+19+19+9+14+7+4+5+20+5+3+20+15+18 = 217%
BULLSHIT – BULLSHIT DETECTOR
103% – 193% = -90%
ASS KISSING – ASS KISSING DETECTOR
118% – 217% = -99%
So, one can conclude with mathematical certainty, that while Hard Work and Knowledge will get you close, Experience will get you there over the 100% mark. It’s funny how how far you can get with just Bullshit and Ass Kissing until you apply the Detectors that we all have internally built in. So now it is proven that Bull Shit and Ass Kissing are only temporary boosts and get you no where in the long run; however, it is now proven “mathematically” that a Good Attitude will always get you the farthest.
Why did I present this in this thread? Because it the same kind of bull shit and utterly deceptive usage of mathematics as what I presented above and doesn’t prove a damn thing!
As a friend of mine has said dozens of times and it’s worth repeating;
“You can use statistics to “prove” just about anything, it just depends on the subjects chosen to participate in the survey, how the survey is conducted, and how the data is compiled; all you have to do is use your imagination.
Consider the following:
A statistical analysis was done to find out what cat turds taste like, and the results showed that cat turds tastes like chicken.
Just because a statistical analysis has been conducted and a conclusion reached does NOT mean that what “The Survey Says” actually represents reality. In the above scenario; dogs were the test subjects, but for some statisticians trying to prove a preconceived notion, that’s irrelevant.”
Wonderful. So why do not studies like the happiness studies using numbers that way automaticallt set of ethics alarms and bullshit detectors?
Jack Marshall said, “So why do not studies like the happiness studies using numbers that way automatically set of ethics alarms and bullshit detectors?”
Ignorance that comes from a dumbed down society. 🙂
“As a friend of mine has said dozens of times and it’s worth repeating;
“You can use statistics to “prove” just about anything, it just depends on the subjects chosen to participate in the survey, how the survey is conducted, and how the data is compiled; all you have to do is use your imagination.”
Actually, Zoltar, I would take issue with this. Statistics cannot and do not prove anything, anytime, ever. The best you can get from statistics is a probability that something (a research result) happened by chance, or that there MIGHT be a relationship between two things. Folks pretty consistently mis-use statistics in the manner that this study does (and you did, hilariously), but for actual proof, you’ll never get it.
dragin_dragon said, “Actually, Zoltar, I would take issue with this. Statistics cannot and do not prove anything, anytime, ever. The best you can get from statistics is a probability that something (a research result) happened by chance, or that there MIGHT be a relationship between two things. Folks pretty consistently mis-use statistics in the manner that this study does…”
You’re taking issue with the wrong thing! People are presenting statistics to “prove” their arguments all the time, in every corner of the media, in politics, in forums all over the internet, it’s everywhere! You should be taking issue with the mis-use of the statistics.
What you failed to take into consideration in the sentences you quoted was the quote marks around the use of the word “prove”.
You’re right, and I apologize. Mu arguments stem from a conversation I have had with my brother-in-law several times. He uses the line to “prove” that experimental science has no idea what it is talking about. He considers himself a “Priest” of Wicken.
dragin_dragon said, “You’re right, and I apologize.”
Accepted; thank you for the effort. 🙂
H-U-M-B-L-E
8+21+16+2+15+5 = 67%
T-A-L-E-N-T-E-D
20+1+15+5+17+20+5+4 = 87%
Humble + Talented = 154%
Seems legit.
P-O-L-I-T-I-C-S
16+15+12+9+20+9+3+19=103%
B-U-L-L-S-H-I-T
2+21+12+12+19+8+9+20 = 103%
Now we have mathematical proof that politics equals bullshit.
Working in the pre-dawn hours here in central North Carolina, the humidity usually skyrockets until the sun comes up.
It has always been my contention that, if people can give 110% effort — and we know, from their statements, that they do so all the time — I believe the humidity often gets to at least 110% as well. It only makes sense, right?
I still don’t think AC could even actually read the equations, let alone understand them.
Well, who could?
Someone with a major in math? My point being, AC defended something he couldn’t read. It’s a hack position to take.
Math doesn’t include happiness equations, because happiness isn’t a number.
The best you can do is is ask people “Are you happy or not”, then quantify which percent of the population self-describes as being happy.
Maybe one could get away with a “On a scale of 1 to 10, how happy are you?”
Give up trying to normalize any of that.
Humble Talent asked, “Someone with a major in math?”
Major in math won’t help when you’re dealing with a statistician trying to prove a preconceived notion with frontier gibberish mathematics generated for the sole purpose of confusing those that do not have a Bachelors in Bull Shit (BS) or a Masters in More Shit (MS) or a Doctorate in in Piled Higher and Deeper (PHD).
It junk math flushing common sense down the toilet.
It’s funny… I was going to say that understanding how they got to their number would prove the junkiness of it, and then I realised what you’re saying we don’t need to, we already know it’s junk. You’re right. The math nerd in me was still interested.
The moment I saw “Yij = happiness of individual…” I knew the entire study was complete hokum, it’s uncompromising bull shit to prove a preconceived notion. You cannot quantify that which is not quantifiable and any attempt to do so in a fabricated equation is junk math.
I not only read them all, I commented specifically on the limitations of my ability to interpret them. I don’t know enough about ordinal regression analysis to comment overmuch on whether or not they used appropriate methods there.
That’s *entirely* beside the point.
We use equations like that in engineering all the time to simply complex phenomenon. Using such equations to advocate policy is straight up “social engineering”….
Can’t find the book, “Understanding Happiness Research,” but there is at least evidence (from a movie) that a book, “Understanding Poetry” exists:
Oh – it was a CHAPTER in a book – not a whole book. My bad. Signature significance for indicating my movie-watching ability, I suspect.
I loved that movie!
…which, unfortunately, depicts an English literature class a long time ago in a galaxy far, far away.