more EVIDENCE  - however basically policymakers in Canada dont give a
sh**.]

http://www.statcan.ca/bsolc/english/bsolc?catno=11F0019M&CHROPG=1

Analytical Studies Branch Research Paper Series

Income and the Outcomes of Children
by Shelley Phipps and Lynn Lethbridge
May 2006

Abstract
This report re-investigates the connection between income and child
well-being for a broad range of
outcomes. The report attempts to address four research questions:

The major results derived from the regression analysis are summarized
below:

1. Higher income is almost always associated with better outcomes for
children. This is true
regardless of the measure of income employed, the assumed functional form
of the
relationship between income and child outcomes, the age of the child, or
the type of child
outcome being studied. It is also apparently true using either the NLSCY or
the Youth in
Transition Survey (YITS)/Programme for International Student Assessment
(PISA) data.

2. The size of the association between income and child outcomes varies
with developmental
domain. Thus, for example, income has particularly strong associations with
cognitive
outcomes (e.g., Peabody Picture Vocabulary Test (PPVT) scores or math and
reading
scores) and behavioural outcomes (e.g., hours spent watching television).
Physical health
outcomes also have quite consistent positive associations with family
income. Associations
are generally smallest with ‘social/emotional’ outcomes (though
hyperactivity is an
exception to this ‘rule’). Again, descriptive evidence from the YITS/PISA
are consistent
with these findings.

3. A three-period average of family equivalent income consistently has the
largest associations
with child outcomes. This is true across almost all kinds of outcomes and
all ages of
children. It is also true for children living in married-couple or
lone-mother families. Thus,
in general, it appears advisable to use an income measure averaged over as
many years as
are available in the data

4. The functional form of the relationship between family income and child
outcomes varies
considerably across developmental domains. Estimating an inappropriate
functional form
may lead a researcher to the (false) impression that income does not
matter, so that it is
important to test a variety of alternatives.

5. It is almost never true that beyond the low-income threshold, income is
unimportant for
children’s outcomes. (Non-nested hypothesis tests reject ‘low-income’
specifications in
favour of continuous income specifications in almost every case.)

6. While over half of the outcomes studied here increase more quickly with
income at lower
than at higher income levels, it is almost always true that there is not a
ceiling above which
income no longer matters for child outcomes.

7. The relationship between incomes and outcomes appears to ‘flatten’ out
toward the linear
for older as compared to younger children. Thus, increases in income at
very low-income
levels are particularly important for the youngest children.

8. We find little evidence of differences in the size of the income/outcome
associations for
children of different ages. However, only a subset of outcomes are measured
identically
across age groups.

9. For middle-aged and for older children, changes in family income appear
to be less
important for child outcomes than levels of family income. However, income
changes are
more important for younger children, particularly if they happen earlier in
life (i.e., between
1994 and 1996 rather than between 1996 and 1998). Income ‘ups and downs’
are
particularly important for child emotion scores, an outcome for which
income levels appear
to play a less important role.

10. Results using the Youth in Transition Survey show a positive
correlation between
socioeconomic status and child outcomes, in general. While income was not
available in this
survey at the time of analysis, direct comparisons by income are not
possible. Therefore a
proxy for income in a socioeconomic status (SES) index is used with
outcomes which can
be categorized into the same categorical domains as the NLSCY. It is
interesting that the
differences across SES quintiles are particularly large for outcomes in the
cognitive and
behavioural domains and somewhat smaller for social and emotional domains
which fits
with results obtained using the NLSCY.

Keywords: labour market participation, family income, cognitive outcomes,
children,

-------------------
Problems/Questions? Send it to Listserv owner: [log in to unmask]


To unsubscribe, send the following message in the text section -- NOT the subject header --  to [log in to unmask]
SIGNOFF SDOH

DO NOT SEND IT BY HITTING THE REPLY BUTTON. THIS SENDS THE MESSAGE TO THE ENTIRE LISTSERV AND STILL DOES NOT REMOVE YOU.

To subscribe to the SDOH list, send the following message to [log in to unmask] in the text section, NOT in the subject header.
SUBSCRIBE SDOH yourfirstname yourlastname

To post a message to all 1000+ subscribers, send it to [log in to unmask]
Include in the Subject, its content, and location and date, if relevant.

For a list of SDOH members, send a request to [log in to unmask]

To receive messages only once a day, send the following message to [log in to unmask]
SET SDOH DIGEST

To view the SDOH archives, go to: https://listserv.yorku.ca/archives/sdoh.html