Half Sigma

The Skinny Man Effect

In the following tables, I compare body mass index (BMI) with family income, using the 1991 National Health Interview Survey. Remember that all data is self-reported, and people lie on surveys. Men, specifically, lie about their heights if they are shorter than 6'0".

Variables
Role	Name	Label	Range	MD	Dataset
Row	hs_bmi(Recoded)	BMI	1.0000-7.0000		2
Column	incfam(Recoded)	family income	1-3		1
Control	sex	sex	1-2		1
Filter	age(35-60)	age	0-99		1

Statistics for sex = 1(male)
Cells contain: -Row percent -N of cases		incfam
		1 0-19,999	2 20,000-49,999	3 50,000+	ROW TOTAL
hs_bmi	1.0000: <20	48.0 118	46.7 115	5.3 13	100.0 246
	2.0000: 20-<22.5	33.9 338	56.6 564	9.4 94	100.0 996
	3.0000: 22.5-<25	26.0 534	61.6 1,265	12.3 253	100.0 2,052
	4.0000: 25-<27.5	24.1 661	63.4 1,740	12.5 343	100.0 2,744
	5.0000: 27.5-<30	26.3 407	62.2 963	11.5 178	100.0 1,548
	6.0000: 30-<35	26.9 335	62.9 784	10.2 127	100.0 1,246
	7.0000: 35+	34.7 140	58.1 234	7.2 29	100.0 403
	COL TOTAL	27.4 2,533	61.3 5,665	11.2 1,037	100.0 9,235

Statistics for sex = 2(female)
Cells contain: -Row percent -N of cases		incfam
		1 0-19,999	2 20,000-49,999	3 50,000+	ROW TOTAL
hs_bmi	1.0000: <20	28.7 307	57.5 616	13.8 148	100.0 1,071
	2.0000: 20-<22.5	27.3 606	62.2 1,380	10.5 233	100.0 2,219
	3.0000: 22.5-<25	29.5 678	60.1 1,384	10.4 239	100.0 2,301
	4.0000: 25-<27.5	34.6 654	56.8 1,072	8.6 162	100.0 1,888
	5.0000: 27.5-<30	44.1 453	48.9 503	7.0 72	100.0 1,028
	6.0000: 30-<35	41.3 595	50.9 733	7.8 113	100.0 1,441
	7.0000: 35+	48.9 391	45.9 367	5.1 41	100.0 799
	COL TOTAL	34.3 3,684	56.3 6,055	9.4 1,008	100.0 10,747

These tables demonstrate what I call the Skinny Man Effect.

For women, the lower their BMI the higher their predicted family income. Remember that this is family income, so it includes her income if she works, and her husband's income if she's married.

But men have a different distribution. It turns out that men who are officially "overweight" with a BMI between 25 and 27.5 have higher incomes than men who have a "healthy" weight. So if you're a man, being thin may be good for your health (if we are to believe the people who make these definitions), but it's bad for your wallet.

You hear over and over again that poverty is a risk factor for obesity, but as we see above, for men, poverty is also a risk factor for having a "healthy" weight.

There is other evidence that men of lower weights have negative outcomes in life. Thus I call it the Skinny Man Effect.

Let's talk about confounding factors for a minute. As age increases, so does BMI. Age also has a positive correlation with income. I used ages 35-60 because it's an age range with less change in these factors. Men have a big increase in both BMI and income as they progress from 18 to 30, but by 35 that effect is leveling off.

Despite the above confounding factors, I am certain that the Skinny Man Effect is real, because I still see it when I plug in a much narrower age range.

Virginity and the Zodiac

Using the General Social Survey, I discovered that 5.8% of male respondents born in the U.S. are virgins. But 7.2% of men born under the sign of Sagittarius are vigins, and 7.0% of Scorpios are virgins.

Have I discovered that there really is something to this astrology thing? I don't think so.

Sagittarii are born between November 23rd and December 22nd. Many school systems have a grade cutoff of December 31st. Children born January 1st or after will be in a lower grade in those school systems. This means that Sagitarrii are likely to be the youngest children in their grade, and thus the least mature physically, emotionally, and intellectually. This puts them at a disadvantage relative to their older classmates and more likely to grow up lacking the social skills necessary to lose their virginity. This is the same effect as the findings about adolescent height and labor market outcomes.

Scorpios are born between October 24th and Novermber 22nd, which explains why they are also more likely to be virgins.

On the other side of the cutoff date, only 5.4% of Aquarii (born January 20th to February 19th) and 4.7% of Pisces (born February 20th to March 20th) are virgins.

The conclusion is that there may be some social benefit to having your male children be among the older children in their grade at school. (This creates a negative sum game when parents try to game the system and get their kids held back a grade. 20% of children will always be the youngest 20% in their grade, and it's not in society's interests for people's entrance into the workforce to be delayed by a year.)

UPDATE

There has been some confusion in the comments about the cutoff date. The most common cutoff date today is actually September 1, but in the past the most common cutoff dates were 12/1, 12/31 and 1/1, and the later cutoff dates probably apply to most of the GSS respondents.

Here is a research paper on this very issue, and there is a chart at the end of the paper showing the years that states changed their cutoff dates.

Liberal women more likely to be sexually harassed?

In my previous post I found that smarter women are more likely to say they have been sexually harassed (at least that's the interpretation based on a loosely worded question).

Following up a tip from Jason Malloy, here I compare political views to whether a woman reports being sexually harassed.

Variables
Role	Name	Label	Range	MD	Dataset
Row	POLVIEWS	THINK OF SELF AS LIBERAL OR CONSERVATIVE	1-7	0,8,9	1
Column	SEXHAR	R EXPERIENCED SEXUAL HARRASSMENT	1-3	0,8,9	1
Weight	WT2004NR	WEIGHT	.35-5.92		1
Filter	SEX(2)	RESPONDENTS SEX(=FEMALE)	1-2		1
Filter	BORN(1)	WAS R BORN IN THIS COUNTRY(=YES)	1-2	0,8,9	1

Frequency Distribution
Cells contain: -Row percent -N of cases		SEXHAR
		1 YES	2 NO	3 NEVER HAVE WORKED	ROW TOTAL
POLVIEWS	1: EXTREMELY LIBERAL	65.5 19	34.5 10	.0 0	100.0 29
	2: LIBERAL	58.3 95	41.7 68	.0 0	100.0 163
	3: SLIGHTLY LIBERAL	49.1 85	49.7 86	1.2 2	100.0 173
	4: MODERATE	40.5 218	58.6 315	.9 5	100.0 538
	5: SLGHTLY CONSERVATIVE	43.9 90	55.1 113	1.0 2	100.0 205
	6: CONSERVATIVE	41.0 102	57.8 144	1.2 3	100.0 249
	7: EXTRMLY CONSERVATIVE	37.5 12	56.2 18	6.2 2	100.0 32
	COL TOTAL	44.7 621	54.3 754	1.0 14	100.0 1,389

Liberal women are far more likely to say that they have been harassed.

It seems pretty clear to me that how a woman interprets the question is strongly related to her political views.

Smart women are more likely to be harassed

As a woman's intelligence increases (as evidenced by her Wordsum score), the likelihood that she has been sexually harassed on the job increases.

1112. Sometimes at work people find themselves the object of sexual advances, propositions, or unwanted sexual discussions from co-workers or supervisors. The advances sometimes involve physical contact and sometimes just involve sexual conversations. Has this ever happened to you?

Statistics for SEX = 2(FEMALE)
Cells contain: -Row percent -N of cases		SEXHAR
		1 YES	2 NO	3 NEVER HAVE WORKED	ROW TOTAL
WORDSUM	0	20.0 1	80.0 4	.0 0	100.0 5
	1	9.1 1	72.7 8	18.2 2	100.0 11
	2	5.0 1	90.0 18	5.0 1	100.0 20
	3	21.4 9	78.6 33	.0 0	100.0 42
	4	35.2 31	62.5 55	2.3 2	100.0 88
	5	38.2 63	61.8 102	.0 0	100.0 165
	6	47.5 105	50.7 112	1.8 4	100.0 221
	7	49.7 80	49.7 80	.6 1	100.0 161
	8	47.5 56	51.7 61	.8 1	100.0 118
	9	53.9 48	44.9 40	1.1 1	100.0 89
	10	63.5 33	36.5 19	.0 0	100.0 52
	COL TOTAL	44.0 428	54.7 532	1.2 12	100.0 972

Are smarter women really being harassed more? Or are smarter women more likely to interpret harassment from the same fact pattern? I suspect the latter.

Women with higher intelligence have read more about sexual harassment and are more likely to have attended college classes taught by liberal professors who increased their awareness of the issue. Women with lower intelligence are more likely to come from lower social classes where they learn to have lower expectations of male behavior.

It's also possible that the question is so complicated that women with lower intelligence just don't understand it. This is a problem I have with a lot of GSS questions. The smart people who write the questions don't seem to understand that many of the respondents are stupid.

UPDATE

I followed this up with a post showing that liberal women are more likely to report being sexually harassed.

Who attends sports events?

In the 1993 General Social Survery, respondents were asked if the attended a sporting event during the last year:

ATTSPRTS 473. Next I'd like to ask about some leisure or recreational activities that people do during their free time. As I read each activity, can you tell me if it is something you have done in the past twelve months? Let's begin with attending an amateur or professional sports event. Did you do that within the past twelve months?

When you compare this with respondents' degree, we see that the more education one has, the more likely one is to have attended a sports event during the last 12 months:

Frequency Distribution
Cells contain: -Row percent -N of cases		ATTSPRTS
		1 YES	2 NO	ROW TOTAL
DEGREE	0: LT HIGH SCHOOL	27.2 74	72.8 198	100.0 272
	1: HIGH SCHOOL	54.8 425	45.2 351	100.0 776
	2: JUNIOR COLLEGE	64.2 61	35.8 34	100.0 95
	3: BACHELOR	69.8 157	30.2 68	100.0 225
	4: GRADUATE	78.8 82	21.2 22	100.0 104
	COL TOTAL	54.3 799	45.7 673	100.0 1,472

It's expensive attending sporting events, so people without college degrees are less likely to be able to afford it. It has also seemed to me that attending sporting events is a popular activity with the upper middle class.

But a very interesting thing happens when you compare this with the respondent's Wordsum score (the ten word vocabulary test which I use as a proxy for IQ):

Frequency Distribution
Cells contain: -Row percent -N of cases		ATTSPRTS
		1 YES	2 NO	ROW TOTAL
WORDSUM	0	.0 0	100.0 3	100.0 3
	1	16.7 2	83.3 10	100.0 12
	2	21.2 7	78.8 26	100.0 33
	3	43.4 23	56.6 30	100.0 53
	4	47.1 49	52.9 55	100.0 104
	5	45.9 61	54.1 72	100.0 133
	6	62.4 136	37.6 82	100.0 218
	7	68.7 103	31.3 47	100.0 150
	8	63.2 67	36.8 39	100.0 106
	9	60.0 51	40.0 34	100.0 85
	10	54.3 25	45.7 21	100.0 46
	COL TOTAL	55.6 524	44.4 419	100.0 943
Means		6.43	5.68	6.10
Std Devs		1.83	2.26	2.07
Unweighted N		524	419	943

It turns out that attendance at sporting events peaks at Wordsum 7 and then falls as the respondent becomes smarter. This is not what one would have expected based on the fact that people with graduate degrees are most likely to have attended a sports event.

My theory is that, as people become smarter than Wordsum 7, watching sports becomes less interesting.

People who earn more think their efforts count for more

I haven't done any GSS analysis in a while, so here's some relatively random analysis.

The following question was asked in 1984:

USCLASS4 71. Here are different opinions about social differences in this country. Please tell me for each one whether you strongly agree, somewhat agree, somewhat disagree, or strongly disagree. READ EACH STATEMENT. CIRCLE ONE CODE FOR EACH. D. What one gets in life hardly depends at all on one's own efforts, but rather on the economic situation, job opportunities, union agreements, and the social services provided by the government.

Not very surprisingly, respondents with higher incomes think that one's efforts count, but people with lower incomes think that one's own efforts don't mean as much.

Frequency Distribution
Cells contain: -Column percent -N of cases		RINCOM82
		1 < $25,000	2 $25,000+	ROW TOTAL
USCLASS4	1: STRONGLY AGREE	6.6 49	4.6 8	6.2 57
	2: SOMEWHAT AGREE	36.1 267	17.8 31	32.6 298
	3: SOMEWHT DISAGREE	38.4 284	38.5 67	38.4 351
	4: STRNGLY DISAGREE	18.8 139	39.1 68	22.7 207
	COL TOTAL	100.0 739	100.0 174	100.0 913

A lot of the questions in the GSS seem to be put there without very much thought. I don't see anything interesting about this question. Maybe I just haven't dug deep enough. But it does teach us a basic concept of human nature. People attribute successes to themselves, but failures to bad luck and forces beyond one's control.

Religiosity, economic views, and children

There has been lots of blog chatter recently about the fact that conservatives have more children than liberals.

In order to better understand this, I ran two regression analyses. The first analysis predicts the respondent's political views as measured on a seven point scale from liberal to conservative. The second analysis predicts how many children the respondent has.

In both cases, the independent variables are (1) EQWLTH, a seven point scale asking the respondent whether he believes the government should reduce income differences between the rich and poor; and (2) BIBLE, which asks if the respondent thinks the Bible is the direct word of God, the inspired word of God, or just a book of fables.

EQWLTH is a measure of economic libertarianism.

BIBLE is a measure of religiosity.

Let's examine the results:

Variables
Role	Name	Label	Range	MD	Dataset
Dependent	POLVIEWS	THINK OF SELF AS LIBERAL OR CONSERVATIVE	1-7	0,8,9	1
Independent	EQWLTH	SHOULD GOVT REDUCE INCOME DIFFERENCES	1-7	0,8,9	1
Independent	BIBLE(1-3)	FEELINGS ABOUT THE BIBLE	1-4	0,8,9	1
Weight	WT2004NR	WEIGHT	.35-5.92		1
Filter	YEAR(1998-2004)	GSS YEAR FOR THIS RESPONDENT	1972-2004		1

Regression Coefficients					Test That Each Coefficient = 0
	B	SE(B)	Beta	SE(Beta)	T-statistic	Probability
EQWLTH	.180	.010	.252	.015	17.191	.000
BIBLE(1-3)	-.404	.030	-.197	.015	-13.430	.000
Constant	4.219	.070			59.843	.000

Variables
Role	Name	Label	Range	MD	Dataset
Dependent	CHILDS	NUMBER OF CHILDREN	0-8	9	1
Independent	EQWLTH	SHOULD GOVT REDUCE INCOME DIFFERENCES	1-7	0,8,9	1
Independent	BIBLE(1-3)	FEELINGS ABOUT THE BIBLE	1-4	0,8,9	1
Weight	WT2004NR	WEIGHT	.35-5.92		1
Filter	YEAR(1998-2004)	GSS YEAR FOR THIS RESPONDENT	1972-2004		1

Regression Coefficients					Test That Each Coefficient = 0
	B	SE(B)	Beta	SE(Beta)	T-statistic	Probability
EQWLTH	.002	.013	.002	.015	.144	.885
BIBLE(1-3)	-.400	.036	-.166	.015	-11.109	.000
Constant	2.522	.084			30.166	.000

DISCUSSION

We see that whether or not a person considers himself liberal or conservative is an equal mix of religion (more religious corresponds with more conservative), and libertarian economic views (more libertarian corresponds with more conservative).

The T-statistic is higher for the EQWLTH question, possibly because the EQWLTH question is a seven point scale while the BIBLE question is only a three point scale. It would be fair to say that religion and economic views are given approximately equal weight when people consider their liberal-conservative orientation.

When we look at how many children the respondent has, the respondent's answer to the EQWLTH question has no correlation with the number of children. But the BIBLE question has about as strong a correlation with number of children as it did with political views.

This means that when people like Steve Sailer make a big deal about the fact that conservatives have more children than liberals, they are not accurately understanding the real correlation.

Being conservative is correlated with more children only because religiosity is correlated with more children. It has nothing to do with "conservative" economic views.

Thus it's misleading to talk about how conservatives have more children than liberals. The true story is that religious people have more children than secular people, and that economic conservatives have no more and no less children than economic liberals.

High IQ doesn’t lead to higher income, part IV

CORRECTION

Dan Morgan, from the blog No Speed Bumps, has pointed out a major discrepency in the Lupinski article. There is indeed a chart showing a much higher percentage of participants having $100K+ income. (Dan Morgan says 37%.)

Perhaps, the authors meant to say that 7.9% of participants earned at least $250,000. It might be possible that 60% of graduates of top ten MBA programs would be making at least $250,000 by the time they are 33. Those investment banks pay a lot of money. You just need the magic of a degree from the right school in order to open the piggy bank.

If the 37% figure is correct, this in no way invalidates the tracks theory of labor market outcomes. The majority of paticipants had graudate degrees of some kind, including highly valuable graduate degrees like MDs and top ten MBAs. If a similar percentage of the law school graduates also attended top law schools, they would also mostly be making more than $100,000.

It's pretty pathetic that a published research paper, signed by so many authors, would have such a glaring error in it. What's the point of reading these published papers if the quality is no better than what you read in blogs?

ORIGINAL POST WHICH IS BASED ON WHAT IS POSSIBLY A TYPO

A research article, Tracking Exceptional Human Capital Over Two Decades by Lubinski, Benbow, Webb & Bleske-Rechek (link to PDF file) was brought to my attention. It was purported to show the fabulous success of those with high IQs, but a closer examination shows just the opposite, and confirms the findings from my previous post on this topic.

This research article tracked people (286 males, 94 females) who “scored in the top 0.01% on cognitive-ability measures” before age 13. This was determined by an early administration of the SAT. These high IQ people were looked at when their average age was 33.6.

Now of course intelligence is primarily a genetic trait, so people identified in their pre-teen years as being super smart are going to stay super smart for the rest of their lives. So the group had great academic success, attended the best colleges, and “doctoral-level degrees (Ph.D., M.D., or J.D.) were earned by 51.7% and 54.3% of male and female TS participants.”

But my interest has been to see if people with high IQs earn more money, and it seems that based on this research paper they don’t.

Interestingly, the paper talks about participants who earned at least $100,000/year without directly telling us how small a percent earned that much money. Well, a little 8th grade algebra was able to ferret out this information. The paper informs us that 20 survey participants had MBAs, that 46.2% of the participants who earned more than $100,000 had MBAs, and that 60% of participants with MBAs reported more than $100,000 of income. Doing the algebra, we discover that only 30 participants had $100,000+ salaries, only 7.9% of the sample.

How does this compare to General Social Survey respondents? I looked at GSS respondents aged 30-37 because this is within four years of the average age of 33.6.

For male GSS respondents who had at least a college degree (N=221), 11.3% reported income of at least $100,000. (For the GSS income group of $90,000 to $109,999, I assumed that half had at least $100.000.) For female respondents with at least a college degree (N=235), that percentage was only 2.3%. Now if we apply these percentages to the male/female breakdown of the participants in the Lubinski paper, we would expect that 9.07% would have incomes of at least $100,000. This is greater than the actual finding of 7.9%.

So we see, the financial success of a group of super-elite high IQ individuals actually lags a group of just average Joes with college degrees.

Why did these super-high IQ individuals have such mediocre career success compared to average college graduates? It all goes back to my tracks theory of labor market outcomes. The people with super-high IQs who used their IQs to obtain valuable degrees (such as MBAs from top schools) which allowed entry to the best career tracks did very well financially, but the vast majority of the sample pursued academic programs that led to lower paying careers.

If one believes that a person’s salary is a measure of that person’s contribution to the economy, then the conclusion is that super-high IQ people are not contributing more to the economy than average college graduates.

Happiness and children for senior citizens

Yesterday I blogged about how child-free women are happier.

Some commenters suggested that children only make you unhappy while you’re raising them, but then they bring great joy to your life when they’re adults.

To help answer this important question, I prepared charts for respondents 65 or older. I figure that by the time you reach that age, your children are independent.

Statistics for MARITAL = 1(MARRIED)
Cells contain: -Row percent -N of cases		HAPPY
		1 VERY HAPPY	2 PRETTY HAPPY	3 NOT TOO HAPPY	ROW TOTAL
CHILDS	0: NONE	46.6 182	46.0 180	7.4 29	100.0 391
	1: ONE	48.6 229	45.1 213	6.4 30	100.0 472
	2: TWO	50.0 438	43.6 381	6.4 56	100.0 875
	3: THREE	44.1 304	49.4 340	6.6 45	100.0 690
	4: FOUR	50.0 230	43.4 200	6.6 30	100.0 461
	5: FIVE	45.0 91	45.5 92	9.5 19	100.0 202
	6: SIX	40.9 47	48.3 55	10.8 12	100.0 114
	7: SEVEN	49.8 38	41.0 31	9.2 7	100.0 76
	8: EIGHT OR MORE	46.3 58	46.0 58	7.7 10	100.0 126
	COL TOTAL	47.5 1,617	45.5 1,550	7.0 239	100.0 3,406

Statistics for MARITAL = 2(WIDOWED)
Cells contain: -Row percent -N of cases		HAPPY
		1 VERY HAPPY	2 PRETTY HAPPY	3 NOT TOO HAPPY	ROW TOTAL
CHILDS	0: NONE	26.5 121	56.8 259	16.7 76	100.0 456
	1: ONE	26.4 138	54.5 286	19.1 100	100.0 525
	2: TWO	26.5 186	57.6 404	15.9 112	100.0 702
	3: THREE	26.4 146	55.2 304	18.4 101	100.0 551
	4: FOUR	28.7 97	53.4 180	17.9 61	100.0 338
	5: FIVE	30.6 57	54.1 101	15.3 29	100.0 188
	6: SIX	29.5 36	52.1 64	18.4 23	100.0 123
	7: SEVEN	23.0 18	57.5 45	19.5 15	100.0 78
	8: EIGHT OR MORE	21.9 32	60.5 89	17.5 26	100.0 147
	COL TOTAL	26.7 831	55.8 1,733	17.5 543	100.0 3,107

As we see from the charts above, children cause a minor increase in happiness for married older respondents.

I thought that children might be especially important for widowed older respondents, because perhaps their children might keep them company or give them emotional support after their spouses die. But very surprisingly, children provide no happiness benefit to the widowed unless they have at least four of them.

Child-free women are happier

Every GSS respondent is asked how happy he or she is.

HAPPY 157. Taken all together, how would you say things are these days--would you say that you are very happy, pretty happy, or not too happy?

Frequency Distribution
Cells contain: -Row percent -N of cases		HAPPY
		1 VERY HAPPY	2 PRETTY HAPPY	3 NOT TOO HAPPY	ROW TOTAL
MARITAL	1: MARRIED	40.4 9,860	51.9 12,670	7.7 1,867	100.0 24,398
	2: WIDOWED	24.6 1,079	56.0 2,459	19.4 851	100.0 4,388
	3: DIVORCED	19.6 956	62.3 3,042	18.1 886	100.0 4,884
	4: SEPARATED	15.8 232	56.2 823	28.0 411	100.0 1,465
	5: NEVER MARRIED	22.6 1,847	63.1 5,169	14.3 1,170	100.0 8,186
	COL TOTAL	32.3 13,974	55.8 24,163	12.0 5,184	100.0 43,321

As we see from the above table, married people are a lot happier. It’s not entirely clear to me whether marriage makes people happier, or if unhappy people are less attractive to the opposite sex and therefore are less likely to get married or to stay married.

Because widowed respondents seem pretty unhappy, the evidence is that marriage does make people happier than they otherwise would be if the were unmarried.

Because marriage is such a huge component of happiness, my future analysis of happiness will be restricted to married respondents.

Frequency Distribution
Cells contain: -Row percent -N of cases		HAPPY
		1 VERY HAPPY	2 PRETTY HAPPY	3 NOT TOO HAPPY	ROW TOTAL
AGE	1: 18-34	38.4 2,701	54.2 3,811	7.3 516	100.0 7,028
	2: 35-64	39.7 5,515	52.4 7,282	7.9 1,102	100.0 13,899
	3: 65-98	47.4 1,620	45.5 1,555	7.1 242	100.0 3,417
	COL TOTAL	40.4 9,836	52.0 12,649	7.6 1,860	100.0 24,345

We see from the above table that senior citizens are happier. My theory is that it’s because they get free money form the government and they don’t have to work. Most people are happier when they don’t have to get yelled at by a boss or worry about losing their job.

Now let’s compare family income (not the respondents’ income like most of my other income analyses) in 1986 dollars for respondents under the age of 65 to happiness:

Frequency Distribution
Cells contain: -Row percent -N of cases		HAPPY
		1 VERY HAPPY	2 PRETTY HAPPY	3 NOT TOO HAPPY	ROW TOTAL
REALINC	1: 0-19999	32.2 1,428	54.6 2,421	13.3 589	100.0 4,438
	2: 20000-39999	38.2 3,168	54.8 4,546	7.1 589	100.0 8,302
	3: 40000-89999	43.3 2,008	51.6 2,390	5.1 236	100.0 4,634
	4: 90000-999999	46.5 971	49.7 1,038	3.8 80	100.0 2,090
	COL TOTAL	38.9 7,574	53.4 10,396	7.7 1,494	100.0 19,464

It turns out that the platitude that “money can’t buy you happiness” is wrong. Higher income leads to greater happiness.

Perhaps I’d be stating the obvious to point out that it sucks to be poor.

Finally, the headline comparison:

Statistics for SEX = 1(MALE)
Cells contain: -Row percent -N of cases		HAPPY
		1 VERY HAPPY	2 PRETTY HAPPY	3 NOT TOO HAPPY	ROW TOTAL
CHILDS	0: NONE	39.9 671	53.3 896	6.8 115	100.0 1,681
	1: ONE	34.7 670	57.0 1,100	8.2 159	100.0 1,929
	2: TWO	40.3 1,429	52.3 1,853	7.4 261	100.0 3,543
	3: THREE	39.8 880	52.4 1,159	7.8 172	100.0 2,212
	4: FOUR	40.3 474	51.4 605	8.3 97	100.0 1,176
	5: FIVE	38.4 202	51.0 268	10.5 55	100.0 525
	6: SIX	34.7 87	52.5 132	12.8 32	100.0 252
	7: SEVEN	39.6 61	46.7 72	13.8 21	100.0 154
	8: EIGHT OR MORE	43.5 80	43.6 80	12.9 24	100.0 184
	COL TOTAL	39.1 4,555	52.9 6,166	8.0 936	100.0 11,657

Statistics for SEX = 2(FEMALE)
Cells contain: -Row percent -N of cases		HAPPY
		1 VERY HAPPY	2 PRETTY HAPPY	3 NOT TOO HAPPY	ROW TOTAL
CHILDS	0: NONE	44.5 780	49.3 865	6.2 109	100.0 1,754
	1: ONE	41.6 902	50.7 1,100	7.7 168	100.0 2,170
	2: TWO	42.3 1,646	51.2 1,991	6.5 253	100.0 3,890
	3: THREE	40.6 982	52.3 1,265	7.0 170	100.0 2,418
	4: FOUR	41.4 515	50.3 626	8.4 104	100.0 1,245
	5: FIVE	40.5 231	51.3 293	8.2 47	100.0 571
	6: SIX	38.6 108	51.1 143	10.3 29	100.0 279
	7: SEVEN	36.4 59	52.3 84	11.3 18	100.0 161
	8: EIGHT OR MORE	33.2 69	54.5 113	12.3 26	100.0 208
	COL TOTAL	41.7 5,292	51.0 6,481	7.3 924	100.0 12,697

For men, one child makes a man unhappy, but having two to four is neutral compared to having none at all. I don’t understand what’s so bad about having just one.

You would have expected that women would be happier if they had children, but the reality is that child-free women are happier. The lesson for women is that they will be happier if the focus on their career and not have children, because money makes a woman happy but children make a woman unhappy.

What exactly is the purpose of having children?